Beruflich Dokumente
Kultur Dokumente
edited by
David N.-S. Hon
Clemson University
Clemson, South Carolina
Nobuo Shiraishi
Kyoto University
Kyoto, Japan
M A R C E L
MARCEL DEKKER,
INC. NEWYORK BASEL
U E K K E R
Library of Congress Cataloging-in-Publication Data
Life and its surroundings are constantly changing within our dynamic world. As we stride
into the new millennium, information technology and biotechnology continue to flourish.
Rapid economic expansion, social development, and high demands for shelter, clothing,
energy, and food for our overpopulated world have resulted in a desperate need for new
and yet functional materials to support society’s infrastructure.
Wood or lignocellulosic-based materials have made a significant contribution to the
quality of living for human beings. With new developments in wood chemistry, scientists
are confident that wood will continue to play an important role in fulfilling the needs of
human beings.
Over the past decade, the trend of emphasizingbio-basedtechnologieshasbeen
observed worldwide. In February 1998, a long-term development project, PlanVCrop-based
Renewable Resources 2020, was implemented among the U.S. Department of Agriculture,
U.S. Department of Energy, and many U.S. companies, agricultural associations, and uni-
versities. The aim of the project was to obtain novel chemicals from plant- and crop-based
renewable resources in order to widen the usage of crops, the yield of which has been
significantly increased through bio-technological advancements. The recent movement of
producing foods by means of genetically manipulated seeds should enhance the effective-
ness of this project. Before the start of this project-which is considered the future of the
petrochemicalindustry-majorchemicalcompanies in the UnitedStates,suchasDow
Chemical,Dupont, and Monsanto,havebeenchanging their strategies in research and
development.Theyhavestrengthened their bio-basedresearch field, trying to yield as
many chemicals as possible from biomass. They are developing production technologies
for ethanol, sorbitol, lysine, tryptophane, citric acid, lactic acid, poly(lactic acid), erythritol,
1,3-propanediol,etc.,frombiomass.Furthermore, in August of 1998 PresidentClinton
issued an executive order, “Developing and Promoting Biobased Products and Bioenergy,”
to further the development of a comprehensive national strategy that includes research.
development, and private sector incentives to stimulate the creation and early adoption of
technology needed to make bio-based products and bio-energy cost-competitive in national
and international markets. Also. there has been research in so-called “green chemistry.”
In this new methodology. biomass is the recommended MW material. Thc importance of
wood and cellulose rescarch is thus rccognizcd.
iii
iv Preface
Since the publication of' the first edition of this book, considerablc advancement i n
various fields ofwood chemistry has been made, as can be attested by many scientific
publications in addition to well-attended international conferences. We contacted the con-
tributors to the first edition, soliciting their opinions on revising and updating the book,
and we received tremendous support from them as well as the publisher. Unfortunately,
and inevitably, several authorswereunableto participate, buttheyrccomrnended their
successors.
Although most of the chapters in this new edition carry the same titles as those i n
the previous edition, they have all been extensively revised and updated. In addition, this
edition includes several new chapters representing important threads in the total fabric of
wood chemistry. These new chapters cover the subjects of chemical synthesis of cellulose,
preservation of wood, preservation of waterlogged wood, biodegradable polymers from
lignocellulosics, recycling of wood and fiber products, and pulping chemistry.
As editors, we feel fortunate to have been able to recruit some of the best talent in
the field to this endeavor. We thank the contributors for their efforts. Any praise for the
content should be addressed to them, and comments and criticisms to us will be welcome.
6. Chemistry of Extractives 2 13
Toshitrki U r ~ r t w ~ ~ r
V
vi Contents
Hiroshi Harada Division of Forest and Biomaterials Science, Graduate School of Ag-
riculture, Kyoto University, Kyoto, Japan
Akira Isogai Department of Biomaterial Science, The University of Tokyo, Tokyo, Japan
Yuan-Zong Lai Faculty of Paper Science and Engineering, SUNY College of Environ-
mental Science and Forestry, Syracuse, New York
vii
viii Contributors
Kazumasa Shimizu Division of Wood Chemistry, Forestry and Forest Products Research
Institute. Ibaraki, Japan
Nobuo Shiraishi Division of Forest and Biomaterials Science. Graduatc School of Ag-
riculture. Kyoto University, Kyoto. Japan
Mariko Yoshioka Division of Forest and Biolnaterials Science, Graduate School of Ag-
riculture, Kyoto University, Kyoto, Japan
Ultrastructure and Formation of Wood
Cell Wall
1
Harada 2 and Fujita
tinctions represent extremes and there are many intermediate arrangements of the vessels.
Variations in arrangements of these vessels with other xylem tissues such as parenchyma
are reflected in the “figure” and “grain” of the wood itself when it is cut from the tree.
The physical properties of wood such as density also result from such arrangements of
the cells.
3. Reaction Wood
Reaction woods that appear on branches or a leaning stem by any force such as a landslide
or snowfall have a peculiar nature. Once reaction wood is formed as a biological response,
the living tree tries topreserve the original position of its stemorbranches.For the
practical use of woods, the reaction woods have not been appreciated very much because
of their different characteristics fromnormalwood in both a physical and a chemical
sense.
The occurrence and nature of reaction woods contrast quite a bit between softwood
and hardwood. In softwood trees, the reaction wood forms at the lower side of a leaning
stem or branches, where the compression stress reacts on the xylem. Therefore, this re-
action wood is generally called “compression wood.” compression woodis heavy and
appears dark brown on account of its highly lignified tracheid walls (see Section II), which
seem to adapt to compression stress. Thus, compression wood is easily distinguished from
normal wood by its dark color. The cambial activity at the lower position of a leaning
stemorbranchacceleratesveryquicklyanddevelops a widercompressionareathan
normal wood on the opposite side. Through the accumulation of compression wood tra-
cheids over many years, a leaning stem will return gradually to the vertical position. The
annual rings of such a stem, however, are conspicuously eccentric.
On the contrary, reaction wood in many species of hardwoods is formed at the upper
side of a leaning stem or branches where the xylem loads the tensile stress. Therefore,
such reaction woods are called “tension wood.” Fibers of tension wood have a slightly
lignified cell wall (see Section 11) that is adapted to the tensile stress just like a bowstring.
It is not so easy to distinguish this area from a normal one on account of its slightly pale
tone, in comparison to the case of compression wood.
Formation
Ultrastructure
and Wall of Cell 3
B. Wood Cells
Wood cells are produced in the vascular cambium from two types of meristematic cells:
the fusiform initial and the ray initial (Fig. l ) . Since cells derived fromthe fusiform initials
that are upright in the stem occupy a major part of xylem, woods show remarkable an-
isotropism. The principal functions of xylem tissue are water conduction from roots to
shoots, the mechanical support of a huge tree body, and a physiological role such as the
storage of starch. Although these functions are common in both softwoods and hardwoods,
the xylem of the latter is more evolved than that of the former, being adapted to each
function.
softwood hardwood
pits
%I
&#
a
cells
.B@'
ray tracheid
FIGURE 1
fusiform
initials
xial parenchyma
ray
initials
\
axial parenchyma c e l l s
parenchyma
p
ray cell
formation. Radial resin canals or latex tubes are formed in the ray tissue of some tropical
hardwoods.
Wood is a natural composite material and a chemical complex of cellulose, lignin, hemi-
celluloses, and extractives [2]. Cellulose is the framework substance, comprising 40-50%
of wood in the form of cellulose microfibrils, whereas hemicelluloses are the matrix sub-
stances present between cellulose microfibrils. Lignin, on the other hand, is the encrusting
substance solidifying the cell wall associated with the matrix substances. The significance
of lignin as the encrusting substance can be demonstrated by examination of the lignin
skeleton created by the acid removal of carbohydrates (Fig. 2).
The roles of these three chemical substances in the cell wall are compared to those
of the constructing materials in the structures made from the reinforced concretein which
cellulose, lignin, and hemicelluloses correspond, respectively, to the iron core, cement,
and buffering material to improve their bonding.
A. Cellulose Microfibrils
The crystalline nature of cellulose in wood has been demonstrated by studies with X-ray
diffractometry and polarization microscopy. This crystalline nature was also confirmed by
the electron diffraction patterns of the secondary walls of wood cells in selected areas [3].
Figure 3a isatransmissionelectronmicrograph of a longitudinalsection of latewood
tracheids of Pinus densifloru, showing the intercellular layer (I), and the S, and S, layers.
The electron diffraction diagram is of a selected area in S2 (Fig. 3b), which is represented
by a small circle. The (101), (loi), and (002) of the equatorial reflections and (040) of
FIGURE 2 Electron micrograph of ultrathin transverse section of earlywood tracheids from Pinus
densgora, showing thedistribution of lignin inthe cell wall, which was skeletonized using the
hydrofluoric acid technique.
6 Fujita and Harada
FIGURE 3 (a) Electron micrograph of ultrathin longitudinal section of tension wood fibers from
Pinus densiporu. (b) The corresponding diffraction diagram taken from the encircled area.
Formation
Ultrastructure
and Wall of Cell 7
the meridional reflection can be seen. It should be noted that crystallographic planes are
based on the Meyer and Misch (1937) model of the unit cell of cellulose I, i n which the
h axis (the fiber axis) is vertical.
I t iswell known that in the wood cell wall, celluloseexists in the form ofthin
threads with an indefinite length. Such threads are called cellulose microfibrils, and they
play an important role in the chemical, physical, and mechanical properties of the wood.
The greenalga, Kdorzia. which is oneform of Chlorophyceae,hasbeenstudied
intensively by microscopists and crystallographers as an excellent material for the ultra-
structural study of cellulose microfibril. Why then is W o t z i a used for the study of the
cellulose microfibril of the wood cell wall? Because the cell walls of Valonia are unlig-
nified, their microfibrils are readily isolated. Furthermore, as described later, Vrtlonicc mi-
crofibrils are approximately 20 nm in width, which is about five times larger than those
of wood, and they are highly crystallized. However, the difference between algal micro-
fibrils such as those of Vhlorli~zand ordinary ones produced by the higher plants also must
be stressed. One of the differences is the selectively uniplaner orientation of algal micro-
fibrils, that is, the ( 101) plane facing the cell surface, while cellulose microfibrils of higher
plants are randomly oriented, although both microfibrils are laid along the cell surface i n
their longitudinal direction [3,4]. The other is the crystallographic heterogeneity in algal
microtibrils as detected by NMR [ S ] , and a triclinic system mixed with an ordinary mono-
clinic system was detected by electron diffraction [6]. The interface between these systems
is not yet shown, although the former amounts to about 50%.
FIGURE 4 Electron micrographs of the cellulose microfibrils of Vuloniu mucrophysa (a) and of
Pinus densifloru holocellulose (b) (disintegration, negatively stained with uranyl acetate), showing
the difference of cellulose microfibril width between wood and Vuloniu.
Formation
Ultrastructure
and Wall of Cell 9
'Reflection examined.
hModewldth.
Source: Ref. 12.
crystalline regions of cellulose. Therefore, the difference in the microfibril width must be
ascribed to that in the size of cellulose crystallites. In addition, the values obtained are
not always equal to the 3.5 nm in elementary fibrils proposed by Muhlethaler [ l l ] .
crystalline nature. It is interesting to note in Fig. 5a that a Vuloniu microfibril does not
have any subunits corresponding to the elementary fibrils [13,17]. Additionally, cellulose
microfibrils appear to be almost square in their cross section in both wood and Vuloniu
[15-171.
FIGURE 8 Lattice images of the cross-sectional face of cellulose microfibrils from Valonia ma-
crophysa, showing the lattice spacings of 0.60, 0.54, and 0.39 nm, respectively.
12 Fujita and Harada
microfibril is rather smaller compared with that of Valonia. In the near future, beam dam-
age at room temperature against wood cellulose microfibrils would be reduced at least 10
times with cryo-electron microscopy.
The cellulose microfibrils of the gelatinous layer of poplar (Populus eurntnictrna)
tension wood in disintegrated samples are found to have many kinks, suggesting that the
cellulose microfibril is highly crystalline [13]. However, the cellulose microfibrils of the
gelatinous layer, about 100 nm in length prepared by ultramicrotome, become shorter than
their original length upon hydrolysis [ 131. As a result, the crystalline regions in the cel-
lulose microfibril of wood cell wall are thought to havesomecrystallinedislocations
caused by chain ends [ 181.
The cellulose microfibrils consist of a core crystalline region of cellulose surrounded
by paracrystalline cellulose and short-chain hemicellulose. Lignin encases them and binds
them into a rigid structure of wood cell wall.
layers of the secondary wall (S), and the warty layer (W) in an earlywood tracheid. The
same layering structure from an earlywood tracheid of Pinus densiflot-a is shown more
clearly in a longitudinal section that was skeletonized by the hydrofluoric acid technique
(Fig. 12).
Figure 16, (pg. 18). shows the texture of the P wall diagrammatically. The microfibril
orientation in the primary wall was interpreted by the multinet growth hypothesis proposed
by Roelofsen [23] and supported for the differentiating conifer tracheids by Wardrop[24].
In the multinet hypothesis, the microfibrils are first deposited transversely to the cell axis
and passively shifted longitudinally during cell extension. From an opposite viewpoint, an
orderedfibrilhypothesiswasproposed by Roland et al. [25] in order to interpret the
crossed polylamellated structure in the primary wall of parenchyma cells. According to
this hypothesis, whether the orientation of microfibrils becomes transverse, oblique, or
longitudinal is determined at the time of deposition of cell wall and may not be changed
thereafter. Recently, Fujii et al. [26] proposed a modified multinet hypothesis of microfi-
brils orientation in the primary wall. The difference between this conceptand Roelofsen’s
theoryisthattheshift of microfibrilorientation during cell extension is made in the
individual lamella and each lamella becomes thin on the outer surface of the P wall due
to extension.
The three layers of the secondary wall, designated S,,S2, and S3,are organized in a
plywood type of construction. The S , or S3,with a large microfibril angle to the cell axis,
is designated as a flat helix, and the S2,with a small angle, as a steep helix (see Fig. 16).
It is also shown that the layers themselves are of lamellae of microfibrils with varying
amounts of shift in orientation, visible in the transmission electron micrograph. The S , is
composed of several lamellae with alternating S and Z helices of microfibril orientation
[28,29], and this structure in the S , is termed “crossed fibrilar texture” [28]. Figure 13 is
F
FIGURE 14 Electron micrograph of ultrathin oblique section of an earlywood tracheid in Pinus
densijffloraand microfibril angles in the secondary wall: top, I. bottom, lumen side.
Formation
Ultrastructure
and of Cell Wall 17
to the tracheid axis from the top S, to the bottom S,. The horizontal line in the upper part
of the figure shows the angles of microfibrillar orientation, the symbols (-) and (+),
respectively, referring to Z and S helices. The gradual changes of the microfibrillar angles
from S, to S, and from S, to S, are shown there.
The helical cellulose microfibril orientationin the S, is typically demonstrated in X-
ray diagrams of wood [32]. The arcs at 0.39, 0.54, and 0.60 nm in the X-ray diagram of
wood show that the cellulose crystallites (microfibrils) lie in a helix around each wood
fiber or tracheid. The microfibril orientations are Z helices in the S, and S helices in the
S,, although S, is the crossed arrangement of S and Z helices.
Preston [32] suggests that the structure with various microfibril angles in the sec-
ondary wall passes through only one cycle, but this may be the brief duration of wall
thickening in higher-plant cell walls compared with that in algae. Roland and Mosiniak
[33] presented a diagram regarding the changesof cellulose microfibril angles in the case
of a secondary wall of tracheids and wood fibers (Fig. 15).
Figure 15 illustrates the case between the S, and S, layer. The change of microfibril
angle is regular and continuous between the S , and S, layers,butitstopsduringthe
deposition of the S , layer. Afterwards the change of microfibril angle reopens toward S,
layer deposition.
The texture of cellulose microfibrils in the P and S walls of softwood tracheids and
hardwood fibers is shown as a schematic diagram in Fig. 16. The thin primary wall (P)
consists of a loose aggregation of microfibrils oriented more or less axially to the cell axis
on the outer surface. The S, layer is a flat helix but with crossed structure, whereas the
S, layer is a steep helix and the S, layer is a flat helix. There are intermediate layers: the
S,,, present between the S , and S, layers; and the S,,, between the S, and S, layers.
The spiral thickening is the ridge of microfibrils that exist on the inner surface of
the S, layer. The spiral thickening is considered part of the S, layer becauseof its continuity
with the S, layer and parallel arrangement to the S, layer microfibrils.
FIGURE 16 Schematic diagram of the microfibril orientation in the primary wall and different
layers of the secondary wall from tracheids and fibers: Po,PI;outer and inner parts of the primary
wall; SlzrS23, intermediate layers between S , and S , and between S2 and S,, respectively.
The warty layer is one of the major structural features of wood cells found by
electron microscopy [34]. It was first foundin softwood tracheids and laterin the tracheids,
vessels, and wood fibers of hardwoods (see Fig. 11). The major chemical constituents of
warts arereported to be lignin and hemicelluloses according to examination by component
removal treatment of ultrathin wood sections [35]. The warts are believed to arise from
the extra wall materials and remains S, layer through
of cytoplasm that are deposited on the
the plasma membrane [36,37]. The warty layer is not found in all softwoods and hard-
woods [30,38]. Parham and Baird [39] have pointed out that the appearance of warts in
wood has a phylogenetic trend. Softwood tracheids and primitive hardwood cells nearly
always have warts, but as the cell types become more advanced or specialized, they be-
come wart-free.
2. Vessels
The texture of cellulose microfibrils in the walls of specialized cells such as vessel ele-
ments and parenchyma cells cannot be readily described as in softwood tracheids and
hardwood fibers. A concept of standardized cell wall organization in vessel elements was,
however, represented by Kishi et al. [40,41]. The microfibrils in the primary wall extend
straight and are arranged parallel to one another within one lamella, and the wall consists
of three parts, P-outer, P-middle, and P-inner, each showing a different microfibril orien-
tation. The microfibrils are oriented transversely with respect to the vessel axis in the P-
outer and are oriented at random in the P-middle. The P-inner consists of a crossed poly-
lamellatedstructure. It isalsoreportedfromthe examination of vessel elements from
nearly 30 Japanese hardwoods with polarizing and electron microscopy that the layered
structure of the secondary wall can be classified into three categories: the typical three-
layered structure, an unlayered structure, and a multilayered structure. The typical three-
layered structure consists of S,, S2,and S3 similar to those of softwood tracheids and
hardwood fibers, although the S, and S3 layers are thicker than those of tracheids and
Formation
Ultrastructure
and Wall of Cell 19
wood fibers. The unlayered structure has only microfibrils, with the orientation of a flat
helix. The multilayered structure has more than four layers, in which microfibril angles to
the vessel axis change. This type of structure contains in some cases the so-called bow-
shaped pattern.
Figure 17 isa TEM micrograph of the transverse sectionof Cinnamomum camphora
and shows the microfibril angle and helix in the part of the bow-shaped pattern appearing
on the vessel wall of the multilayered type of structure. As shown in Fig. 17, the pattern
results from the progressive changes of microfibrillar orientation in the wall from 90" to
0" and from 0" to 90".
3. Parenchyma Cells
In spite of the fact that parenchyma cells had been generally considered to have only
primary wall, thoseof wood are reported sometimes to develop secondary wall, in addition
to complicated primary wall. It is evident from recent studies that ray and axial paren-
chyma cells in both softwoods and hardwoods have variations or complexities in their
wall structure that are not observed in the cell walls of tracheids and wood fibers.
In softwoods, the cell wall structure of the ray parenchyma cells was divided into
five categories by Fujikawa and Ishida [42]. However, as shown in Fig. 18, it is funda-
mentally classified into two types; the firsttype consists of the primary walland protective
90
FIGURE 17 Electron micrograph of ultrathin oblique sectionof a vessel wall stained withKMnO.,
from Cinnamomum camphora, showing a bow-shaped pattern (a) and the microfibril angles and
helices (b).
20 Fujita and Harada
L-
I
l
I P
i
I
S1 i S2 1
I
FIGURE 18 Schematicdiagram of themicrofibrilorientation in the cell wall of softwood ray
parenchyma cell: (a) the first type; (b) the second type. (From Ref. 41.)
layer (Fig. 18a), and the second type consists of the primary wall, secondary wall, and
protective layer (Fig. 18b) [42]. However, the protective layer and a random arrangement
of microfibrils is omitted in this figure. The P, appears with microfibrils of almost parallel
orientation to the ray cell axis, the P, with the network appearance of microfibrils, and
the P3 with several crossed polylamellate at microfibrillar angles of 30-60". It is interesting
to note that the ray parenchymacell wall in thediploxylem of Pinus develops in two
stages: that is, the primary wall and inner protective layer are fornled in the sapwood, and
just before the heartwood is developed, the secondary wall and protective layer are de-
posited. In the axial parenchyma cells of softwood, the cell wall texture is very similar to
that of ray parenchyma cells, except that the microfibrils are arranged in a flat helix with
respect to the cell axis in the P , .
In hardwoods, the primary wall ofray parenchyma cells has the so-called polyla-
mellated structure proposed by Chafe and Chauret 1431. It was pointed out by Chafe and
Chauret [43] that an isotropic layer and protective layer characterize the layered structure
of the secondary wall of xylem parenchyma cells in hardwoods. According to examinations
of thechemicalcomponents of these two layers using aseries of treatments on serial
ultrathin sections, both a protective layer and an isotropic layer are rich in hemicelluloses
and contain some pectic substances and cellulose microfibrils, but they have little lignin
at the first stage of their developing process and become lignin-rich after the deposition
of the inner secondary walls on them [44]. Consequently, both layers are considered the
Ultrastructureand Formation
Wall of Cell 21
same in their origin and are called “amorphous layer” by Fujii et al. [M].Figure 19 shows
electron micrographs of transverse sections by ray parenchyma cell from Tiliu juponicu;
Fig. 19a shows cell walls skeletonized with hydrofluoric acid, while Fig. 19b shows so-
diumchloride-treatedcellwalls.Blackzonesshow an amorphous layer indicating the
presence of much lignin (Fig. 19a), but these disappear through delignification as seen in
Fig.19b.
It has been reportedbyFujii et al. [45] fromtheexamination of ray and axial
parenchyma cell walls from about 50 species of Japanese hardwoods that the secondary
wall is composed of a lignified cellulosic layer (CL) and an amorphous layer (AL) and
that the cell wall structure can be classified into three types according to the presenceand
organization of these two kinds of layers. Figure 20 is a schematic diagram of the cell
wall organization of hardwood ray parenchyma cells: (1) 3CL-type, (2) 3CL+AL-type,
(3) 3CL+AL+IL-type. CL refers to the lignified cellulosic layer that is similar to the
ordinary wood cell wall, whereas ICL refers to the lignified cellulosic layer inside the
amorphous layer (AL). The 3CL-type wall structuremay be considered thestandard struc-
ture of parenchyma cells of hardwoods, whereas the 3CL+AL-type wall structure occurs
in cells that have extensive pit contact with vessels.
FIGURE 19 Electron micrographs of ultrathin cross section of the ray parenchyma cell from Tilia
japonica, showing the amorphous layer (AL) of the secondary wall: (a) delignified cell wall; (b)
cell wall skeletonized using hydrofluoric acid treatment.
Harada22 and Fujita
3CL AL
1-q .: .............,..,..........:..-
x:
ICL
(a) (C)
FIGURE 20 Schematic diagram of the cell wall organization of hardwood ray parenchyma cell,
showing three types of wall structure: (a) 3CL; (b) 3CL AL,(c) 3CL + AL ICL. + +
cellulose of the G layer is highly crystalline. Its microfibrils are oriented parallel to the
longitudinal axis of the fiber and the G layer is easily separated from the remainderof the
fiber wall. Another structural feature of the tension wood fiber wall is that the G layer
deposits on any one of the normal three secondary wall layers, S,, S2,and S,. The sec-
ondary wall of the tension wood fiber consists of three types, that is, S, + +
G , S, S, +
G , and S, + Sz + S, + G, depending on the wood species or part within a stem. Con-
sequently, the G layer is called “the S, layer” when we refer to the S,, Sz,and S, layers.
1. Pit Structure
Pits are gaps in the secondary wall of wood cells. There are two types of pits: bordered
pits and simple pits. Generally, pits are present as pairs between two adjacent cells: bor-
dered pit pairs, simple, and half-bordered pit pairs. In softwood, the pit border region of
the cell wall is composed of border thickening (BT), S,, S2,and S, from the outer part of
the cell wall as shown in Fig. 24. The presence of BT and thicker S, are features of the
pit border wall. The microfibrils circle at theBT and sweep around thepit at the individual
layers S,, Sz,and S,. In softwood bordered pit pairs, many species show a thickening at
the center of the pit membrane. The torusis suspended from fine cellulosic strands to form
a margin around the torus as shown in Fig. 25. The margin consists of an open net of
FIGURE 24 Schematic diagram of pit border organization in bordered pits of softwood tracheids.
BT, initial pit border.
Formation
Ultrastructure
and of Cell Wall 25
2. Vesture Pits
In hardwoods, the pit chamber and pit apertures that are decorated by outgrowths of wall
material are known as vestured pits. The outer growths of vestured pits are constructed
chemically of lignin, hemicelluloses, and a little pectin [35].The shape and size of the
outgrowths of vestured pits are variable. The development of vestured pit outgrowths is
regarded as similar to that of warts.
R. ..
l!
I
f
FIGURE 26 (a) Light micrograph around the cambial zone ( C ) ,phloem (Ph), and enlarging xylem
(E) from a transverse section of Robinia pseudoacacia. Most fusiform cambial cells are undergoing
periclinal division, except for a trace experiencing anticlinal division (arrow).(b) Electron micro-
graph of fusiform and ray cambial cells. (c) Cytoplasmic feature of enlarging cells.
27
28 Fujlta and Harada
FIGURE 26 Continued
For instance, the differentiation of tracheids implies their maturing process from birth at
the cambial zone to death after the secondary wall formation, by which both water-con-
ducting functions and mechanical properties are given to the tracheid. The method of
differentiation of parenchyma cells is quite different from that of tracheids, because they
have only the primary wall or an underdeveloped secondary wall on the primary wall.
They may be already functioning at thecambial zoneand have the ability to redifferentiate.
Therefore, their differentiation is not addressed here.
First of all, the differentiationof softwood tracheids, which is the most basic process
of wood cell formation, will be discussed in detail. When a specimen block around the
cambial zone is taken from the stem of a living tree and then a transverse section is
observed under a light microscope, it is noticed that cells are piling up on a radial row
from the mature phloem to the mature xylem through the cambial zone (Fig. 27a). The
differentiating zoneof tracheids is located between the cambial zone and the mature xylem
area.
If the whole life of a particular tracheid from birth to death could be traced in situ
in a tree stem, the tracheid differentiation would be clearly elucidated. However, it is really
impossible to do so because the cells must be fixed with some reagent to preserve their
cytoplasmic structure. Regrettably, their dynamic cell actions evolve into static phase by
fixation. Therefore, the differentiating process of a tracheid must be deduced from the
static cell structure of a series of differentiating tracheids. From this point of view, the
differentiating zone of earlywood tracheids is favored for the precise examinationof their
differentiation. In the spring, the production of tracheids from the cambial zone is very
D
3
Q
n
0
7
3
s
0
3
FIGURE 27 (a) Light micrograph of the cambial zone (C) and the derivative tracheids in five differentiating stages (RE, S,,
S,, S3, and F) between phloem (Ph) and mature xylem (MX)from Cryprorneriujuponicu. (b) Enlarged view of S, depositing
cells.
h)
W
Harada30 and Fujlta
FIGURE 27 Continued
microfibrils (see Section IV). The secondary wall, which is still porous and flexible after
the deposition of hemicelluloses, is encrusted with lignin and becomes very rigid. The
lignification of the secondary wall, which willbe called “S-lignification” in contrast to
“I-lignification,” is the most active after the S, stage,namely, in thefinal (F) stage of
differentiation, although its initiation can be detected already during the S, stage. In this
F stage some decorative elements such as warts or helical thickenings are added on the
inner surface of the wall. After the wall layers develop, tracheids lose their cytoplasm by
autolysis. Amorphous substances that have embedded the pit membrane also dissolve en-
zymatically sometime in the F stage. Tracheid differentiation is completed as this point
and water conduction is achieved in the mature xylem (MX).
The differentiation of vessel elements is characterized by enormousexpansion in
both the radial and tangential directions. Although the developing stages of tracheids can-
notbe applied directly to those of vessel elementsdue to a different secondary wall
structure, the relationship of enlargement to secondary wall thickening and lignification is
consideredsimilar to the sequence of tracheid differentiation. Needless to say, the for-
mation of perforation pores is completed by the disappearance of the membrane itself,
apart from the removal of only an embedding substance in the bordered pit pairs.
On the contrary, the differentiation of wood fibers is characterized by the remarkable
elongation in cell length that occurs at cell tips [50], and the other properties of differ-
entiation are quite similar to those of tracheids. In hardwood, although the differentiation
of both vessel elements and wood fibers proceeds simultaneously, vessel elements differ-
entiate faster than wood fibers.
How long a wood cell needs for its differentiation is also an important question. The
time requirement for differentiation has been deduced by several methods, but the results
are conflicting. A detailed timerequirementwascalculated for young trees of several
softwoods by means of periodic inclinations for the internal date marking on the xylem.
By these markings and the cell numbers contained in each differentiating stage, a time
requirementofabout three weeks for passingthrough the five developingstages of a
tracheid (RE, S , , S?, S,, and F) was calculated [51].
uole. In fact, it can be pointed out by arealneasurements that the vacuole is the best
developed of the cells at the RE stage, when cells are just expanding (Fig. 28a).
A nucleous is located around the central position of a fusiform cell in the longitudinal
direction [SO], but on the transverse plane it is still pushed to one side of the cell lumen
by vacuolation (Fig. 26c). In the cytoplasm, ordinary cell organelles such as Golgi bodies
(Go),rough and smooth endoplasmic reticula (r-ER and S-ER). mitochondria (M), plastids
(P), small vesicles (v), ribosomes, microtubules, and so on, are contained in a very narrow
cytoplasmic region, although the occurrence of these cell organelles except microtubules
between plasma membrane (Pm) and tonoplast (T) is not so abundant during the difl’er-
entiation of tracheids, wood tibers, or vessel elements. On the contrary, the cytoplasm of
differentiating ray and axial parenchyma cells is crowded with many cell organelles. Es-
pecially, starch grains in the plastids and lipid droplets are very abundant, and r-ER are
also well developed, whereas microtubules are very scarce. Ray cambial cells and mature
ray cells arc almost identical to the differentiating parenchyma cells in their cytoplasmic
features(Fig.26b). However, the number and size of starchgrains and lipid droplets
contained in mature parenchyma cells change during a year 152-541.
Thecytoplasmicfeatures of tracheidschange 21 little both i n quality and quantity
according to their differentiation. The increase and decrease of the cytoplasmic area and
its constituents of cell organelles were revealed by the combined use of light microscopy
(Fig. 28a) and electron microscopy (Fig. 2%) on the differentiating zones of normal and
compression woods of Cqptmtwricr juponicu.
Areas of cell outline (A,,,,,,,;,,),cytoplasmicsurface (A,,,,,,,,,),and central vacuole
(A,,,,,,,,,) were measured on an enlarged light micrograph of the transversesections o f
differentiating tracheids using a digitized system connected to a computer (Fig. 28a). The
nxasuretnent was performedalongthedifferentiatingtracheids, which were numbered
from the initiation of the S , stage, and about 30 radial rows were surveyed. These radial
rows of tracheids were sectioned at random in their longitudinal direction, so that the
average value of tracheids of the same cell number reflects the makeup of volume i n each
region. Areas of cell wall (A,v,J and cytoplasm (A,,.,,,,,,,,,,,)
can be calculated by finding the
remainder between those of the cell outline, the cytoplasmic surface. and the central vac-
uole, respectively. These values are diagramed in Figs. 29 and 30. I t should be noted that
the cytoplasmic volume of both the normal and compression woods has two peaks during
tracheid differentiation. The earlier peak i n both cases is at the intermediating phase from
the S , stage to the S, stage, whereas the later one is located just prior to the initiation of
the S , or F stage. On the contrary, during the S2 thickening, the cytoplasm is rather poor.
Following this, proportions i n RE, S , , early S?, middle S,, late S,, and S, stages (Fig. 29)
show the relative constituents of major cell organelles surveyed by electron microscopy
(Fig. 2%). The general change in these cell organelles can be grasped by inultiplying the
relative value by the total area of cytoplasm diagramed in Fig. 29.
In addition to the changes in these cell organelles, the plasma membrane, important
to the transportation of materials in and out of the cytoplasm, is always observed during
tracheid differentiation and disappears after the development of the cell wall.
The cytoplasmic features of differentiating wood fibers and vessel elements are also
similar to those of tracheid differentiation, although the vacuolation of vessel elements is
Inore extreme. In some species, such as acer or black locust, the living wood tibers are
formed during the later period of a growing season. Their protoplast remains after the
development of a cell wall and stores many starch grains in the cytoplasm for several
years. Therefore, the mature xylem i n the sapwood is composed of ray and axial paren-
chymacells and sometimes the living wood fibers as the cells have aprotoplast.The
Formation
Ultrastructure
and of Cell Wall 33
8 5
.. B
FIGURE 28 Light (a) and electron (b) micrographs of transverse sections of S? depositing tra-
cheids from Cryptomeria japonica. Areas of some cell organelles, for instance, Golgi bodies (Go),
were measured with electron micrographs such as those shown in (b).
Fujita and Harada
r
6
1
0
0
c e l l number
C RE SI s?r Sam Sar S3 F
FIGURE 29 Changes of cytoplasmic volume during the differentiation of normal wood tracheids
and proportions of some cell organelles at the stages of RE, S , , early S2, middle S,, late S?, S,, and
F in Cryptonzeriu jqoniccr (see Figs. 28a and 28b).
FIGURE 30 A change of cytoplasmic volume during the differentiation of compression wood (see
Fig. 28a).
Formation
Ultrastructure
and Wall of Cell 35
cytoplasmic features of these living cells are affected by the season, and also some of
them seem to be specialized in their cell shape and cytoplasm. That is, the cells surround-
ing a vessel, particularly those directly contacted, become envelope-shaped and are very
rich in Golgi bodies, r- and S-ERs, ribosomes, and mitochondria common to cells of active
phase. On the other hand, their storage function seems to decay. These vessel-associated
parenchyma cells are shown to concern the transportation of materials with vessel lumens
[52,53] and also the formation of tyloses or gum that plugs the vessel lumen [55-571.
There is no doubt that cell walls are formed by the actions of cell organelles contained in
each cell, even though some precursors of wall materials such as sugars may be supplied
by the intercellular transport system.Therefore, cell wallformation is realized by the
careful observation of cytoplasm that is undergoing cell wall development. It is also very
important to select proper plant materials for precise examination of cell wall formation,
becausegeneral plant cellsbearmanyphysiologicalfunctions in addition to cell wall
formation. Moreover, the cell wall is composed of several types of chemical materials that
are supposed to be metabolized by different cell actions, and their deposition on the wall
may overlap. These complicated factorsare the major reason that the formation mechanism
of plant cell walls has not yet been explained clearly, in spite of many investigations.
Differentiating wood cells such as tracheids are very useful materials from this point
of view. That is, they construct a very thick secondary wall, of which the ultrastructure
and chemical components have been examined in detail, and the general sequence of cell
wall formation can be traced through the series of differentiating cells along a radial row.
Besides, the cell organellespossessed by thesecells are concernedonlywith cell wall
formation, except vacuolation for the turgor pressure. In addition, if the depositing phase
of individual wall materials such as cellulose, lignin, or hemicelluloses can be detected
separately in differentiation, the relationship of cell organelleswith the metabolism of
those materials would be grasped more clearly.
A. CelluloseMicrofibrilDeposition
Cellulose is the mostbasic cell wall material in thewhole plant and it constructsthe
framework structure of cell walls in the form of crystalline microfibrils as mentioned in
Section 11. The formation has been studied using various plant cells from lower plants
such as fungi or algae, to higher plants. Plasma membranes located just inside developing
cell walls seem to be the most important cell organelles in relation to cellulose microfibril
deposition. Although cross-sectional structures composed of unit membraneshadbeen
observed by ordinary electron microscopic methods such as chemical fixation and ultrathin
sectioning, faceviewsalongthemembranebecamepossiblewith the development of
freeze-fracture or etching methods coupled with replication. Small particles on the outer
surface of plasma membranes had been reported in various plant cells. The epoch-making
discovery, however, was the characteristic assembly of granules located in the interior of
the plasma membrane and revealed on the fractured surface in green algae such as Oocys-
tis, Myclusteriu, or Vcloniu [58-601. Interesting structures have been reported using mainly
single or naked cells such as algae [62], actobacteria [63,64], and cotton fiber [61], which
can be frozen rapidly.
Harada36 and Fujita
Small granule assemblies at the tips of microfibrils are called the terminal complex.
These granules are considered to be the enzyme for the polymerization of cellulose mol-
ecules and their alignment i n the conlplex in relation to crystallization [64]. In under-
evolvcd algae such as vrtlonicr, the assemblies are large and linear, corresponding to their
thick microfibrils [62]. On the other hand, evolved plants have small groups called rosettes
[6 11. Thus, the form of the complex is considered to be related to the shape of the cellulose
microfibrils and also to the evolution of plants. The polymerization and crystallization of
cellulose microfibrils have been surveyed in detail using AcetoDactor- . x - y l i t z ~ ( ~which
~/, pro-
duces a thin cellulosic thread [6S] and has various mutants (661.
The sequence of cellulose synthesis described above has not been traced in differ-
entiating wood cells yet, because the freeze-fracture method is difficult to apply to them.
However, cellulosic frameworks of wood cell walls are supposed to be constructed by a
similar way. perhaps by a rosette. Although the freeze-fracture method isvery effective
for visualizing characteristic structures such as the terminal complex on the membrane,
the overall structure of differentiating wood cells depositing cellulose microfibrils must be
examined by ordinary sectioning methods. Especially in wood cells depositing secondary
wall layers, a control lnechanism for microfibrillar orientation is a very interesting view-
point. Also, as the cellulose deposition is accompanied by the synthesis and accumulation
of hernicclluloses and lignin, actions of various cell organelles must be traced in detail.
Hence, the deposition phase of cellulose microfibrils i n differentiating tracheids can
be traced in both normal and reaction woods. The phase can be detected by the increment
of cell wall thickness (671, by means of autoradiography [68-70] (Fig. 3 l ) , and by chem-
ical analysis of selectively collected rnaterials in some developing stagesof tracheids [ 7 I -
731 and wood fibers [74](Fig. 32). The results obtained by these methodsshow that
cellulose microtibrils are supplied to the wall mainly in the early and middle phases of
the S, and S2 deposition stages. In addition to these deposition stages of cellulose micro-
fibrils, most noticeable were the deposition of the G layer in the tension wood fibers and
the S, thickening stage in the compression wood tracheids. This stage is composed of the
deposition of cellulose microfibrils and is followed by the depositing stages of hemicel-
luloses and lignin.
Compared with other differentiating stages of tracheids and wood fibers, the cyto-
plasm of cells forming the G layer isvery poor in its activity due to the fact that the
region between the plasma membrane and the tonoplast is very narrow and cell organclles
are rare there (Fig. 3%) 1751. The exceptionally abundant cell organelle in the cytoplasm
is microtubules (MT). They are regularly distributed just inside the plasmamembrane
(Pm), keeping a constant space of approximately 8 nm to the inner membrane and also
between themselves (Figs. 33a and 33b). They are exactly oriented parallel to the depos-
iting cellulose microfibrils in the stages of the G layer as well as the S , and S2 layers (Fig.
33b). The diameter of the microtubules is approximately 23 nm, and their numbers increase
up to 20 per I p m of the plasma membrane, as calculated by their transverse direction.
This abundant distribution results in the covering of about 40% of the cytoplasmic surface
(Fig. 33b). A feasible link between the microtubules and the inner layer of plasma mem-
brane is also discernible (Fig. 3321).These characteristics strongly suggest that microtubules
and plasma membrane comprise the outermost complex of cytoplasm. On the other hand,
there are only traces of Golgi bodies, S- and r-ERs, and the vesicles derived from them in
the cytoplasm, in spite of the very active synthesis of cellulose microfibrils in this phase
of the cell.
On the contrary, in the beginning of the S, thickening stage of compression wood
tracheids, in which cellulose microfibrils are supplied to the wall at 45" to the cell axis,
Formation
Ultrastructure
and of Cell Wall 37
FIGURE 31 Serial light microscopic autoradiographs of “before section treatment” (a) and “after
sectiontreatments”(b)withsodiumchloriteandhot 1.3% H,SO, from the differentiating com-
pression wood tracheids in Cryptomeria japonica administered with 3H-glucose. Silver grains in (b)
show the specific incorporation of radioactivity only on the inner surface of S , and S2 thickening
tracheids, which reflects the deposition of cellulose by way of “apposition.” Removed activity can
be detected in the intercellular layer of cells in the
S , stage (arrows) and in the preexisting secondary
wall of cells in the late S2stage (cells marked by an asterisk)by a comparison between (a) and (b)
that implies lignin and hemicelluloses are supplied to the wall by wall of “intussusception.”
Harada38 and Fujita
cytoplasm isvery dense andwide. A similar complex between the microtubules and plasma
membrane, however, is still observed [76]; microtubules are very abundant beneath the
plasma membrane, having a link to it (Fig. 34a). The direction of microtubules is also in
this case parallel with depositing cellulose microfibrils. Various cell organelles in the cy-
toplasm, such as Golgi bodies or ER, were shown to be involved in the synthesisof lignin
precursors for the succeeding lignin deposition into the S, layer (see the next section).
The characteristic appearance of microtubules is always applied to the cells depositing
other wall layers such as S , , S2,and S3 without exception. It is interesting to note that the
reorientation of microtubules precedes that of depositing cellulose microfibrils in the tran-
sition from the completion of a wall layer to the initiation of the next layer (Fig. 35).
Moreover, the treatment of colchicine in which microtubule construction is obstructed by
the formation of conjugation with microtubule protein “tubulin” resulted in a remarkable
disturbance of depositing cellulose microfibrils (Fig. 36) [77,78]. These results clearly
show that although microtubules present inside the plasma membrane cannot readily syn-
thesize long and rigid microfibrils, these are involved in a great extent in the control of
depositing cellulose microfibril orientation.
Formation
Ultrastructure
and of Cell Wall 39
FIGURE 32 Continued
B. LigninDeposition
Lignin is a very important cell wall component, particularly in wood cells, for the en-
hancement of the physical properties of cell walls and also for sealing the wall from
prevention of waterleaks.Fortunately,lignincan be detected under an ordinary light
microscope with the use of several stains such as a Wiesner reagent and Maule color
reaction. Ultraviolet microscopy is especially useful for the quantitative analysis of lignin
distribution in the cell wall [79] and more useful for studying the types of lignins present
in the cell wall [80,81]. Needless to say, transmission electron microscopy coupled with
potassium permanganate staining [82] or hydrofluoric acid treatment [83], electron probe
microanalysis [84], and autoradiography [85-901 are also very useful for the observation
of lignin from various points of view.
The lignification of tracheid walls is generally known to last for a long period, from
the S, stage to theF stage [79]. During this period, the lignification starts at the cell comer,
spreads into the intercellular layer, and extends centripetally to the secondarywall. Lignin
deposition, however, should be examined more closely in relation to the deposition of
cellulose and hemicelluloses on each wall layer. This was attempted in the differentiating
tracheids of compression wood, which were convenient for separating thelignification of
the I region and S region because of their conspicuously highly lignified secondary wall,
especiallyattheouterregion of the S, layer [67]. It has been shownthatthelignin
deposition can be separated into two lignification stages, namely, I- and S-lignification.
The former is active only during the early stage of secondary wall thickening, mainly at
40 Fujita and Harada
FIGURE 33 Electron micrographs of the cytoplasm-cell wall region of a transverse section from
a tension wood fiber of Populus euramericana depositmg G layer (a) and of an obliquely sliced
section from a S,-depositing fiber (b).
the S , stage, and is soon finished. The shape of I-lignification seems to stop the enlarge-
ment of cell size and adheres firmly between neighboring cells. On the other hand, the
latter proceeds mainly after the development of a secondary wall framework, even though
it begins at the middle phase of S , thickening. At any rate, lignin precursors permeate
deeply into the cellulose microfibril framework of both primary and secondary walls and
accumulate by way of “intussusception.” These two types of lignification were also ap-
plied in the differentiationof normal wood tracheids[86,91,92]. Moreover, when the speed
Formation
Ultrastructure
and Wall of Cell 41
. . ..... .,v*
(b)
FIGURE 33 Continued
of lignin accumulation and the distribution of peroxidase were compared between the two
regions [86], I-lignin seemed to be richer in the “condensed-type lignin” caused by the
bulk polymerization than the S-lignin. This would be so because lignification proceeds
with the higher content of lignin monomers and peroxidasein a rather large space without
microfibrils. This assumption was confirmed by selectively labeled precursors coupled with
light microscopic autoradiography [89,90].
When the lignifying cells are observed from the viewpoint of cytology, the cytoplasm
is wider and denser than that of cells depositing cellulose microfibrils. Especially in the
compression wood tracheids, an enormous amountof lignin precursors must be synthesized
Harada42 and Fujita
Q
t
in their cytoplasm and thentransported from the cytoplasm to the wall. The area of
cytoplasm becomes wider at the S , stage and also at the transition from Szto F (Fig. 30),
where the cytoplasm becomes rich in Golgi bodies (Go) and ER (Fig. 34a). Although
small vesicles (v) are produced mainly from Golgi bodies, they do not move to the cy-
toplasmic surfaceyet. These small vesicles increasein number and grow larger, occupying
of Cell Wall
Ultrastructure and Formation 43
FIGURE 34 Continued
the largest part of the cytoplasm during the following late phase of the S, stage. S-ligni-
fication at the F stage is characterized by the active fusion of the well-developed vesicles
to the plasma membrane and by the release of the vesicle inclusion to the wall area,
namely, exocytocis. The cytoplasmic area resultsin the formationof an empty region after
lignification (Fig. 34b). This sequence indicates thatlignin precursors are synthesized and
stored in the vesicles that havebeen derived from Golgi bodies and on occasion from ER.
In fact, the process was proven by autoradiography using tritiated lignin precursors [ S S ] .
These sequences were also examined inboth lignifications at the I and S regions of
normal woodtracheids(Figs. 32aand32b) [86]. InS-lignification, S-ER seems to be
related to the lignification in addition to the Golgi bodies (Fig. 29), whereas I-lignification
Harada44 and Fujita
;S1
FIGURE 35 Electron micrograph of an obliquely sliced cytoplasm-cell wall region from a tension
wood fiber of Populus euramericana, which is just traveling from S2to the G layer. Many micro-
tubules (MT) are oriented parallel to the fiber axis, although several (large arrowheads) are still
oblique. Fine striations in the cell wall (small arrowheads) show the deposition of the S layer, and
cellulose microfibrils of the G layer cannot be detected yet.
is performed mainly by the action of Golgi vesicles, similar to the case of compression
wood tracheids. The lignin of compression wood tracheids is generally known to be rich
in the condensed-type lignin. The cytoplasmic features of these lignifying cells seem to
be consistent with the types of lignin suggested by Takabe et al. [86]. That is, the I- and
S-lignins of compression woods and also the I-lignin of normal wood are metabolized
mainly by Golgi bodies and the derivative vesicles, being richin the condensed-type lignin,
Formation
Ultrastructure
and of Cell Wall 45
whereas the S-lignin of normal wood is synthesized through the cooperation of Golgi
bodies, S-ER, and their vesicles, resulting in noncondensed-type lignin.
It is interesting to note that the cytoplasmic regions becomewider, corresponding to
both lignifications in the I- and S-regions in both normal and compression woods (Figs.
29 and 30). In addition, the peak of the S-lignification of compression wood is bigger
than that of normal wood. The tendency of these peaks is to respond to the absolute
amount of lignin that will be supplied to the separate wall regions. The precursors of lignin
are most likely synthesized in the cytoplasm and stored temporarily, and then released
from the cytoplasm to the wall, whereas the cytoplasm may possibly be rather narrow
during the active depositing phase of cellulose microfibrils.
C. HemicelluloseDeposition
Examination of hemicellulose deposition is divided into two groups. One covers the mi-
croscopic observations [68-70,931; the other is the chemical analyses of the tissues or
cell walls collected selectively [71-741. Although the use of microscopy is a prerequisite
for the observation of the microlevel localization of hemicelluloses, the specific staining
method has not been improved enough to be applied on each hemicellulose and even
mixed ones, being difficult to distinguish from cellulose or lignin. If any specific radio-
Harada46 and Fujita
active precursor of each hemicellulose can be applied to the differentiating xylem, auto-
radiography will provide invaluable information on the deposition of hemicelluloses. A
similar effect, although applied only to the total hemicelluloses, was achieved by a com-
bined technique involving autoradiography and the removal of hemicelluloses from the
tissue [93] or sections [53] that had been administered with tritiated glucose as the general
source of cell wall materials. It becomesclear by these methods that hemicelluloses,
although not so deeply as in the cases of lignin, accumulate in the preexisting framework
of cellulose microfibrils by way of “intussusception” (Fig. 31). The depositing phase also
intermediates the deposition of cellulose microfibrils with lignin accumulation.
The deposition of each hemicellulose must be traced by the sugar analyses of tra-
cheids or wood fibers selectively collected from the differentiating zone according to their
development. This technique was achieved qualitatively by Meier et al. [7 l ] and improved
quantitatively by Takabe et al. [73]. Judging from the content of polysaccharides in wood
cell walls and the sugar constituents of hemicelluloses, glucose, mannose, xylose, arabi-
nose, and galactose are, respectively, reflected in cellulose, glucomannan, arabino-glucono-
xylan, and galactoglucomannan. As shown in Fig. 37a, mannose is supplied to the wall
just after the cellulose microfibril deposition, followed by the deposition of xylose in the
tracheid differentiation. In the case of wood fibers (Fig. 37b), xylose deposition follows
directly the deposition of cellulose microfibrils. On the contrary, galactose and arabinose
seemto be supplied to the walltogether in both stages of I- and S-lignification. The
disagreement between the depositing manners of xylose and arabinose, namely, arabinose
showing more affinity for lignin than xylose, may suggest that a chain of arabino-glucu-
rono-xylan is not polymerized at one time but separately. That is, in concert with lignifi-
cation, arabinose may be added, possibly as a side chain of the backbone of xylan already
deposited.
I 1 . . . . . . . . a ’ . L
0’
1 2 3 4 5 6 7 8 9 1 0 1 1 1 2
i r d c t l o nn u m b e r f rdctlonnumber
REFERENCES
I. D. A.Kribs. Bot. Gn:.. Y4:547 (193.5).
2. H. A. Core, W.A. C M . and A. C. Day, Wood Structure r r t d Idmtijiccrtiorz. 2nd ed.. Syracuse
Univ. Press. New York, p. 30 ( I 979).
3. J. F. Revol and D. A. 1. Goring. Wood Sei.. 14: 120 (1982).
4. J. Sugiyama. H. Chanzy, and J. F. Revol. Plmtcr, /Y3:260 (1994).
S . H. Yamamoto and F. Horii. M~rcronlolrcrrlc.~,
26: 13 13( 1993).
6. J. Sugiyama, R. Vuong. and H. Chanzy. Mrrcror?loleclt/c.s,24:4168 ( 1991).
7. A. Frey-Wyssling. Scirrm,, //Y:80 (1954).
8. A. hey-Wyssling and K. Muhlethalcr. Mrrkrortlol. Clwrn., 62:25 (1963).
9. R. D. Preston and J. Cronshaw. Nature, 18/:248 (1958).
I O . K. Miihlethaler. Bioh. Z. S c h ~ ~ i :For.stwreitf..
. 30:55 (1960).
I I . K. Miihlethnler. i n Ccllltlrrr Ultrtrstrrrcturc~of Wooc!\ Plat1t.s (W.A. C W . Jr., ed.).Syracuse
Univ.Press,NewYork. p. 191 ( 1965).
12. V. Bnlnshov and R . D. Preston, Ncrtrrw. 176:64 ( 1955).
13. H. Haradn and T. Goto, i n Ce//lrlo.sc, t r t d Other Ntrt/trrr/ PnlynrcJr .Sy.stcr/r.s (R. M. Brown. Jr..
ed.). Plenum Publishing, New York. p. 383 (1982).
14. A. N. J. Heyn. J . Cell Biol.. 2Y: I81 (1966).
I S . J. Sugiyama. H. Haroda. Y. Fujiyoshi. and N. Uyeda. Pltrr~ttr.Ihh: I61 ( 1985).
16. J. Sugiyomn. H. Harada. Y. Fujiyoshi. and N. Uycdo. Moku:tri Gtrkktrishi. 31:6 I ( 1985).
17. J. F. Revol. J . Mrrtrt: Sci. Lett.. 4 : 1347 ( 1985).
18. K. Muhlethaler. J. P o / w w r Sci.. C2N:305 (1969).
19. A. J . Kcrr ancl D. A. I. Goring. Wood Sci.. Y: I36 ( 1977).
20. K.Rue1 and F. Barnoud, ISWPC Stockholm. I : I I (1981).
31. Y. Kataoka. S. Saiki. and M. Fujita. Mokrr:rri Gnkkctishi, 38:327 (1992).
32. T. Kerr and 1. W. Bailey. ./. A r d d Arborotrr/r~,/5:327 ( 1934).
33. P. A. Roelotken. i n A d 1 w r x ~ c . sirr Bo/trrlictr/ Rc,srtrrch. Vol. 2 (R. D. Preston. d . ) . Academic
Prcss. L,ondon a n c l NcwYork. p. 69 ( 1959).
24. A. B. W d r o p . A r ~ s t r d ..I. Hot.. h:299 ( 1958).
25. J. C. Rolnnd, B. Vam. and D. Reis. J . Cc,// Sci.. / Y : 1 3 9 (1975).
36. T. Fujii.Pl1.D.thesis.KyotoUniversity.Kyoto. Japan. 1981.
48 Fujita and Harada
73. K. Takabe, M. Fujita, H. Harada. and H. Saiki. Mok~czoi GNkkaishi, 29:183 (1983).
74. K. Takabe, M. Fujita, K. Tanaka, and H. Harada, Bull. Kyoto CJniv. Forests. .56:234 ( 1984).
7s. M. Fujita, H. Saiki, and H. Harada, Mokuzui Gakkaishi,20:147 (1974).
76. M. Fujita, H. Saiki, and H. Harada, Moktczai Gnkkaishi, 24:355 (1978).
77. J. D. Pickett-Heaps, D e ~ dBiol.. /5:206 ( 1967).
78. M.FijitaandH.Harada.“Colchicinetreatmentandmicrolibrilorientationindifferentiating
compression wood tracheids,” Proc.. 27th A n r l d Meetirrg of Jup,crrwsc~Wood Resectr-chSociety.
p. 311 (1977).
79. A. B. Wardrop, TAPPI, 40:22S (l9S7).
80. B. J . Fcrgus and D. A. I . Goring, H o l ~ f o r . s c l ~ ~ c24:
r ~ 118
g , (1970).
81. A. Yoshinaga, M. Fujita, and S. Saiki, Mokuzcli Gnkknishi, 38:629 (1992).
82. M. Mauer and D. Fengel, Hol;for.schurlg. 44:4S3 (1990).
83. I . H. Clark, TAPPI, 45:310 (1962).
84. S. Sakaand D. A. I. Gorgin.in Biosy1lthesi.s crrltl Riorlesmrltrtiorl of’ Wood Cort~pot~c,r~t.s(T.
Higuchi. ed.), Academic Press, New York, p. S 1 (1983.
85. M. Fujita and H. Harada, Mokuzai Gakknishi, 25:89 (1979).
86. K. Takabe, M. Fujita. and H. Harada. Mokuzcti Gtrkknishi, 31:613 (1985).
87. N. Terashima. K. Fukushima, and K. Takabe, Holifi)r.sc~hwzg,40(Suppl.): 101 ( 1986).
88. K. Fukushima and N. Terashima, J . Wood Cllern. T e c h o l . , /0:413 (1990).
89. N. Terashima and K. Fukushima. Wood Sci. fi~cl~nol., 22:2S9 (1988).
90. K. Fukushima and N. Terashima, Hol;fi~rsc.hur~g. 45:87 (1991).
91. M. Fujita, K. Takabe, and H. Harada, ISWPC (Tsukuba), !:l4 (1983).
92. K. Takabe, M. Fujita, H. Harada, and H. Saiki, Mokuzai Gczkkaishi, 27:813 (1981).
93. P. M. Ray, J . Cell Biol.. 3.5:660 ( 1967).
94. D. Fengel and G. Wegener, Wood, Walter de Gruyter. Berlin, 1984.
This Page Intentionally Left Blank
Chemical Composition and Distribution
Shiro Saka
Kyoto University, Kyoto, Japan
1. INTRODUCTION
51
52 Saka
wall layers. Instead, the gelatinous layer (G layer) is usually deposited adjacent to the cell
lumen (Fig. 2c). The G layer contains little or no lignin and consists mainly of cellulose
microfibrils oriented parallel to the fiber axis.
FIGURE 2 Continued. (c) Cross section of tension wood fibers in Enoki [Celtis sinensis Pers vat.
japonica (Planch) Nakai]. Transmission electron micrograph of KMn0,-stained ultrathin section.
Note a nonstained G layer deposited following the stained S2 and S3 layers. (Courtesy of the late
Prof. Emer. H. Harada, Kyoto University, Kyoto, Japan.)
also contains varying quantitiesof extractives that are always more abundant
in heartwood
than sapwood.
The chemical composition of reaction wood differs from that of normal wood. Table
3 shows a comparison made by Timell [ 1l ] of the average chemical composition of normal
and compression woods of many conifers. Pronounced compression wood contains, on
average, 39% lignin and 30% cellulose, compared to 30% and 42% for normal wood,
Cellulose 45 42 45 48 51
Lignin 24 19 22 21 24
Glucuronoxylan 25 24 35 26 19
Glucomannan 4 3 3 3 4
Pectin, starch, 2 1 4 4 2 "
ash, etc.
"All values in percent of extractive-free wood.
Source: Ref. 9.
omposition
Chemical 55
Cellulose 42 41 41 41 41
Lignin 27 29 29 33 31
Glucuronoarabinoxylan 9 13 9 7 14
Galactoglucomannan 18 18 18 16 12
Pectin, starch, ash, etc. 2 1 3 3 2
respectively. The content of galactoglucomannan is only 9%, half that in normal wood.
The amount of xylan, on the other hand,is the same in thetwotissues.Compression
woodcontains 2% of a 1,3-linked glucanand 10% of a galactan,bothpresentinonly
trace amounts in normal wood.
With regard to chemical composition of tension wood, the most characteristic feature
is that it contains less lignin andxylan described as pentosan, but has much more cellulose
56 Saka
Lignin 30 39
Cellulose 42 30
Galactoglucomannan 18 9
1.3-Glucan Trace 2
Galactan Trace IO
Glucuronoarabinoxylan 8 8
Other polysaccharides 2 2
"All values in percent of extractive-free wood.
Source: Ref. I 1.
and galactose residues than normal woods (Table 4) [ 121. Furthermore, higher ash and
uronicacidcontentsarereportedfortension woodthanin the side woodof Japanese
beech (Fugus crenuru Blume) [ 131. The higher cellulose content in tension wood is due
to the presence of a gelatinous layer, which is often quite thick and unlignified (Fig. 2c).
Lignin 40 28 40 40
Cellulose 35 42 35 30
Galactoglucomannon 9 20 II 9
1,3-Glucan 2 - 2 2
Galactan Tracc Trace Trace IO
Glucuronoarabinoxylan 11 8 10 7
Pectin 2 1 I I
Other polysaccharides I 1 I 1
in the tracheids. Overall, the results in Table 5 are in good accord with those by Perilii
[ 161 and Perilii and Heitto [ 171. However, these authors indicated more xylan in ray cells
than normal tracheids i n Scots pine (Pirzus sylvesfris L.). The lower content of xylan in
Table 5 is due possibly to partial removal during chlorite delignification.
Also included in Table 5 are the chemical compositions of tracheids and ray cells
from red pine compression wood [ 181. Although compression wood tracheids have a dif-
ferent chemicalcompositionfromnormalwood tracheids, the ray cells in compression
wood are chemicallyindistinguishablefromthose in normalwood.Comparedwith the
compression wood tracheids, the ray cells have the same contents of galactoglucomannan,
1,3-glucan, and xylan. However, the galactan typical of grosscompressionwood is
missing.
TABLE 6 Polysaccharide Composition of Earlywood and Latewood from Normal Scots Pine
and Balsam Fir Compression Wood
Balsam fir compression
wood’ pine” Scots Normal
Latewood
Earlywood
Latewood
Earlywood
Polysaccharide (%l (%) (%) (%)
FIGURE 4 Holocellulose skeleton of two tracheids in compression wood of tamarack [Lark lur-
kina (Du Roi) K. Koch]. Note the low concentration of polysaccharides inthe S,(L) layerand
absence of substances in the middle lamella region. Transmission electron micrograph of a cross
section. (Courtesy of Prof. Emer. W. A. C M , Jr., State University of New York, Syracuse, W.)
of polysaccharides may obscure the data, overall the results obtained by this method are
in good agreement with the holocellulose distribution inferred from the lignin skeleton
seen in Figs. 4 and 5 [33]. Some other methods also have been proposed; Parameswaran
and Liese [34] have given an excellent review of these studies.
For the localization of pectin, Albersheim et al. [35] have used the hydroxylamine-
iron method developed by McCready et al. [36,37] for the quantitative measurement of
pectin. With this method, Parameswaran and Liese [34] have found a homogeneous dis-
tribution of pectin across the secondary wall. The middle lamella also was found to be
highest in concentration. The use of ruthenium red and alcian blue also is proposed for
staining pectin substances [38-401.
For quantitative determinationof the polysaccharide distribution, the microdissection
technique has often been used. One of the oldest is Bailey’s work in 1936 [41] for the
pentosan content determinationof the middle lamella in Douglas fir. In 1959, Meier [42,43]
adopted a similar technique for hardwood fibers (Betula verrucosa Ehrh.) and softwood
tracheids (Pinus sylvestris L. and Picea abies Karst.) at different stages of development
60 Saka
rides between the compound middle lamella and the secondary wall, a finding that is not
in agreement with Meier's results [42]. It appears that a treatment of slight delignification
may cause a partial dissolution as well as redistribution of hemicelluloses. The arabinose
and galactose contents for the compound middle lamella were foundto be 7.3% and 7.6%,
values that are considerably lower than those from nonlignified wood [6].
More recently, Whitinget al. [49]developedanothermethod of preparing wood
tissue fractions from the compound middle lamella and secondary wall of black spruce
(Piceu nzariana Mill.) by taking advantage of the difference in density ( p ) between lignin
( p = 1.4 g/mL)andpolysaccharide ( p zz 1.5 g/mL).Themost significant finding after
analyses of carbohydrates for these wood tissue fractions [20] was that the concentrations
of celluloseandglucomannan are smaller in the middlelamellathan in the secondary
wall, whereas the concentrations of other polysaccharides are more or less the same in
both the secondary wall and the middle lamella regions. Compared to the previous methods
by Meier [42] or Hardell and Westermark [48], the method of Whiting et al. [49] is more
reliable, due to only the physical treatment of specimens without introducing any chemical
changes.
U
0.2 0.3 0.4 0.5 0.6 0.7
FIGURE 6 Polysaccharide content versus the lignin content for thevarious tissue fractions of
black spruce (Picecc tnnriat~nMill.). (From Ref. 20.)
62 Saka
a sample collected by Hardell and Westermark must be to some extent contaminated with
the secondary wall fractions. It is of interest to note that both the results of Meier for the
compound middle lamella and of Whiting and Goring for the true middle lamella show
glucomannan to be the lowest among hemicelluloses in the middle lamella region.
For overall trendsof the carbohydrate distribution across the cell wall, Meier [42,43]
has indicated that, although the cellulose content is very low in the ML (middle lamella)
and P (primary wall) regions, arabinan is almost completely confined to M + P regions,
and galactan is almost completely confined to M + +
P S , regions. Glucomannan, how-
ever, increases from M + P to the S , layer in softwoods and remains at a rather constant
low level in hardwoods. Glucuronoxylan in hardwoods has a higher concentration in the
secondary wall than in the M P. +
Figure 7 shows the distribution of polysaccharides across the woodcellwall of
Cryptomeria tracheids obtained by Takabe [U] through the technique of Meier [42]. In-
terestingly, cellulose is rich in the middle of the S , layer, whereas hemicelluloses of glu-
curonoarabinoxylan and galactoglucomannan are abundant in the S , and outer parts of the
S, and S3 layers. The warty layer (W) is composed mainly of galactoglucomannan.
...
...
m Glucuronoarabrnoxylan
0 blactoglucomannan
0 tcllulorc
FIGURE 7 The distribution of polysaccharides across the wood cell wall of tracheids in Crypto-
meria japonica D. Don. (From Ref. 44.)
64 Saka
Percent
GALACTAN ARABINAN
CELLULOSE ARAEINO-
GLUCURONO-
GALACTO- XYLAN
GLUCO-
MANNAN
content of arabinan and galactan would be due to high pectin content. It is reported that
chemical composition of the primary wall is the same in normal and compression woods
[5 1,521.
V. DISTRIBUTIONOF LIGNIN
A. Introduction
Unlike polysaccharides, a number of reliable methods can be used to study the distribution
of lignin in wood. One of the oldest procedures is selective staining, followed by study
under the light microscope[53]. Although some doubt exists as to this specificity for lignin
[54,55], potassiumpermanganatestaining [56] has been usedextensivelyforstudying
lignin distribution by electron microscopy [57-601.
Also reported were studies by electron microscopy of lignin skeletons (Fig.5 ) created
by the carbohydrate removal by brown-rot fungi [61] or concentrated hydrofluoric acid
[32,62-641. Although some alteration of the lignin through condensation may result and
the possible presence of residual carbohydrates may obscure the data, overall the results
obtained by this method are in reasonable agreement with those from potassium perman-
ganate staining [59].
Although the above methods are useful in elucidating the presence of lignin in the
various morphological regions of wood, they can provide only qualitative evaluation of
thelignindistributionacrossthecellwall. For quantitativevisualization of thelignin
distribution, ultraviolet (UV) microscopy with thin sections of wood has provided good
results. This method was initiated by Lange [65], who estimated the weightconcentration
of lignin to be, respectively, 16% and 73% for the secondary wall and compound middle
lamella of Norway spruce tracheids. This result was in excellent agreement with Bailey’s
value of 71% for the Douglas fir middle lamella fractions obtained by a direct analytical
method [66]. Previous to Bailey’s work, Ritter [67] had concluded that approximately75%
of the lignin in wood is located in the middle lamella, with the other 25% being located
in the secondary wall. Apparently, a distinct difference exists between the results by Lange-
Bailey [65,66] and Ritter [67]. However, considerable confusion has appeared in the lit-
Composition
Chemical 65
erature. Some of this confusion could be due to the use of the symbol o/o to denote both
the percentage fraction of the total wood lignin contained in a particular morphological
region and the lignin content of that region. Therefore, g/g. i.e., g of lignin/g of cell wall
substance. is used in this chapter to denote lignin concentration. The symbol o/o is then
reserved for the proportion of total lignin in a particular morphological region.
Later. Goring and co-workers [68-701 refined the UV microscopy method through
a preparation of the thin section (0.5 p m ) to avoid errors caused by nonparallel illumi-
nation. Goring et al. then determinedquantitatively the distribution of lignin in wood
[50,70-751 and proved that the result by Lange-Bailey[65,66] was correct.They also
proved that the conclusiondrawn by Berlyn and Mark [76] is correct. that the middle
lamella region can contain at most 40% of the total lignin in wood due to its small volume
fraction of wood.
In addition to these, they discovered that different lignins occur in different types of
cells and different cell wall regions of wood [72-741. More recently, through the use of
UV microscopy, Yang and Goring [77.78] have found that the secondary wall lignin of
softwoodscontains twice as many phenolic groupsas the middlelamella.This finding
was later confirmed by Whiting and Goring [79] from a study of the secondary wall and
middle lamella fractions.
Of other methods for the quantitative assay of the lignin distribution, Lange and
Kjaer [SO] proposed the use of interference microscopy, and Boutelje [811 later refined
this technique.More recently, Saka et al. [82-861 developeda new techniqueforthe
quantitative determination of the lignin distribution in wood. The method involves a spe-
cific bromination for lignin in a nonaqueous system (CHCI,). Bromine concentrations in
the various morphological regions of wood are then determined by electron microscopy
(TEM or SEM) coupled with energy-dispersive X-ray analysis (EDXA). By knowing the
lignin reactivity toward bromination, the distribution of lignin can be determined for var-
ious morphological regions of wood. Figure 9 shows the direct comparison made between
two techniques of UV microscopy and EDXA measurement in bromination [S61 over the
1 .oo
Earlywood Latewood
I
0
.-C
.- SECONDARY
WALL
l
I I I
I I 1
0
15 10 5 5 10
Cell number
FIGURE 9 Variation of lignin concentrationsacrosstheearlywood/latewood boundary of black
spruce measured by U V microscopy ( 0 ) and the EDXA technique ( 0 ) .(From Ref. 86.)
66 Saka
Earlywood S 87 72 0.23
ML 9 16 0.50
ML,, 4 12 0.85
Latewood S 94 82 0.22
ML 4 IO 0.60
ML,, 2 8 1 .oo
FIGURE 11 Scanning electron micrograph (a) of brominated latewood trachids in loblolly pine
(0.5-pm section). The distribution map (b) of Br-L X-rays was taken of the same area as the scanning
electron micrograph. The distribution of bromine (c) was taken along the line across the double cell
wall. (From Ref. 83.)
Chemical Composition and Distribution 69
"From Ref. 9 1.
'From Ref. 73.
'Fiber/fiber.
dFiber/vesscl.
'Fibedray.
'Ray/ray.
70 Saka
parison, the results obtained by Fergus and Goring [73] through UV spectral analysis are
also included. It is indicated by both methods that the fiber secondary wall (S2) contains
predominantly syringyl residues, whereas the vessel secondary wall (S?) consists mostly
of guaiacyl residues.
The study by UV-EDXA [91] revealed that the ray parenchyma cell contains about
equalproportions of guaiacylandsyringylresidues in lignin. However,apredominant
amount of syringyl-type lignin was found by UV spectral analysis [73], as in the fiber
secondary wall.
For the cell corner middle lamella (ML,,), 80-100% of the lignin was found to be
guaiacyl residue, with the remaining 0-20% being syringyl residue by the UV-EDXA
technique [91]. This result is not in agreement with the data by UV spectral analysis [73].
However, it supports the later suggestion of Musha and Goring [75]that the middle lamella
lignin consists entirely of guaiacyl residues.
Itis therefore apparent in Table 13 that, in hardwoods, the ratio of guaiacyland
syringyl residues in lignin varies in different morphological regions. These findings have
been supported by several investigators; Wolter et al. [93] have shown that the vessels in
aspen callus cultures contain a pure guaiacyl lignin. Kirk et al. [94] found that the fungal
degradation of lignin in birch wood was consistent withthe presence of syringyl-rich lignin
in the fiber walls. Furthermore,Yamasakiet al. [95] isolated syringyl-rich lignin from
several hardwoods. Hardell et al. [96] fractionated birch wood to determine the syringyl
and guaiacyl ratio, and indicated that lignins in both the middle lamella and vessel sec-
ondary wall are rich in guaiacyl units, whereas the ratio of syringyl/guaiacyl residues is
high in the fiber and ray cell. Cho et al. [97] studied the filmlike substance isolated from
the fines of birch in which a high proportion of the compound middle lamella was rec-
ognized.This material wasfound to possessalow ratio of syringyl to guaiacyl units.
Terashima et al. [98] administered 'H-labeled guaiacyl and syringyl model compounds to
magnolia shoots and determined their location in the growing cell wall by microautora-
diography. They found that the vessel wall, cell comer, and compound middle lamella
were lignified by the deposition of guaiacyl-type lignin, and the fiber wall was composed
of syringyl-guaiacyl lignin.
Recently, UV and visible-light microscopic spectrophotometry have been combined
with the Maule color reaction for detecting syringyl lignin by Yoshinaga [99], and this
method has been extended to taxonomic studies of the distribution of hardwood lignins
[ 100- 1021.
Table 14 shows the distribution of lignin in white birch wood as determined by UV-
EDXA [91].Forcomparison, the results obtainedearlier by UV microscopy 1721 are
included. For the fiber secondary wall, the lignin concentration in the S, layer is slightly
lower than in either the S , or S2 layer. However, its difference is so small that the lignin
may be considered to be distributed uniformly across the secondary wall. The vessel walls
also reveal a uniform distribution of lignin, but the concentration is about 1.9 times higher
than that of the fiber walls, which in turn is higher than that of ray parenchyma cells. The
cell comer middle lamella (ML,,) associated with fibers and vessels has the highest lignin
concentration. In spite of sufficient analytical resolution by the EDXA system, the middle
lamella between cell corner areas (ML) was 10-30% lower in concentration than the cell
corner middle lamella (ML,,). It is of interest to note that the lignin concentration in the
middle lamella regions of hardwoods is lower than that of softwoods, as seen in Tables 9
and 14.
A comparison of the data made between UV-EDXA and UV microscopy techniques
indicates that lignin concentrations in fiber and vessel secondary walls are in agreement
Composition
Chemical 71
with each other. However, the lignin concentration in the ray parenchyma cells byUV-
EDXA[91] is nearly half aslow as the dataobtained by UV microscopyalone[72].
Although the middle lamella between two cell corners (ML) of fibers and vessels revealed
similar values by these two techniques, the concentration in the cell corner middle lamella
(ML,,) was lower by the UV-EDXA technique [91]. The observed discrepancies are due
probably to the uncertainty in estimating the guaiacyl/syringyl ratio, as the analysis is
made by UV microscopy alone.
Essential
Major Constituent
C Ca 754 B 0.9 Ag 0.23 Au 0.04
0 K 865 Mn 19.3 AI 5.4 0.02
Ga
H Mg 171 Fe 2.6 Ba 20.2 In 0.03
N Na 23 MO 0.005 CO 0.0 1 La 0.04
P Si - Cu 2.5 Cr 0.05 Li 0.13
S Zn 0.9 Ni 0.1 1 0.003
Sn
c1 Pb 0.12 v 0.001
Rb 2.0 Zr 0.002
Sr 10.2
Ti 0.11
For seven species, Young and Guinn [ 1091 have determined the distribution of 12 inorganic
elements in various tissue areas of a tree such as the roots, bark, wood, and leaves. The
results indicated that both total ash content and concentration of each element vary sig-
nificantly within and between the species. Therefore, unlike major cell wall components
such as cellulose and lignin, the content of inorganic constituents varies to a great extent
with the environmental conditions under which the tree has grown [ 105,1131. Little has
been published regarding the morphological distribution of elements in the cell [ 1 19- 1221.
By microincineration, Lange [ 1 191 found that mineral constituents of Swedish spruce are
deposited predominantly in the compound middle lamella. Zicherman and Thomas [ 1201
also have pointed out that careful ashing of microtome sections of loblolly pine (Pinus
r n e h L.), followed by electron microscopic observations, gives an ash residue distributed
throughout the cell wall and concentrated in the compound middle lamella and S, layer.
Wultsch [ 1211 stated that manganese is concentrated in ray cells, and Bergstrom [ 1221
reported that the phosphoruscontent is highest in the cambiumandadjacentxylem
portions.
Saka and Goring [ 1151 have studied the distribution of inorganic constituents from
the pith to the outer ring of black spruce (Picea nznriann Mill.) by means of TEM-EDXA.
The TEM-EDXA technique is a useful tool fordetectinganyelementaboveneonand
recently above boron in the periodic table. Figure 12 shows seven morphological regions
of the tracheids, ray tracheids, andrayparenchymacells investigated. Thedarkcircle
indicates the location of the analysis and its diameter corresponds to the resolution of
analysis (400 nm). Detected were 15 different elements, such as Na, Mg, AI, S, Cl, K,
Ca, Cr, Fe, Ni, Cu, Zn, and Pb, above neon in the periodic table. The secondary walls of
tracheids, ray tracheids, and ray parenchyma cells usually contain detectable concentrations
of only four elements: sulfur, chlorine, potassium, and calcium. In contrast, almost all the
elements were found to be localized and concentrated in the torus and half-bordered pit
membrane regions (Fig. 13). The total content of inorganic constituents decreased in the
order of torus (2%) > half-bordered pit membrane (1%) > middle lamella (0.4%) > ray
parenchyma cell wall (0.3%) > tracheid secondary wall (0.1-0.15%). The total content of
inorganic constituents was higherin earlywood than latewoodfor any of the morphological
omposition
Chemical and Distribution 73
FIGURE 12 Transmission electron micrographs of a cross section of black spruce showing the
seven different morphological regions. All micrographs were takenat the same magnification.
S , = secondary wall of the tracheid
CC = cell comer middle lamella surrounded by tracheids
TT = tours in an intertracheid pit pair
SR = secondary wall of the ray parenchyma cell
M = a half-bordered pit membrane between ray parenchyma cell and tracheid
SRT = secondary wall of the ray tracheid
TRT= torus in an intertracheid pit pair between ray tracheid and tracheid
74 Saka
FIGURE 13 EDXA spectra from the tracheid secondary wall and tracheid torus in black spruce.
(From Ref. 115.)
regions studied. This is probably because the earlywood tracheids that have large lumens
and abundant pits are the major water-conducting tissues, whereas thick-walled latewood
tracheids with fewer pits may act as a physical or mechanical support for the wood.
Bailey and Reeve [l231 have recently used imaging microprobe secondary ion mass
spectrometry (SIMS) to determine the distribution of the trace elements in black spruce
(Picea rnariana Mill.). This imaging microprobe S N S technique is a powerful tool for
detecting inorganic elements with high spatial resolution and high sensitivity. Their overall
findings correlatewell with results from the TEM-EDXA studyby Saka and Goring[ 1151.
However, due to its higher sensitivity compared with the EDXA technique, the distribution
of the elements within the cell wall could be more clearly demonstrated. Figure 14 is one
example in which some elements are visualized and concentrated in the middle lamella
region.
Recently, Saka and Mimori [l241 have studied the distribution of inorganic constit-
uents of Japanese birch wood (Betula platyphylla Sukatchev var. Japonika Hara) by the
SEM-EDXA technique with thin sections. Figure 15 shows six morphological regions of
the fibers, vessels, and ray parenchyma cells investigated. The dark circle corresponds to
the resolution of analysis (800 nm). Detected were 11 different elements: Na, Mg, Al, Si,
P, S, Cl, K, Ca, Fe, and Zn. The secondary walls of wood fibers, vessels, and ray paren-
chyma cells usually contained detectable concentrations of three elements, S, Cl, and Ca,
while, in the amorphous layer of ray parenchyma cell and pit membrane between vessel
and ray parenchyma cell, almost all of the detected elements were found to be localized
and concentrated (Fig. 16). The total content of inorganic constituents decreased in the
order amorphous layer (0.68%) > fiber middle lamella (0.54%) > vessel middle lamella
Composition
Chemical 75
FIGURE 14 Ion image and its intensity for Ca. Fe, and Mn from a tangential section of a double
cell wall of black spruce heartwood. (Courtesy
of Prof. D. W. Reeve, Universityof Toronto, Toronto,
Ontario, Canada.)
(0.48%) > ray parenchema cell wall (0.15%) > fibersecondary wall (0.14%) > vessel
secondary wall (0.10%).This observed trend is basically the same as found
in black spruce
by Saka and Goring [115].
In the previous sections, current knowledge of the distribution of cell wall constituents
was described. In this section, therefore, how these constituents construct and organize the
cell wall structure is discussed.
In wood cell walls, cellulose acts as the structural
framework in the formof cellulose
microfibrils, while hemicellulose is the matrix substance present between these microfi-
brils. Lignin, on the other hand, is the encrusting substance binding the wood cells together
and giving rigidity to the cell wall. Generally, the S2 layer increases with increasing wall
thickness, whereas the S , and S , remain fairly constant. Because of its greater thickness,
the S2 layer is largely responsible for the physical and mechanical properties of the cell
walls.
Figure 17 shows the relationship for softwoods between the lignin content and mi-
crofibrillar angle (e) in the tracheid S2 layer determined by the X-ray diffraction method.
Since the majority of the lignin in softwoods is in the tracheid S , layer [50], the whole
lignin content of wood must be closely correlated to the lignin concentration in the S ,
layer of the tracheid. Thus, from Fig. 17, the lignin concentration in theS , layer increases
76 Saka
FIGURE 15 Scanning electron micrographs of a transverse section of the Japanese birch wood
showing the six different morphological regions considered in this study.
Fs = secondarywall of thewoodfiber
F,, = cell comer middle lamella surrounded by wood fibers
Vs = secondarywall of thevessel
VML= cell comer middle lamella surrounded by vessel and wood fibers
Rs = secondarywall of therayparenchymacell
RA, = amorphous layer in the ray parenchyma cell
with increasing microfibrillar angleof the tracheid S2 layer [ 1251. The biosynthetic origin
of this relationship is not known. However, it does suggest that, in order to construct the
enforced plywood type of structure shown in Fig. 1, the three major chemical constituents
of wood mutually interact and strengthen each other to make up a natural supercomposite
material.
Figure 18 shows such an ultrastructural arrangement of cellulose microfibrils, hemi-
cellulose, and lignin in wood cell walls as proposed by Harada and CBtC [126]; around
the core of cellulose microfibrils, paracrystalline regions of cellulose are thought to exist,
which are associated with hemicellulose and lignin. Lignin encases them and binds them
into the rigid structure of the wood cell wall.
At the molecular level of arrangement of the chemical composition, the presence of
a chemical bond between lignin and carbohydrate has been proved to be a lignin-carbo-
hydrate complex (LCC) [l271 which is considered to be a compatibilizer-like substance
localized at the interface between hydrophobic macromolecules of lignin and hydrophilic
carbohydrates, by enhancing the physical and mechanical properties of wood [128].
Chemical Composition and Distribution 77
D- . 4 I I
FIGURE 16 Scanning electron micrograph (a) of a cross section of Japanese birch. The arrow
shows the location of the EDXA analysis at the pit membrane between vessel and ray parenchyma
cells from which the EDXA spectrum (b) was obtained.
45
2 40
Y
c
c
2 35
S
c
*g30
3
25
20
0 10 5 020 4 0 3 0
Microfibrillar angle Cel
VIII. CONCLUDINGREMARKS
Knowledge of the chemical composition of woodis essential for studying the physical
and chemical properties of wood. However, it can provide nothing but the average of the
cell wall constituents. For a better understanding of wood properties, more detailed infor-
mation is required about their distribution across the wood cell wall. However, in spite of
a variety of methods proposed, all the methods have drawbacks and thus some discrepancy
exists among investigators. A good method for resolving such discrepancies would be to
separatevarioustypes of tissues physicallywithoutintroducinganychemicalchanges
[49,96]. Analysis of the separated tissues could then provide definitive information on the
distribution of the cell wall constituents at the various morphological regions of wood.
REFERENCES
1. A.B.Wardropand D. E. Bland, in Biochemistry of Wood (K. Kratzland G. Billek, eds.),
Pergamon Press, London, p. 92 (1959).
2. H. Harada, Mokuzai Gakkaishi, 30:513 (1984).
3. F.F.P. Kollmann and W.A. C M , Jr., Principles of Wood Science and Technology, Vol. I,
Solid Wood, Springer-Verlag, Berlin, p. 43 (1968).
4. B. L. Browning, The Chemistty of Wood, Wiley-Interscience, New York, p. 57 (1963).
5. T. E. Timell, in Cellular Ultrastructure of Woody Plants (W. A. CBtC, Jr., ed.), Syracuse Univ.
Press, New York, p. 127 (1965).
6. H. Meier, in Biosynthesis and Biodegradation of Wood Components (T. Higuchi, ed.), Aca-
demic Press, New York, p. 43 (1985).
7. F.F.P. Kollmann and W. A. CBtC, Jr., Principles of Wood Science and Technology, Vol. I,
Solid Wood, Springer-Verlag, Berlin, p. 55 (1968).
8. W. A. C M , Jr., in Recent Advances in Phytochemistry, Vol. 11 (F. A.Loewusand V. C.
Runeckles, eds.), Plenum Press, New York, p. 1 (1977).
9. T. E. Timell, Wood Sci. Technol., 1:45 (1967).
10. H. Higuchi, KASEAA, 13:206 (1975).
11. T. E. Timell, Wood Sci. Technol., 16233 (1982).
12. G. Schwerin, Holdorsch., 12:43 (1958).
13. M. Fujii, J. Azuma, F. Tanaka, A. Kato, and T. Koshijima, Wood Res., 68:8 (1982).
14. G. C. Hoffmann and T. E. Timell, Tappi, 55:733 (1972).
Chemical Composition and Distribution 79
108. H, E. Young, P. N. Carpenter. and R. A. Altenberger. Tech. Bull. No. 20, Univ. Maine, Maine
Agric. Exp. Sta. (1965).
109. H. E. Young and V. P. Guinn, Rrppi, 49:190 (1966).
1 IO. M. L. Harder and D. W. Einspahr, Tappi. 63:110 (1980).
111. H . E. Young. Forest Prod. J., 2 / 5 6 (1971).
112. A. J . Panshin and C. de Zceuw, Te.srbook of Wood Techrwlogy, 3rd ed., Vol. l , McGraw-Hill,
New York, p. 73 (1970).
113. B. L. Browning. The Chenlisfry of Wood, Wiley-Interscience, New York, p. 355 (1963).
114. J. Savard, J. Nicolle, and A. M. Andre, Analyse chimique des bois tropicaux, Centre Tech-
nique Forestier Tropical. Nogent-sur-Marne (1960).
1 15. S. Saka and D. A. I. Goring, Mokuzcri Gakkcrishi, 29:648 (1983).
116. E. T. Choong, G. Abdullah, and J. Kowalczuk, LSU Wood Uti1i:utiorl Notes No. 24, (1976).
117. B. E. Cutter, E. A. McGinnes, Jr.. and D. H. McKown, Wood Fiber, 12:72 (1980).
118. H. Wazny and J. Wazny. Hol: Rol~-Werkst.,22:299 (1964).
119. P. W. Lange. Pulp. Ptrper Mcrg. Carl., 59:2 10 (1958).
120. J. B. Zicherman and R . J. Thomas, firppi, 54:1727 (1971).
121. F. Wultsch, P ~ p e rFc11x. 4: 128 ( 1944).
122. H. Bergstriim, S ~ w z s kPcrpper.stidn.. 62:160 (1959).
123. J. H. E. Bailey and D. W. Reeve, J . Pulp Prrper Sci., 20:J83 (1994).
124. S. Saka and R. Mimori, M o k w a i Gakknishi, 40:88 (1994).
125. S. Saka and M. Tsuji, Cellulose Clwm Teclmol.. 2/:225 (1987).
126. H. Harada and W. A. C6t6, Jr.. B i o s p t h e s i s and Biodegradation of Wood Co,nponer~f.s,
p.
20, Academic Press, New York (1985).
127. T. Koshijima, Wood Res. Tech. Nores, 19:1 1 (1984).
128. S. Takase, N. Shiraishi. and M. Takahama, Wood Processing and Utilizcrtion, p. 243, Wiley,
Chichester. U.K. (1989).
This Page Intentionally Left Blank
Structure of Cellulose: Recent
Developments in Its Characterization
Furnitaka Horii
Kyoto University, Kyoto, Japan
1. INTRODUCTION
The first Laue X-ray diffraction photographs of native cellulose, such as ramie and bam-
boo, were taken by Nishikawa and Ono in 1913 [l]. Since then a lot of effort has been
madetocharacterize the crystal structure of nativecellulose,mainlyusing the X-ray
diffraction technique. Nevertheless, the detailed structure has not been clarified as yet, and
the reason for the difficulty has only recently been understood reasonably. Although it was
assumed in the traditional analyses that native cellulose was composed of single homo-
geneous crystals, this assumption has been found not to be fulfilled in most specimens
from different native sources: native cellulose crystals have been confirmed to be com-
posites of at least two types of crystal allomorphs. This chapter deals with the process of
theconfirmation of such a new crystal structuremodelfornativecellulosesomewhat
historically. Further recent developments relating to this model of crystallization are also
described by focusing attention on cellulose biogenesis by a bacterium. For other progress
in characterization of cellulosic materials, refer to recent review articles [2-41.
Since the first observation of CP/MAS I3C NMR spectra of native celluloses in 1980 [5,6],
it has been recognized that there are great differences in multiplicities of the Cl and C4
resonance lines between two groups, the bacterial-Vrtlorziu group and the cotton-ramie
group. Figure 1 shows CP/MAS I3C NMR spectra of the crystalline components of these
native celluloses [7,8], which were selectively measured by using longer "C spin-lattice
relaxation times (TIC)of the components[9,10]. Asis readily seen,cotton and ramie
celluloses give almost the same crystalline spectrum, whereas bacterial and klonia cel-
luloses produce a different type of crystalline spectrum. The most prominent features of
these two types of crystalline spectra appear in the C1 resonances; a predominant doublet
with a weak central singlet is observed for cotton and ramie celluloses, while a triplet
composed of an enhanced singlet and a minor doublet appears for bacterial and Valonia
celluloses. These fine splittings, including those in the C4 and C6 lines, should be due to
83
a4 Horii
IIIIIIII,,I,,,II,,,,1,,,~I,,,,l,,,,I,,,,I,,,,I,,,,I,,I,I,,,,I,,,,I,
120 110 10 0 90 80 70 60 50
P m from TMS
FIGURE 1 CP/MAS "C NMR spectra of the crystalline components included in differcnt native
celluloses: (a) cotton; (h) ramie; (c) bacterial; (d) Vcrlorzicc celluloses 171.
surements of IR spectra about 40 years ago [ 1 l]; the absorption bands due to stretching
of OH and CH groups were markedly different between the bacterial-Vulonia group and
the cotton-ramie group. A separate electron diffraction study [ 121 also proposed that the
so-called eight-chain unit cell was appropriate to interpret the diffraction pattern observed
for Vuloniu cellulose, whereas the diffraction spots for ramie and cotton celluloses were
usually indexed in terms of the Meyer-Misch-type two-chain unit cell. In the 1970s more
detailed structural analyses were also carried out by wide-angle X-ray diffractometry [ 13-
151. However, these important efforts did not lead in a straightforward way to the idea
that native cellulose crystals are composed not of single pure crystals but of a mixture of
crystals with different crystal forms.
In 1984,AtallaandVanderHart [ 16,171 proposed by their careful elucidation of
CPMAS l3C NMR spectra of different native celluloses that native cellulose crystals are
composites of two allomorphs that are referred to as cellulose I, and cellulose I,. This
proposal, which should be estimated as an epoch-making contribution to the enhancement
ofnew effortstoward the structural characterization of nativecellulose,wasmade by
showing that the individual resonance lines were obtained by appropriate linear combi-
nations of the spectra of the regenerated cellulose I that contains almost the pure I, form
and bacterial cellulose that is rich in the I, form. Figure 2 shows schematically the spectra
of the C, and C, carbons thus obtained for celluloses 1, and I, [16-181. The I,, form has
singlets for the C l and C6 carbons and a doublet for the C4 carbon, whereas the I, form
has doublets for the C l , C4, and C6 carbons. The multiplicities in the I, and I, spectra
are reasonably interpreted by the structural inequivalences in single unit cells, in contrast
to those in the whole spectra as described in Fig. 1.
However, the novel structural model thus proposed, which is hereafter referred to as
the composite crystal model, was not readily established, because the relative intensities
of the C l and C4 triplets could not be fully interpreted by this model for different native
celluloses [7], suggestingthe existence of some exceptions depending onthe native sources
of celluloses. For example, Fig. 3 shows the results of lineshape analysis of the C4 res-
onance lines for native celluloses with higher contents of cellulose I,, [ 191. The most upfield
line, which would be described by a single Lorentzian curve according to the composite
crystal model, are found in this case to be composed of two Lorentzians with different
intensities for these cellulose samples. In such a situation we need additional experimental
evidence to support the composite crystal model.
c1 c4
cellulose I a
I
I
I I I
I I I
I I I
I
I
I
I
i
I h! I
t
cellulose I g
The first treatment was developed originally from the steam explosion of wood, by
carefully elucidating the experimental fact that CP/MAS "C NMR spectra of Japanese
birch and cypress were greatly changed before and after steam explosion at 255°C [24].
However, drastic degradation of cellulose occurs with saturated steam at 255°C when the
samples are not wrapped with glass fiber sheets. Such degradation Inay be induced by the
effect of hydrogen ions that are produced by the much higher level of the dissociation of
water at higher temperatures. Considering this situation, influences of the pH values of
aqueous nledia were investigated at different temperatures. Finally, annealing was found
Structure of Cellulose 87
, . 1 . , 1 , I . I , l i , I , I . I
92 90 R8 X6 92 90 88 X6 92 90 86
ppm from TMS
FIGURE 3 Lineshape analyses of the C4 resonance lines for Valonia mncrophyscr (a), Cladophonr
(b), and bacterial (c) celluloses [19].
c1
C 2.3.5
l
Y r
100
60 80
ppm from T M S
0. 1 1 1 1 1 I I 1 I
t
1 0
Annealing temperature /'C
FIGURE5 The mass fraction of cellulose I, as a function of annealing temperature when annealed
in 0.1 N NaOH aqueous solution: (A) under saturated steam pressure; (0)under 5 kbar.
fraction diagram becomes completely symmetric for the sample annealed at 260°C, indi-
cating monoclinic character. It was also found that all spots shownin Fig. 6C are indexed
with a two-chain P2, unit cell with a = 0.792 nm, b = 0.822 nm, c = 1.036 nm, and y =
97.3", whilethecorrespondingspotsshown in Fig. 6Aareindexedwithatwo-chain
triclinic unit cell with a = 0.954 nm, b = 0.825 nm, c = 1.036 nm, a = 90°, p = 57.0",
y = 96.6". It is therefore concluded that cellulosesI, and I, should be assigned to the two-
chaintriclinicandmonoclinicphases,respectively.However, the more stable form of
cellulose I, can be assigned to the one-chain triclinic phase, as described later.
Similar structural changes by annealing are also recognized by FT-IR spectroscopy.
Figure 8 shows the FT-IR spectrum of Rhizoclonium cellulose before and after annealing
in the 0.1 N NaOH aqueous solution [27]. Two absorption bands are clearly observed at
3240 and 3270 cm" in the OH stretching region and also at 750 and 710 cm" in the
CH2 rocking region for the original sample, but the former bands almost disappear after
annealing. This fact clearly indicates that the former bands are assignable to cellulose I,,
whereas the latter bands can be ascribed to cellulose I,. Moreover, it may be suggested
FIGURE 6 Electron diffraction diagrams of Vuloniu cellulose before andafter its annealing: A,
original; B, annealed at 240°C; C, annealed at 260°C [26].
90 Horii
mer i d i a n
I l I I I
0 .2 .4
R
FIGURE 7 Densitometer traces of the third layers of the electron diffraction diagrams for Valonia
cellulose annealed at different temperatures: (a) control; (b) 220°C; (c) 240°C; (d) 260°C.
4000 3600
3200
2800
2400 1000 800 600 400
wavenumber (cm-') wavenumber
(cm-')
FIGURE 8 FT-IR spectra o f original (A) and annealed (B) Khi:oc.lorliltr?! cellulose 127)
Structure of Cellulose 91
Since this equation is a sort of calibration curve, the mass fraction f , (=f r"")
of cellulose
I, can be determined by FT-IR spectroscopy using a specimen of the order of milligrams.
valonia, bacterial
cotton, ramie
I-
L-
- annealed
tunicate
[21,30]. In contrast, it is still impossible to obtain the pure l,, form in nature and also by
any artificial method at present.
As for the characterization of woods, there is normally a difficulty in exact I3C NMR
measurements of the crystalline component because of the low crystallinity and the co-
existence of hemicelluloses and lignin. However, our recent CP/MAS "C NMR analysis
has clearly revealed that normally lignified woodcellulose in Populus muxomowiczii,
which belongs to hard woods, is cotton-ramie type, the mass fraction of cellulose I, being
estimated to be about 0.8 1311. This result seems to be in conflict with our preliminary
results for Japanese cypress and birch [24], in which cellulose I,, was assumed to be rather
dominant. Recently, Newman measured separately the CP/MAS "C NMR spectrum of the
cellulose component as a result of the removal of the contributions from hemicelluloses
and lignin by using the difference in ' H spin-lattice relaxation time TlpHin the rotating
frame for different hardwoods and softwoods [32]. Since the C1 and C6 resonance lines
still seemed to contain unidentified contributions, the relative peak intensities of the most
downfield and upfield lines of the C4 triplet were used to estimate the relative proportions
of the I,, and I, forms. It was then concluded that the I,, fraction for softwoods was at
almost the same level as for Vulonia-bacterialcelluloses,while the fractions for hard-
woodsweresimilar to those for the cotton-ramie group. This conclusion may suggest
that the crystallization of cellulose in woods will be affected in the presence of hemicel-
luloses and lignin, possibly resulting in the difference in fractions of I, and I, forms. A
possible stress-induced crystallization of cellulose I,, which may be also induced in woods
by the coexistence of hemicelluloses and lignin in the hybrid composites, will be described
in the case of bacterial cellulose in a later section.
. c
B Ti4
0
110
@ O O
FIGURE 10 A series of spot electron diffractograms presenting the triclinic features, which were obtained for the single
microfibril of Microdictyon cellulose. (From Ref. 33, with permission of the American Chemical Society,Washington, DC.)
P
1
A
B 0
.
.
0 0
. . .
0 . . 0
FIGURE 11 A series of spot electron diffractogramspresenting the monoclinic features, which were obtained for the single
microfibril of Microdicryon cellulose. (From Ref. 33, with permission of the American Chemical Society, Washington, DC.) I
-.
P
Structure of Cellulose 95
cell, in good accord with the crystal transformation from cellulose I,, to I, as described
above.
Figure 12 shows the crystal structure models proposed for the one-chain triclinic and
two-chain monoclinic phases on the basis of the microdiffraction analysis described above
[33]. The monoclinic unit cell is the so-called Meyer-Misch-type cell, and the central chain
in the cell is shifted downward by c/4 with respect to the comer chains. Moreover, this
chain is rotated by 7.4” relative to the (200) plane, while there is no such rotation for the
comer chains in the corresponding planes [33]. In contrast, when the central chain and the
corner chains at the right side are shifted upward respectively by cl4 and 2d4 with respect
to the corner chains at theleft side, one type of two-chaintriclinic unit cell can be obtained.
Moreover, when the central chain is identical with the comer chains without any rotation
with respect to the (200) plane, then another type of one-chain triclinic unit cell can be
defined. This is the case for another allomorph of cellulose, cellulose I,, as shown in Fig.
12. As for the relative position of neighboring chains in the (1-10) plane of the one-chain
triclinic unit cell, there are two possibilities, “parallel up” and“paralleldown.”Here,
“up” or “down” indicates that the z coordinate of 0 5 is larger or smaller than the coor-
dinate of C5, respectively.
A molecular dynamic simulation in the crystalline environment was carried out by
using the structure models of celluloses I, and I, shown in Fig. 12 as starting structures
[37]. The program used was GROMOS 87 equipped with the appropriate force field. The
triclinic phase was simulated under the periodic boundary condition for 4 X 6 X 3 (a X
b X c ) unit cells with 2016 atoms, while the monoclinic phase was simulated for 3 X 3
X 3 unit cells with 15 12 atoms. It was found that the I, phase is energetically lower by
8.7 kJ mol-’ cellobiose-’ than the I, phase, in agreement with the almost complete crystal
transformation from cellulose I,, to I, at higher temperatures [20,21]. Such higher stability
in the I, form is due to intraplanar electrostatic interactions, particularly in the (200) plane.
Moreover, the rotation of the chain in the (200) plane increases from 7.4” to I l S ” , resulting
Monoclinic
Two-Chain
Triclinic
One-Chain
FIGURE 12 Crystal structure models of celluloses I, and I, assignable to the one-chain triclinic
and two chain monoclinic crystals, respectively [33].
96 Horii
Even after the establishment of the composite crystal model for native cellulose, there still
remains a question to be answered: how are celluloses I,, and I, crystallized in nature?
The recent electron microdiffraction analysis described above also revealed that the I,, and
I, phases are alternatively locatedwitha periodicity of about 100 nm along the single
microfibril for Microdictyorz [33]. In the case of bacterial cellulose, such a periodic struc-
ture cannot be observed and almost one series of diffraction spots assignable to the I,,
phase is obtained along the single microfibril [39,40]. Since the mass fraction of cellulose
I, is 0.37, this phase must be located in the thin central core area along the microfibril,
as will be described later. These facts suggest that the mode of distribution of the two
allomorphs in single microfibrils may change from sample to sample and thus the crys-
tallization of the allomorphs also depends on the conditions for the production of different
cellulose samples.When the hemicellulosesand ligin coexistduring the crystallization
process, effects of these materials should be well evaluated (411. In this situation, it will
be very important to investigate the crystallization process in each system producing prob-
ably different composite crystals in nature. Here the case of the bacterial cellulose system
is described in some detail, because this system has been considerablyinvestigated hitherto
142-441.
FIGURE 13 Transmission electron micrographs of the negatively stained twisting ribbon assembly
(a) and splayed microfibrils (b) produced in the presence of 1.0% CMC [19].
J l l l l ~ l l l l l f l l l l l l l l l ~ ~ l f I I ~ I ~ ~ ~ 1 1 ~ ~ I 1 ~ ~ ~ I I ~ I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 [ 1 1 1 1 l * ~
100 80 60
ppm f r o m T M S
FIGURE 15 CP/MAS "C NMR spectra for bacterial celluloses cultured in the presence of dif-
ferent polymeric additives: (a) control; (b) 2.0% XG; (c) 2.5% CMC; (d) 25% PVA; (e) 10% PEG
[481.
100 Horii
1
0 0.5 1.o 1.5
Concentration of CMC l wt%
FIGURE 16 Mass fraction of cellulose I,, in bacterial cellulose cultured in the presence of CMC
with various degrees of polymerization (DP) and degrees of substitution (DS) versus concentration
of CMC: ( 0 ) control; ( 0 ) DP = 630, DS = 0.65; (A)DP = 80, DS = 0.57; ( 0 ) DP = 40, DS =
0.57; ( 0 ) DP = 80, DS = 1.43 [29].
cellulose [40,50]. It was therefore suggested that cellulose I,, is preferably crystallized in
larger-size microfibrils, whereas cellulose I, is dominantly produced in thinner microfibrils.
The mechanism of the preferential crystallization of celluloses I, and I, in microfibrils
with different sizes will be discussed after the description of the crystallization of cellulose
I1 in the bacterial cellulose system.
of celluloses I, and I, as well as cellulose II. Here, experimental results at the initial stage
are briefly described, because investigation has just started with this system.
Figure 17 shows transmission electron micrographs of the negatively stained band
materialspreparedbyincubationfor 3 h in Hestrin-Schramm’s medium at 4°C on an
electron microscope grid [53]. As is clearly observed in Fig. 17a, the band material is
extruded out perpendicularly to the long axis. This band material seems to be composed
of small irregular granules loosely linked to form strandlike structures in lying
the direction
of band extrusion, as pointed out in the previous paper [51]. However, when the micro-
graph is enlarged by 7.5 times as shown in Fig. 17b, most strands seem to be composed
of irregularly coiled or folded thin fibrils. Since each strand is basically extruded from
each pore on the surface of the bacterium in agreement with the previous finding [51],
each strand will correspond to the subelementary fibril with an irregular coiled or folded
appearance. In some areas parallel orientation of several units of the subelementary fibril
are observed, but almost no discrete electron diffraction due to the crystalline entities
could be detected for this band material by selected-area electron diffractometry.
In contrast, three discrete diffractions are obtained on the equatorial line for the band
material detached from the bacterium cell, as shown in Fig. 18 [53]. These three diffrac-
tions are well indexed as (1-10). (110), and (020) by using the unit cell of cellulose 11
[13], in good accord with the previous result [51]. Although they are considerably arced,
molecular chain axes are confirmed to be almost perpendicular to the extrusion direction
of the band bacterial. It is therefore concluded that cellulose 11 is partly formed in the
native band material, but it is very important to obtain more informationabout the detailed
structure of the band material, including the existence of folding structureof subelementary
fibrils, which was suggested previously [51,52], to propose the mechanism of the crystal-
lization of cellulose 11.
In this bacterial system, another important finding is that the band material and the
normalribbonassemblyaresequentiallyproducedwhentheincubationtemperatureis
changed between 4°C and 28°C [53]. Figure 19 shows the drastic change in the cellulose
production from the band material to the normal ribbon assembly as observed by electron
microscopy. In this case a TEM grid with a culture drop containing Acetobacter xylinum
that had been kept at 4°C for 5 h was transferred into an incubator regulated at 28°C and
the culture was carried out there for 7 min. It is clearly found that the band material,
FIGURE 17 Transmission electron micrographs of the negatively stained native band material that
was cultured at 4°C [53].
102 Horii
FIGURE 18 Selected-area electron diffraction pattern of the native band material detached from
the bacterial cell. Three equatorial reflections are ascribed to those of (l-lo), (110). and (020) planes
of cellulose II [53].
which was formed at 4"C, is linked sequentially to the normal ribbon assembly that is
produced at 28°C.
It is very difficult at present to describe how the production of the band material
or the normal ribbon assembly is controlled in this system. However, the movement of
the bacterium may be associated with the formation of these structures. Previous light-
microscopic observation [42,44,54] revealed that the bacterial cell rotates around its lon-
gitudinal axis as it is propelled forward by the elongating ribbon during the production of
the ribbon assembly. We also observed similar specific movements of the cells in the
culture medium by a differential-interference light microscope with a TV system [55]. In
contrast, their movements are not specifically similar to the well-known irregular move-
ments of Escherichia coli, when the cellulose is not produced. More careful observation
of the movements of Acetohacter xylinum is necessary at 4 and 28°C to elucidate the
processes of the different fibrillar structures.
Here, E is the Young's modulus of the plate along the longitudinal direction, r is the half-
width of the plate, and 8 is the twisting angle around the A axis, which is defined for a
given length L of the plate. Equation (2) clearly indicates that shear stress F, which is
proportional to r2, is induced along the longitudinal direction of the plate, when the plate
undergoes twisting around the long axis.
The normal ribbon assembly is not a homogeneous thin plate as shown in Fig. 20,
but the constituting microfibrils with somewhat different sizes are closely connected with
each other, probably through hydrogen bonding. Moreover, the twist of the ribbon will be
produced during the aggregation of the microfibrils. Nevertheless, when the normal ribbon
assemblyundergoes twisting around the longitudinal axis before crystallization, shear
104 Horii
A
P' '
I
-r -
stress may be induced along the longitudinal axis in proportion to the squares of the half
width of the ribbon. Under such shear stress, cellulose I,, will be preferentially produced
through so-called stress-induced crystallization, because the orientation of the molecular
chains in the I,, form seems to be organized by the shear stress along the chain axis. That
is, the cellulose chains in a given ( 1 10) plane are shifted upward or downward by c/4
against the corresponding chains in the upward or downward successive (110) planes, as
shown in Fig. 12. In addition, cellulose I, will be crystallized mainly in the regions along
both edges of the ribbon assembly, because the shear stress is higher in those regions.
In contrast, cellulose chains are packed in a more stable form in the unit cell for
cellulose I,, asshown in Fig. 12. This type of crystal may beformedunder less shear
stress or without shear stress, which may be referred to as stress-jree crystcrllization. Ac-
cording to the structural model shown in Fig. 20, cellulose 1, will be crystallized in the
central core region because the shear stress is less or free in this region. When the ribbon
assembly is partly disordered or splayed into somewhat thinner microfibrils, such a part
may be also composed of cellulose 1, because the twisting induces much less shear stress
for disordered or smaller-size microfibrils. The dependence of the mass fraction of cellu-
lose I, on the size of microfibrils described in the previous section is also well interpreted
in terms of the structural model shown in Fig. 20.
In the case of Microdicfyon cellulose, in which 1, and I, crystals appear alternately
along a single microfibril as describedabove [331, another crystallization mechanism
should be proposed: the so-called two-step orientationcrystallization [ 191. Before the
initialization of the crystallization, the cellulose chains are highly oriented in each micro-
fibril just as in the liquid crystalline state, but they are not fully extended to the molecular
chain length. Therefore, after the crystallization of cellulose I,, under some different stress
(stress-induced ctystallization), the residual part of such chains left in the noncrystalline
state will be relaxed in length as a result of the full extension of the crystalline chains.
This willlead to the relaxation of the stress, andthen anothertype of crystallization,
Structure of Cellulose 105
stress-free crystallization, may occur to form cellulose I, in the central part of the non-
crystalline region. Thissuccessivetwo-step orientation crystallization will producethe
alternate I, and I, crystals along the single microfibril.
Finally, it should be noted that cellulose I can be crystallized in vitro after a cellulase-
catalyzedpolymerizationofP-cellobiosyl fluoride substratemonomer in acetonitrile/
acetate buffer [56].Under the normal condition employing unpurified cellulases, cellulose
I1 is crystallized in a reaction medium after the polymerization as expected in analogy of
the crystallization of cellulose from the solution [57]. In contrast to this fact, the cellulose
I allomorph was confirmed by selected-area electron diffractometry for fibrous materials
produced by using substantially purified cellulaseenzyme,although the discrimination
between the I, and I, forms was not achieved [56].Such an interesting result may be due
to the parallel orientation of extended cellulose chains with the same polarity, which is
assumedto be organized as a result of a micellaraggregation of the partially purified
enzyme and the substrate in the nonaqueous/aqueous solvent system. More detailed in-
vestigations of this system will also contribute to the exact interpretation of the crystalli-
zation mechanism of native cellulose.
VII. CONCLUDINGREMARKS
As described above, the composite crystal model has been established in native cellulose,
although there may be minor exceptions depending on the sources of cellulose. Neverthe-
less, the detailed structure, such as the chain conformation, hydrogen bonding, and mo-
lecular packing, has not yet been clarified in both allomorphs because precise structural
analyses of the intensities of X-ray or electron diffraction diagrams have not been per-
formed systematically so far. The situation is more serious for cellulose I,, because no
pure I, specimen is available yet either in nature or by an artificial method. As a better
way, the analysis will be made for specimens with higher contentsof the I,, form, assuming
the linear combination of diffraction intensities of the I, and I, forms. In that case it should
be taken account that the I, form is frequently subjected to some modification depending
on the source of cellulose as shown in Fig. 3.
As for the crystallization processes of celluloses I, and I, as well as cellulose I1 in
nature, more detailed observations are necessary for the microfibril or subelementary mi-
crofibril structure at the level of several nanometers to several tens of nanometers by
transmissionelectronmicroscopyandalso for the overall movementsof Acerobacter
xylinurn by high-performance light microscopy. Information about the enzymatic degra-
dation, acid hydrolysis, alkaline treatments, etc., for microfibrils, which is not described
here because of space limitations, will be also very helpful for understanding the structures
of microfibrils andsubelementary microfibrils. Through these investigations ourunder-
standing of the structure of native cellulose will be greatly advanced in the next 5 or 10
years.
ACKNOWLEDGMENTS
The author thanks Dr. Asako Hirai and Hiroyuki Yamamoto for their constant cooperation
throughout the work described here. He is also grateful to Prof. Masaki Tsuji and Prof.
Junji Sugiyama for their kind cooperative contributions to structural analyses by electron
microscopy.
106 Horii
REFERENCES
1 . S. Nishikawaand S. Ono, Proc. Tokyo Math.-Phys. Soc., 7131 (1913).
2. K. Okamura, in Wood and Cellulosic Chemistry (D. N.-S. Hon and N . Shiraishi, eds.), Marcel
Dekker, New York and Basel, p. 89 (1991).
3. L. M. J. Kroon-Batenburg, and J. Kroon, Carbohydrate in Europe, p. 15 (1994).
4. A.C.O’Sullivan, Cellulose, 4: 173 (1997).
5 . R. H. Atalla, J. C. Cast, D. W. Sindorf, V. J . Bartuska, and G. E. Maciel, J . Ant.Chern. Soc.,
102:3249 (1980).
6. W. L. Earland D. L. VanderHart, J. Am. Chem. Soc., /02:3251 (1980).
7. F. Horii, A. Hirai, and R. Kitamaru, Macromolecules, 20:2117 (1987).
8. F. Horii, in Nuclear Magnetic Resonance in Agriculture (Pfeffer, P. E., Gerasimowicz, W. V.,
eds.), CRC Press, Boca Raton, FL, chap. 10 (1989).
9. F. Horii, A. Hirai, and R. Kitamaru, J. Carbohydr.Chem., 3:641 (1984).
10. A. Hirai, F. Horii,andR.Kitamaru, CelluloseChem.Technol., 24:703 (1990).
11. H. J. Marrinan and J. Mann, J. PolymerSci., XXI:301 (1956).
12. G. Honjo and M.Watanabe, Macromolecules, /81:326 (1958).
13. A. Sarko and R. Muggi, Macromolecules, 7486 (1974).
14. K. H. Gardnerand J. Blackwell, Biopolymers, 13:1975(1974).
15. C. Woodcock and A. Sarko, Macromolecules, 13:1183(1980).
16. R. H. Atalla and D. L. VanderHart, Science. 223:283 (1984).
17. D. L. VanderHart andR. H. Atalla, Macromolecules, /7:1465 (1984).
18. H. Yamamoto and F. Horii, Macromolecules, 26: 13 13 ( 1993).
19. F. Horii, H. Yamamoto, and A. Hirai, Mucromol. Symp., 20: l97 (1997).
20. F. Horii, H. Yamamoto, R. Kitarnaru, M. Tanahashi, and T. Higuchi, Macron~olecules,20:2946
(1987).
21. H. Yamamoto, F. Horii, and H. Odani, Macromolecules, 22:4130 (1989).
22. A. Hirai, F. Horii, and R . Kitamaru, Macromolecules, 20:1440(1987).
23. H. Chanzy, B. Henrissat, M. Vincendon, S. F. Tanner, and P. S. Belton, Cnrbohydr. Res., 160:
1 (1987).
24. M. Tanahashi, T. Goto, F. Horii, A. Hirai, and T. Higuchi, Mokuzai Gakkaishi, 35:654 (1989).
25. E. M.Debzi, H. Chanzy, J. Sugiyama, P. Tekely, and G. Excoffier, Macromolecules, 246816
(1991).
26. J. Sugiyama, T. Okano, H. Yamamoto, and F. Horii, Macromolecules, 23:3196 (1990).
27. J. Sugiyama, J. Persson, and H. Chanzy, Macromolecules, 24:2461(1991).
28. J. H.Wiley and R. H. Atalla, in The Structure of Cellulose (R. H. Atalla, ed.), ACS Symp.
Ser. 340, American Chemical Society, Washington, DC, p. 151 (1987).
29. H. Yamamoto, F. Horii, and A.Hirai, Cellulose, 3:229 (1996).
30. P. S. Beleon, S. F. Tanner, N. Cartier, and H. Chanzy, Macromolecules, 22:1615 (1989).
31. M.Wada, T. Okano, J. Sugiyama, and F. Horii, Cellulose, 2:223 (1995).
32. R.H.Newman, J . WoodChem.Technol., 14:451(1994).
33. J. Sugiyama, R. Vuong, and H. Chanzy, Macromolecules, 24:4168 (1991).
34. K. H. Meyerand L. Misch, Helv. Chim. Acta, 11534 (1937).
35. D. P. Miller and A. Li, in Cellulose and Wood, Chemistty and Technology (C. Schuerch, ed.),
Wiley, New York, p. 139 (1989).
36. R. P. Millaneand T. V. Narasaiah, CelluloseandWood:Chetnistty und Technology (C.
Schuerch, ed.), Wiley, New York, p. 39 (1989).
37. A. P. Heiner, J. Sugiyama, and 0. Teleman, Curbohydr. Res., 273:207 (1995).
38. B. J. Hardy, and A. Sarko, Polymer; 37:1833(1996).
39. F. Horii, A. Hirai, H. Yamamoto, and J. Sugiyama, Preprinfs of the3rdAnnual Meeting of
the Cellulose Society of Japan, published by the Cellulose Society of Japan (% Institute for
Chemical Research, Kyoto University), p. 1 1 (1996).
40. A. Hirai, F. Horii, M. Tsuji, J . Sugiyama, and H. Yamamoto, Proc. 1st Int. Symp. Inst.Chem.
Res., Kyoto Univ. (ICRIS ’96), p. 132 ( 1 996).
Structure of Cellulose 107
1. INTRODUCTION
Lignin exists as one of the essential wood components, ranging in amount from 10% to
30%. It is thought that lignin is a polymer formed by the enzymatic dehydrogenation of
phenylpropanesfollowed by radical coupling.Softwood lignin is composedmainly of
guaiacyl units originating from the predominant precursor, rrans-coniferyl alcohol (l),
while hardwood lignin is composed of both guaiacyl andsyringyl units derived from trans-
coniferyl (L) and trans-sinapyl (2) alcohols, respectively. Grass lignin contains p-hydrox-
yphenyl units derived from trans-p-coumaryl alcohol (3),besides units originating from
the foregoing two precursors. However, strictly speaking, almost all plants consist more
or less of all three units, namely, guaiacyl, syringyl, and p-hydroxyphenyl moieties. Lignin
has no optical activity, in contrast to other compounds, because the radicals formed by
enzymatic dehydrogenation couple with one another at random to give the lignin polymer.
OH OH OH
109
San0
110 and Sakakibara
problem. Information on the degradation products from protolignin could provide direct
evidence about lignin structure. For this purpose, the following procedures for bringing
about lignin degradation are effective: catalytic hydrogenolysis, controlled hydrolysis, and
degradation by thioacetolysis or thiacidolysis followed by reduction. As complementary
procedures, oxidations of lignin with KMnO,, nitrobenzene under alkaline conditions, and
acidolysis are available. The quantitative data for various functional groups and linkage
types in protolignins are also essential to understanding lignin structure. Ultraviolet (UV),
infrared (IR), and nuclear magnetic resonance (NMR) spectroscopic techniques, particu-
larly when used in conjunction with chemical modification, have contributed to estimating
the frequencies of functional groups and linkage types.
Some lignin structural models could be proposed from these combined studies, but
they do not represent a strict molecular structure such as those for other natural polymers
such as cellulose and proteins. This may be unavoidable, however, in view of the bioge-
netic mechanism governing the formation of lignin.
II. SPECTROSCOPY
A. Ultraviolet(UV)Spectra
Ligninshows a strongabsorptionspectrum in the UV region, becauseof its aromatic
nature. The lignin spectrum of a typical softwood in Fig. 1 [ l ] has two maxima at 205
and 280 nm, shoulders at 230 and 330 to 340 nm, and a minimum at 260 nm. In general,
softwood lignin shows a maximum at 280 to 285 nm, and hardwood lignin at 274 to 276
nm [ l ] . As the method is verysimple, UV spectrophotometric investigations are used
extensively to characterize lignin preparations.
Aulin-Erdtman [2-51 applied UV spectra to estimate effectively the amounts of
certain functional groups, especially the phenolic hydroxyls of lignin, by means of differ-
FIGURE 1 UV spectra of pine and beech lignins and A E curve for pine lignin. (From Ref. 1.)
Chemistry of Lignin 111
OH OH
W
0
, H“
.bCH3
Pew [6] found that C-4-substituted guaiacyl and related compounds with unconju-
gatedsidechainsshow similar andalmost identical curves, as shown in Fig. 2 . These
model compounds haveultraviolet absorption maxima at 280 nm, but unlike that for lignin,
the curve tills abruptly to zero at 300 nm and nearly to zero at 250 nm. The filling in of
the spectral trough at 250 nm in lignin can be explained by the presence of biphenyl-
linked units (Fig.2). Until then, biphenyl units in lignin molecules had generallybeen
assumed to be minor contributors, butPew [7 I presumed that coniferous lignin may contain
considerable biphenyl-linked units.
A simpleand rapid method for the determination of phenolichydroxylgroups in
lignin preparations was developed by Goldschmid [8] based on the A&, method. The phenol
content of the sample is calculated from the absorptivity maximum of the resulting dif-
ference curve, and the molar absorptivity maximum of model phenols is determined i n
the samemanner. The 300-nm maximum of the differencecurves is characteristic of
phenolic hydroxyl groups without con.jugation. The method is suitable for routine appli-
112 Sakakibara and San0
1.0
0.8
E 0.6
m
-2
S1
2
0.4
0.2
0
230240250260270280290 300 320340360400
A-mp
FIGURE 2 UV spectra of spruce cellulytic enzyme lignin and of C4substituted unconjugated
guaiacyl compounds. (From Ref. 6.)
cation to technical softwood lignin preparations. Phenylcoumaran (g) (Arna 281 nm) un-
dergoes dehydration during acidolysis to give a phenylcoumarone structure (B), which
has a strong absorption band at maximum 310 nm. By studying the spectral curves of
these compounds, Adlerand Lundquist [9] estimated the contentof phenylcoumaran units
in lignin.
Schdning and Johansson [lo] studied the W absorption of lignin from pulp waste
liquor and concluded that acid-soluble lignin components should be determined at 205
nm, because the absorption maximum at 280 nm is influenced by degradation products
from carbohydrates, such as furfural. It was found that wood samples from spruce and
pine contain 0.2%acid-soluble lignin, from birch and eucalyptus 3-4%, from wheat straw
2%, and from bamboo 1.5%. Klason lignins, especially those from hardwood, must be
corrected for the acid-soluble lignin content.
Wegener et al. [l11 found that hexafluoropropanol is an excellent solvent for UV
and IR spectroscopy of lignins, because the absorption maximum of water-insoluble sam-
ples near 200 nm can be recorded exactly and evaluated quantitatively, due to the high
UV transmittanceproperties of thissolventwithoutinterferencefromthedegradation
products of polysaccharides.
B. Infrared(IR) Spectra
Infrared (R) spectroscopy has been used often for the characterization of lignin because
the technique is simpleand the sample to be studied does not need to be dissolved in any
solvents and is required only in very small quantity. Qpical IR spectra of soft- and hard-
woods are shown in Fig. 3. Various lignin preparations can be easily compared using this
technique. Assignment of various absorption bands of lignins in IR spectra has been stud-
Chemistry of Lignin 113
I I I I 1 I
I so0 1600 1100 120c 1000 800
Wavenurnber (cm")
FIGURE 3 FTIR spectra of MWLs from oak, birch, and spruce. Legend: oak (Q~erc-uscrispcln
Blurne), birch (Berrrla plntyphylln Sukatchev var. jcrponicn Hara), and spruce (Pinus glehnii Mast.).
ied by a number of lignin investigators, using a variety of methods. The major absorption
band frequencies and the most probable assignment of each band in guaiacyl and guaiacyl-
syringyl lignins are shown in Table I , which has been somewhat supplemented from that
summarized by Hergert 1 1 21.
Kolboe and Ellefsen [l31 used IR spectroscopy as an independent method for esti-
mating lignin content and coumaran groups. The absorption at 1515 cm" was chosen for
the determination of lignin, because this region is assigned to aromatic skeletal vibrations.
The lignin content was estimated by the difference spectrum between the original wood
and holocellulose at ISIS cm-', giving 28-29%. Thatis in agreement with the value
generally accepted. Further, the absorption band at 1495 cm" was assigned to the cou-
tnaran ring, and it was estimated to contribute about 5% of the total phenylpropane units.
Sarkanen et al. 1141 comparcd the spectra of specifically deuterated guaiacyl and syringyl
models with thosc of undeuterated lignin models, to give several new band assignments.
From these deuteration studies, the hitherto unidentified bands at 1450 to 1420 cm" were
considered to be associated with the ring-stretching modes strongly coupled with the C-
H in-plane deformation similar to the l SO0 c m - ' band. The 1340- l380 and 1250- l l SO
cm" bandswere usually assigned to phenolic hydroxyl groups. The latter 1250- I 150
cm I band may be described as an "0-H in-plane deformation with ring-stretching char-
acter," and the 1340- 1380 cm ' band a s having a "ring-stretching with 0 - H bending
character." The 1240 cm ' band scctncd to have the most pronounced methoxyl character,
114 and Sakakibara San0
Band (cm"')"
Hardwood Softwood
lignin Assignment
3450-3400
3425-3400 stretching0-H
2940
2920 C-H stretching
methylene
ingroups
and
methyl
2880 2875-2850
2820
1715-1710
1715 Carbonyl
stretching in unconjugated
and
ketone
conjugated carboxylic groups
1675-1660 1675-1660 Carbonylstretching in conjugated
ketone
groups
1650-1630 sh Carbonyl stretching in y-lactone
1595 1605 skeletalAromatic vibrations
1505 1515-1510
1495 Coumaran ring
1470- 1460 1470- 1460 C-H deformations (asymmetric)
1430 1425 Aromatic skeletal vibrations
1370 1370- 1365 C-H deformations (symmetric)
1330- 1325 Syringyl ring breathing with C-0 stretching
1235-1230
1270 1275 Guaiacyl ring breathing with C-0 stretching
1230
1 l40 1145 C-H inplane deformation in guaiacyl
1130 C-H inplane deformation in syringyl
1085 1085 C-0 deformation in secondary alcohol and aliphatic
ether
1035 1030 C-H in-plane
deformation in
C-0
guaiacyl,
and
deformation in primary alcohol
970 970 (trans)
deformation
out-of-plane
=CH
915 855 Aromatic C-H out-of-plane deformation
860 815 sh
750-770 sh 750 sh
"sh: shoulder.
Source: Ref. 12.
because on methylation, the intensity of the 1240 cm" band increases at the expense of
the 1275 cm" band.
MWLs from the sapwood of several wood species have been characterized by ana-
lytical and spectral methods [15]. The results indicate that the basic lignin structure of
conifers is almost the same, and the differences between wood species seem to be the
differences in the identity and amount of ester groups ( l 7 15 cm- ') related to various
lignins. The IR spectra of lignins isolated from whole wood samples that include heart-
woodshowedan extra absorptionband at 1630 cm", whichwas totally absent in the
MWL from sapwood. Also, the IR spectrum of larch lignin showed a weak peak from a
conjugated carbonyl group at 1670 cm-'. These extra bandsthat may originate from poly-
phenols were considered to be indicative of the chemical modification of lignin during the
heartwood formation process. It was also demonstrated that the infrared spectrum is useful
Chemistry of Lignin 115
0 0
;
A
HCCl3
1
I
HMD
8.48 7.81 7.50 5.18 5.74 6.28 9.95
8 7 4 5 6 7 8 9 10 6
FIGURE4 ‘H NMR spectrum of acetylated spruce MWL (60 MHz, solvent: chloroform-d). (From
Ref. 18.)
San0
116 and Sakakibara
The presence of highly shielded protons in the range 6 1 SS-0.38 is noticeable. These
protons are clearly due to methyl or methylene groups that are not attached directly to
oxygen functions, carbonylgroups,aromaticsystems,orotherdeshieldinggroups,but
their structural origin seems to be hydrocarbon contaminant because lignin does not in-
clude these groups.
Lenz [l91 studied ‘H NMR spectra of bothunderivatizedandacetylated lignins,
using various deuterated solvents. Twelve lignin preparations, including MWL, kraft and
soda lignins, and dioxane lignin, were examined. It was found that alkali and acidolysis
lignin preparations from both hard- and softwoods showed marked differences in the de-
gree of condensation of aromatic rings, phenolic and aliphatic hydroxyl groups, and the
number of highly shielded aliphatic protons. Spruce MWL gave a proton distribution close
to that found by Ludwig [IS]. Bland and Sternhell [20] obtained from ‘H NMR spectra
estimates of the fraction of protons attached directly to aromatic nuclei in some lignin
preparations from Pinus radiata and Eucalyptus regnans. They estimated the frequencies
of “condensed” units in the lignin molecule; those in methanol and acetylated lignin from
E. regnans and methanol lignin from l? radiata are 0.60, 0.66, and 0.71/OCH3, respec-
tively. However, it should be noted that the condensed units here are aromatic rings in
which at least one carbon atom, not necessarily the 5-carbon alone, is linked directly to
another carbon atom outside the ring.
According to Morohoshi et al. [21], NMR spectra indicate that the degrees of con-
densed units in lignins from normal wood and compression wood of Abies sachalinensis
are 0.48 and 0.79, respectively. Also,from the results of the other analytical data, the
structure of compression wood lignin is generally composed of more condensed units.
Horisaki et al. [22] estimated the frequencies of “condensed units” in lignins using
methoxyl content, the ratio of guaiacyl and syringyl rings, and proton numbers assigned
to aromatic rings, ethylene groups, and benzyl alcohol groups obtained by the NMR (500
MHz) of acetylated lignins and their hydrogenation derivative-containing internal standard
(p-nitrobenzaldehyde) for quantitative analysis. They foundthat the contents of condensed
units and benzyl alcohol groups are 0.35 and 0.29/CCJin spruce MWL, and 0.30 and 0.451
C, in birch MWL.
The formylprotons in the lignin moleculecould not bedetectedwithlower-fre-
quency (60-MHz) NMR spectrometers [ 17-19]. Later, Lundquist and Olsson [23] studied
the formyl groups in spruce lignin at 270 MHz and found that the signals in the range 6
9.6- 10.0 were essentially due to the protons of aldehyde and vanillin units. By integration,
the amount of coniferyl aldehyde units was determined to be 4%. MWLs from birch 1241
and spruce [25] were further studied. The spectra are shown in Fig. 5. The use of a 270-
MHz instrument and new techniques have greatly increased the amount of available struc-
tural information. Signals in the spectra of acetylated birch and spruce lignins are sum-
marized in Table 2. The peak at 6 308 of birch lignin was attributed to H,, in p-p structures,
and integration suggests that around 0.05/C,,C3 are involved in @-pstructures. The peak
at 6 5.44 was assigned to H in p-5 structures or noncyclic benzyl aryl ethers, and 6 4.60
(H,) and 6.01 (HJ were essentially attributed to p-0-4 structures. Integration of the 6.01
peaksuggested that 40-50% of the units are attached to an adjacent unit by a p-0-4
linkage. In the case of spruce MWL, the peak at 6 2.28 (aromatic acetate) was found to
correspond to 0.26 phenolic groups/C,J (although certain phenol groups in biphenyl struc-
tures are not included in this estimate [ 17,181. From the peak at a 6 2.62 attributed to H,,
in p-p structures, 0.01 to 0.02/C,C, were estimated for dihydrofuran or open-type p-p
units. Later, however,from the results of studies usingdeuterioacetonesolutions [261,
higher values of 0.02 to 0.03/C,C3 were obtained. The integral of the 6 5.49 peak, H,, i n
Chemistry of Lignin 117
FIGURE 5 Spectral range 6 8.0- 10.5 (400 MHz, solvent DMSO-d,) of MWL from spruce. Notes:
the assignments of the peaks are indicated in the figure (the peak at 6 8.14 is due to an identified
contaminant). The dashed lines indicate the baselines used in the connection with quantitative es-
timates. (From Ref. 27.)
TABLE 2 Assignments of Signals in the 'H NMR Spectra of Acetylated Birch and Spruce
MWLs in Chloroform-d
S unitslppm
1.26 contaminantHydrocarbon
2.01,
1.95, 2.02 Aliphatic
acetate
2.13 Aliphatic
acetate
(including
aromatic
acetate in biphenyl
structure)
2.28-2.29 acetate Aromatic
2.94 Unknown
absent
peak
(the
is in acetate
the of lignin
reduced
with
NaBH,)
3.08 Hp in p-p structures
3.76-3.8 I Protons in methoxyl
groups
4.18,
4.27-4.28 H, in several
structures
4.39 H, primarily in p - 0 - 4 (erythro) and p-5 structures
4.43 H,structures
in several
4.60 H, in p - 0 - 4 structures
(birch)
4.65 H,, in p - 0 - 4 structures
(spruce)
including
methylene
protons in
cinnamyl alcohol units
4.70 H, in p-p structures(birch)
including
methylene
protons in
cinnamyl alcohol units
5.44 H, in p-5 structures
noncyclic
and benzyl
ethers
aryl
(birch)
5.49 H,, in p-5 structuresnoncyclic
and benzylaryl
ethers
and H,, in
aryloxypropiophenones (spruce)
6.0 1-6.06 H,, in p - 0 - 4 and p-l structures,and vinyl protons
6.93-6.94 Aromatic and vinyl protons
7.4 1 Aromatic
protons in benzaldehydeunits, vinyl protons
carbon
theon
atom adjacent to aromatic rings in cinnamaldehyde units (spruce)
7.50 Aromatic
protons
located
ortho to carbonyl
groups
(birch)
7.53 Aromatic
carbonyl
protons
groups
located
(spruce)
to
ortho
9.64 protons
Formyl in cinnamaldehyde units
9.84-9.86 Formyl
protons in benzaldehyde
units
On the basis of the model compounds, the various signals of the "C NMR spectra
of beech and spruce MWLs were assigned [29]. The spectra of DHP, spruce and beech
ligninsareshown in Figs. 7 and 8, respectively. The numbering 1-40 of the peaksis
convenient for comparing these spectra. The assignment o f . these carbon atoms is sum-
marized in Table 3. These chemicalshiftscan be divided roughly intothreereg' 'Ions:
carbonylcarbonsappearat 6 200- 160 (with respect t o TMS as a standard), Cl -C6
aromatic carbons, C,, and C, of double bonds on the side chain, at 6 160- 1 0 0 , and other
saturated side chains C,,, C,,, and C, at 6 90 to 20. Methoxyl carbons always appear in
the narrow range of 6 56.3 t 0.2. From these peak assignments, it was concluded that
the content of phenylcoumaran units i n spruce lignin is much more than that in beech
lignin and presumed that the content of dioxabicyclo-octane units [pinoresinol (3) and/
or syringaresinol (g)] of beech lignin may be more than that of sprucc lignin. The "C
NMRspectra of thedehydrogenationpolymers."Zulauf"-and"Zutropf"-DHPs.were
compared with thosc o f sprucelignin 1301, indicating that the conditions of formation
Chemistry of Lignin 119
I I I
6 9 8 7 6 5 4 3 2 1 0
FIGURE 6 ' H NMR spectra (270 MHz) of acetylated MWLs from birch and spruce wood. (From
Refs. 24 and 25.)
affect the constitution of DHPs. Spruce lignin differs from Zulauf-DHP in a lower content
of pinoresinol units and cinnamyl alcohol groups, and in a greater content of P-aryl ether
units, a-carbonyl groups, and etherified guaiacyl residues. Nimz and Ltidemann 1311 also
investigatedacetylatedlignins and DHPs. Acetylation makes it not only easy to assign
signals, but also has the advantage of increasing the solubility of lignin preparations (Fig.
9). Gagnaire and Robert (321 studied a DHP polymer model of lignin that was synthesized
by enzymatic dehydrogenation of coniferyl alcohol enriched to 90% in "C on the benzylic
position. Gated proton decoupling and selective proton irradiation were used to facilitate
assignment of the difficult "C, signals of the DHP. C,, atoms of various structural units,
such as vanillin, vanillic acid,coniferylalcohol,cinnamaldehyde. @ - S , p-p, and p-0-4
dilignol units, and the C,, involved in benzyl etherbonds in p-0-4 dilignolstructures,
were assigned.
Obst and Ralph [33] have tried to determine the relative syringyl/guaiacyl ratios for
hardwood lignins, and the following rcsults wereobtained. The syringyUguaiacyl inte-
grated peak area ratio for red oak fiber MWL is I .69, and that for white birch fiber MWL
is 2.28. Lapicrer and Monties 1341 estimatedeasily the ratios of syringyl/guaiacylfor
hardwood lignins using the signal intensities of C-2 plus C-6 for each from the conven-
tional NMR spectra.
120 Sakakibara and San0
3
19 22 33 34 40
6
a
3 38 5 24 29 33 34
2
2a
FIGURE 8
L 4a
TABLE 3 "C-Chemical Shifts (6) in ppmfrom TMS and Relative Intensities of Acetylated
Spruce MWL (Fig. 7) and Beech MWL (Fig. 9)
~
Beech, Spruce,
Signal PPm PPm
no. (intensity) (intensity) Assignment
TABLE 3 Continued
~~~~~~~ ~~ ~ ~ ~ ~
Spruce, Beech,
Signal PPm PPm
no. (intensity) (intensity) Assignment
R1
6I
OR
0~~
A :R,=H. R~=Ac
B : R,=H, R H k y l
C : Rl=OMc, RFAC
D : R I S M C , Rl-
(R = alkyl or Ac)
A A
Ace t y 1a t e d
SPRUCE MWL
"quant i t a t i v e "
170 150
Acetylated
SPRUCE MWL
"routine"
Wood
species Reference
formula
Methoxy-free
C, formula
Spruce
Beech
OR
(6) R= H or alkyl (z,
San0
124 and Sakakibara
aldehyde function was elucidated by Adler et al. [40]. From this color reaction, Adler et
al. [42] estimated coniferyl aldehyde groups to be 2-2.5% in spruce lignin and 3-4% in
BL of western hemlock. Later, however, the A&,, method indicated that 3-4% [42] and
3% [45] cinnamyl alcohol groups are present in spruce MWL.
The existence of a small amount of coniferyl alcohol groups i n the lignin molecules
was demonstrated with a color reaction exploited by Lindgren and Mikawa [44]. Coniferyl
alcohol (1)and its 4-0-methyl ethers (8) react nitrosodimethylnitrile after tosylation, giving
the p-dimethylaminoanilide of styrylglyoxalnitrile (y), which isredin color (maximum
475 nm), via intermediate (Fig. 11). Coniferyl alcohol groups in spruce lignin were esti-
mated to have the same content as cinnanlyaldehyde moieties (2%) by this color reaction.
The presence of cinnamic acid-type side chains in wood lignins is negligible, even
to a lesser extent than in glass lignins [45,46].
Glycerolsidechains in lignin molecules are still beingdebated,buttwo p-0-4
dilignols with glycerol side chainswere isolated by hydrolysis with dioxaneandwater
[47,48], extraction with water [49], and also by treatment with metallic sodium in liquid
ammonia [50].A model experimentsubstantiated the fact that the glycerol side chain could
not beformedfroma p-0-4 structure by hydrolysisunder neutral or acidic conditions
[51]. Higuchi et al. [52] detected small amounts (0.03-0.6%) of three arylglycerols in the
enzymatic dehydrogenation mixture of p-hydroxycinnamyl alcohols. This finding supports
the contention that glycerol side chains can exist in the native lignin macromolecules, but
the presence in lignins is rather insignificant because of formation of only small of for-
maldehyde by periodate oxidation 153,541.
Dilignolsbearingo-propanolsidechains (-CH,-CH,CH,OH) (m) havebeen
isolated from the dioxane-water hydrolysate of protolignin [55]. The aglycones of gly-
cosides (11)and (g) that were isolated as extractives from scotch pine (Pinus .syhwrris)
were optically active [56]. On the other hand, the hydrolysis product (m) is not optically
active, suggesting that it may originate from lignin.
The NMR spectra of MWLs show unknown proton signals in the higher-field region
(6 0.38-1.58), and Ludwiget al. [IS] estimated the content of correspondingproton
content to be 0.2-O.4/C,C3. The protons in this range may correspond to those of y-methyl
and methylene. Compound (11) may represent one of these side chains. Presumably, some
disproportionation may occur, or a reducing mechanism similar to that in the case of wood
extractives may be operative during biogenesis.
R7 O O M e
Chemistry of Lignin 125
OMe
OMe
bo"
OMe
FH20H
I
OMe
OMe
aniline
0 Ce,pAmethylanin&lidc of
styrylglyoxylnimlc ,red (mm.4751x11)
known. Adler and Marton [59] determined carbonyl groups spectrophotometrically through
the reduction of various aldehydes and ketones in guaiacylpropane structures with sodium
borohydride in alkaline solution. The Acr changes in the adsorption spectra were treated
by means of borohydride reduction curves, where the carbonyl compounds are reduced to
the corresponding alcohols with different rates. Then, the A&,. curves of methylated and
unmethylated spruce MWLs were qualitatively and quantitatively analyzed by Comparison
with the A&, curves of modelcompounds. The results allowed the values for various
carbonyl contents per OCH, in MWL to be estimated as shown in Fig. 12. In addition to
these carbonyl groups (in total, 0.09-0.1 I/OCH,), unconjugated carbonyls may exist by
as much as the same amount. The types of the unconjugated carbonyl groups in the lignin
molecule have not been elucidated completely as yet, but some of this is due to glycer-
aldehyde-P-aryl ether substructure which the side chains have been eliminated. Examples
of and As,, curves are shown in Figs. 13 and14.
C.!'&
OH
I
OMe
/o
OMe
n
' OMe
OH
m
&/o
N
' OMe carbonyl
V
unconjugated
& x10 - 3
NaBH4
0.01N NaOH
I I
0
(b) 250 300 350 my
by spectrophotometry (Fig. IS). From these results, it was estimated that 30% of all guaia-
cy1 units have freephenolichydroxyls.Thestructural units with afree phenolic group
react with I-nitrosonaphthol to give compounds with a maximum absorption at SOS nm.
Okay [67] has determined free phenolic hydroxyl groups in lignins by the application of
this reaction. However, the values obtained by these methods do n o t always agree with
each other, because they are not equally effective with all types of phenolic hydroxyls.
For instancc, the AE, method cannot be used for determining hindered phenolic hydroxyls.
It has been claimed that the potentiometric titration with sodium colamine in ethylenedi-
amine could be used todetermine all freephenolicgroups (621. but i n fact, the values
obtained are somewhat larger (0.33-0.34/OCH, in spruce MWL) than they should be
because of the cleavage of benzyl aryl ethers. Furthermore, the oxidation with periodate
Chemistry of Lignin 127
4,O
I UV absorption (neutral)
log E.
3.0
2.0
1 .o
[65] is applicable only to guaiacyl lignins (0.30/OCH3 in spruce MWL). The 'H NMR
approach is simple and can be used to determine the protons of phenolic acetoxyl groups
in acetylated lignins, but the accuracy is not satisfactory (0.27/OCH3 [ 191 and 0.29/OCH3
[ 181 in spruce MWL). Furthermore, Li and Lundquist [27] estimated by the NMR spec-
trometry of lignins at 400 and 500 MHz using DMSO-d, that phenolic groups in spruce
MWL are 0.24/Cy, of which about 20% originate from phenolic groups in biphenyl and
diary1 ether substructures, and those in birch MWL are 0. I WCy, in which the proportion
in guaiacyl and syringyl units is 3:2.
Robert and Brunow [68] have estimated the phenolic hydroxyl groups in MWL by
"C NMR. Chang et al. [69] have determined the phenolic hydroxyls in cellulolytic enzyme
lignins fromsweetgumandspruceandobtainedsomewhatlowervaluesthanthose in
.Q
OH
OMe
-Q
N~OJ
0
0
+ CH30H
u1) W
nces Method
Phenol-OH/OCH,
E. Condensed Units
Potassium nitrosodisulphonate, known as Fremy's salt, oxidizes p-substituted phenols to
o-quinones. For example, 4-propylguaiacol (e)
is oxidized to methoxy-5-propyl-o-qui-
none (16)
- (Fig. 16) 1741. The guaiacyl units (B)possessing an unsubstituted 5-position
are called "uncondensed," and the units that carry C-C or ether bonds at this position
are called "condensed." Adler and Lundquist applied this oxidation to estimate uncon-
densed units in lignin. The o-quinone (16)formed can be quantitatively determined by
means of spectrophotometry. Hereby, it was found that 0.15-0.1 8 units per methoxyl in
MWL were uncondensed phenolic units, corresponding to 50-60% of the 0.30 phenolic
TABLE 6 Numbers of Hydroxyl Groups per OCH, Calculated from the Intensities of Acetyl
Signals in Acetylated Samples and Those Obtained by Acetylation and Aminolysis
Total
Phenolic
Secondary
Primary
e OH Sample OH OH OH
c-c-c c-c-c
OMe
OH 0
ui, 0
FIGURE 16 Determination of condensed units by oxidation of guaiacyl nucleus with Fremy’s Salt.
units present; i.e., 0.12-0.15 per methoxyl were presumed to be condensed units from the
difference, corresponding to 40-50%. However, as the oxidation with Fremy’s salt can
onlybeapplied to the units witha free phenolichydroxyl, no information is obtained
about etherified units. By means of ‘H NMR spectroscopy, about 45-50% condensed units
[18,21] and less amounts of them (35% and 30% for spruce and birch MWLs) [22] have
been estimated. The latter were calculated by combination of the amounts of OCH, de-
termined by the bromine method and of aromatic protons by ‘H NMR of hydrogenated
MWLs using p-nitrobenzaldehyde as an internal standard.
F. p-5 LinkedUnits
P-5 linkage units are represented by the phenylcoumaran structure (B).Dehydrodiconi-
feryl alcohol (B) was isolated by the 0.5% HCl-methanol treatment of spruce wood by
Freudenberg et al. [75], and phenylcoumaran lignols were isolated through dioxane-water
hydrolysis [76,135].
Adler et al. [77] found that dihydrodehydrodiconiferyl alcohol (g) and its phenyl
methyl ether are converted to the corresponding phenylcoumarone (B)in 90% yield after
20 h of heating in 0.2 M HCI (Fig. 17). The phenylcoumarone with its stilbenoid conju-
gationshowsastrongabsorption at 310 nm,and the phenoliccompoundgivesa Asi
maximum at 338 nm. This conversion provides characteristic difference curves. Quanti-
tative evaluation of the spectra indicated that the spruce MWL contained 0.1 1 dimeric
structures per methoxyl that can be converted into phenylcoumarone moieties on acido-
lysis. When the MWL was methylated with diazomethane, the number of phenylcouma-
rone moieties formed on acidolysis was about 0.08. This means that the MWL contains
about 0.03 P-5-condensedbut not ring-closeddimeric units (E), which,however,are
converted into phenylcoumarone moieties closure on acidolysis. Therefore, the number of
phenylcoumaran moieties (1- 8) in the MWL is calculated to be 0.08 per methoxyl.
CH2CH2CH20H
OMe
OMe ‘acidolysif
H20.2Oh.rdlux)
*
(O.2N Ha in dioxanc-
8-
c-
OMe
0
I
OMe
OH OH
0 uz)
c-c-c c-c-c
OMe I OMe
$L0
OH
CHOR
@OM.
OH
OH
phenolic hydroxyls ( 2 0 ) have been determined with the quinone monochloroimide (2)
color reaction by Gierer [79,80] (Fig. 18). Adler et al. [81,82] found that p-alkoxybenzyl
alcohol (3) is oxidized to the corresponding aryl ketone ( 2 5 ) withdichlorodicyan-p-
quinone (2), which is reduced to (B)
(Fig. 19).The a-ketone (g) formed was determined
spectrophotometrically, and a value of 0.16/OCH3 was obtained. Thus, p-hydroxybenzyl
alcohol units constitute 0.05/OCH3, esterified benzyl alcohol 0.10/OCH3, and benzyl ether
(except cyclic ether) 0.06/OCH3. Higuchi et al. [83] estimated a-aryl ethers to be 0.07-
0.09/C,C3 from the results of the acidolysis of MWLs of bamboo, beech, and Thuja stun-
dishii. Freudenberg et al. [84] estimated these groups by cleaving the benzyl ethers with
OMe OMe
OH 0
quimemono- indophenol
chlorimidc A mal. 6x)nm
NC Cl
OMe OMe
0 OH
OMe OMe
0 W (21)
dichlomdicyan-pquinonc
sodium colamine in ethylene diamine solution. Besides this reaction, cleavage with meth-
anolic hydrochloric acid after methylation and sulfonation were also applied [85].From
the results, it was estimated that p-hydroxybenzyl aryl ethers amounted to about 0.041
OCH, and p-alkoxybenzylaryl ethers 0.06-0.09/OCH3. These results are summarized in
Table 7. The benzyl alcohol units have also been determined semiquantitatively from NMR
spectra as 0.33/OCH, [ 181, 0.32 [19], 0.31 [68], or 0.29 [22] in spruce MWL and 0.45
[22] in birch MWL. Gagnaire and Robert [32] have estimated the benzyl alcohol content
to be 0.31/OCH3 from the "C NMR of DHP.
H. p-0-4 LinkedUnits
Arylglycerol-p-aryl ether units (g) and (g) belong to the most important substructures
in lignin molecules. It is well known that Hibbert's ketones are formed from these linkages
in lignin during alcoholysis (Fig. 20). Adler et al. [86] found that acidolysis liberated about
0.3 phenolicOWOCH, in MWL andreportedasimilarnumberofterminalC-methyl
groups characteristic of the side chains of Hibbert's ketones (31-34). The results indicated
that p-0-4 units (2-2) may be 25-30% of all phenylpropanes. From 'H NMR spectra,
Lundquist estimated later that 40-50% of birch lignin units [24] and 30-50% of spruce
lignin units [25] are attached to an adjacent unit by a p-0-4 linkage. Miksche et al. have
studied the oxidation products of lignins and have estimated the content of p-0-4 units
in birchandspruceMWLstobe0.62 [87] and 0.49-O.51/C,C3 [88], respectively, by
multistage oxidations.
TABLE 7 BenzylAlcoholandBenzylEtherContents in
Spruce Lignin
z Hydroxyl
~~
cH20pb
I
H?--0
HC-0-C
I
HC HC2 0-p oO
I
HC-OH
b
CH2OI
C
I
CH
?leo
II
- O b
OMe OMe
OH
0 0 0
H,Oo OMe
0
' 0.30/C&
OH OH
1. p-l LinkedUnits
Nimz [S91 first isolated diarylpropanediols (g) from the degradation products of spruce
and beech protolignins by mild hydrolysis. The mechanism of formation of the p-1 linked
units (g) has been proposed by Lundquist and Miksche [90] as shown in Fig. 2 1. Phenoxy
radicals (2)and (x)
maycoupletogive the cyclohexadienone (E),which maybe
cleaved, giving the p- 1 linked compound (B) and glyceraldehyde-2-aryl ether (S).Lund-
quist and Miksche estimated the content of this aldehyde (0.3%) from the yield of meth-
ylglyoxal during acidolysis of lignin.
F d O H
R18,.""'
J. BiphenylStructures
Aulin-Erdtman [91] found that the mostobvious effect ofincreasing the pH of lignin
solutions was a higherabsorption in the UV spectrumabove 300 nm.Thisabsorption
band is characteristic of biphenyl structures. The number of hydroxy-biphenyl units in
Chemistry of Lignin 133
CH2OH
I
H?-OAr
H?
CH20H
I
*
CH20H
I
HC-OAr
I
HrH I
~ H O HOMe
%,H'
-
0
- OMe
+ 0-CHOH
0
OMe
CH
Q 0O M e
up,
- H20
Hys;H
KMe
I
CH0
I
+ HC-OAr
I
CH20H
OH
ue, 0
black spruce BL was estimated using a twice-reduced difference curve at 325-340 nm.
The results obtainedindicated that there are 0.05/OCH3 biphenyl units in BL. Pew [7]
estimated a still higher value, 0.25/C6C,, for spruce MWL. Miksche et al. estimated values
of 0.045 for birch lignin [87] and 0.095-0.11/C,C3 for spruce lignin [88] from oxidation
product yields. Nimz [38] estimated 0.O23/C,C3 for beech lignin. Between these estimated
values there are considerable differences. From the permanganate oxidation products [92],
in additionto the mainbiphenyl unit, 5-5 (e),
5-6 (g), 5-1 (Q), and 6-6 (g) type
biphenyl units are presumed to be present, but have not been isolated from the products
of hydrogenolysis and hydrolysis, indicating that they may give rise to minor products.
COOH
Me0
OMe
OMe
OMe
0-5 linkage (m) from the hydrogenolysis products of larch compression wood lignin.
Miksche et al. estimated the frequency of 4-0-5 linked units from the yield of perman-
ganate oxidation products to be 0.035-0.04 for spruce lignin [88] and 0.065/C6C3for birch
lignin [87]. On the other hand, Nimz et al. [95] reported 0.O15/C,C3 for beech lignin.
L. p-p LinkedUnits
p-p type structures are involved in lignans represented by pinoresinol (44) and syringa-
resinol (e). Of unitslinked in this way, (@ wasfound to be formed by enzymatic
dehydrogenation of coniferyl alcohol (L) [97]. However, with the exception of the finding
that pinoresinol is detected in the products formed by room-temperature methanolysis from
spruce protolignin [75], isolation of that compound from lignin degradation products has
failed. Syringaresinol,on the otherhand,was isolated from the productsof the mild
hydrolysis of beech wood by Nimz [98]. Omori et al. [99] also isolated syringaresinol and
episyringaresinol from Fraxinus mandshurica by hydrolysis with dioxane and water. Fur-
ther, a dioxa-bicyclo-octane composed of guaiacyl and syringyl units (S) [48], dimethox-
ylariciresinal [ 1001, and a trilignol involving the syringaresinol moiety [99] were isolated,
as will be described later. Nimz [95] also isolated compounds with the isolariciresinol ring
andwith the tetrahydrofuran ring substituted byguaiacyland syringyl units (m). In
general, p-p linked units are involved much more in hardwood lignin than softwood lignin.
Lundquist [25] estimated a low content for pinoresinol units (0.02-0.03/OCH3) in spruce
MWL from its 'H NMR spectrum. Miksche et al. also estimated 0.03-0.05/C,C3 p-@units
for birch lignin [87] and 0.O2/C,C3 forspruce lignin [88].OgiyamaandKondo [ 1011
estimated the content of pinoresinol structures as 0.05-0.10/OCH3 for softwood lignin
from the yield of the di-y-lactone formed by nitric acid oxidation.
R2
HO$..( OMe
H0
W. DEGRADATION
A. Oxidation
1. Alkaline Nitrobenzene
Freudenberg [l041 was the first to report that lignin provides a high yield of vanillin (E)
by the alkalinenitrobenzeneoxidation. The yield of 20-28% (E) fromspruce lignin
proved the aromatic nature of lignin. Later, Leopold [ 1051 studied these oxidation products
from spruce wood in detail, also deducing the presence of p-hydroxyphenyl units in soft-
wood lignin (Table 8). Leopold [ 1061 and Pew [ 1071 demonstrated that the side-chain
structure has considerable effect on the yields of oxidationproductsfrommodelcom-
pounds. Units having side chains substituted with a hydroxyl at the a-position and vinyl-
type guaiacyl units give high yields of vanillin, and those with a-carbonyl groups increase
the yield of vanillic acid (S).Units bearing alkyl substituents atthe 0-position to phenolic
hydroxyl exhibit the highest resistance to oxidation. Substituents at the a-carbon of the
side chains and cyclic ether structures such as pinoresinol are also difficult to oxidize. The
products obtained by alkaline nitrobenzene oxidation are summarized in Fig. 23. Brink et
al. [108,109] investigated the oxidation products (methylated) of white firin detail and
reported various compounds besides the products.
MeO
Me0
OH OH
MeO OMe
OH OH 0 OH
m W dioxepin
TABLE 8 NitrobenzeneOxidationProducts
from Spruce Wood
Vanillin ( 5 4 ) 27.5
p-Hydroxybenzaldehyde (S) 0.25
Syringaldehyde ( g ) 0.06
Dehydrodivanillin ( g ) 0.80
Vanillic acid (g) 4.8
acid Syringic (a) 0.02
5-Formylvanillic
acid (63) 0.1
5-Carboxbvanillin (9) 1.2
Dehydrodivanillic
acid (E) 0.03
Acetoguaiacone (g) 0.05
2. PermanganateOxidation
Freudenderg et al. [92,93] heated spruce lignin or spruce wood with 70% aqueous potas-
sium hydroxide in order to bring about hydrolytic cleavage of ether linkages and subse-
quently protected the phenolic groups liberated by methylation. Permanganate oxidation
of the methylated products at pH 6-7 gave veratric acid ( 7 4 ) in a yield of about 8% of
the lignin and minor amounts of isohemipinic and dehydrodiveratric acids (g,%). Fur-
thermore, they isolated and identified 19 methoxy-substitutedbenzenecarboxylic acids.
CH0 COOH 71 F2
R1 R2 M e O m O M e
OH
Rl=H, R2=OMe
(U) R,=RZ=OMe
(12, Rl=Rz=H
0 Rl=CHO R2=OMe
I
0 R]=COOH, R2=OMe
W Rl=CHzOH. RZ=OMe
R
l
OH
R=H
(hl) R=CHzOH
161)R=COCH3
c66)R=CHzCH2CH20H
(42) R=CH2COCH3
Miksche and co-workers [ 1 10,ll l] found that considerably higher yields of the aromatic
carboxylic acids were obtained if the oxidation was carried out by a mixture of sodium
periodate and permanganate in aqueous t-butanol with sodium hydroxide at 82°C. Since
the product mixture contained appreciable amounts of phenylglyoxylic acids, however, the
latter acids were degraded to the corresponding benzoic acids in a secondary oxidation
step, consisting of brief treatment with alkaline hydrogen peroxide. The latter method gave
well-reproducible results as shown in Table 9.
If wood or isolated lignin were methylated and oxidized, the resulting aromatic acids
reflected the units in lignin which carried a free phenolic hydroxyl group. Preheating with
alkali convertednonphenolic units intophenolic. This is, andfrom the increase in the
yields of aromatic acids, therefore, the proportion of etherified units could be estimated
[94]. Alkaline cleavage of the ether linkages was performed under conditions for kraft
cooking or oxidation with alkaline cupric oxide. The mixture of benzoic acids was finally
methylated, and the resulting mixture of methyl esters was assigned by gas chromatog-
raphy and a combination of gas chromatography and mass spectrometry for quantitative
estimation and structural identification, respectively.
Table 9 shows the yields of the major acids obtained on degradation of a methylated
spruce MWL and of the same lignin which had been subjected to ether cleavage prior to
methylation [ 1 1 l]. The results obtained with methylated kraft lignin prepared from spruce
wood meal is also included in Table 9, indicating the good reproducibility of the method,
and spruce MWL is structurally very similar to the lignin in the wood. Also, pretreatment
with NaOH/CuO gave considerably higher yields of most of the aromatic acids than pre-
treatment under kraft cooking conditions.
A total of 40 aromatic acids has been identified in the reaction mixture obtained on
oxidation of methylatedspruceMWLwhich had notbeensubjected to ether cleavage
[ 112,1131. From spruce lignin, hemipinic acid (z),biphenyls (4J-g,@-@), diphenyl
ethers (87-90), andbenzenepolycarboxylic acids weredetectedbesides veratric (B),
isohemipinic (E), and metahemipinic acids (z),
as summarized in Fig. 24. More abundant
aromatic acids were obtained from softwoods were (z,2), (T),(g),and (@-g), and
MWL methylated
2.0 0.75 0.70.25 11.2 1.1 I S5
RH
COOH
R1 OMe R2 RQ C O O H
OMe
(HL) R=H
0 R=OMe
0 R=COOH
COOH COOH
I
Me0 Me0 Q
R=H
@&)R=OMe OMe
OMe OMe R OMe
@ &l
Rl=Rz=H COOH OMe OMe
(81) RI-+Me ,Rz=H
(8e) R=H
(86) RI=RflMe
@Q) R=OMe
Me0
OMe
ceu
COOH
COOHA
koM
I I
COOH COOH COOH
Me0 V O OMe
M eMe0 ‘’ ’ Me0
/ OMe
OMe OMe OMe OMe OMe
0 W W
COOH
COOH COOH
Me0
Me0
OMe 0
OMe
0 OMe 0 OMe
The methoxyhydroquinone ring in these acids may be have been formed by reduction of
a methoxy-p-quinone moiety which seems to be the hydrolysis product of a 2,4-cyclo-
hexadienone diary1 ketal structure such as (3) being formed by dehydrogenative coupling
[ 1131. It points to the presence in lignin of a methoxyhydroquinone.
B. Hydrogenolysis
Previously, the hydrogenolysis of lignin wasstudiedtoproducechemicalsandalsoto
obtain structural information. Recently, protolignins have been subjected again to catalytic
YHOH
YHOH
R
OH R=OMe
Cez, R I = R 2 = b = H or OH R=H
R3=H or OMe OH
0 OH
CH2
CH20H
Me0Q O M e OMe
OH
0 MeO’ OMe (ep) R=H or OMe
OCH2CH2OH MeOQ O M e
0 OH
umz,
CH2R
I YH2OH
y 2
HC-
H&”O“CH
I
OMe
MeO OMe
OH OH
OMe MeO OMe
cuL1)R=H or OH
OH OH W
UM) R=H or OH
FIGURE 25 Products obtained by mild catalytic hydrogenolysis of protolignins with copper chro-
mium oxide at about 240°C.
San0
140 and Sakakibara
Me0 R3
OH OH
uQ6)R,=R2=H or OH
R3=H or OMe
OMe
Q O M e
OH
R=H or CHIOH
m6
oc'o"- y 2 HOH27 OC'O, ?H2
H?- 7H2 H?-
YH
CH2 CH2
uu
OH W
FIGURE 25 Continued
hydrogenolysis under conditions at about 240°C with copper chromium oxide, leading to
the isolation of various dimeric and trimeric compounds (94-111) besides a substantial
amount of monomers [102,103,114-117,1211. The lignols isolated are summarized in Fig.
25. Hydrogenolysis cleaves most of aryl-alkyl ethers, but a few of these linkages remain
intact as seen in compounds (95- R, = OH, - 96). Compound (96)
- has an a-hydroxyl that
Chemistry of Lignin 141
R
I
CH CH20H
II I
CH l
HOH27 Q O M e
HOH2C
HC-0
I I
R
$$,
CHOH
OMe
I
HC- 0
HC-0 CHOH OH OH
R=H or OMe 0 Rl=CHCHCHzOH
Rz=H
G OH
O M e R$ OMe u1z) R,=CHCHCHO
R2=OMe
0 R,=CHzCH?CHzOH
Rz=OMe
FIGURE 26 Products obtained by mild hydrolysis of protolignins with aqueous dioxane at 180°C.
San0
142 and Sakakibara
has quite exceptionally remained unaffected. Compounds (m) [ 102,1141 and (m) [ 1031
should give metahemipinic acid (2) by permanganate oxidation of methylated lignins.
Compound (m) was isolated from the lignin of larch compression wood, but its occur-
rence in normalwood lignin isalsoprobable. Itcanbeconsidered that compounds
(101,102) could be derived by reductive cleavage from pinoresinol (e) and (E), re-
spectively. However, a model experiment indicated that the alkyl-alkyl ether of the tetra-
hydrofuran ring is very stable toward hydrogenolysis, and the starting material (g) was
almost completely recovered [ 1151. This fact suggests that compounds (E) may not be
derived from compound (e) after all, and these linkagepatterns may exist independently
in lignin molecules. Compound (g) was isolated from the hydrogenolysis products of
hardwood protolignin [ l 161, indicating that the alkyl-alkyl ether is fairly stable against the
reductive cleavage. Compound (96)contains a 7-0-4 linkage [ 1171 so far not known. It
is, however, very probable that a p-0-4 linkage can be enzymatically rearranged to a y-
0 - 4 linkage. The existence of 7-0-4 linked units waspostulatedatan early stage by
Freudenbeng [ 1191 in 1933 without experimental support, but later he abandoned the idea.
Subsequently, in 1960, Brauns et al. [l201 proposed the same substructure. However, this
linkage pattern may be a minor one in the lignin macromolecule. Compound (F), which
was isolated from hardwood protolignin [ 1211, has a heterocyclein the molecule involving
a 4,5-dihydroxy-3-methoxyphenylmoiety.
Compounds ( I 10,111) are trilignols with a y-lactone. They have no optical activity.
The IR spectra of MWLs show a small shoulder at 1760 cm”, indicating the existence
of y-lactones. The facts indicate that p-hydroxy cinnamic acids are also involved in the
radical coupling scheme after enzymatic dehydrogenation. Nimz [95] cleaved protolignin
from beech wood with thioacetic acid using boron trifluoride as a catalyst followed by
reductionwithRaney nickel toproducemanydimericcompounds.Theproductswere
somewhat different from those formed by hydrogenolysis.
C. MildHydrolysis
1. Hydrolysis withWater
Nimz [125-1291 percolated extractive-free wood powder with water at 100°C for several
weeks (“mild hydrolysis”), showing that beech wood loses about 40% of lignin, while
only 20% of lignin in spruce wood goes into solution. From hydrolysis products of spruce
were isolated eight dilignols, two diastereoisomeric trilignols, and one tetralignol: guaia-
cylglycerol-p-coniferyl ether (e) [ 1261, a trilignol involving p-0-4 and p-1 linkages
(123)[127], a tetralignol involving one 6-1 and two p-0-4 linkages (125)[ 1271, guaia-
(e)
cylglycerol-p-guaiacylglycerol ether [ 1261, and guaiacylglycerol-p-coniferylalde-
hyde (M)[ 1281, and from beech were isolated syringaresinol (g) [98] and three diar-
ylpropanediols (115)[89,123].
OH
OH
M e o ~ o M e
MeoQoMeI
H2?/"CH +
Me0QOMe
OH
OH
0
OH
OH
M e O A O M e M e o ~ o M e
CH20H
I
'CH H2qA0"CH II
I CH20H HC-
I CH
l
6
HC- CH
I I
$/C" I
OMe
J +H20
H o H q G O H
HOHC
I
R Q OH
O M e
OMe
+
Meo6
H2?C
,"H
HC-
I
I
CH
I
O+C-~-CH2
of monolignols. A compound (M)with an w-propanol side chain supports that the ex-
istence of such reduced sidechains in the lignin structure is probable,considering the
highly shielded signals in their NMR spectra, as discussed before 18,191. The formation
of cornpound (E), which has eliminated an aromatic ring, may be explained i n two
ways: ( a ) ring closure during hydrolysis or (b) displacement of the diarylpropanediol dur-
144 Sakakibara and Sano
OH OH
OMe
OMe
(m
FIGURE 28 DegradationofcY,P-diarylethers (3 and g)
by “mild hydrolysis”using 50%
aqueous dioxane at 180°C for 20 min. (From Ref. 139.)
ing coupling ofradicals (Fig. 28) formed eitherin wood in situ or secondarly by homolysis
of phenolic P-ethers in lignin, as shown in Fig. 29 [137,141].
As described before, the degradation products from hydrogenolysis and hydrolysis
provide much important information about lignin structure and supportthe theory of lignin
formation by enzymatic dehydrogenation of cinnamyl alcohols. Hydrolysis mainly cleaves
the a-ethers of side chains in the lignin molecule in spite of poor model experiments, but
nonphenolic P-ether cleavage occurs hardly at all [58,139]. The model experiments with
p-hydroxyarylglycerol-a$-diary1 and -&aryl ethers [ 138- 1411 have indicated that homo-
lysis occurs to a slight extent during hydrolysis, resulting in subsequent coupling of the
radicals formed. But Sakakibara [l651 has stated that the formation of artificial products
from protolignin by “mild hydrolysis” is negligible because of the following fact: De-
hydrodiconiferyl alcohol (M)and pinoresinol (e) and dehydrodiguaiacylpropane (106,
R,=OCH,, R,=R,=H) that were found in the model experiment [138-1401 could not
bedetectedeven in traces in the hydrolysatefromspruceprotolignins in spite of the
proposed mechanism of compound (m) shown in Fig. 27. The homolytic degradation of
phenolic p-0-4 linkages will be summarized in the following section because of getting
many informations on homolytic cleavage of phenolic P-ethers under the conditions similar
to those for “mild hydrolysis” and “catalytic hydrogenolysis” of protolignins.
3. Homolysis
With the aim of finding a method which would degrade lignin without involving simul-
taneous condensation reactions, Nimz [ 125- 1291 and Sakakibara [ 130- 1391 subjected
extractive-freesoftwoodandhardwoodto“mildhydrolysis”under neutral or slightly
acidic conditions followed by a percolation with hot water and a mild hydrolysis with
Chemistry of Lignin 145
CH20H CH20H
l H A - 0 9 I
HC-
+ QOMe
0.
OMe Rb
OH 0
Ra
uz&, (24.0%)
CH20H
l
CH
II
m+RaorRb
OMe
0- OH
RC (1) (2.6%)
OH
.OMe CHsCH-CH20H
CKOH I
2xRa +
CH20H
H { - 0 9 FH20H
j_\
OMe
Ra+Rc - OMe
+
H$-0
Q O M e
Q O M e 0
' w(3.656)
Ra + Rb + RC - polymerichydrolysisprodUCts of
aqueous dioxane at 1 80”C,respectively. Separately, they have isolated and identified many
monolignol-to-trilignol hydrolysis products in small amounts, of which most compounds
are identical to each other. The hydrolysis products due to cleavage of the benzyl aryl
ether linkages have been thought to give valuable information on the end groups in lignin
and the biosynthesis process of lignin, since it is probable that in the “mild hydrolysis”
the hydrolytic cleavage of benzyl aryl ether linkages in lignin is the main reaction in spite
of deficient experimental evidence.
To obtain definitive information for the reaction mechanism of lignin under the con-
ditions of “mild hydrolysis,” some lignin model compounds were subjected to mild hy-
drolysis. The hydrolysisofmodelcompounds(126,127)forphenolicandnonphenolic
a$-diary1 ethers were carried out in 50% aqueous dioxane at 125-180°C for 20- 120 min
[138-1411. The phenolic a-aryl ether bond in (126)was cleavaged by 41% and 83% at
140°C for 20 min and 120 min, and by 100% at 180°C for 20 min, and the nonphenolic
a-aryl ether bond in (m) by only 39% at 180°C for 20 min [139]. The latter was more
resistant to mildhydrolysisthan the former. The nonphenolicP-ethercompoundwas
recovered quantitatively from the reaction mixture of (g). New dimeric and trimeric
compounds were obtained from the reaction mixture of (3) [138].
Guaiacylglycerol-P-guaiacyl ether (m) was subjected to “mild hydrolysis” by two
procedures, that is, 50% aqueous dioxane (pH 3.54) at 180°C for 20 min, and also water
(pH 3.54) at 110°C for 48 h in place of the percolation procedures. Thin-layer chromat-
ograms of reaction products obtained from the former were completelyidentical with those
from the latter, indicating that the cleavage of the P-ether according to the two procedures
proceeds by the same mechanism [140].
From the reaction mixture obtained by the former procedure of (m), the starting
material(24.0%),coniferylalcohol (e)
(1)(2.6%),pinoresinol (0.8%), 1,2-diguaiacyl-
1,3-propanediol (E) (1.4%), dehydrodiconiferyl alcohol (m) (4.5%), and two trimeric
compounds (1 29,130) were isolated and identified besides substantial amounts of unknown
polymerizedmaterials [l401 as illustrated in Fig. 29. The compound (B) (3.8%) was
composed of phenylcoumaran and P-ether moieties, whereas (m) (3.6%) had two P-aryl
ether links. Gel filtration of the reaction mixtureshowed that their molecularweights
increase with increasing reaction time, demonstrating that the degradation and polymeri-
zation of lignins take place simultaneously by “mild hydrolysis.”
The polymerizationtookplaceduring the heat-up time up to 18O”C, and became
predominent for 120 min to form more stabilized polylignols in large quantity. Both the
relative absorbance of the “mild hydrolysis” products at A,,, about 280 nm for UV (neu-
tral), and at A,,, about 300 nm for ionization differential spectrum ( A s i ) increase with
increasing the reaction time, respectively, when measured immediately after the hydrolysis.
In 348 h after the hydrolysis both of them decrease strikingly, and the absorbance at A,,,,,,
about 255 nm for UV (neutral) increased reversely compared to every one when measured
immediately. The UV spectrum for the reaction products, which was measured 348 h after
“mild hydrolysis” of (m) for 120 min, were very similar to that in spruce MWL except
for slightly high absorbance at A 360 nm for A&,, due to phenolic conjugated carbonyl
groups. The results obtained by UV analyses show that more reactive sites such as qui-
nonemethides are present among the “mild hydrolysis” products of the model compounds,
MWL, and wood-in-situ lignin [140].
When 1-guaiacyl-2-guaiacoxy- 1 -propene-3-01 (131) was subjected to mild hydrolysis
with 50% aqueous dioxane at 180°C for 20 min, Hibbert’s monomers (3 1-34) were formed
in addition to the starting material (15.6%) (Fig. 30). This implies that phenolic arylgly-
II CHO COOH
CH
I OMe
Ribkrt’s ketones +
(3.l-U I
QOMe
OH (Total about 16%) OH
(L31) (15.6%)
FIGURE 30 Degradation of I-guaiacyl-2-guaiacoxy- 1-propene (131)by “mild hydrolysis” using 50% aqueous di-
oxane (pH 3.54) at 180°C for 20 min [%: yield of the amount of (=)I. (From Ref. 140.)
Sano
148 and Sakakibara
cerol-P-ethers in lignin are never degraded via corresponding enol-ethers under the “mild
hydrolysis” conditions [ 1401.
From the “mild hydrolysis” products of syringylglycerol-P-syringylether (133)by
aqueous dioxane [ 14 l], syringol ( 1 4.5%), syringaldehyde (2)(1.1 YO),sinapyl aldehyde
(134)( I . I %), 1-syringyl-2-syringoxy-3-hydroxypropanone-1 (136)(0.9%), D,L-syringa-
resinol (g) (5.6%), 2-formyl-3-hydroxymethyl- 1,4-bis-syringyl- 1,3-butadiene (3)
(1.3%), and 1,2-syringyl- 1,3-propanediol (m) (7.1 %) were isolated besides the starting
material (33.3%) and polymeric compounds as shown in Fig. 3 1 . Sinapyl alcohol (2) and
syringylglycerol-P-sinapylether (m), which were not detected amongthe products, might
be too thermolabile to exist in the reaction mixture, even if they were formed during the
reaction process. The thermolabilities of (2,137) are apparent from the fact that they are
”
Me0G O MOMe
OH
e
-H*O
0 H ? - 0 9
0
CH
OMe
OMe
-6
Me0
'?H
0
+M
+I[.
e O Q O M x e O Q O M e
0.
Kb
OH
(I4 S%)
OH
Meo60Me
CH20H CH
I
HC-?H
l
"t
M e o ~ o M e
H+CH
7H2OH
CH
II
.CH
2XRa"+ I
CH2OH
CH HC-o-kH,
Me0O 0O M e Me0G O M e
Me0
Q OMe
CHo Me0
OH
Me0
OH
0
U)(7 1%)
FIGURE 31 Continued
ing to the total content of coniferyl alcohol end groups (about 2%) in softwood lignin.
This means that a considerable part of the coniferyl alcohol must originate from sources
other than a-ether end group substructures in lignin. The stability of coniferyl alcohol (1)
toward “mild hydrolysis” in 50% dioxane (pH 7) at 130°C demonstrated that coniferyl
alcohol is degraded or polymerized fairly rapidly even at 130°C. And guaiacylglycerol-P-
guaiacyl ether ( 128), which was degraded by 80% in the same medium after 40 min at
1 8O”C, a f f o r d e d 0 at a 20% mole yield. No Hibbert’s ketones (31-34) could be detected
in the reaction mixtures even when about 50% of ( 128) was degraded in 50% dioxane at
a pH as low as 2.7 and 160”C, indicating that a c i d o 5 i s does not occur by “mild hydro-
lysis.” Based on the above results, Lundquist et al. [ 1441 concluded that the use of “mild
hydrolysis”asa tool in lignin analysis leads to erroneous and confusing results if the
homolytic cleavage of the @-ethers is neglected.
It is well-known that phenolic P-ethers in lignin are prone to undergo reactions via
ionic intermediates under acidic and basic conditions. However, it has been reported re-
cently that the P-etherbond i n syringylglycerol-P-syringylether (133)is subjected to
homolysis under conditions similar to those in a soda cook [ 1451. Furthermore, Sipila et
al.[l461 allowed syringylglycerol-P-syringylether (m) to stand in aqueousdioxane
solutions at pHof 4-7 at roomtemperatureandfoundsyringaresinol (G),P-l (E:
R,=R2=OCH,) and 2,6-dimethoxy-p-quinones, giving further evidence for the existence
of homolytic P-aryl ether cleavage in delignification of hardwoods.
Sano [ 147,1481 subjected extractive-free oak wood meal, guaiacylglycerol-P-guaia-
cy1 ether (m), and its 4-0-methyl ether to “solvolysis” in p-cresol-water ( 1 : 1 ) at 180°C
for 30 min in order to explain the delignification mechanism of wood lignin by solvolysis
pulping with aqueous phenol at elevated temperatures. Fromreaction products of the wood,
six compounds (143-148) were isolated and identified, as the same cresolated compounds
Chemistry of Lignin 151
KMe
OMe +
Q O M e
OH
20 h
OH
OH W (35%)
+ condensed products
(30%)
with guaiacyl and syringyl rings also obtained, respectively, from the reaction products of
guaiacylglycerol-P-guaiacyl ether (128)and sinapyl alcohol (2) treated under the same
conditions. The reaction mechanism by phenolysis has been illustrated asfollows(Fig.
33). The cleavage of phenolic @-ethers in lignin proceeds via homolysis to form two free-
radicals, which are trapped by the phenols as a solvent and/or other active intermediates
to give p-cinnamyl alcohols (1-3) from end groups and new phenol groups in lignin. The
latter are further cleavaged in reaction sequences of the “peeling” type to lead the exten-
sivedepolymerization of lignin.p-Hydroxycinnamylalcohols (1-3) aredehydrated to
extended quinonemethides (E), which are condensed with the phenols by ionic reactions.
The resulting resonance-stabilized phenolated compound (m)are oxidized with the phe-
nol radicals and/or lignin radicals to the corresponding radicals, followed by radical cou-
pling. Interestingly, it may be noted that no syringaresinol (G)was detected among the
solvolysis products of oak wood in spite of the inactive compound toward the solvolysis
reaction [ 147,1481. Lin et al. [ 1501 recently reported that (m) is subjected to homolysis
under conditions for the liquefaction of wood with phenol at elevated temperatures.
Thus, homolytic cleavages of phenolic P-arylethers in wood lignin can occur at
elevated temperatures in many technical processes, for example, high-yield pulping and
steam hydrolysis [ 1491, and in thedegradationproceduresfor structural studies under
San0
152 and Sakakibara
coupling
Radical
coupling
Radical
Products Products
neutral conditions at pH 2-9 and elevated temperatures, for example, mild hydrolysis and
hydrogenolysis to yield p-hydroxycinnamyl alcohols and their radicals, and end phenoxy1
radicals in lignin, followed by formation of condensed polymers and chromophores, which
influence the brightness and brightness stability of wood and high-yield pulps, and orga-
nosolv pulpings.
The mild catalytic hydrogenolysis of protolignins has been studied extensively to
obtain structural information. Sakakibara et al. [99-1171 subjected protolignins to catalytic
hydrogenolysis in 60-90% dioxane containing a catalyst, copper chromium oxide,at 220-
240°C for 1 h. The hydrogenolysis of protolignins appears to proceed as follows: proto-
lignins are depolymerized and solubilized in aqueous dioxane at 220-240°C higher than
those for “mild hydrolysis,” then subjected to catalytic hydrogenation and hydrogenolysis
to form “stabilized” hydrogenolysis products. The reaction mechanism for the solubili-
zation and hydrogenolysis of protolignin needs clarification in details to apply the catalytic
hydrogenolysis as a meaningful tool in the elucidation of lignin structures. On treatment
of 1,l-diphenyl-2-picrylhydrazinewith aqueous dioxane at 180°C for 30 min, the hydrazyl
radical is formed in large amounts, but not at below 140°C [ 1381, indicating that phenolic
hydroxyl groups may be converted to phenoxy radicals at elevated temperatures.
Chemistry of Lignin 153
OH
CH -
II
OMe
’ O O C H 3
0
FIGURE 33 Continued
HCO
l
y 2 H
+ +C,
0
I
Q O M e um
OH OH
W
I
1
CH20H CH3
I
CH3
I
CH3 CH3
6. 6 6 6 6
I I
c=o c=o c=o
l - 1 F=" I
H OH
66
I
OH
uiz, 0
1l
OH
II
OH
0
FIGURE 34 Monomeric products obtained by acidolysis of spruce MWL. (From Ref. 124.)
formed from model compounds (128) were detected, the predominating ketol (31)being
obtained in yields of S-6% of the lignin. In addition, the presence of small amounts of
homovanillin (x) and formaldehyde (E) was demonstrated. The side reaction can be
regarded as areverse Prins reaction of the benzyliumion intermediate. Theseresults
constitute clear evidence of the substructures of the guaiacylglycerol-P-aryl ether type.
The monomer fraction contained small amounts of ketol (m), coniferyl aldehyde Q), and
p-coumaraldehyde (E).
In a similar monomeric fraction obtained from the acidolysis of birch MWL, a num-
ber of syringyl analogs were detected in addition to most of the compounds shown in Fig.
34 [ 1521. The yields of the syringyl monomers were higher than those of the guaiacyl-
monomers, although the ratio of syringyVguaiacyl is about 1 : 1 in birch. This is due to the
fact that some of the guaiacyl units are linked to an adjacent unit by S - S , p-S, and S-0-4
bonds. which cannot occur in syringyl units.
From the dimeric fraction obtainedfromspruce lignin, compounds (45,154- 160)
were isolated and identified as summarized in Fig. 35. With the exception of trace con-
stituents ( l S5,160), the dimeric fraction of birch lignin gave the same guaiacyl compounds,
"
CHpOH CHpOH
I
c=o
I
c=0 ?H ?"
I I
qH2
+HC
OM
II.
Ff
+OMe HC-CH3
I
6 GOMe
H3C, OMe
7-0
OH OMe OH OH OH
QOMe
OH ui8) u.33
OH
0 w UIZ)
(m) (rn) 'CH2
'
HpC'
HC-
I
AH
I
Me0 0 OH
(m
Q O M e
OH
HC-YH
I
HpC-, /CH
O I
(41,
Me0Q O M e
OH
FIGURE 35 Dimeric products obtained by acidolysis of spruce MWL. (From Ref. 124.)
156 Sakakibara and San0
the corresponding syringyl analogs with one or two syringyl nuclei, and, furthermore, the
stereoisomeric compounds D,L-syringaresinol ( S ) and D,L-epi-syringaresinol (E).
The phenylcoumarone (154)and the stilbene (155)originate from a lignin substruc-
ture as shown in Fig. 36. The phenylcoumarone (%), which is formed in much higher
yield than the latter, has a characteristic and very strong UV absorption 11531. This per-
mitted its quantitative estimation, which indicated that about 10% of the C, units in the
spruce lignin are connected to anadjacent unit byan a-0-4 aswellas a p-5 linkage,
giving rise to a phenylcoumaran system. The acidolytic conversion of a hydroxylmethyl-
substituted phenykoumdraninto a methyl-substitutedphenylcoumarone is readily ex-
plained by a sequence of ring opening, allylic rearrangement, and recyclization.
The dimeric compounds (156-159)all exhibit only one side chain per two guaiacyl
residues. It was postulated that these compounds could arise from a 1,2-diguaiacyl-1,3-
propanediol substructure (E)incorporated into lignin by acid-hydrolyzable linkages (Fig.
37). A plausible mechanism for their formation has been presented [ 1541. Compound (E)
and related compoundscarryingone or twosyringyl nuclei were isolated from“mild
hydrolysis” products of softwood and hardwood, respectively, [89,129,133,134], and (3)
was also detected among the low-molecular productsof coniferyl alcohol dehydrogenation.
The biogenesis of the lignin substructure (E)can be visualized as shown in Fig. 21. A
p-l coupling between a coniferyl alcohol radical and the radical of a p-hydroxybenzyl
alcohol end group forms (E)and a glyceraldehyde-2-aryl ether group (c). Experimental
evidence for the presence of glyceraldehyde end groups of type ( 4 0 ) was provided by the
detection of pyruvaldehyde (150,methylglyoxal) in the acidolysis mixtures from spruce
and birch MWL. The mechanism of the acidolytic formation of pyruvaldehyde is presented
in Fig. 38. Colorimetric determination of the aldehyde formed on acidolysis of spruce and
birch MWL, as well as of model compounds (g), indicated that only about 2% of the C,
units of lignin were bound to glyceraldehyde as shown by formula (S). The results of
‘H-NMR studies on the aldehyde groups present in spruce MWL point to a similar value
for the amount of glyceraldehyde groups. This figure, however, appears very low in view
of the fair yields of degradation products of the p-1 type obtained by thioacetolysis of
s?
r
6
2.
a
a
L
acidolysis
Q O M e Y O M e m
OH OH
OH
w (1121 w
1,3-diolunits (2).
FIGURE 37 Acidolysis of 1,2-diarylpropane- (From Ref. 154.)
I
c-c-c
beech wood. It has been proposed by Sarkanen that the unconjugated carbonyl groups in
spruceMWL,whichamount to about 10 carbonylgroupsper 1 0 0 C, units, mightbe
regarded as being present in glyceraldehyde groups. This proposal, however, does notfind
support in the analytical results mentioned above. In Fig. 39, the substructures which have
been disclosed by the acidolysis procedure are summarized. Of these structures, the ar-
ylglycerol-P-aryl ether structure undoubtedly is the most abundant one. As already men-
tioned, C, units which are linked to an adjacent unit by forming a phenylcoumaran system
occur in spruce lignin in an amount of about 10%.D,L-syringaresinol ( g )and its stereo-
isomer D,L-episyringaresinol (m) have also been found among the acidolysis products
from birch. The corresponding guaiacyl compounds (S), however, could not be found in
spruce or birch acidolysis mixtures. If pinoresinol (44) structures are present in lignin in
t;
C, C 0’
c:
C
I
H2COH G O M e MeoQoMe
H,C*O-?H
HC-0
Chemistry of Lignin 159
appreciable amounts, one must assume that they are linked to adjacent units by acid-stable
bonds, i,e, 5-5 and 5-0-4 bonds, to an unexpectedly great extent. The "NMR spectrum
of spruce lignin also indicates a very low content of pinoresinol (e)
structures. However,
experimental evidence in favor of the occurrence of such structures in conifer lignin has
been presented. A further acidolysis product exhibiting coupling, namely, D,L-divanillyl-
tetrahydrofuran (E), was isolated in small amounts from spruce lignin acidolysis. The
compound differs from the other dimeric acidolysis products in possessing a lower degree
of oxidation. Degradation of beech wood by thioacidolysis afforded the syringyl analog.
Its formation seems to involve an oxido-reduction process, but it remains open what sub-
structures (@) in lignins originated from [154].
13 F, Cl I$X)SI I NaOl I
CH2R
CHZOH l
I
HqSo
R' OMe
OH
OH
R=H or OMe
K'=llorOlI
(40.7)
FIGURE 40 Degradation of lignin by thioacetolysis with thioacetic acid. (From Ref. 95.)
160 Sakakibara and San0
R
OH
OMe
&Q'16
R OMe
/ OMe
OH H $ oOH
O M e
R=H or OMe OH OH
R=H or OMe
CH3 CH3
HC
I
Me0 OMeMeO
OMe OMe
Me0 R OMe
OH OH
OH OH
\ R=H or OMe , R'=H2or =O
/
u6i, (0.5%)
CH3
I
CH3
I
C
Me0 OMe
R
OHOH OH
OH
or OH
R=H R=OMe R=OCH3. R=OMe
R=H or OH (0.1%)
c166)(0.45%) (142) (0.4%)
0 (0.3%)
FIGURE41 Dilignols obtained from beech protolignin by thioacetolysis (% of lignin). (From Ref.
95.)
Chemistry of Lignin 161
methods, especially acidolysis and oxidative degradation cited above. The a-p-junction
(165) found in some compounds (0.5%) was assumed by Nimz and Das [95] to be present
i n e e c h lignin, although it cannot be the result of a dehydrogenation. Its formation and
that of some related structures is assumed to be due to a proton-catalyzed polymerization
of coniferyl alcohol (1)or coniferyl alcohol end groups caused by the natural acidity of
the cell sap. However, we must also point to the possibility mentioned by Nimz [ 1801 that
(1)andsinapylalcohol (2) may arise during the degradation process, analogous to the
formation of (L) as an intermediate in kraft cooking of spruce lignin. The heating with
alkali in the last degradation step may cause Michael-type additions of the p-C atom of
coniferyl alcohol to the a-position of quinonemethide structures whichalsocanbe as-
sumed to be intermediates. The a-p-linked products (m) may accordingly be artifacts.
On the basis of the yields of crude and pure degradation products, Nimz has cal-
culated the frequencies of the various bond types in beech lignin and has also proposed a
structural scheme for this lignin (Fig. 46).
Recently,a new acid degradationmethod, thioacidolysis (solvolysis in dioxane-
ethanethiol with boron trifluoride etherate) has been studied by means of a reproducible
and mild routine procedure to obtain detailed structural information about lignin [ 156-
1591. The acid degradation of lignin was composed of two consecutive thioacidolysis and
desulfuration of thioacidolysis products over Raney nickel as illustrated in Fig. 42, which
is similar in principle to that proposed for thioacetolysis (Fig. 40). However, thioacidolysis
is carried out using a few milligrams of sample in dioxane at 100°C for 4 h instead of
thioacetolysis performed at 20°C for 1 week. Reaction conditions for the former seem to
be more advantageous as a tool for structural studies of lignin than those for the latter.
The thioethylated monolignols and the desulfurated dilignols of spruce MWL and wood
are shown in Fig. 38 [1591. Among the thioethylated monomers from spruce MWL, the
compounds (170,171), whichwereformedfromuncondensed p - 0 - 4 linked units, were
obtained in a total yield of 9 3 % of the monolignols. They reflect the higher content of
uncondensed p - 0 - 4 linkages in the MWL products (m) (m) and from coniferyl alcohol
end groups were obtained in a total yield of 6.6% based on the monomers. The compounds
(m) (m), and orginating from coniferyl aldehyde end groups and from dihydroconiferyl
alcoholendgroups, respectively, weredetected in trace amounts.The yields of main
dilignols obtained from spruce MWL and wood are shown in Fig. 43 and characterize the
various types of condensed linkages in softwood lignin. From their yields, it can be con-
cluded that p-S, 5-5, and p-1 linkages are present as major types of condensed interunit
linkages in softwood lignin. In addition, the dimerswithdiphenylether (181)(4-0-S),
phenylisocoumaran (p-5), and tetrahydrofuran (M)(p-p) structures, of which the struc-
tures were assigned only from their fragmentation patterns, were detected in trace amounts,
and their linkages seem to be minor types among condensed linkages. However, the latter
two dilignols have been assigned to be compounds with different structures in two papers
reported by the same authors [ 159.1, indicating that they need to be isolated and identified.
The total area of GC peaks due to the dilignols assigned accounts for more than 90% of
all the peaks corresponding to dilignols in the chromatograms of spruce MWL and spruce
protolignin. The total amount of the dilignols is about 30 mole% of the main thioacidolysis
monomers obtained from both of the lignins. The relative importance of condensed inter-
unit linkagesimplied by the mole ratio of the thioacidolysis dilignols is almostequal
between spruce MWL and sprucewood lignin in situ.
The yields of compounds which were characterized as monolignolsanddilignols
among the thioacidolysis mixture were only 40-S0% of lignin [ 1593, which may reflect
the limitation of thioacidolysis results to the degradable part of lignin.
San0
162 and Sakakibara
H2COH HpCOH
I H2yH H27OH I
HCSEt HCOR2
HCOR2
l
HCORp
- I
6z; 6E":;61
I @,R1
OMe F 6 OR
_I)
OMe9F3
OR
BF3 E(SH
OH OR OH
H2COH H2COH
I
COR2
I It !$p I
I
'OMe
OR OR OR OR OH
H OH
In order to use the interesting thioacidolysis method as a routine procedure for the
characterization of the total structure of lignins, it is necessary that the thioacidolysis
productscontaining trilignols to oligolignols be clearly characterizedandacid-derived
condensation by thioacidolysis clarified.
Me OMe OMe
OH
OH OH OH OH OH
0 0 uz2, 0 uze, W
Me0
OMe
OH
Meor
CH3
H0
R=H.
p5OH
\
/series OMe
OH
p5 series
or CH20H
H O H z C e Me
u&+M=33%)
p3OH
HOH2CfMe
\senes
OMe
p5 senes
u8L,
i Me0
OH
p p series
Rl=CH3 or CH20H
R2=H 01 CH3
FIGURE 43 ThioethylatedmonomersandRaneynickel-desulfurateddimersobtainedbythioac-
idolysis of spruce MWL (% of dimers). (From Ref. 158.)
R2
OMe
F E - 0 - 0
I 8-Q
6" 6 A
B
C
6 C
"8
F' C
I
FC
?
C-
F
F
go
l
I
D E F
F FC
F F cI $ I
C?-
I C
G H
0
G(traces)
F
F FC
6-6 I
11 are to be regarded as approximate rather than accurate, although mostofthem are fairly
reproducible. Some uncertainty is attached to the values (0.02-0.15) given for bond type
C in spruce MWL in spite of differing estimation methods. Sarkanen [ 1721 has stated that
p-1 units are main substructures in endwise lignin for the middlelamellaregion rather
than in bulk lignin for the cell wall. However, Lapierre et al. [l581 have reported that p-
Chemistry of Lignin 165
1 units are almost as frequent as in spruce in-situ lignin and MWL preparations on the
basis of the results obtained by thioacidolysis.
A: a-0-4 (open) 6
B: p - 0 - 4 34-39 22-28 60 65b
p-0-4" 2
c: p-l 7 15
D: p-5 6 6
E: 5-5 4.5 2.3 4.5
F: P-p 3 5
P-B 2
p-P and L Y - ~ ~ 0.5
G: 4-0-5, 4-0-1 1 5.5 6.5 1.5
H: C(a)-2, C(a)-6 1-1.5 0.5- 1 1 .5-2.5
I: a-P 2.5
number of these units has been arbitrarily selected because of the lack of adequate quan-
titative data.
D. Heterogeneity of Protolignin
These structural modelsareonlyaverage pictures for lignin structures that may have
different chemical configurations in the different morphological regions of the cell wall.
Fergus and Goring [ 17 I ] indicated varying amounts ofsyringyl- and guaiacylpropaneunits
in the various cell wall layers and middle lamella of birch wood by means of UV spec-
trophotometry. Matsukura et al. [ 1721 showed by oxidation and alcoholysis of spruce wood
that the lignin polymer has a heterogeneous structure. Furthermore, it has been pointed
out that lignin in the middle lamella region may possess more of the nature of an endwise
Chemistry of Lignin 167
H$?-0“ bH
W O M e
FHOH
FHOH
CH20HI CH20H
7HpOH
HV-
-0
OMe
Me0
MeO
MeO
MeO OMe
OMe
Me0
OMe
Me0
MeO
OMe
-0 0- -0 OH
polymer than that permeating the polysaccharide matrix in the S2 layer [ 1731. Compression
wood contains not only more lignin but also more condensed-type units than normal wood
lignin [21]. The reason may be explained by the fact that the outer S, layer of compression
wood is highly lignified, whereas the middle lamella is not completely lignified [ 1741.
The controversysurroundinglignin-carbohydratecomplexes(LCC)hasbeende-
bated for a long time. In spite of numerous studies, the question of whether the association
between lignin andcarbohydrates is physicalorchemical in naturehas not yetbeen
resolved. However, the evidence for chemicalbondsbetweenthemhasbeengrowing
gradually. For instance, enzymatic hydrolysis of LCC preparations gives concentrated LCC
Chemistry of Lignin 169
fractions that indicate the existence of covalent bonds between lignin and carbohydrates,
because enzymes do not cleave lignin-carbohydrate linkages [ 175- 1801. The structure of
protolignin must take into consideration the presence of LCC linkages. Some typical forms
of LCC linkages have been suggestedby Freudenberg et al., who isolated phenylpropane-
cane sugar compounds from the mixture arising from the simultaneous enzymatic dehy-
drogenation of coniferyl alcohol and cane sugar [ 181, l 821. Finally, a word should be
mentioned about Brauns lignin (BL), which Brauns [ 1831 first isolated from black spruce
by extraction with ethanol in 1939. Since then, BL has been consideredto be natural lignin
and has been used for several basic studies of lignin. However, it is not clear whether BL
is a true lignin or not, as Freudenberg [ 1841 pointed out that BL may be a fraction of
resinous material.
Recently, various new lignans have been isolated [ 1601. They consist of dimeric,
trimeric, and tetrameric phenylpropanes that are very similar to the lignols from degra-
dation product mixtures, as already mentioned, except for optical activity and some other
details. The existence of monomeric, oligomeric, and polymeric phenylpropanes in the
lignan fraction suggests that these constitute a continuous spectrum of lignans. In conclu-
sion, BL can be considered a polymeric fraction of lignans and not a true lignin.
VI. OUTLOOK
by a novel process with less amounts of energy and environmental pollution to use as
wood chemicals of high value.
REFERENCES
1. T. Terashima, in Chemistry of Lignin (J. Nakano, ed.), Yumi Publishers, Tokyo, p. 70 (1978)
(in Japanese).
2.G.Aulin-Erdtman, Svensk Pupperstidn., 55:745-749(1952).
3.G.Aulin-Erdtman, SvenslcPapperstidn., 56:91-101(1953).
4.G.Aulin-ErdtmanandL.Heghom, Svensk Puppersridn., 60:671-681(1957).
5. G. Aulin-Erdtman and R. Sanden, Acra Chem. Scand., 22: 1187- 1209 (1968).
6. J. C.Pew, J . Org. Chem., 28:1048-1054(1963).
7.J.C.Pew, Nature. 193:250-252(1962).
8. 0. Goldschmid, Anal. Chem., 26:1421-1423(1954).
9. E. Adler and K. Lundquist, Actu Chem. Scand., 17: 13-26 (1963).
10.A.G.SchdningandC.Johansson, SvenskPupperstidn., 68:607-613(1965).
11.G.Wegener,M.Przyklenk,andD.Fengel, Holdorsch., 37303-307 (1983).
12. H. L. Hergert, in Lignins (K. V. Sarkanen and C. H. Ludwid, eds.), Wiley-Interscience, New
York, pp. 267-297 (1971).
13. S. Kolboeand 0.Ellefsen, Tuppi, 45:163-166(1962).
14.K.V. Sarkanen,H.-M.Chang,andB.Ericsson, Tuppi, 50572-575 (1967).
15. V. Sarkanen, H.-M. Chang, and G. G. Allan, Tuppi, 50583-587 (1967).
16. K. V. Sarkanen, H.-M. Chang, and G.G. Allan, Tappi, 50587-590 (1967).
17. C . H. Ludwig, B. J. Nist, and J. L. McCarthy, J. Am. Chem. Soc., 80:1186-l196 (1964).
18. C. H. Ludwig, B. J. Nist, and J. L. McCarthy, J. Am. Chem. Soc., 8 0 1 196- 1202 ( 1964).
19. B. L.Lenz, Tuppi, 51511-519(1968).
20. D. E.Blandand S. Sternhell, Ausrral. J . Chern., 18:401-410(1965).
21.N.MorohoshiandA.Sakakibara, Mokuzui Gakkaishi, 17:393-399(1971).
22. K. Horisaki, K. Shimatani, Y. Sano, and T. Sasaya, in 37th Lignin Symp., Kyoto, Japan, pp.
125- 128 (1992).
23. K. Lundquistand T. Olsson, Acrn Chem. Scond., B31:788-792(1977).
24. K. Lundquist, Actu Chem. Scand., B33:27-30(1979).
25.K.Lundquist, Actu Chem. Scand., B34:21-26(1980).
26. K. Lundquist, A. Paterson, and L. Ramsey, Acta Chem. Scand., B37:734-736(1983).
27. S. LiandK.Lundquist, Nord. Pulp Paper Res. J., 9:191-195 (1994).
28. R. M.Ede,G.Brunow,L.K.Simola,and J. Lemmetyinen, Holdorsch., 44:95-101(1990).
29.H.-D.LtidemannandH.Nimz, Mulcrornol. Chem., 175:2409-2422 (1974).
30.H.Nimz, I. Mogharab,andH.-D.Ltidemann, Makromol.Chern., 175:2563-2575(1974).
31. H.NimzandH.-D.Ltidemann, Hol$orsch., 30:33-40(1976).
32.D.GaganireandD.Robert, Makromol.Chem., 178:1477-1495(1977).
33. J. R. ObstandJ.Ralph, Holdorsch., 37297-302 (1983).
34.C.Lapien-eandB.Monties, Holdorsch., 39:367-368(1985).
35.A.Bjdrkmanand B . Person, Svensk Pupperstidn., 60:158-169(1957).
36. K. Freudenberg and A. C. Neish, in Molecular Biochemistry and Biophysics, Vol. 2, Springer-
Verlag,Berlin-Heidelberg,p.113(1968).
37. D. Fengel,G.Wegener,and J. Feckel, Holdorsch., 3551-57 (1981).
38.H.Nimz, Holgorsch., 20:105-109(1966).
39.A.Sakakibara, Rec. Adv. Phytochem., 11:117-139(1977).
40. E. Adler,K.J.Bjorkvist,and S. Haggroth, Acta Chem. Scund., 293-94 (1948).
41. E. AdlerandL. R. Ellmer, Acra Chem. Scand., 2339-840 (1948).
42. E. Adlerand J. Marton, ActaChem. Scand., 15:357-369(1961).
43. J. Martonand E. Adler, Acra Chem. Scand., 15:370-383(1961).
Chemistry of Lignin 171
143. H. Nimz, U. Tschirner, and M. Roth, in 1983 Int. Symp. on Wood and Pulping Chernistv,
Tsukuba, Japan, Vol. I , p. 90 ( I 983).
144. U. Westermark, B. Samuelsson, and K. Lundquist, Res. Chem. Intermed., 21:343-352 (1995).
145. K. M. J. Barrow, R. E. Ede,and I. D. Suckling, in 7th Int. Syntp. on Wood and Pulping
Chemistry, Vol. I , pp. 73-81 (1993).
146. J . Sipil,G. Brunow, and P. Tunninen, in 7th Int. Symp. on Wood ctnd Pulping Chemistry,
Vol. 3, p. 12 ( 1 993).
147. Y. Sano and A. Sakakibara, Mokuzai Gakkaishi, 31: 109-1 I8 (1984).
148. Y. Sano, Mokuzai Gakkaishi, 35:813-819(1989).
149. M.Bardet, D. Robert,and K. Lundquist, Svensk Papperstdn., 6:61 (1985).
150. L. Lin, M. Yoshioka, Y. Yao, and N. Shiraishi, in Abstr. 46th Annual Meeting of the Japan
Wood Research Society, Kumamoto, Japan, p. 278 (1996).
l 5 l. K. Lundquist and R. Lundtgren, Acta Chern. Scand., 26:2005-2023 (1972).
152. K. Lundquist, Acta Chern. Scand., 27:2597(1973).
153. E. Adler and K. Lundquist, Acta Chenz. Scand., 17:13-26 (1963).
154. K. Lundquist and G. E. Miksche. Tetrcthedron Lett., 2131 (1965).
155. H. Nimz, Chem. Ber., 102:799-810(1969).
156. C. Lapierre, B. Monties, and C. Rolando, J . Wood Chem.Technol., 5:277-292 (1985).
157. C. Lapierre and B. Monties, Holiforsch., 40:47-50(1986).
158. C. Lapierre, B.Pollet, and B. Monties, Holiforsch., 4 5 - 6 8 (1991).
159. N. Terashima,R. H. Atalla, S. A. Ralph, L. L.Landucci, C. Lapierre,and B. Monties,
Holiforsch., 50:9- 14 (1 996).
160. A. Sakakibara, T. Sasaya, K. Miki, and H. Takahashi, Holzj-orsch., 41:1-11 (1987).
161. M. Erickson, S. Larsson, and G. E. Miksche, ActaChem. Scund., 27:903-914 (1973).
162. S. Larsson and G. E. Miksche, Acta Chern. Scand., 25:647-662 (1971).
163. K. Freudenberg, Holiforsch., 18:3-9(1964).
164. K. Freudenberg, Hokforsch., 18:166- 168 (1 964).
165. K. Freudenberg, Science, 148595-600(1965).
166. K. Freudenberg and A. C. Neish, in Molecular Biology, Biochemistry and Biophysics, Vol. 2,
Springer-Verlag, Berlin-Heidelberg, p. 103 (1968).
167. E. Adler, Wood Sci. Technol., /1:169-218(1977).
168. W. Glasser, in Pulp and Paper, 3rd ed., Vol I (J. P. Casey, ed.), Wiley, New York, pp. 39-
I l l (1980).
169. A. Sakakibara, Wood Sci.Technol., 1489- 1 0 0 (1980).
170. H. Nimz, Angew. Chem., 13313-321(1974).
171. B. J. Fergus and D. A. 1. Goring, Post-Grad. Res. Rep. No. 12, McGill University, Montreal,
Quebec, Canada (1968).
172. M. Matsukura and A. Sakakibara, Mokuzai Gakkaishi, 15:35-39 (1969).
173. K. V. Sarkanen, in Lignins (K. V. Sarkanenand C. H. Ludwid, eds.), Wiley-Interscience,
New York, p. 154 ( 1 97 1 ).
174. W. A. Cote, B. W. Simon, and T. E. Timmell, Svensk Papperstidn., 69547-558 (1966).
175. K. P. Kringstad andC. W. Chang, Tappi, 52:2382-2385(1969).
176. F. Yaku, Y. Yamada, and T. Koshijima, Holiforsch., 30: 148-156 (1976).
177. 0. Eriksson and B. 0. Lindgren, Svensk Papperstidn., 80:59-63(1977).
178. R. Simonson, Svensk Papperstidn., 74:153-165(1971).
179. D. FengelandM. Przyklenk. Svensk Papperstidn., 78:617-620 (1975).
180. 0. Eriksson and D. A. I. Goring, Wood Sci.Technol., 14:267-279 (1980).
18I . K. Freudenberg and G. Grion, Chrrn. Ber., 92: 1355- 1363 (1959).
182. K. FreudenbergandJ. M. Harkin, Chem. Ber, 93:2814-1819(1960).
183. E. F. Brauns, J . Am.Chem. Soc., 61:2120-2127(1939).
184. K. Freudenberg, Angew. Chenz.,68:84-92(1956).
This Page Intentionally Left Blank
Chemistry of Cell Wall Polysaccharides
1. INTRODUCTION
Cellulose and hemicellulose are major cell wall polysaccharides in woody plants which
consist mainlyof lignified secondaryxylem.Theirchemicalstructureshavebeenwell
characterized [ 1-41. Recent studies have revealed important biological and physiologi-
cal functions of the primary cell wall polysaccharides in plant growth and regulation
[5-71. Significant differences exist in polysaccharide composition of the primary cell
walls and the lignified secondary walls. In this chapter, the structure and functions of
cell wall polysaccharides, particularly pectin and hemicellulose, in growing tissue are
discussed.
The primary cell wall consists of two phases, a microfibrillar phase and a matrix phase.
As shown in Table I , the primary cell walls consist of pectin, hemicellulose, cellulose,
and a small amount of glycoproteins and phenolics. Polysaccharide compositions of pri-
mary cell walls are very similar in hardwoods(angiosperms)andsoftwoods(gymno-
sperms) (Table 2) [8], although the secondary wall polysaccharide compositions of these
two divisions are different [4].
Traditionally, polysaccharides are classified as follows: pectins are polysaccharides
extracted from cell walls by hot water, ammonium oxalate, weak acid, or chelating rea-
gents; hemicelluloses are not extracted by weak acids but are by relatively strong alkali.
The wall residue remaining after alkali extraction is mainly cellulose. This classification
based on extraction procedure is convenient to use, but it is not always appropriate. For
instance, pectic polysaccharides can be defined as galacturonic acid-containing polysac-
charides [9]. Using this structural definition, pectic polysaccharides are extracted with both
chelating reagents and with alkali. The a-cellulose fraction also contains residual galac-
turonic acid residues. The subdivision of matrix polysaccharides into pectin, hemicellulose,
and cellulose is, however, useful (see Section IV).
175
176 lshii and Shimizu
Phase Components
The most commonly occurring sugars found in cell wall polysaccharides are shown in Fig.
1. The carbonyl (aldehydo or keto) carbons of reducing sugars form acetal or ketal link-
ages, called glycosidic linkages. A glycosidically linked sugar is called a glycosyl residue;
e.g., 4-linked glucosyl residues are glucosyl residues glycosidically linked at C- 1 and also
have another glycosyl residue attached to them at C-4. In this chapter the following ab-
breviations are used: Glc = glucose; Gal = galactose; Man = mannose; Xyl = xylose; Api
= apiose; Ara = arabinose; Rha = rhamnose; Fuc = fucose; GlcA = glucuronic acid; GalA
= galacturonic acid; AceA = aceric acid; Kdo = 3-deoxy-~-manno-octulosonic acid; Dha
= 3-deoxy-~-lyxo-2-heptulosanic acid; p = pyranose ring form; f = furanose ring form.
Xylem differentiating zones of woody plants, commonly called cambial tissues, contain
primarywallsandmiddle lamella. Collection of cambial tissues for chemical analysis
requires many young trees and much time-consuming work. The cambial tissues of aspen
(Populus tretnuloides), basswood (Tilia umericarzn) [ 10- 141, and sugi (Cryptotneria ja-
p o n i c a ) [ 151 were isolated and their cell wall polysaccharides were well characterized.
a
9
147
Pectin Homogalacturonan 15
Rhamnogalacturonan-I
Rhamnogalacturonan-I1
15
5 j2Z
"From Ref. 5.
"From Ref. 10.
'From Ref. 8.
"From Ref. 15.
178 lshii and Shimizu
bH OH OH bH
p-D-Glucopyranose a-L-Rhamnopyranose a-L-Fucopyranose
OH
a-L-Arablnofuranose p-D-Xylopyranose p-D-Mannopyranose
OH
I I I
OH OH OH
p-D-Galactopyranose a-D-Galactopyranosyluronicactd p-D-Glucopyranosyl-uro
nic acid
'0
CH2OH
I
64
OHOH
CHpOH
p-D-Aplofuranose
HO-C
OH
2-keto3-deoxy-D-manno-
H
O H
COOH
3-dWxY-D-l~~o-2-hep
tulosaric acid (Dha)
OH
p-L-Aceric acld
Suspension-cultured cells are also a good source for preparing primary cell walls. Alber-
sheim and his co-workers [ 161 have used suspension-cultured sycamore (Acer pseudopla-
ranus) cells for studying the structures and functions of primary cell walls. Suspension-
cultured cells of woody plants can be easily obtained from young leaves, seedlings, and
cambial tissues. Procedures for preparation of cell walls from suspension-cultured cells
(Fig. 2 ) and isolation and fractionation of polysaccharides are described byYork et al.
[ 161, while alternative procedures for preparing cell wall materials from fresh tissues have
Chemistry of Cell Wall Polysaccharides 179
h
5
.e
aJ
3
c
Ir
0
M
0 3
2
v
Y
v)
aJ
M
v)
:a
8 2&
B
180 lshii and Shimizu
been proposed by Selvendran and O’Neill [ 171 (Fig. 3). The sequential extraction of pectin
and hemicellulose from cell wall material is shown in Figs. 4 and 5, respectively.
These approaches gave an enormous amount of information on the structure of in-
dividual polysaccharides, but they also have their limitations. One notable disadvantage is
that certain bonds in the wall are destroyed in order to extract its components [ 18,191,
which may make it difficult to observe important covalent bonds of the polysaccharides.
This drawback might be partly overcome by comparing different extraction methods and
especially by comparing chemical and enzymatic extraction methods. A second limitation
is that cell wallpolysaccharidesareoftenheterogeneouswithrespectto their primary
structure and molecular size. Extraction and purification of polysaccharides may result in
only partial recovery of a particular type of molecule, and the recovered material may not
be representative of the whole. This ultimately leads to a distorted picture of the overall
structure of the component.
Despite the many limitations of studying cell wall polysaccharides, extremely pow-
erful tools for analyzing small amounts of complex carbohydrates are currently available.
Theseinclude capillary gas-liquidchromatography-massspectrometry (capillary GC-
MS), liquid chromatography-mass spectrometry (LC-MS), fast-atom-bombardment mass
Fresh tissue
Homogenize in 1.5%(w/v) SDS
(2.5 fresh weight)
containing 5 mM Na,S20s at
5-10°C until homogeous
Filter and wash (x 2 ) in 0.5% (w/v)
SDS containing 3 mM dm-3Na2SzOS
t
Supernatant: containing
Residue
cell contents and water-
soluble cell wall
components
+
Residue
+
Supernatant:containing
water-soluble pectic
substances
1
(1) Extract with 90%(v/v) DMSO
to remove starch
(2) Wash in distilled water, and dialyse
FIGURE 3 Procedure f o r (he preparation of cell wall material from fresh tissue. (From R d . 17.)
Chemistry of Cell Wall Polysaccharides 181
Residue SIN
filter through glassfiber
paper, dialyse, concentrate
t
Residue
SIN
Filter, dialyseand concentrate
CDTA-(2) soluble
Na2C03-(1)
Stir with 100 mL 50 mM
Na2C03+ 2 0 mM NaBH,
(0,-free unde Ar)
for 2 0 h at 1'C
Centrifuge and wash withHzO
t
Residue
UN
1
Filter, adjustpH to 5 with
acetic acid, dialyse
and
concentrate
Na2C03-(l)-(cold)-soluble
Na2C03-(2)
Re-extract with 50 mM
Na2C03 as above for
2 h at 20°C
Centrifuge and wash
SIN
Extractwithalkali(Fig. 5) Filter,etc.,asabove
Na2C03-(2)-(room
temperature)-soluble
FIGURE 4 Extraction of pectic polysaccharides from cell wall material. (From Ref. 17.)
182 lshii and Shimizu
Depectinated
CWM
Stir with 75-100 mL
02-free 0.5M KOH +
10 mMNaBH,
under Ar for 2 h at 20°C
Filter under suction
v
Residue Filtrate,
adjust pH to 5 with
acetic acid, dialyse
and
concentrate
0.5 M KOH-soluble
v
Residue Filtrate,
adjust pH to 5 with
acetic acid, dialyse
and
concentrate
1 M KOH-soluble
t
Filtrate,
Residue
treat as above
a-Cellulose
4 M KOH-soluble
FIGURE 5 Extraction of hemicellulose from depectinated cell wall material. (From Ref. 17.)
Several methods are available for the analysis of intact cell walls. Solid-state NMR
and Raman spectroscopy give useful information. Immunocytochemical study using gold-
labeled antibodies and wall-degrading enzyme also yields information about the distribu-
tion of cell wall components.
A. MonosaccharideComposition
1. Acid Hydrolysis
Plant cell wall polysaccharides contain at least 12 monosaccharides which are linked by
a variety of glycosyl linkages. Theirsusceptibility to acid hydrolysis varies [20]. Therefore,
a single hydrolysisprocedurecannotprovidequantitativehydrolysis of everyglycosyl
linkage. Furthermore, the monosaccharides released by acid have different acid stabilities.
Neutral noncellulosic polysaccharides can be hydrolyzed quantitatively using 1 M H,SO,
for 2.5 h at 100°C, or 2 M trifluoroacetic acid (TFA) for 2 h at 120°C [21]. The cellulose
Polysaccharide
-
NeutraVAcidic NeutraiIAcidic
I I
Partially I
Carboxyl-reduce
I
Methylate
degrade I I
(chemiclally, I
Methylate I
Carboxyl-reduce
enzymatically)
Oligosaccharides
Partially degrade
Gel permeation
Ion-exchange
I Reduce
Realkylate
Alkylatedoligoglycosyl alditol
HPLC
l
Purified oligosaccharides I Reversed-phase HPLC
Hnarysls
'"C-NMR
is first degraded with 72% H,SO, for 2 h at 20°C and then hydrolyzed completely by 1
M acid for 4 h at 100°C [ 171. TFA is more convenient than H,SO, because it is volatile
and gives higher recovery of neutral and acidic sugar residues than H,SO,. Ketoses are
very acid-labile sugars and degrade during hydrolysis of polysaccharides. Kdo and Dha
can be released by 1 M acetic acid at 40°C for 16 h and 2 M TFA at 120°C for 5 min,
respectively [9]. Kdo and Dha are specific components of rhamnogalacturonan-I1 (RG-II),
which are present in all higher plants [9].
2. AlditolAcetatesMethod
Neutral sugars released by acid hydrolysis are reduced with sodium borohydride (NaBH,),
acetylated, and analyzed as alditol acetates by gas-liquid chromatography (GLC) [ 161.
The alditol acetates are well separated from each other by GLC and determined quanti-
tatively (Fig. 7).
3. TrimethylsilylEthers of MethylGlycosides
Polysaccharides that contain acidic sugarresidues are subjected to methanolysisand
methyl glycosides of neutral sugars and also methyl ester of uronic acid residues, which
are converted into trimethylsilyl (TMS) derivatives. TMS derivatives can be analyzed by
GLC with capillary column [l61 (Fig. 8).
B. Glycosyl-LinkageAnalysis
Methylation analysis gives information about the glycosyl linkages of the polysaccharides.
The procedure includes permethylation of all free hydroxyl groups, hydrolysis, reduction
of partially methylated monosaccharides, and acetylation. The Hakomori method, using
sodium dimethylsulfinyl anion (dimsyl anion) in dimethylsulfoxide, is commonly used for
permethylation [22]. Potassium dimsyl anion is more easily prepared [23]. Blakeney and
Stone [24] prepared the lithium carbanion from butyllithium, which is available in a much
purer form than the alkali metal hydrides, thus providing a clearer chromatogram. Ciucanu
and Kerek [25] reported a new methylation method in which powdered sodium hydroxide
and iodomethane are added to carbohydrate in dimethylsulfoxide. Methylation with sodium
hydroxide and iodomethane in SOz-diethylamino-dimethylsulfoxide is useful for cellulose
and lignified samples [26].
Acidic polysaccharides have to be reduced either before or after methylation to an-
alyze their uronic acids as partially methylated alditol acetates (PMAAs). The uronic acids
are reduced with a deuterated reagent so that they can be distinguished by mass spectrom-
etry. Water-soluble polysaccharides whose carboxyl acids have been activated with car-
bodiimide can be reduced with sodium borohydride and subjected to methylation analysis.
Alternatively, the methylated polysaccharides are readily reduced by refluxing with lithium
aluminumdeutride in dichloromethaneether [17]. York et al. reported that lithium tri-
ethylborodeutride in tetrahydrofuran is highly effective onmethyl esterified methylated
polysaccharides [ 161. Methylated polysaccharides are hydrolyzed with 2 M TFA for I h
at 120"C, reducedwithsodiumborodeutride,andacetylated.PMAAs are separatedon
capillary GLC columns. PMAAs can be easily identified from the pattern of fragment ions
produced by electron-impactmassspectrometry[27].ThePMAAsderivedfrom arabi-
noxylan are shown in Fig. 9.
Chemistry of Cell Wall Polysaccharides 185
X
x
I I I I I 1 I
Time (min)
C
0
f E
FIGURE 8 GLC of TMS derivatives of neutral and acidic sugars. (From Ref. 16.)
1. Homogalacturonan
Homogalacturonan is a homopolymer of a- 1,4-linked galacturonic acid (Fig. 10). Oligo-
galacturonides having DP 2- 12 were prepared from cell walls of soybean by partial acid
hydrolysis and isolated by ion-exchange chromatography [3l]. Recently, oligogalacturon-
ides (DP 5-20) were separated by high-performance ion-exchange chromatography on a
Carbo Pac PA1 column (Dionex) (Fig. 11).
The carboxyl groupsof many of the galacturonosyl residues of pectic polysaccharides
in cell walls are partly esterified [32], but the distribution of the methyl esters is unknown.
Calacturonosyl residues are partially acetylated at 0-3 position [33].
Homogalacturonanshave structural roles in the plant cell walls, and oligogalact-
uronidesderivedfromthehomogalacturonanhave biological activities [34],including
elicitation of defense responses, influence on plant growth and development, and promo-
tion of rapid changes in ion flux (Table 3).
2. Rhamnogalacturonan I (RG-I)
RG-I was solubilized from suspension-cultured sycamore cell walls after treatment with
EPG[9].Glycosylcomposition analysis of RC-Ishowed that arabinose,galactose,ga-
lacturonic acid, and rhamnose are its major monosaccharide constituents (Table 4). Small
amounts of fucoseare present. Glycosyl-linkageanalysis,followingreduction of the
methyl esterified galacturonic acid residue, established that RC-I is a complex, branched
Chemistry of Cell Wall Polysaccharides 187
a $ " o ~ o ~ o \
OH OH 9
OH
+ G
J
Methylation
Me0
OH
J o Q g o COOMe \
MeO
M
OMe OMe
Reductbn of uronate
AcM hydrolysls
t
Reduction
Acetylation
AIS{-::
CHDOAc CHDOAc CHDOAC CHDOAc
M&$"" M e 0 i o M e
AcO OAC OAc
CH20Me
CH20Ac CH20AC CH20Ac
-+4)-a-~GalpA-(l+4)-a-~GalpA-(l-+4)-a-~-GalpA-(l+4)-a-~-GalpA-(l+4)-a-~GalpA-(l-+
FIGURE 10 Structure of homogalacturonan. Galacturonic acid residues are partly esterified.
540-
480-
420 -
l
- 360-
B
11
al
U7
6 300-
0 6
f 240-
I
U
180-
0
120-
60 -
0-
Polysaccharides
Chemistry
Wall of Cell 189
‘Dp (degree of polymerization) range of oligogalacturonldes that show the designated biological activity.
hOrder-of-magnitude estimatton of the concentratton of oligogalacturonldes that give the half-maximum biolog-
ical response. The concentration is included only where purified oligogalacturonides are assayed.
‘Numbers in parentheses represent dp of most active oligogalactouronide.
dND = dp of active oligogalacturonides not determined.
Source: Ref. 34.
3. Rhamnogalacturonan I1 (RG-11)
RC-I1 is a polysaccharide composed of a 1,4-linked a-D-galacturonic acid backbone with
both keto-sugars (i.e., ketoses) and aldehyde sugars (i.e., aldoses) in its oligosaccharide
190 lshii and Shimizu
Rha 16 9 9
Fuc 2 1 4
Ara 32 35 30
XY 1 4 0 2
Gal 31 43 21
Glc 0 0 6
GalA 10 26
side chains. RG-I1 was for the first time isolated from the walls of suspension-cultured
sycamore cells with EPG hydrolysis followedby size-exclusion chromatography [23]. The
polysaccharide is composed of about 11 glycosyl residues and has extremely complicated
glycosyl-linkage compositions (Tables 6 and 7) [9]. Aceric acid (3-C-carboxyl-S-deoxy-~-
xylose) (Fig. 1 ) was identified for the first time in nature as a component of RG-I1 [42].
Kdo and Dha (Fig. l), which are acid-labile molecules, were also identified as integral
components of RG-I1 [9]. Partial acid hydrolysis of the RG-I1 gave several oligosaccha-
rides (Fig. 14) [9,43]. Organization in RG-I1 of the oligosaccharides was elucidated by
sophisticated sugar linkage analyses. Partial acid hydrolysis of methylated, carboxyl-re-
duced. and remethylated RG-11, in combination with selective deuteriomethyl labeling of
those hydroxyl groups exposed by the partial acid hydrolyses, showed the points of at-
tachment of oligosaccharides to the a-1A-linked galacturonosyl residues [9,44] (Fig. 15).
The exact galacturonosyl residues in the homogalacturonan (Fig. 14-1) to which oligosac-
charides 1, 2, 3, and 4 (Fig.14-2-5) are attachedhave not beendetermined. RG-I1is
present in Douglas fir [S], sugi [38], rice (Oyvzra sativa) [4S], onion (Allium cepa] [46],
kiwi fruit (Actinidin deliciosa) [47], bamboo shoot [48], and all other higher plants that
have been examined [S].
Very recently a borate-RG-I1 complex was isolated from radish (Rq3hanu.s sativus)
roots 1491, sugar beet pulp [SO], sycamore [SI], red wine [ S I ] , and bamboo shoot 1521.
Boron (B) is knownto be an essential micronutrient for all higher plants [ S ] . Boron-
deficiencysymptoms first appear at growingpointsand are characterized by cell wall
abnormalities (541. The finding that B selectively binds RG-I1 in pectin to form a cross-
link (Fig. 16) implies that B-RG-I1 complex plays an essential role in cell wall architecture
and cell formation.
A pectic polysaccharide containing galacturonosyl and xylosyl residues (xylogalact-
uronan) (Fig. 17) was isolated from mountain pine pollen 1.551 and modified hairy regions
of apple [ 5 6 ] . Xylogalacturonan was solubilized when the hairy region of polysaccharides
was treated with rhamnogalacturonan hydrolase, which cleaves the backbone of RG-I but
does not cleave xylogalacturonan. This result suggests that xylogalacturonan and RG-I is
connected covalently in the wall.
Polysaccharides
Chemistry
Wall of Cell 191
TABLE 5 Glycosyl-Linkage Composition of RC-I Isolated from the Cell Walls of Sycamore,
Douglas Fir, and Maize
Sycamore
Percent branched rhamnosyl residues
Glycosyl
48% linkage" fir Maize
Rha T 0 3.0 0.5 0 0.4
2 7.8 15.2 1.3 1.1 5.7
4 0 2.8 0.6 0 0
2,3 0 1.6 0 t 0
2,4 8.O 8.O 10.6 2.8 11.8
239 0.6 1.1 2.0 0.5 2.1
Fuc T 1.4 0.6 2.4 t 0
3,4 0 0 0 0.5 0
Ara T 9.5 2.2 2.8 13.4 14.9
2 2.2 0 0.4 t 0.8
3 2.2 0.9 2.6 17.2 2.5
5 11.2 3.4 8.0 19.8 13.0
23 1 .O l .4 3.2 6.8 l .4
3s 3.5 1.1 5.9 7.4 6.1
XYI T 2.0 0.7 0.3 t 1.1
4 0 0 0 t 1 .O
2.4 0 0 0 t 1.1
Gal T 6.3 8.1 6.6 15. I 14.2
2 0.6 1.o 2.9 0.8 0.2
3 2.7 2.2 3.9 4.2 3.5
4 8.4 2.9 11.3 4.9 4.8
6 7.5 4.4 8.1 1.o 4.8
2,4 6.3 0 6.6 15.1 14.2
2.6 1.2 0.2 2.5 0 1.9
3.4 0 0.4 1.4 1.2 0.7
3,6 l .2 0.6 6.5 1.1 2.6
4.6 2.4 1.5 0.8 1.1 0
GlcA T 0 0 0 1.6 0
GalA T 1.6 4.4 0.6 0' 0'
4 15.2 30.6 12.5 0 0
2,4 1.o 0.2 0.2 0 0
3,4 0 1.1 1.9 0 0
"Values expressed as mol%.
Source: Ref. 9.
4. Arabinans
Arabinans have been isolated from the tissues and cell walls of many plants [7]. It still
remains unknown whether arabinans exist in growing tissue as separate homopolymer, or
as covalently linked side chains of RG-I. Arabinans are highly branched molecules com-
posed of a-lS-linked arabinofuranosyl residues that are more frequently substituted at 0-
3 than at 0-2 (Fig. 18). Sycamore RG-I contains arabinosyl oligosaccharides (DP 2-20)
attached to 0-4 of the 2-linked rhamnosyl residues in the backbone [36,58].
192 lshii and Shimizu
(1)
+4)-a-~-GalpA(1-.2)-a-L-Rhap-(l--4)-a-D-GalpA(l-r2)-a-L-Rhap(l-
a-L-Araf-(1+5)-a-L-Araf-( 1-+2)-a-L-Araf-(l-r3)-p-~-GaIp(1-.4)-Rhamnitol
Araf-[Araq,,-Rhamnitol
Galp-[Galp],,-Rhamnitol
FIGURE 12 Structure of rhamnogalacturonan-l: ( 1 ) backbone structure and (2) several side chains.
5. Galactans
Some primary cell walls contain p-IP-linked galactans (Fig. 19) [7]. Galactose-containing
oligosaccharidesattached to 0 - 4 of 2,4-linkedrhamnosylresidueswere isolated from
sycamore RG-I [59,60] and tobacco RG-I [61].
6. Arabinogalactan
Two types of arabinogalactan were isolated from plants (Fig. 20) [7]. Arabinogalactan I
is a polysaccharide composed of a p- 1,4-linked galactosyl backbone that is substituted at
0-3 with short a-i,S-linked arabinosyl side chains. Arabinogalactan I1 is found in gym-
nosperms, especially in larches 1621. It is a highly branched polysaccharide containing p-
3, p-6-, and P-3,6-linked galactose with various amounts of arabinosyl, galactouronosyl,
andglucuronosyl residues. Suspension-cultured plant cellshavebeenfound to secrete
arabinogalactan protein into the culture medium 191. The polysaccharide portion is very
similar to arabinogalactan 11. This polysaccharide may be covalently attached to hydroxy-
proline-rich proteins [62,63].
B. Xyloglucan
Xyloglucan is the principal hemicellulose of the primary cell walls of dicotyledonous
plants. It was first isolated and characterized from tamarind (7hmarindu.s indicu) seeds.
The polysaccharidesformabluecomplexwith iodine. Thestructureandfunction of
primary cell wall xyloglucan have been reviewed extensively [7,64,65]. The basic structure
of this cell wall polysaccharide consists of a backbone of p-1,4-linkedD-ghCOSYl residues,
with D-xylosyl side chains a-linked to 0 - 6 of some of the glucosyl residues. Some of the
xylosyl side chains are extended by the addition of D - G a l j , ~ - F u c + 2 - a - ~ - G a l + to 0-
Chemistry of Cell Wall Polysaccharides 193
H.OH
b
' * O H
c=c-
H
Lo
(4)
OCH3
FIGURE 13 Feruloyl oligosaccharides isolated from spinach and sugar beet cell walls.
194 lshii and Shimizu
ycamore
linkagebGlycosyl
Rha T' 8.6 6.6 5 .O
2 t t 1.7
5.9 3 5.7 6.3
3.1 3.72 3 4.5
Fuc
5.52MeFuc 5.3 T 4.8
Apiose 9.0 3' 10.9 10.2
Ara 5.8 Tf 6.1 6.3
TP 0 0 4.9
2, 5 .O 4.2 0
2MeXyl T 4.5 7.3 4.1
Gal T 5.24.9 6.8
5.6 2,4 6.5 5.5
3,4 0 0 2.8
GalA 10.0 10.2
T 10.3
4 6.08.8 7.9
3.13,4 7.3 7.6
2.82,4 4.6 4.2
2,3,4 1 .S 2.7 1.5
GlcA 6.3 6.3 6.7
"5-LinkedKdo and 5-linked Dha are also present in these preparations and account for
-5% of the material.
hValues expressed as mol%.
Source: Ref. 9.
Chemistry of Cell Wall Polysaccharides 195
a-D-GalpA-(l~4)-(a-D-GalpA]5.,-(1-4)-~-GalpA
(1)
a-D-Galp
1 1
& 4
2 3
~-D-GalpA-(l+4)-a-~-Fu~p(l+4)-p-~-Rhap(l+d)-Apif
3 2
t t
1 1
2Me a-D-Xylp a-D-GalpA
(2)
2Me a-L-Fucp
1
t
2
a-D-Galp(l-+2)-p-~-Acef-(l+3)-p-~-Rhap(l+3')-Apif
4
t
1
a-L-Arap
2
t
1
a-L-Rhap
(3)
a-L-Rhap(l+5)-D-KDOp
(4)
p-L-Araf-(l-.5)-D-DHAp
(5)
+4)-a-~GalpA-(l+4)-a-~GalpA-(l-t4)-a-DGalpA-(l+4)-a-~GalpA-(l+4)-a-~GalpA-(l+
3 3
t t
1 1
P-DXYlP
~5)-a-~-AraF(l+5)-a-~-AraF(1-15)-a-L-Araf(1+5)-a-~-AraF(l-15)-a-~-AraF(1+5)-a-~-Araf-(l~
3 3 2
t 1‘ t
1 1 1
-L-Araf a-L-Araf
+4)-p”alp-(1~4)-p-D-Galp-(l-t4)-p-PGalp-(l~4)-p-D-Galp(l+4)-(3-D-Galp-(1-14)-~-D-Galp-~l+
3 3
t t
1 1
a-L-Araf-(l+5)-a-L-Araf-(l+5)-a-L-Araf a-L-Araf
Arabinogalactan I
~3)-~-r>Galp(l~3)-~r>Galp(l-13)-~-r>Galp(l~3)-~-r>Galp~l~3~-~-r>Galp~l-13~-(3-~Galp(l~3)-
6 6 6
t t t
1 1 1
S-c-Galp a-~-Araf-(l-13)-B-r>Galp a-L-Araf-(l-13)-(3-r>.Galp
6 6 6
t 7 7
1 1 1
a-L-AraC(1-13)-(3-r>Galp a-L-Araf
a-~-AraF(1+3)-p-r>Galp
6 6
t 7
1 1
L-Araf a-L-AraF(1+3)-p-r>.Galp
6
t
1
a-L-Araf
Arabinogalactan II
Abbreviations Abbreviations
2.
%
Structure Old New Structure Old New .z
Glc+Glc-tGlc-tGlc XG10 XLFG Gb+Glc+Glc+Glucitol XG7-01 XXXGol
s
t xt I xt I t t t 2-
Xyl
f 7
Gal Gal
t
Xyl Xyl
Glc+Glc+Glc
t t
Xyl
XXG s-
Fuc Xyl Xyl 8
3
v)
GIc+G Ic-t GIc-tGlc XG9 XXFG GlcjGlc FG P)
7 t t t c)
c)
Xyl Xyl x I 3
7 x71
Gal
e5
Gal
1 1' 8
Fuc Fuc
Glc-tGlc-tGlc+Glc XG9n XLLG GIc+G Ic-tG Ic-t GIc-Gc-tG I c j GIc-t Glc XXXGXXXG
t t t t t t t t t
Xyl xI xI Xyl Xyl Xyl Xyl Xyl Xyl
7 7
Gal Gal Ara
J
Glc+Gb+Glc-tGb XG8 XXLG G1c-t Glc-tG b+G Ic-t Gk+G Ic-tG Ic+G Ic XXFGAXXG
t t t t t t t t f
Xyl Xyl xI Xyl Xyl xI Xyl Xyl Xyl
7
Gal
?
Gal
t
Glc+Glc+G k+Glc XLXG Fuc
t 7 1 '
x
Xyl
fGalI Xyl
~ ~ ~
a-L-Fucp
(1) 1
1
2
@D-Galp
1
I
FIGURE 22 Structure of xyloglucan repeating units of (1) XXFG and (2) XXXG.
wall would probably limit the self-association of cellulose fibers and might provide sites
for cross-linking of cellulose fibers [64].
Xyloglucan appears to havea regulatory function as well asstructuralfunctions
[7,64,65]. Xyloglucan oligosaccharides (XXFG) inhibited auxin-induced elongation of pea
stem segments at about 10” M [78,79]. Related oligosaccharides lacking a fucosyl residue
(XXLG and XXXG) are ineffectual. Xyloglucan oligosaccharides can promote the elon-
gation of pea stem segments in the absence of 2,4-D [80].
/
OCH,
OH
OH
- f 1,11
Cellulose Microfibril
FIGURE 24 Hydrogen bonds between cellulose microfibril and xyloglucan. (From Ref. 64.)
C . Xylans
Xylans are major hemicelluloses in the primary cell wall of monocots and are found in
smaller amounts in the primary cell walls of dicots. Secondary walls of dicotyledonous
plants contain a significant amount of xylan. The basic structure of xylan has been re-
viewed [4]. Xylans have a backbone of /3-1,4-linked xylose residues (Fig. 25). The back-
bone is substituted by varioussidechainsattachedthrough 0 - 2 or 0-3 of the xylosyl
residues. Terminal arabinofuranosyl residues are usually attached to 0-3 of the 4-linked
xylosyl residues. The backbone is substituted by a-linked 4-O-methyl-/3-~-glucopyranosyl
uronic acid on 0 - 2 of xylosyl residues and acetyl esters on 0 - 2 or 0-3. The degree of
side-chain substitution determines the degree of solubility of the xylan and its ability to
bind to cellulose. Xylans having a high degree of side chains are more water soluble and
bind less tightly to cellulose, whereas molecules with fewer side chains are less water
soluble and bind to cellulose tightly.
Primary cell walls of gramineous monocots contain esterified ferulic and p-coumaric
acids [S l]. Three feruloyl and two p-coumaroyl arabinoxylan oligosaccharides were iso-
lated from bamboo shoot cell walls (Fig. 26) [76,82-841. Feruloylation and p-coumaro-
ylation occur at 0 - 5 of the arabinofuranosyl side chain of xylan. It has been hypothesized
that the feruloyl esters are subjected to peroxidase-catalyzed coupling (Fig. 27) to yield a
diferuloyl group, thereby cross-linking the xylan molecules. Such “lateral” cross-linking
of polysaccharides could have profound effects on the physical properties of the cell wall
and thus on its ability to grow and to resist enzymatic digestion [28,81].
Feruloyl arabinoxylan oligosaccharides appear to have regulatory functions just like
xyloglucan oligosaccharides. The feruloyl oligosaccharides inhibited auxin-stimulated and
gibberellin-inducedelongationgrowth of rice cells[85,86]. An arabinoxylanoligosac-
charide has no inhibitory effect. Ferulic acid itself has weak inhibitory effect. These results
indicate that the feruloyl substituent of feruloyl oligosaccharides is necessary for the in-
hibitory effect, but the glycosyl portion of feruloyl oligosaccharide is also important for
increasing this inhibitory activity.
202 lshii and Shimizu
UH
OH
I
OH
H
*' CH,O
0
A- (3)
OH
FIGURE 26 Continued
E. Glucomannans
Glucomannans are major hemicelluloses of the secondary cell walls of gymnosperms, as
well as being a minor component of angiosperm secondary walls. This has beenwell
summarized in the reviews [2-41.
A. Covalent Linkages
1. GlycosidicLinkage
There is Some evidence for covalent bonding between xylan and pectin. Kat0 and Nevins
isolated arabinoxylan-rhamnogalacturonancomplex from maize cell walls [881.
h)
0
P
Polysaccharide
I
Polysaccharide
I
H202 2H20
P
Peroxidase
H&O Ferulate Diferulate
i
OH
0 ‘0
I
Polysaccharide
nl
3
P
Chemistry of Cell Wall Polysaccharides 205
Cleaving reagents
mple Possible
Cross-link stated)(aq. unless
(a) Covalent
Glycosidic Arabinogalactan .RC-I Endoglycanases, hot acid, dry HF
Xylan. xyloglucan
Xylan. RC
Ester (uronoyl) Pectin. cellulose Esterases, Na,CO,, NaOH,
MeOH/NaOMe
Borate di-ester RC-II'RG-I1 Weak acid
Ester (uronoyl, etc.) Feruloyl .pectin Esterases, NaOH, MeOWNaOMe
Feruloyl .arabinoxylan
Feruloyl . xyloglucan
Phenolic coupling ( 1) Extensin. (Tyr-Tyr) . NaCIOz a pH 4 and 70°C
(1) Ether extensin (diferulate also cleaved by
(2) Biphenyl (2) Pectinpectin NaOH, etc.)
Xyloglucan * xyloglucan
Arabinoxylan arabinoxylan
Ether PS. (ether). feruloyl-PS NaC102?BBr,?
Disulfide Cystine (R-SS-R') Dithiothreitol, mercaptoethanol
(b)Noncovalent
Hydrogen bond Hemicellulose. cellulose MMNO, KOH (urea, guanidinium
thiocyanate, and heat are not
very effective)
Ionic bond Extensin. pectin Salts (LaCI, > CaClz > NaCI),
acids, alkalis
Calcium bridge Chelating reagents, e.g., EGTA,
CDTA, EDTA, oxalate,
hexametaphosphate; low pH
Hydrophobic Gelling of pectin (also Organic solvents
interaction involves H-bonds)
van der Waals (Many) (Reagents that change molecular
bonds conformation?)
Lectin bond Lectin. PS Sugar hapten; denaturation
2. Diferuloyl Cross-Link
The occurrence of ferulic and p-coumaric acids ester-linked to arabinoxylans in grasses
[87], to pectic polysaccharides in spinach [39-411 and sugar beet [40], and to xyloglucan
in bambooshoot [76] is well characterizedasdescribed in the previous section. The
possibility of covalent linkages betweenesterified ferulic acid on wall polysaccharides was
first proposed in 1971 by Geissman and Neukon [89]. The ferulic acid residues on feruloyl
arabinoxylan from wheat flour have been cross-linked with peroxidase and hydrogen per-
oxidase to make a gel. This demonstrated that a dehydrogenative coupling between two
esterified ferulic acid residues on arabinoxylan to form dehydrodiferulic acid had occurred
(Fig. 27). Sugar beet pectin that contains ferulic acid esterified to arabinose and galactose
residues [40] also make gel following peroxidase-catalyzed, oxidative cross-linking [90].
206 lshii and Shimizu
The phenolic coupling has been invoked to explain termination of cell expansion. Small
amounts of diferulate have been detected by alkaline hydrolysis of cell walls. The goal of
detecting an oligosaccharide fragment, cross-linked by a diferuloyl bridge, was achieved
in 1991 [91]. The linkage group was isolated and characterized from bamboo shoot ara-
binoxylan, providing definitive evidence for the existence of diferuloyl ester cross-link
(Fig. 28). Ralph et al. [92] isolated a series of ferulic acid dehydrodimers in addition to
5-5 coupled dehydrodiferulate from saponified grass cell walls (Fig. 29). These dehydro-
dimers (8-5, 8-0-4, and 8-8) also are involved in cross-linking of polysaccharides in cell
walls. Isolation ofwallfragmentscontainingthesedehydrodimers is animportant
challenge.
Other possibilities for dimerization of phenolic acid substituents of polysaccharides
exist. A series of homo- and heterocyclodimers of the cyclobutane type, formed by head-
to-tail or head-to-head association of ester-linked p-coumaric acid and ferulic acids, were
isolated [93] (Fig. 30).
A similar peroxidase-catalyzed cross-link may occur between tyrosine residues of
extensin (Fig. 31). The phenolic ether linkage of isotyrosine is known to form intramo-
lecularly within extensin, and may also occur intermolecularly [28].
3. Ester Bonds
Cold Na2C03,which hydrolyzes ester bonds but does not cause p elimination-degradation,
solubilized pectin that is not solubilized by chelatingreagent or EPC [9]. Theseester
bonds may be methyl esterified galacturonosyl residues, diferuloyl bridges, or ester bonds
between uronic acids and neutral sugars. Although intermolecular ester bonds have been
proposed [94], their identification has not yet been achieved.
FIGURE 28 A diferuloyl arabinoxylan hexasaccharide isolated from bamboo shoot cell walls.
Chemistry of Cell Wall Polysaccharides 207
HOv 0
“ O v O OH
MS
0
B. NoncovalentLinkages
As cell wall polysaccharides are polyhydroxylic, many hydrogen bonds form in the walls.
Multiple hydrogen bondsare present within cellulose microfibrils. Hydrogen bondsare
probably responsible for the incorporation of xylan, xyloglucan, and glucomannan into the
cell wall. The hydrogen bonding between cellulose and xyloglucan, and between cellulose
and xylan, has been demonstrated in vitro.
Ionicbonds will be formed in the case of polymers that contain charged groups.
Homogalacturonan and 4-O-methylglucuronoxylan have negative charges, while extensin
has a positive charge. Ionic binding may occur between these charged polymers in the
wall. Negatively charged galacturonic acid residues in homogalacturonan and RG-I can
form cross-linking with Ca’+ to form an “egg-box” [95]. About 15-20 contiguousga-
208 lshii and Shimizu
OH
H0
lacturonic acid residues are needed in each chain to make a stable complex. Methyl es-
terification in homogalacturonan and rhamnosyl residues in RG-I interrupt the concerted
binding.Furthermore, acetylation occurs at 0 - 3 of galacturonicacidresidues in homo-
galacturonan [333 and at 0-2 and 0-3 of galacturonic acid residues of RG-I, respectively
[35-371. Therefore,egg-boxcross-linkage is likely to be limited in the wall, although
isolated pectin and pectin in jam and jellies are known to give rise to gel-like structures
in vitro.
The cell wall has a variety of components that assemble to form an extremely complicated
structure. Several wall models have been proposed. An early cell wall model was proposed
by Albersheim and co-workers [96]. This model contained covalent links between xylo-
glucan and RG-I and hydrogen bonds between cellulose and xyloglucan, leadingto indirect
cross-linking of cellulose microfibrils through a series of hydrogen bonds and covalent
bonds in the matrix. Some of the details of this model have been disproven by subsequent
chemical analysis [ 181. Capita and Gibeaut [97] and McCann and Robert [98] have pro-
posed structural models for primary cell walls basedon this new information. It is probably
impossible to describe all cell wall components and their interactions with a single and
simple model. However, the existence of two principal polysaccharide networks in the
growing cell wall have been proposed: a tension-resistant load-bearing cellulose/xyloglu-
can network and a compression-resistant pectic polysaccharide network.
walls. This chapter also discusses the functions of primarycell walls. The recent discovery
that oligosaccharide fragments derived from cell wall polymers in growing tissues can act
as potent and specific regulators of gene expression is of importance to the biochemistry
of all living systems [see refs. 34,99,100].
REFERENCES
1. G. 0. Aspinal, Advan. Carbohyd,: Chem., 14: 429 (1959).
2. T. E. Timell, Advan. Carbohyd,: Chem., 19: 247 (1964).
3. T. E. Timell, Advan. Carbohydr: Chem., 20: 409 (1965).
4. K. Shimizu, in Wood and Cellulosic Chemistry (D. N.-S. Hon and N. Shiraishi, eds.), Marcel
Dekker, New York, p. 177 (1991).
5. A. G.Darvill, M. McNeil, P. Albersheim, and D. P. Delmer, in The Biochemistry ofplants,
1 (N. E. Tolbert, ed.), Academic Press, New York, p. 91 (1980).
6. P. Albersheim and A. G. Darvill, Sci. Am., 253(3):58 (1985).
7. M. McNeil, A. G. Darvill, S. C. Fry, and P. Albersheim, Ann. Rev. Biochem., 53: 625 (1984).
8. J. R. Thomas, M. McNeil, A. G. Darvill, and P. Albersheim, Plant Physiol., 83: 659 (1987).
9. M. A. O’Neill, A. G. Darvill, and P. Albersheim, Methods Plant Biochem., 2: 415 (1990).
10. B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 39 (1978).
11. B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 51 (1978).
12. B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 63 (1978).
13. B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 79 (1978).
14. B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 137 (1978).
15. Y. Edashige, T. Ishii, T. Hiroi, and T. Fujii, Hol~orchung,49: 197 (1995).
16. W. S. York, A. G. Darvill, M. McNeil, T. T. Stevenson, and P. Albersheim, Methods Erzzymol.,
118: 3 (1986).
17. R. R. Selvendran and M. A. O’Neill, Methods Biochem. Anal., 118: 3 (1986).
18. P. M. Dey and K. Brinson, Advan. Carbohyd,: Chem. Biochem., 42: 265 (1987).
19. C. Brettand K. Weldron,in Topics in PlantPhysiology, 2, Unwin Hyman, London, p. 4
( 1990).
20. G. 0. Aspinal, in Techniques of Chemistry, 4 (K. W. Bentley and G. W. Kirly, eds.), Wiley-
Interscience, New York, p. 379 (1973).
21. P. Albersheim, D. J. Nevins, P. D. English, and A. Karr, Carbohyd,: Res., 5: 340 (1967).
22. S. Hakomori, J. Biochem. (Tokyo), 55: 205 (1964).
23. A. G. Darvill, M. McNeil, and P. Albersheim, Plant Physiol., 62: 418 (1978).
24. A. B. Blakeney and B. A. Stone, Carbohyd,: Res., 140: 3 19 (1985).
25. I. Ciucanu and F. Kerek, Carbohydr: Res., 133: 105 (1984).
26. A. Isogai, A. Ishizu, J. Nakano, S. Eda, and K. Kato, Carbohyd,: Res., 138: 99 (1985).
27. P.-E. Jansson, L. Kenne, H. Liedgren, B. Lindberg, and J. Lonngren, Chern. Commun. Univ.
Stockholm, No. 8 ( 1976).
28. S. C . Fry, Ann. Rev. Plant Physiol., 37: 165 (1986).
29. X. Qi, B. X. Behrens, P. R. West, and A. J. Mort, Plant Physiol., 108: 169 1 ( 1995).
30. A. Bacic, P. J. Harris, and B. A. Stone, in Biochemistry of Plunts, 13 (J. Preiss, ed.), Academic
Press, San Diego, CA, p. 297 (1988).
31. E. Nothnagel, M. McNeil, and P. Albersheim, Plunt Physiol., 71: 916 (1983).
32. G. 0. Aspinal, Polysaccharides, Pergaman Press, Oxford, U.K. (1970).
33. T. Ishii, Plant Physiol., 113: 1265 (1997).
34. A. Darvill, C. Auger, C. Bergmann, R. W. Carlson, J.-J. Cheng, S. Eberhard, M. G. Hahn,
V.-M. L6, V. Marfi, B. Meyer, D. Mohnen, M. A. O’Neill, M. D. Spiro, H. van Halbeek, W.
S. York, and P. Albersheim, Glycobiology,2: 181 (1992).
35. P. Komalarilas and A. Mort, Curbohydr: Res., 189, 261 (1989).
36. P. Lerouge, M. O’Neill, A. G. Darvill, and P. Albersheim. Cnrbohyd,: Res., 243, 359 (1993).
Chemistry of Cell Wall Polysaccharides 211
Toshiaki Umezawa
Kyoto University, Kyoto, Japan
1. INTRODUCTION
Extractives are the wood constituents which can be extracted with neutral solvents. They
are obtained by extracting wood meal with organic solvents or water or by steam distil-
lation, and some are obtained as exudates from wounded trees.
The amountofextractives is small,generallyupto 5-10% in the woodin the
temperate zone. However, in some tropical woods relatively high amounts of extractives
are found [l].
Amongwood species, differences of chemical structures of three major cell wall
components, cellulose, hemicellulose, and lignin, are few. However, a great diversity in
extractive composition is found throughout wood species. Although the extractives are low
in concentration compared with those of the cell wall polymers, this fraction characterizes
eachwoodspecieschemically.Mostcomponents of woodextractives are classified as
secondary metabolites, and the distribution of specific compounds is restricted in certain
wood species. This feature provides the basis of chemotaxonomy of woody plants.
Furthermore, individual compounds are often found in specific tissues of individual
trees, and their amounts can vary from season to season even in the same tissue. Many
phenolic compounds are accumulated in heartwood, whereas they are found only in trace
amounts in the corresponding sapwood.
Extractives are the predominant contributors to woodcolor, fragrance, and durability.
Extractives also influencethe pulping, drying, adhesion, hygroscopicity, and acoustic prop-
erties of wood.Manyextractiveshave specific biological activities, andvariouswoods
have been used as sources of crude drugs and medicines for centuries.
Recent progress of the biosynthesis of secondary metabolites of woody plants in-
cluding heartwood components [2] has suggested the possibility of biotechnological con-
trol of their biosynthesis. This may lead to biotechnological production of biologically
active extractives, and to elucidating molecular mechanisms of heartwood formation.
Progress in the chemistry and biochemistry of natural products including wood ex-
tractives has been reviewed in a number of articles, such as in Natural Product Reports.
Thisjournalincludesregularreviews of the relevant literature publishedduringwell-
defined periods with respect to individual topics in the fields.
213
214 Umezawa
Lignans and neolignans are phenylpropanoids that occur in many plants including soft-
woods, hardwoods, and medicinal plants [3-51.
The term “lignan” was introduced by Haworth to describe a group of phenylpro-
panoid dimers, where the phenylpropane units were linked by the central carbon (C8) of
their side chains [6]. Gottlieb coined “neolignan” for compounds containing two phen-
ylpropane units that are linked otherwise than C8-C8’ [7]. Later, neolignans were redefined
as the dimers of allyl- or propenylphenyls, while lignans were regarded as the dimers of
cinnamyl alcohols [8]. However, in this review Haworth’s definition of lignans [6] and
Gottlieb’s former definition of neolignans [7] will be used, because the original definitions
are being applied by most researchers.
Tri- and tetramers of phenylpropanoid units are referred to as sesqui- and dilignans,
respectively.
Lignans are classified into several subgroups:dibenzylbutanes,dibenzylbutyrolac-
tones, furans, furofurans, aryltetralins, arylnaphthalenes, and dibenzocyclooctadienes [3,5]
(Fig. I). Lignans often occur as glycosides. Examples of neolignans are shown in Fig. 2.
Chemical structures of lignans and neolignans are similar to that of lignins. Thus
(+)-pinoresin01 (5) and the corresponding (-)-enantiomer are typical lignans, while this
structure is also involved in a lignin molecule as a substructure (i.e., pinoresinol or p-p’
substructure) [9]. Also the structures of neolignans are similar to lignin substructures of
p-0-4 (arylglycerol p-aryl ether), 5-5’ (biphenyl), and p-5 (phenylcoumaran) types [IO].
In spite of the structural similarities, lignins differ sharply from lignans and neolignans in
terms of optical activity. The former is optically inactive, while the latter is optically active,
suggesting the difference in stereochemical mechanisms in their biosyntheses [9].
Following the first example of an in-vitro enantioselective formation of an optically
pure lignan, (-)-secoisolariciresinol (l),with an enzyme preparation from Forsythia in-
termedia [ 1 l], much investigation has been done for enzyme systems in lignan biosynthesis
of Forsythia spp. [9,12,13]. Figure 3a shows the conversion of coniferyl alcohol (14) into
lignans with Forsythia enzymes. Recently, cDNA cloning of pinoresinol/lariciresinol re-
ductase which catalyzes reduction of (+)-pinoresin01 (5) and (+)-lariciresinol (4) to (+)-
lariciresinol (4) and (-)-secoisolariciresinol (l),respectively, has been reported [ 141. In
addition, enantioselective coupling of two coniferyl alcohol radicals giving rise to (+)-
pinoresinol (S) was reported [15]. The coupling was highly enantioselective only in the
presence of dirigent protein isolated from Forsythia plant. Formation of (+)-secoisolari-
ciresinol (lS), the oppositeenantiomerto the one in Forsythia spp.,with an enzyme
preparation from Arcfiurn lappa petioles, has been reported (Fig. 3b) [16]. On the other
hand, an enzyme preparation from seeds of A. lappa was recently found to catalyze en-
antioselective formation of (-)-secoisolariciresinol (1) from coniferyl alcohol (14), indi-
cating that two enzyme preparations from different organs of a single plant species can
catalyze the selective formation of different enantiomers of a lignan [ 171.
The structural similarity with lignins suggests that neolignans maybe synthesized
by enantioselective coupling of two phenolic phenylpropane units. Recently, two examples
of enantioselectiveformation of neolignansfrom coniferyl alcohol (14) werereported
[18,191.
Lignanshavesuch biological activities asantitumor[podophyllotoxin (7) and
steganacin (9)], antimitotic [podophyllotoxin (7)],antioxidant (nordihydroguaiaretic acid
andsesaminol), and antiviral [podophyllotoxin (7) againstcytomegalovirusandherpes
simplex l-virus, and (-)-arctigenin (3) against human immunodeficiency virus], etc. [3-
Chemistry of Extractives 215
Dibenzylbutane
Dibenzylbutyro- Furan
lactone
OH
H0
OCH3
-Secoiso-
(-) R=E (+) -Lariciresinol 4
lariciresinol (-) "atairesinol 2
1 Re83
(-) -Arctigenin 3
E'urofuran Aryltetralin
OCH3 OCH3
R=E Podophyllo- Diphyllin 8
(+) -Pinoresin01 5 toxin 7
R=OC83
(+) -Syringaresinol 6
Dibenzocyclo-
Glycoside
octadiene
Steganacin 9 Arctiin 10
OCH3
Kadsurenone 11 Magnololl2
Eonokiol13
%"
Coniferyl (+) -Pino- (+) -Larici-
alcohol 14 resinol 5
I resinol 4
J
(b)
H3CO
H O T : :
- \
0
OCH3 Coniferyl O C H ~ (+) -Secoiso-
OH alcohol 14 OH lariciresinoll5
FIGURE 3 Enzymatic lignan formation.
5,201. Antagonism toward the platelet-activating factor (veraguensin) and inhibitory activ-
ities toward certain enzymes have also been detected in many lignans [3-5,201. Among
the biologically active lignans, antitumor podophyllotoxin (7)has attracted particular in-
terest [see also Section XLC]. Mammalian lignans are known to be produced from plant
lignans such as secoisolariciresinol diglucoside by the action of bacterial flora in the colon
of human or animals. The mammalian lignans and their precursor secoisolariciresinol di-
glucoside have a protective effect on the promotion stage of mammary tumorigenesis [21],
and are receiving widespread interest.
A neolignan, kadsurenone (ll),has antagonistic activity to platelet-activating factor
[22].Honokiol (13) is antibacterial and antifungal, and magnolol (12) has antimicrobial
and muscle-relaxant activities [23,241.
111. NORLIGNANS
OH
Hinokiresinol 16 Agatharesinol 17 Sequirin
(Sequirin-C) 18
Sugiresinol
Hydroxysugiresinol
Cryptoresinol 21
( Sequirin-A) 19 (Sequirin-B) 20
HdHaWO Pueroside-A
H O f/l o H
cis-Hinokiresinol
H0 OH OH 22 23
\3
OH 2,3-Bis(p-hydroxy-phe
Yateresinol
Sequirin-D
nyl)-2-cyclo-pentene-
24 25 l-one
26
FIGURE 4 Examples of norlignans.
ylethane (C,-C,) units connected via C8-C7' [e.g., hinokiresinol (16)] and in some cases
C8-C8' [e.g., yateresinol (24)] and C9-C8' [e.g., sequirin-D (25)](Fig. 4).
Most norlignans have been isolated from softwoods belonging to Cupressaceae, Tax-
odiaceae, and Araucariaceae [25-271, while some norlignans occur in herbaceous plants.
cis-Hinokiresinol (23) was isolated from Anen7arrl1et~a crspI7ndt.loide.s (Liliaceae) 1281, and
two norlignan glycosides, pueroside-A (22) and -B, were isolated from PuercrriLI lohtc1
(Leguminosae) [29] (Fig. 4).
No concrete experimental evidence has been reported for the origin of the two ar-
omatic nuclei of a norlignan molecule. Based on structural considerations, several possible
schemes have been proposed for the biosynthesis of norlignans, involving coupling of two
phenylpropane units followed by loss of one carbon atom 125,301.
Norlignans are of particular interest with respect to heartwood coloration. The col-
oration of heartwood of Japanese cedar (Cty>tornc~ria japonica) is due to polymerization
218 Umezawa
of norlignans, e.g., hydroxysugiresinol (20) and sequirin-C (18) [31,32], andthat of hinoki
cypress (Charnaecyparis obtusa) is related to hinokiresinol (16) [33].
Inhibitory activity on cyclic adenosine-3’,5’-monophosphatephosphodiesterase was
reported for cis-hinoki resinol (23) [28]. Sugiresinol (19) and hydroxysugiresinol (20) have
inhibitory effect on polymerization of methyl methacrylate [34].
IV. FLAVONOIDS
Qfy
Flavanone
0
Flavone Cbalcone
H \ I OI S H \ O S
OH 0 OH 0 R 0
Sakuranetin 27 Apigenin 28 R=on, Chalconaringenin
29
R=H, Isoliquiritigenin
30
Dihydroflavonol
(Flavanonol) Flavonol Aurone
OH
0
H o e : :
I I
' OH
0 OH
OH 0
Dihydro- Quercetin 32 Sulfuretin 33
kaempferol 31
Flavan-3-01 Flavan-3,I-diol
(Catechin) (Leucoanthocyanidin) Anthocyanidin
OH OH
OH
H \o d : : H o e OHo H
OH
OH OH OH OH
(+)-Catechin 34 Leucodelphinidin Pelargonidin
35 36
Isoflavonoid Neoflavonoid
%
Daidzein 37
p-Coumaroyl-CoA 39
Flavonoids
HOOC>SCoA
0
Malonyl-CoA 40
Many genes of the enzymes involved in flavonoid synthesis have been cloned, and
mechanisms of the gene expression have been investigated intensively [47,48]. Thus, fla-
vonoids are one of the best-understood groups of plant secondary metabolites, especially
in terms of biosynthesis. It should be noted that extensive genetic information available
about flavonoids, especially anthocyanin pigments of flowers, has accelerated significantly
the biosynthetic studies in this field.
V. STILBENES
I
OH Lunularin Batatasin I
45
VI. DIARYLHEPTANOIDS
Diarylheptanoids are composed of two phenyl rings connected with a C , carbon chain
(Fig. 8). Many of this type of compound are isolated from plants belonging to the Betu-
laceae and Zingiberaceae. Besides these two families, the occurrence of diarylheptanoids
in the following species was also reported: Centrolobium spp. (Leguminosae), Myrica spp.
(Myricaceae), A c e r spp. (Aceraceae), and Garuga spp. (Burseraceae). Recently the chem-
istry and biological activity of diarylheptanoids have been reviewed [66,67].
OH OH 0 0
Ilannokinol 47 Curcumin 48
H0 OH
P l a t y p h y l l o s i d a 49
"'OH
H 0 OH
Asadanin 50 Acerogenin A 51
Two different results of feeding experiments have been reported regarding biosyn-
thetic precursors of diarylheptanoids. One suggested that one aromatic ring of curcumin
(48) is derived from a cinnamate unit and the other from acetate (or malonate), based on
administration of [ I4C]acetate, [I4C]malonate, and [14C]phenylalanine to Curcuma longa
[68]. Other studies suggestedthat diarylheptanoids, acerogeninA (51), and platyphylloside
(49) were derived from two phenylpropane units and one acetate (or malonate)unit, based
on administration of [“C]cinnamate, [14C]phenylalanine, [ I4C]acetate, and [ 14C]malonate
to Acer nikoense [acerogenin A (51)] [69], and of [14C]cinnamate and [I4C]malonate to
Betula platyphylla [platyphylloside (49)] [70].
VII. ISOPRENOIDS
lsoprenoids is the generic name of compounds composed of isoprene (C,H,) units con-
nected linearly or cyclically. Isoprenoids consist of terpenoids, steroids, and tropolones.
Since the chemistry of terpenoids, steroids, and tropolones has been developed indepen-
dently in spite of their close relationship in biosynthesis, the three classes are usually
treated separately. Terpenoids are divided into monoterpenes (C,,,), sesquiterpenes (C,,),
diterpenes (C?”), sesterterpenes(C,,), triterpenes (C,,,), tetraterpenes (C,,), and polyterpenes
(C,,,), depending on the number of the constituent isoprene (C,) units. Each subclass of
terpenoids is further classified into many groups of different carbon skeletons.
The terpenoid compounds are generally elaborated via the mevalonate pathway as
outlined in Fig. 9. Although the mevalonate pathway has generally been accepted, a novel
pathway concerning the early steps of isoprenoid biosynthesis toward isopentenyl pyro-
phosphate (56) has recently been demonstrated [71]. The novel pathway, which involves
the condensation of a triose phosphate with activated acetaldehyde, has been characterized
in several different bacteria [71,72]. In addition,Eisenreichetal.haveshown that the
taxane carbon skeleton is not of mevalonate origin in Taxus chinensis [73] (see also Section
X1.C).
Isoprenoids are the largest group among plant secondary metabolites, and occur in
a huge number of plants including woody plants. It is beyond the scope of this book to
list all the plants producing this class of compounds and to describe their biosynthetic
schemes. Comprehensive lists of the compounds, their biosynthesis, and biological activ-
ities are summarized elsewhere [2,37,74-771.
A. Terpenoids
Specific fragrances of different woods are usually due to the composition of monoterpenes
and volatile sesquiterpenes. They can be easily separated from wood by steam distillation,
and the oily substance obtained is called “essential oil.” Turpentine, essential oil from
Pinus spp., is obtained by steam distillation of exudates from pine trees (oleoresin); the
residue is gumrosin,which is composedmainly of diterpene acids (rosin acids), e.g.,
abietic acid (91). Turpentines obtained from pine wood and those recovered from kraft
pulp waste liquor are called wood turpentine and sulfate turpentine, respectively. Rosins
are used for sizing of papers.
Monoterpenes are derivedfromgeranylpyrophosphate (58). They are subdivided
into acyclic and cyclic monoterpenes (Fig. IO) a-Pinene (64) and @-pinene (65) are major
components of turpentine.
Chemistry of Extractives 223
y SCoA y SCOA
Acetyl-coA Acetyl-coA
52 52
Ki,SCOA Aceto-
acetyl-coA
0 ,o 53
3-Eydro~y-3-methyl-
glutaryl-CoA
54
Mevalonic acid55
- JJPP pyrophospxate
1 56 Isopenten
1
uopp Dimethylallyl
pyrophosphate 57
Sesqui-
.
Farnesyl '
pyrophosphate 59
terpenoids
Squalene \ Tri-
terpenoids
cDi-
terpenoids
G e r m lgeranyl 'Phytoene \ Tetra-
pyropxosphate 60 terpenoids
Sester-
terpenoids
Getan lfarnesyl
pyropiosphate 61
Polyterpenoids
Sesquiterpenes are derived fromfarnesyl pyrophosphate (59), and constitute the larg-
est class of terpenoids [74]. Some 120 distinct skeletal types of sesquiterpenes are known
[74]. Figure 11 depicts the important sesquiterpene skeletal types from acyclic (e.g., far-
nesane) to tricyclic (e.g., thujopsane), which are often encountered as wood constituents.
Qpical examples of each type of sesquiterpenes are also shown under the corresponding
skeletal types in Fig. 11.
Diterpenes are derived from geranylgeranyl pyrophosphate (60), and some 130 dis-
tinct skeletal types are reported [74]. Figure 12 shows the typical diterpene skeletal types
224 Urnezawa
(- ) -Limonene 1,B-Cineol
67 68
with corresponding examples. A phytane diterpene, plaunotol (87), has anti-ulcer activities
[78]. ent-Gibberellane diterpenes, gibberellins, are important plant hormones.
Occurrence of sesterterpenes in higher plants is highly limited.
Triterpenes are elaborated from squalene formed via tail-to-tail coupling (coupling
between pyrophosphate ends) of two farnesyl pyrophosphate units 121. Two major classes
of nonsteroidal triterpenes are tetracyclic and pentacyclic [74]. Examples of typical skeletal
types of this class are shown in Fig. 13: lanostane, dammarane,euphane,limonoids
[tetranor (C,,) compounds], quassinoids (mainly Czo compounds), lupane, and oleanane.
Figure 13 also shows examples of compounds belonging to each skeletal type. Oleananes
often occur as aglycons (sapogenins) of saponins [76].
B. Steroids
Steroids, which are also derived from squalene, are compounds with cyclopentanoperhy-
drophenanthrene skeleton and their congeners elaborated from them. The basic structure
of steroids is shown in Fig. 14. Positions on the same side as the angular methyl (18- and
19-CHJ are denoted as p, and those on the opposite side are denoted as a. Substituents
on the 8, 9, 10, 13, 14, and 17 positions of steroids are projected to Sp, 9a, lop, 13P,
14a, and 17P positions, respectively.
p-Sitosterol (or sitosterol) (108) (Fig. 14) is widely distributed in the plant kingdom.
a- and P-Ecdysones isolated from silk worm (Bombyx rnori) have steroid skeletons, and
are well known as molting hormones. Compounds with similar structure to the ecdysones
were isolated from plants (e.g., Podocarpus and Taxus spp.), and are referred to as phy-
toecdysones [e.g., ponasterone A (109)] (Fig. 14) [79].
C. Tropolones
Tropolones are nonbenzenoidaromaticcompoundshaving a seven-memberedenolone
structure. They occur in Cupressaceae plants and exhibit antimicrobial activity. Examples
Chemistry of Extractives 225
B i s a b o lFaanren e s a n e
Caryophyllane
Germacrane Cuparane
OH
of this class are shown in Fig. 15a. Hinokitiol(P-thujaplicin) (111) was isolated from
Charnaecypuris ruiwunmsis by Nozoe [80,81]. a,P, and y-Thujaplicins, (110), (lll),and
(112), were isolated from Thuju plicufu by Erdtman [82].
Tropolones are composed of 10 or 15 carbon atoms, and they have been regarded
as mevalonate origin, i.e., a subclass of isoprenoids. Recently this has been supported by
feedingexperiments.[“C]Mevalonic acid wasfound to beincorporatedinto hinokitiol
(111) in suspensioncultures of Cupressus lusitunicu, suggesting that hinokitiol (111) is
elaborated via the mevalonate pathway [83].
226 Umezawa
q
y-Acora- P-Chamigrene a-Cedrene Thujopsene
diene 83 84 85
82
FIGURE 11 Continued
On the other hand, the tropolone structure and aromatic ring of colchicine (114) (Fig.
15b),which is analkaloidfrom Colchicum autumnale, arederivedfromtyrosineand
phenylalanine, respectively [84]. Thus colchicine (114) is a phenylpropanoid compound
butnot an isoprenoid. In addition,tropolonesoccurring in microorganismsarebiosyn-
thesized via the acetate-malonate pathway [85].
VIII. QUINONES
Various types of quinones occur in many plant families, and most of them are benzoqui-
nones, naphthoquinones, or anthraquinones. Most of the quinones found in nature are p -
quinones, but o-quinones also exist [86]. Spica1 examples of this class are shown in Fig.
16. Quinones are biosynthesized via various pathways, i.e., the shikimate, the mevalonate,
and the acetate-malonate (polyketide) pathways [2].
Quinones are pigments and have various biological activities. Juglone (116), which
occurs in black walnut (Juglans nigra), is skin-irritating [87-891. This compound is also
well known as an allelochemical [90]. Tectoquinone (117) and related compounds have
strongantitermite activity [91]. Mansonone A (118) and its congenercause allergies
[89,92].
Chemistry of Extractives 227
Phytane
LJwwLL
OH
HO-
Phytol 86
mPlaunotol87
OH
Abietane Nagilactone
&
p!?
COOH
A
acid 92 lactone A 94
91 93
IX. TANNINS
Tannins are water-soluble phenolic compounds having molecular weights between500 and
3000. Besides giving the usual phenolic reactions, they have special properties such as
the ability to precipitate alkaloids, gelatin, and other proteins [93,94]. This class of com-
pounds also has high astringency, and gives blue or green coloration with femc chloride.
Tanninsare distributed widely in wood, bark, andleaves of many plants. Tannins are
classified into hydrolyzable and condensed tannins.
228 Umezawa
COOH
Taxane
10-Deacetyl- Taxol 99
baccatin I11
98
FIGURE 12 Continued
Hydrolyzable tannins (Fig. 17a) are esters of an aliphatic poly01 and phenolic acids
(Fig. 17b), and can be hydrolyzed into the components. As shown in Fig. 17a, galloyl,
hexahydroxydiphenoyl, and depside galloyl groups are esterified to the polyol, generally
D-glucose. Hydrolyzable tannins that give gallic acid (120) by hydrolysis are referred to
as gallotannins, while compounds that afford ellagic acid (122) are referred to as ellagi-
tannins. Hexahydroxydiphenic acid (121) is lactonized to give rise to ellagic acid (122)
in the hydrolysis. Some ellagitannins [e.g., casuarinin (119)] have C-glycosidic structures.
Two different pathways have been proposed for the biosynthesis of the phenolic unit,
gallic acid (120). One is P-oxidation of the side chain of cinnamates to give gallic acid
(120), while the other is direct conversion of 3-dehydroshikimate to gallic acid (120) [95].
Recently, it has been shown that formation of gallic acid (120) via cinnamic acids can be
ruled out as a major pathway in the fungus Pllycornyce.7 blakesleeanus and in young leaves
of Rhus typhirza, and gallic acid (120) is probably formed from an early intermediate of
the shikimate pathway, most probably 3-dehydroshikimate [96]. In the pathways to hy-
drolyzable tannins, the first specific intermediate is P-glucogallin ( l-O-galloyl-P-D-glu-
cose), which is formed from UDP-glucose and gallic acid (120). Then, P-glucogallin is
Chemistry of Extractives 229
H3CO"
$$
0
OH
,8
H0 H
5P-Steroid
5a-Steroid
p-sitosterol Ponasterone A
108 109
Nootkatin 113
H3coq
H&O
H3C0
. 0
CH3
OCH3
Colchicine 114
FIGURE 15 Examples of tropolones.
Chemistry of Extractives 231
H3C0 OCH3
0 0 0
2,C-Dimethoxy- J'uglone TectoquinoneMansonone A
118 p-benzoquinone
117 16 1
115
l
R= yo CO
/OR
OH
RO&&&%' Depside
OH OH galloyl
OR group
Hexahydroxy-
diphenoyl
Hydrolysable tannin - group
OH
.n 119
(b)
OH
Gallic acid Haxahydroxy- Ellagic acid
120 dipbenic 122
acid
121
metal ions and with other molecules such as proteins, polysaccharides, and alkaloids [54].
These properties underlie their biological activities as well as their industrial applications.
Tannins exhibit various biological and pharmacological activities, e.g., bacteriocidal
action, inhibition of HIV replication, as well as astringency [37,54]. Inhibitory effects of
condensed tannins on the activities of Streptococcus sobrins glucosyltransferases involved
in dental caries formation were also reported [99]. Recently, the role of plant phenolics in
diets is attracting considerable interest, since epidemiological investigations suggested that
the consumption of beverages containing plant phenolics (such as tannins and flavonoids),
in particular, green tea and red wines, reduced the risk of certain degenerative diseases
[54,100]. Recent study demonstratedthat the iron-chelating properties of polyphenols con-
tribute to limit the growth of microorganisms [ l o l l .
The ability of tannins to complex proteins has long been utilized for tanning agents
in leather manufacturing,while that to complexmetal ions hasbeenappliedtodyeing
[37,54].Condensed tannins, especiallywattle (ormimosa) tannins, havebeenused to
produceadhesives for wood-basedmaterialssuchas particle board and plywood.The
tannin adhesives have been successfully commercialized especially in South Africa, Aus-
tralia, and New Zealand [ 102- 1041.
Many reviews of chemical properties, biological significance, and commercial sig-
nificance of tannins are available in a recent book [105]. Mechanisms of tannin-protein
complexation will be described briefly in Section X1.B.
X. OTHER COMPOUNDS
A. Sugars
D-Glucose and D-fructose are found, along with sucrose, in the sapwood of woody plants,
while L-arabinose is found in heartwood [ 1061. Larches (Larix spp.) contain significant
amounts (IO-25%) of a water-extractable arabinogalactan [ 1071.
B. Glyceridesand Waxes
Glycerides are esters of glycerol and long-chain fatty acids. Among the glycerides, triesters
(triglycerides) are dominant. Triglycerides of pine woods cause pitch trouble in ground
wood pulping. Waxes are complex mixtures of aliphatic compounds, and a majority of
these compounds are wax esters composed of fatty acids and fatty alcohols, hydrocarbons
and derivatives, long-chain fatty acids, and long-chain fatty alcohols [ 1081.
C. MonomericAromaticCompounds
Phenylpropanoid monomers are distributed widely in plants, including woody plants. Con-
iferin (124) and syringin (125) (Fig. 18), aglycons of which are precursors of lignins and
lignans, were isolated from many woody plants [109,110].
A soil bacterium, Agrohacteriumtumefaciens, can initiate the neoplastic disease
called crown gall on dicotyledonous plants. Tumor-inducing plasmid of the bacterium is
used to construct a useful vector to produce transgenic plants. Virulence gene expression
Chemistry of Extractives 233
OH
R=E
Coni feryl OH
alcohol 14 R=H
R=OCH3 Coniferin 124
Sinapyl R=OCE3
alcohol 123 Syringin 125
Umbelliferone Aesculetin
130 131
R
Urushiol 132
R=ClSH25-31
FIGURE 18 Examples of monomeric aromatic compounds.
of the bacterium is activated by coniferyl alcohol (14) and sinapyl alcohol (123) [ l I l ]
(Fig. 18) as well as acetosyringone (3’,5’-dimethoxy-4’-hydroxyacetophenone) [ 1 121.
Various phenylpropanoid monomers are components of essential oils, for example
(Fig. 18), anethole (126) (star aniseed oil), eugenol (127) (clove oil), safrole (128) (sas-
safras oil), and cinnamaldehyde (129) (cassia oil), which have been used for spice and
perfume.
Coumarins constitute another class of phenylpropanoids which have 2H- 1 -benzo-
pyran-2-one structure. They are distributed widely in plants, particularly Umbelliferae and
Rutaceae [ 113- 1 151. Two examples, umbelliferone (130) and aesculetin (131), are shown
in Fig. 18. Biological activities of coumarins are reviewed in the literature [37].
Urushiol (132) (Fig. 18) is a major component of urushi (Japanese lacquer), milky
exudate from urushi (Rhus vernicijlua), which has been used to produce japan (Japanese
lacquer ware) [ 116,1171.
234 Urnezawa
"+,/
OCH3 N
H
/
Berberine 133 Quinine 134 -tine ,I
135 OCH3
OCH3
H% H3COys
0 OH
9 H
' HH0 H 3 c 0H 3 H"
~ C: Ow
0 OCH3
OCH3
Yohimbine
Strychnine
Reserpine OCH3
136
Camptothecin
D. Alkaloids
Although the occurrence of alkaloids is much less than other wood extractives, most of
them are of considerable interest due to their biological activity.
Examples of alkaloids from woody plants are as follows [37,118- 1201 (Fig. 19):
berberine (133) from Phellodendron arnurense, which is antibacterial; quinine (134) from
Cinchona pubescens, which is employed for the treatment of malaria; emetine (135) from
Cephaelisipecacuanha, which is anamebicide;yohimbine (136) from Pausinystalia
yohimbe, which is a selective inhibitor of the presynaptic a-2-adrenergic receptors; strych-
nine (137) from S t r y c h n o s n u - v o m i c a , a very toxic alkaloid; reserpine (138) from Rau-
wolfia serpentina, which is antihypertensive; and camptothecin (139) from Camprotheca
acurninata. which is antitumor.
Extractives of wood influence various properties of wood, e.g., color, fragrance, and du-
rability. Some extractives have injurious effects on human health[89]. Troubles in pulping
processes and adhesion in production of wood-based materials are sometimes due to ex-
tractives. They are described in detail in a book [ 1211 and outlined byKaiin the first
edition of this book [122]. Therefore, several topics in these fields are described briefly
in this section.
Chemistry of Extractives 235
A. AcousticProperties
Recently, acoustic properties and internal friction (loss tangent, tan S) of several woods
and a cane have been observed to be strongly influenced by extractives. The term tan S
is an indication of decrement of vibration of solid materials, and materials with lower tan
S exhibit higher sound radiation.
Methanol extraction of heartwood specimens of western red cedar (Thuja plicata),
which is usedfor the top plate of the guitar, increased tan S values by 15.3-36.9%
[ 123,1241. The same effect of methanol extraction was observed for rosewood (Dalbergia
spp.), black cherry(Prunus serorina), and padauk (Pterocarpus indicus),whereas no effects
of heartwood contents or methanol extractives were detected in bubinga (Guibourtia de-
meusei) [ 123,1241.
Pernambuco (Guilandinaechinata syn. Caesalpiniaechinata) andcane (Arundo
donax) have been used for violin bows and reeds of woodwind instruments, respectively.
Water extracts from pernambuco reduced tan S value [ 1251, while in the case of air-dried
cane, water extracts increased tan S value [126]. It was suggested that the water extracts
of the cane consisted mainly of oligoglucans [126].
B. Tannin-ProteinComplexation
The ability to complex with proteins is one of the most important features of tannins, and
the mechanisms for tannin-protein complexation have long been studied. These studies
clarified effects of environmental factors (such as pH, temperature, and ionic strength) on
the process, and the following three principal features of tannin structure and properties
which are important in the complexation with protein have been established: molecular
size, conformational flexibility, and water solubility of the tannin. [54,94,127-1291. The
precipitation ability increases with an increase of the degree of polymerization of con-
densed tannin [130], while in the case of hydrolyzable tannin this ability is enhanced with
the addition of galloyl ester group [ 13 l]. Selective interaction of the condensed tannin
from Sorghum bicolor with various proteins was also investigated, and it was reported
that proline-rich and flexible proteins have high affinity for the polyphenol [ 1321.
As for the mode of interaction between tannins and proteins, involvement of hydro-
gen bonding and hydrophobic interaction was suggested [ 1271. Although the relative im-
portance of these two types of interactions remains uncertain, the initial association of
protein-polyphenol by hydrophobic interaction followed by reinforcement of the associ-
ation by hydrogen bonding was proposed [ 1271.
Since tannins have complicated structures, synthetic tannin models which have def-
inite structures are useful to elucidate mechanisms of tannin-protein complexation. Based
onexperimentswithaseries of synthesizedcondensedtanninmodels, the distribution
pattern of the phenolic hydroxyl group in the tannin molecule, but not the existence of o-
dihydroxyphenyl groups, was found to be important for higher protein-precipitating ca-
pacity [ 133,1341.
Two mechanisms were proposed for tannin-protein co-precipitation [94,127]. One
is a “cross-linking mechanism,” in which one tannin molecule binds more than two protein
molecules simultaneously to form an aggregate to be precipitated. The other is a “two-
stage precipitation mechanism,” which consistsof initial complexation of tannin molecules
to a protein molecule to form a complex, followed by aggregation of the complexes to
give precipitates. Recently, a series of hydrolyzable tannin models has been synthesized,
and studies withthe models suggested that the two-stage mechanism is involved in tannin-
protein co-precipitation [129,135,136].
236 Umezawa
Pinus taedcl and 50% Pinus virginiuna) with the fungus reduced ether-extractable pitch
components by 22% compared with chips seasoned naturally for 2 weeks. GC-MS analysis
of pitch components in fresh chips and in fungus-treated chips showed significant declines
in the concentrations of triglycerides, fatty acids, and resin acids [ 1531.
REFERENCES
1. N. Migita, Y. Yonezawa, and T. Kondo (eds.), Wood Chemistry (in Japanese), Kyoritsu Shup-
pan, Tokyo ( 1 968).
2. T. Higuchi, Biochemistry and Molecular Biology of Wood, Springer-Verlag, Berlin (1997).
3. C. B. S. Rao (ed.), Chemistry of lignans, AndhraUniversityPress,AndhraPradesh,India
( 1978).
4. W. D. MacRaeandG. H. N.Towers, Phytochemistry,23: 1207 (1984).
5. D. C. Ayres and J. D. Loike. Lignans:Chemical,BiologicalandClinicalProperties, Cam-
bridge University Press, Cambridge, U.K. (1990).
6. R.D.Haworth, Ann. Rep. frog. Chem., 33: 266 (1936).
7. 0. R.Gottlieb, Phytochernistry, 11, IS37 (1972).
8. 0. R. Gottlieb, Fortscht:Chem. Org. Nuturst., 35, 1 (1978).
9. T. Umezawa, in Biochetnistr?:andMolecularBiology of Wood, (T. Higuchi, ed.), Springer-
Verlag, Berlin, pp. 181 - 194 ( 1997).
IO. 0. R. Gottlieb, in Chemistry of lignuns (C. B. S. Rao, ed.), Andhra University Press, Andhra
Pradesh, India, pp. 277-305 ( I 978).
1I. T. Umezawa,L. B. Davin,and N. G.Lewis, Biochem.Biophys.Res.Cotnmun.. 171: 1008
(1990).
12. A. Chu, A. Dinkova, L. B. Davin, D. L. Bedgar, and N. G. Lewis, J . B i d . Chern., 268: 27026
( 1993).
13. P. W. Park,H.-B.Wang,L. B. Davin, and N. G. Lewis, Tetrahedron Lett., 35: 4731 (1994).
14. A. T. Dinkova-Kostova, D. R. Gang, L.B. Davin, D. L.Bedgar, A. Chu, and N. G. Lewis,
J. B i d . Chern., 271: 29473 (1996).
IS. L. B. Davin, H.-B. Wang, A. L. Crowell, D. L.Bedgar, D. M. Martin, S. Sarkanen, and N.
G. Lewis, Sciencc~,275: 362 (1997).
16. T. Umezawa and M. Shimada, Biosci.Biotechnol.Biochem.. 60: 736 (1996).
17. S. Suzuki, T. Umezawa,andM.Shimada, Biosci.Biofechnol.Biochem., 62: 1468(1998).
18. J . D. Orr and D. G. Lynn, Plant Physiol., 98: 343 (1 992).
19. T. Katayama and Y. Kado, J . Wood Sci., 44: 244 ( 1998).
20. T. Umezawa, MokuzaiCakkaishi. 42: 91 1 (1996).
21. L. U. Thompson, M. M. Seidl, S. E. Rickard, L. J. Orcheson, and H. H. S. Fong, Nutrition
and Cancer, 26: 159 ( I 996).
22. T. Y. Shen, S.-B. Hwang, M. N. Chang, T. W. Doebber, M.-H. T. Lam, M. S. Wu, X. Wang,
G. Q. Han, and R. Z . Li, Proc. N r d . Acad. Sci. USA, 82: 672 (1985).
23. A. M.Clark, F. S. El-Feraly, and W.-S. Li, J . Pharmaceut. Sci., 70: 951 (1981).
24. K. Watanabe,H.Watanabe, Y. Goto, M. Yamaguchi, N. Yamamoto, and K. Hagino, Plantu
Med., 49: 103 (1983).
25. D. A. Whiting, Naturcrl Prod.Rep., 4 : 499(1987).
26. D. A.Whiting, NaturalProd. Rep., 7 349(1990).
27. R. S. Ward, Nuturd Prod. Rep., 12: 183 (1995).
28. T. Nikaido, T. Ohmoto, H. Noguchi, T. Kinoshita, H. Saitoh, and U. Sankawa, Plantcl Med.,
43: 18 (1981).
29. J. Kinjo, J. Furusawa,and T. Nohara, Tetrahedron L e f t . , 26: 6101 (1985).
30. A. J . Birch and A. J. Liepa, in Chemistr~y ofLignans (C. B. S. Rao, ed.), Andhra University
Press, Andhra Pradesh, India, pp. 307-327 (1978).
3I . Y. Kai, H. Kuroda, and F. Teratani, Mokuzai Gakkaishi. 18: 3 1 S ( I 972).
238 Umezawa
98. T. Matsuo and S. Ito, Agric. Biol. Chem., 42: 1637 (1978).
99. T. Mitsunaga, I. Abe, M. Kontani, H. Ono, and T. Tanaka, J. Wood Chem. Technol., 1 7 327
( 1997).
100. A. L. Waterhouse, Chem. Ind.. 1995: 338.
101. I. Mila, A. Scalbert, and D. Expert, Phytochemistry, 42: 155 1 (1996).
102. A. Pizzi, in Polyphenolic Phenomena (A. Scarbert, ed.), INRA, Paris, pp. 267-274 (1993).
103. Y. Yazaki, Proc. 18th Symp. on WoodAdhesion, Tokyo,pp. 36-48 (1997).
104. H. L. Hergert, in AdhesivesfromRenewableResources,ACS Synzp. Ser: 385 (R. W. Hem-
ingway and A. H. Conner, eds.), American Chemical Society, Washington, DC, pp. 155-171
(1989).
105. R. W. Hemingway and P. E. Laks (eds.), Plant Po/yphenols, Plenum Press, New York (1992).
106. J. N.BeMiller,in NaturalProducts of WoodyPlants I , (J. W. Rowe, ed.), Springer-Verlag,
Berlin, pp. 155-1 78 (1989).
107. T. E. Timell, Wood Sci. Technol., I : 45(1967).
108. P. E. Kolattukudy and K. E. Espelie, in NaturalProducts of WoodyPlants I (J. W. Rowe,
ed.), Springer-Verlag, Berlin, pp. 304-367 (1989).
109. M. Terazawa, H. Okuyama, and M. Miyake, Mokuzui Gakkuishi, 30: 322 (1984).
110. M. Terazawa and M, Miyake, Mokuzai Gakkaishi, 30: 329 (1984).
1 1 1. P. A. Spencer and G. H. N. Towers, Phytochemistry, 2 7 2781 (1988).
112. S. E. Stachel, E. Messens, M. van Montagu, and P. Zambryski, Nature, 318: 624 (1985).
113. R. D. H.Murray, Forrschr: Chem.Org.Nuturst., 35: 199 (1978).
114. R. D. H. Murray, Nuturul Prod.Rep., 12: 477, (1995).
115. 0. Theander and L. N. Lundgren, in Natural Products of Woody Plunts I (J. W. Rowe, ed.),
Springer-Verlag, Berlin, pp. 369-399 (1989).
116. W. E. Hillis, in NaturalProducts o f WoodyPlants I (J. W. Rowe,ed.), Springer-Verlag,
Berlin, pp. I - 13 ( 1989).
117. W. E. Hillis, Heartrt'ood and Tree Exudates, Springer-Verlag.Berlin, pp. 59-60 (1987).
1 1 8. C. W. W. Beecher, N. R. Farnsworth, and C. Gyllenhaal, in Nuturul Products of Woody Plants
II (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 1059-1 164 (1989).
119. S. I. Sakai, N. Aimi, E. Yamanaka, and K. Yamaguchi, in Natural Products of Woody Plunts
I (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 200-257 (1989).
120. S. Tachibana, Mokuzai Gakkaishi, 41: 967(1995).
12 I . J. W. Rowe (ed.), Natrtrd Products of Woody Plants. I and 11, Springer-Verlag, Berlin (1989).
122. Y. Kai, in Wood andCellulosicChemistry (D.N.-S. Hon and N. Shiraishi,eds.),Marcel
Dekker, New York, pp. 2 15-255 ( 199I ).
123. H. Yano, Holiforsch.,48: 491 (1994).
124. H. Yano, H. Kajita and K. Minato, J . Acous?. Soc. Am., Y6: 3380 (1994).
125. M. Matsunaga, M. Sugiyama, K. Minato, and M. Norimoto, Holiforsch., "TO: 51 1 (1996).
126. E. Obataya, T. Umezawa, F. Nakatsubo, and M. Norimoto. Holz,forsch., 53: 63 (1990).
127. C. M. Spencer, Y. Cai, R. Martin, S. H. Gaffney, P. N. Goulding, D. Magnolato, T. H. Lilley,
and E. Haslam, Phytochemistry, 27: 2397 (1988).
128. G. Luck. H. Liao, N. J . Murray, H. R. Grimmer, E. E. Warminski, M. P. Williamson, T. H.
Lilley, and E. Haslam, Phytochemistry, 37: 357 ( 1994).
129. H. Kawamotoand F. Nakatsubo, Recent Res. Devel. Phytochem., I : 515(1997).
130. E. C.Bate-Smith, Phytochemistry, 12: 907(1973).
13 I . E. Haslam, Biochem. J.. 139: 285 (1974).
132. A. E. Hagerman and L. G. Butler, J. B i d . Chenl., 256: 4494(1981).
133. H. Kawamoto, F. Nakatsubo,and K. Murakami, J. Wood Chenz. Technol.. 1 0 59 (1990).
134. H. Kawamoto, F. Nakatsubo,andK.Murakami, J. Wood Chem. Technol.. l o : 401 (1990).
135. H. Kawamoto, F. Nakatsubo, and K. Murakami. Ph?,toc/lerni.stt:v,40: 1503 ( 1995).
136. H. Kawamoto, F. Nakatsubo,and K. Murakami, Phytochemistry, 41: 1427 (1996).
137. G.Appendino, Nrtturul Prod.Rep.,12: 349 (1995).
138. N. Vidensek, P. Lim, A. Campbell, and c. Carlson, J . Nuturul Prod., 53: 1609 (1990).
Chemistry of Extractives 241
Chemistry of Bark
Kokki Sakai
Kyushu University, Fukuoka, Japan
1. INTRODUCTION
Tree bark usually refers to all tissues external to and surrounding the vascular cambium.
It occupies a much smaller volume than the wood of a mature tree stem because fewer
bark cellsare produced than wood cells and also because the outermost bark cellsare
continuously discarded in most tree species, while wood cells are retained and thus ac-
cumulate as the tree grows.
In spite of its small volume, the bark plays important roles in a living tree. Tree
barks often have developed complex anatomy and/or chemical compositions in order to
manifest or maintain three main functions: (1) nutrient transport from the leaves to the
rest of the tree, (2) protecting the sensitiveinnercambiumfromdesiccation,and (3)
shielding from the environment as the primary defense of the tree against wildfire, me-
chanicalinjuriescaused by heavy wind, and attacks by phytopathogens,phytophagous
insects, larger animals, and so on.
The objective of this chapter is to provide the new information about the chemical
composition and utilization of tree barks, mainly since 1986. The basic knowledge of bark
chemistry and important findings in this field up until 1985 were well stated by Laks in
the first edition of this book [ 11.
The formation and anatomy of bark are described very briefly, as quite detailed discussions
of them were presented by Laks [ l ] .
The epidermis. the cortex, and the primary phloem are produced during longitudinal
growth from the apical meristem located at the apices of growing roots and branches. The
secondary phloem and periderm are formed during radial growth of the tree. Accordingly,
the primary phloem remains in only the outermost region of the bark of young trees. In
most tree species the outer bark cracks and peels off as the tree grows, due to the suc-
cessive formation of periderms within the bark and growth of the xylem. Thus the mature
bark consists mainly of secondary phloem and periderm, a group of tissues including the
phellogen or cork cambium, the phelloderm, and the phellem or cork tissue. As the sec-
243
244 Sakai
ondary phloem thickness, new periderm is formed within the phloem and any cells to the
exterior of the periderm soon die.
Most tree barks, therefore, have two zones, the inner bark that contains some living
cells and the outer bark or rhytidome that does not contain any living cells. These regions
are sometimes compared to the sapwood and heartwood in the xylem, respectively. Re-
cently, Trockenbrodt [2] surveyed and discussed the terminology used in bark anatomy,
andsuggestedterms for the tissue zones as illustrated in Fig. 1 . In the barkchemistry
field, however, the terms “outer bark” and “inner bark” have been used for “rhytidome”
and “secondary phloem up to the last-formed living secondary phloem,” respectively.
As described above, bark consists of the inner bark and outer bark zones. It is therefore
preferable to determine the chemical composition accuratelyfor these two zones separately.
However, chemical analyses have often been made with the whole bark because separation
between the zones is sometimes cumbersome and time consuming, and because for most
applications the whole bark is to be utilized. It has been known that chemical analyses by
standard methods for wood often give erroneous results due to the presence and variability
of suberin and high-molecular-weight tannins which are rarely or never found in wood.
Surprisingly, some erroneous data were published evenin 1991 [ 3 ] .Preliminary extractions
or corrections for these substances are very important for the accurate analysis of the bark
of trees.
Some of the general and comprehensive reviews onthe chemical composition of tree
barks were referred to by Laks [ l ] . More recent examples of bark analysis are listed in
Table 1 to show how different the chemical compositions of the tree bark are from those
of wood. In general, bark contains much more extractives, slightly less lignin, and smaller
amounts of holocellulose as compared to wood of the same tree.
Nonextractable components in the bark consist of polysaccharides (cellulose, hemi-
cellulose,and pectic substances),phenolicpolymers (lignin andhigh-molecular-weight
tannins), and cross-linked polyesters (suberin and cutin). Significant parts of hemicellulose
/- rhytidomc
secondary phlocm u p to
thc last fonncd pc’idclnd
living sccondaryphlocm
calnbiunl
xylcm
FIGURE 1 Suggested tcrms for the tissue zones resulting from rhytidomc formation. Rcproduccd
with permission of IAWA.
Chemistry of Bark 245
Components
1 % NaOH
Species Extract.
Lignin
Ash Holocellul. Cellulose Pentosan soh. Ref.
and pectins can be removed during the extraction of tannins with alkaline solutions. This
solubility also causes the difficulty in accurate analyses of the bark components. Recent
studies on nonextractables in the bark are very limited.
A lipidic biopolymer, suberin, is the most important structural component of the cork
cell walls, which play the physiologically important role of thermal and hydric insulation
in bark tissues. Suberin composition of Cedr-us libani bark was studied, and it was noted
that a,w-dibasic acids and w-hydroxy-monobasic acids were the two dominated by hexa-
decane- I , 16-dioic acid and 16-hydroxyhexadecanoic acid as the major constituents in the
suberin monomers of the bark of this species [ 1 l ] . Besides these acids, smaller amounts
of long-chain primary alcohols, alkanoic acids, and ferulic acid were observed in C. libani
bark, as are often observed with other species. Thus, the composition of the suberin mon-
omer of the C. libnni bark was quite different from that of spruce and pine, where unsat-
urated C,,-acids (hydroxyoctadecenoic acid and octadecenedioic acid) were major mono-
mers [ l ] . There were trials to isolate suberin from the bark without chemical action. Perra
et al. extracted lignin and suberin from bark of Fngus sylvaticcr milled in a vibratory ball
246 Sakai
IV.BARKEXTRACTIVES
Generally, bark contains much larger quantities of extractives than wood of the same tree,
as seen from Table 1. Bark chemical components can be roughly fractionated by a se-
quence of extractions starting with a nonpolar solvent, such as petroleum ether, in many
instancesfollowed by diethyl ether, ethanol, and water. Lesspolarcompoundssuchas
waxes, resin, lipids, higher fatty acids, phytosterols, and terpenes are extracted mainly with
non- or less polar solvents. On the other hand, ethanol and water dissolve relatively polar
substances, such as flavonoids, phenolics, their glycosides, condensed tannins, sugars, etc.
Acetone-water mixture is the preferred solvent for extraction of condensed tannins [ 15b].
Ethyl acetate is sometimes used for extraction of low-molecular-weight tannins or oligo-
meric proanthocyanidins. Aqueous NaOH solutions (e.g., 1%) have been often used for
extraction of polymericcondensed tannins. However, attention should be given to the
alkaline rearrangement resulting in partial or total rearrangement of tannins of the pro-
cyanidin class to “phenolic acid,” which have less phloroglucinol functionality [ 161. Sa-
ponification of ester bonds in cutin and suberin can also occur in aqueous alkali [ 141.
The yields of sequential extracts from barks were compiled as tables by Laks [ I ] for
various softwood tree species: Pseudotsuga menziesii, Pinus tueda, P. echinattr, l? virgi-
niana, P. elliottii, P. sylvestris, P. brutia, l? btuksiana, P. corltortu, P. nigru, P. hulepensis,
l? radiata,Piceueizgelnmnni, P. abies,Larixkaempferi, L. eurolepsis, L. deciducr,Abies
balsurrtea, and Tsuga canadensis. Also compiled are the yields from barks of some hard-
wood species: Bruguierugymrtwrrhi:a.Rhizoplzera stylosa, Fugus sylvutica, Betula ver-
I ’ U C O S ~ I ,B. allegl~uniensis,B. papyrifera, A1rur.s rubra, Populus tremuloides, Quercus d b a ,
and Q. rubra. Extractives yields recently published for barks of some species are listed in
Table 2.
The extractives content can vary according to a number of factors, including envi-
ronmental, genetic, and seasonal variations. Seasonal fluctuation of extractive contents was
investigated with barksof Quercus rohur; Piceacrbies, and Pirius.sdvestris grown in
Germany [ 171. No season-related differences were observed in the amount of cyclohexane
extract. Amounts of ethanol extract, NaOH cxtract,and water extract increased to different
Chemistry of Bark 247
extents from spring to winter. Of the three extracts, an increase in ethanol extract content
was the most extensive. The extractives yield also depends on the aging between felling
the tree and carrying out the analysis. Tree age influences the quantity and quality of bark
extractives. In freshly cut P o ~ ~ u l u s t r m u r l o i d e sthe
, higher up the tree,thegreater the
amounts of bark extractives, probably due to the fact that the more inner bark regions are
towards the top of the tree [ 201. Young Pinus rtrdiata bark (12 years old) provided ex-
tractives in a significantly higher yield (32.9%) than that (29.4%) from 30-year-old bark
of the same species [2l]. In the case of Rhizoplzoru mucmmta, however, barks of older
trees gave higher extractives and tannin contents [22].
AcO RA
FIGURE 4 Structures of taxol and some taxane diterpenoids obtained from bark of Erxus spp.
R‘
Margolonc R=HZ,R’=Mc,R”=COOH
Margolonone R=O, R=Mc, R”=COOH
lsomargolononc R=O, R’=COOH, R”=Me
FIGURE 6 Diterpenoids from the stem bark of Azcldirachfn indica.
which is regarded as a bitter tonic, astringent, and as being useful in fever, thirst, nausea,
vomiting, and skin diseases [37]. These diterpenoids possessed different antibacterial ac-
tivities against Klebsiella oxytoca, Staphylococcus epidermidis, and Serratia lutea. Super-
critical carbon dioxide was used to extract cis-abienol efficiently from the bark of Abies
saccharinensis [38]. This extract has a tobacco odor and oviposition-deterrent activity to
some aphids and worms.
Betulinand related triterpenes with the lupaneskeletonarecommon in barks of
Betula species. Theirstructures are shown in Fig.7. However, their contents are very
different from species to species, as shown in Table 5 [ 1,39-411. White-barked birches,
B. verrucose in Europeand B.plutiphyllcr var. japonica in Japan,contained betulin to
about 25% of their outer bark weight. On the other hand, much less betulin was observed
in the barks of dark-barked birches, B. ermani in Japan and B. nigrll in North America.
Especially in the latter species, betulin and lupeol were no longer the predominant triter-
pene components, and 3-O-acetyloleanolic acid predominated [40]. Triterpene contents in
inner bark and root bark were very much smaller than those in outer bark. It is noteworthy
that all the triterpenes isolated were caffeoylated in the root bark of B. ermcmii [41]. Betulin
oligosaccharides were enzymatically synthesized by using cyclodextrin glycosyl transfer-
ase withan aim to increase its biological activity [42].Unsaponifiables in the barks of
Fclgus crenata, Cyclobanopsis glauca, and Castanopsiscuspiciata wereclaimedtobe
responsible for “pitch trouble” during kraft pulping process, as lots of pitch specks were
observed on pulp sheets prepared from wood chips contaminated with their barks. These
barks have larger contents of unsaponifiables than wood 1431.
B. Low-Molecular-WeightPhenolics
Laks [ l ] dealt withlow-molecularphenolicextractivesincluding flavonoids, lignans,
phenylpropanoids,andothersimplephenolicsfound i n tree barks until 1985and cited
somereviewsandbooksconcerningthem. In this section onlymore recent results are
referred to.
Phenolic compounds are present in relatively large amounts in coniferous tree barks.
Thebarks of somehardwoodspecies,such as Q w m l s , Euctrlyptus, Accrcicl. and Soli.v,
also contain large amounts of phenolic extractives, as exemplified in Table 6, while some
of otherhardwoodspecies, Betulcl spp., for instance, containonly negligible amounts
of phenolic extractives in their outer barks. Itis likely that phenolic constituents i n the
252 Sakai
CH,OH
CH 2OH
CH0
CO2H
CO2Mc
CO ?H
C02Mc
CO 2H
CH3
CH 3
CH3
CH0
AC: CH,C&
inner bark and root bark are largely glycosylated so that they can be translocated in the
phloem sap.
“0,1, and R denote outer bark, inner bark, and root bark, respectively.
hIncludes methyl oleanolate and 3-0-acetyloleanolic acid.
revealed the presence of gallic, protocatechuic, vanillic, and ellagic acids, and protocate-
chuic aldehyde, together with some flavonoids and ellagitannins [46].Some ellagic acid
derivatives, methylated and/or glycosylated, were isolated from the stem bark of Diplo-
panax stachvanthus [47].
The bitter-tasting bark of Prunus gra-yancl contains phenylethanoid (C,-C,) glucosides
and their caffeate esters, together with a tannin-related 3,4,S-trimethoxybenzoyl-glucoside.
The latter two had a strong bitter taste [48].Arylpropanoids (C,-C,) biologically synthe-
sized from the amino acid phenylalanine are very common in the plant kingdom and are
constituents of the naturally important phenolic polymers, lignins and condensed tannins.
Arylbutanoid (C,-C,) glycosidesareknowntobe present in the inner bark of Betula
pendula 1491 and in the stem bark of A c e r nicoense [SO]. Bioactive styryl-lactones, gon-
iofufuranone, goniopypyranone, and 8-acetylgoniotriol (Fig. 8) were isolated from the stem
bark of Goniothalarnus giganteus [Sl]. Thesecanberegarded as phenylheptanoids
(C,-C,).
Isorhapontin, one of the stilbenes (C(,-C2-C,),was found to be present in very high
content (13%) in the bark of Picea engelmanni. This content was dramatically different
I
254 Sakai
from those in other softwood barks from five genera, suggesting isorhapontin to be a good
taxonomic marker [S2].
Phenyl propane dimers (C&,-C$,) linked at the P-positions are referred to as lig-
nans, and ones linked at a position different from the P are sometimes called neolignans.
A book of lignans in general [S31 and a review of lignans from woody plants [S41 was
published recently. In the root bark of A c c r n / / ~ r ~ ~ ~ r rsrn/ico.su.s,
nc~x an empiricaloriental
medicinal plant. five lignans, two coumarins, and three other phenolic compounds were
identifed. Seven of these compounds were glycosides and one was caffeate. The predom-
inant constituent was syringaresinol diglucoside 1551. A neolignan rhamnoside and a lignan
xyloside were isolated from the inner bark of Ostrya japonica [57]. Betula ermanii inner
bark contains two lignan glycosides and a dicaffeoylated lignan [41]. Contents of lignans
were compared between the wood and bark of Magnolia kobus var. borealis, which has
been used as a material for Oriental medicine blends [58].The bark contained much larger
amounts of lignans, especially permethylated lignans, magnolin, and yangambin. The bark
of Ocotca catharinensis contains I O neolignans(average yield 0.07%) of thebicy-
clo[3.2. lloctanoid and the hydrobenzofuranoid structural types [59]. These neolignans are
also present in much larger contents in the bark than in wood of the same species.
A series of diarylheptanoid (C,-C,-C,) glycosides together with the aglycon, platy-
phyllone, was identified in the inner bark of Betula pendula [49], and a diarylheptanoid
glycoside platyphylloside was found to be present at high levels in B. pendula (20-60
mg/g dry bark) but at low levels in B. pubescens (<OS mg/g drybark). Based on this
observation, a simple precipitation method to distinguish between these species, which are
hard to distinguish morphologically, has been developed [60]. Quite large yields of cyclic
diarylheptanoids, myricanol, and myricanone, in addition to tannins described below, were
obtained from the bark of Myrica esculenatea or the hairy bayberry tree common in south
China [61].
Glucosides of hydroquinone derivatives were found in the fresh barks of Quercus
acutissima [63], Mallotus japonica [64], and Ostrya japonica [56]. Naphthoquinones, is-
odiospyrin and plumbagin from the barks of Diospiros trees in Japan, exhibited antifungal
activities against some phytopathogenic fungi [62].
R5 R5
Catechin(CAT) 2R :3s H OH OH OH OH OH H
Epicatcchin (ECT) 2R:3R OH H OH OH OH OH H
ent-Catechin (entCAT)* 2S:3R OH H OH OH OH OH H
Galloatechin (GCT) 2R : 3 s H OH H OH OH OH OH
Fisetinidol (FIS) 2R :3s H OH OH OH OH OH H
Robinctinidol (ROB) 2R :3s H OH OH OH OH OH OH
256 Sakai
quercetin,myricetin,and their glycosides found in the bark. After that, some excellent
reviewsandbooksonflavonoids and related compoundswerepublished[15b,66,67],
though they are not confined to the bark components. The flavan-3-01s catechin and epi-
catechin are very commonly present in the bark, probably as precursors to and by-products
of the biosynthesis of condensed tannins. Flavan-3-01 glycosides were thought to be rare.
However, some were recently found in the inner bark of different species, for example,
catechin 3-O-a-~-rhamnopyranosidein Quercus miyagii bark [68]. catechin 3-o-p-D-gh-
copyranoside and 3-O-[au-~-rhamnopyranosyl-( 1 +6)-P-~-gulucopyranosyl]-catechin in
Quercus marilandica bark 1691, and catechin 7-O-~-D-glucopyranosidein the inner barks
of Betula pubescens 1701 and Pseudotsugu menziessii [71]. The latter also contained epi-
catechin 7-O-p-D-glucopyranoside, catechin 4'-O-P-~-glucopyranoside, and 3'-O-methyl-
epicatechin 7-0-P-D-glucopyranoside as flavan-3-01 glycosides [71]. The position 3 of the
flavanols is sometimes galloylated, and secretagogue activity of 3-0-galloyl-epicatechin
was reported [72]. Unstable gallates of catechin esterified at the 3',4' and/or 7 positions
were recently found in Acacia gerrardi bark (731. Gallic acid, catechin, and epicatechin
showed inhibition of chemiluminescence production by the activated human polymorpho-
nuclear leukocytes, which is a measure of oxidative burst [74].
There has been an expolsive growth in the understanding of the chemistry of plant
proanthocyanidins over the past two decades, due to the development of techniques in
chromatographic isolation andspectrometric structural analysis, including 'H- and "C-
NMR and MS [66,75].Manyproanthocyanidindimersand trimers and sometetramers
have been isolated from barks of many species in recent years. The nomenclature system
commonly accepted for the proanthocyanidins is outlined as follows 1761: (1) The fun-
damental structural units are defined in terms of the monomeric flavan-3-01s. The names
catechin, epicatechin, fisetinidol, etc., are reserved for units with 2R absolute configuration.
Flavan-3-01 units with the 2s configuration are distinguished by the "ent" (namely, en-
antio) prefix. ( 2 ) The interflavonoid bond at C-4 is indicated by an arrow and its config-
uration by the ap nomenclature in parentheses, as (4p -+8). Structures of some proan-
thocyanidin oligomers are shown in Fig. 10. Their flavanol units are often linked by (4a
-+ 8) bond in 2R: 3 s units, such as catechin, or (4p + 8) in 2R:3R units, such as epi-
catechin. In some cases there are ( 4 a + 6) or (4p + 6) bonds, too. Some of them have
0-acyl (galloyl usually) or 0-glycosyl bond at position 3. Basedon the structuresof
proanthocyanidin oligomers. structures of polymeric proanthocyanidins or condensed tan-
nins have been constructed.
Some of the proanthocyanidin oligomers have physiological or biological activities.
The bark of Psrudotsuga menziessii contains an epicatechin tetramer (compound 5 in Fig.
10) that is a potential inhibitor of rat brain protein kinase C as well as bovine heart cyclic
AMP-dependent protein kinase [77]. From the bark of Byrsonima crassifoliu, which hsa
beenusedmedicinally by the MixeIndians to treat gastrointestinal disorders and skin
infections, a proanthocyanidin trimer (4 in Fig. 10) and related ent-epicatechin-based olig-
omers were isolated as guided by a nematodicidal activity assay [78]. The stem bark of
Stqphonodendron adstringens, a medicinal plant grown in Brazil, contains a variety of
3-0-acylated prodelphinidins (epigallocatechin-based proanthocyanidins) [79]. Antitermite
activity was observed with a low-molecular-weight proanthocyanidin fraction from Acacia
mearnsii bark, consisting of dimeric and trimeric prorobinetinidins [SO].
C. HydrolyzableTannins
In general, galloyl and the related hexahydroxydiphenoylesters of glucosehavebeen
referred to as gallotannins and ellagitannins, respectively, because they are easily hydro-
Chemistry of Bark 257
H0 H0
OH
OH
OH
1 R=H 3
2 R=Gk
H0
H0
OH
OH
4 OH
5
NO. Structure plant spccics Ref.
l CAT-(4a+8)-CAT many General
2 ECT-(4R+8)-CAT Gcncral many
3 ~-O-G/C-CAT-(~U.+X)-CAT Quercus
mutihndicu [69]
~-~-G~I-ECT-(~B+~)-~-~-G~I-C~~-ECT-(~U.+X)-
cruss4~~,iu ,7Kl
cnt-ECT
5 ECT(4lJ+8)-ECT-(4l~+X)-ECT-(4~~8)-ECT Pseudotsugu menziesii [ 771
lyzed by acid to give gallic acid and ellagic acid. A recent book and a review of Haslam
[81,82] gave an extensive discussion of hydrolyzable tannins. Description in this section
is confined to bark tannins, although tremendous numbers of hydrolyzable tannins have
been isolated during the past decade from various materials, including fruits, leaves, and
wood parts of different plant species [84].
Recently, Nonaka [83] suggested that some of the hydrolyzable tannins can be called
“complex tannins” based on their structural features, which have a flavonoid, flavanol, or
proanthocyanidin moiety in addition to the more common structural constituents, polyols
(mostly D-glucose) and the gallic acid derivatives. Quite a few novel gallotannins, ellag-
itannins, and complex tannins have been found and structure-determined in the barks of
Quercus acutissima [85,89], Q. miyagii [85,89], Q. stenophylla [85,89], Q. rnongolicu var.
grosseserrata [85,89,90,86], Q. denata [89], Q. petraea [91], Castanea crenata [85], Cas-
tanopsis cuspitada var. sieboldii [89], Cercidiphyllurn japonica [92], Mallotus japonicus
[93,94,96], M. philippinensis [96], Pterocatya streptera [95], Eucalyptus viminalis [95],
Hammamelis virginicrna [88], Anogreissus acurninata var. lanceolata [87], Psidium gua-
juva [97], and Platycarya strobilacea [98]. A few examples of hydrolyzable tannins are
shown in Fig. 11.
It is noteworthy that the configurations at the glucose C-l position of representative
C-glycosidic ellagitannins, 1S for castalagin and 1R for vescalagin [99], were revised to
the opposite configurations, 1R and lS, respectively, on the basis of NMR spectroscopy
[ 1001. Accordingly, the configurations at this position must be revised for all C-glycosidic
tannins known before 1990, including ones in Haslam’s book [81] and stenophyllanin B
shown in the first edition [l].
D. Condensed Tannins
Condensed tannins distribute more widely than hydrolyzable tannins in the plant kingdom
and are mostly polymeric proanthocyanidins with structures closely related to the oligo-
meric structures depicted in Fig. 10. Somebooksandreviewsdealingwithcondensed
tannins have been published recently [ 15b,81,101- 1031. The principal types of condensed
tannins and variations in their hydroxylation patterns of the A- and B-rings are shown in
Fig. 12. Their chemical characteristics are strongly influenced by hydroxylation patterns
of the A-ring; thus procyanidins and prodelphinidins with a phloroglucinol type A-ring
tend to react more easily than profisetinidins and prorobinetinidins having the resorcinolic
A-ring, with both nucleophilic and electrophilic reagents. Molecular weight or degree of
polymerization (DP) of these compounds also affects their nature; thus condensed tannins
of low DP are soluble in water and some polar solvents, whereas high-DP tannins are not
extractable in neutral solvents. Large parts of polyphenols in the conifer barks are solu-
bilized by alkaline extraction following neutral solvent extractions. The extracts formerly
called “phenolic acids” are now known to be artifacts produced from rearrangements in
high-DP procyanidins in the presence of alkali [ 1,161. In this connection, Yazaki et al.
[l041 reported the detrimental effect ofalkalinetreatmentof the extractsfrom Pinus
radiata barkon their Stiasny values, which is a reliable estimate of the polyflavanoids
reactive to formaldehyde.
Characteristics of condensed tannins from the barks of Pinus taeda, a number of
Japanese tree species, Acacia mearnsii, andsomeotherconiferousspecieshavebeen
studied as described in the first edition [l]. More recently, the structure of procyanidin-
type polymers of Douglas fir (Pseudotsuga menziessii) inner bark was studied using deg-
radation with phloroglucinol as an analytical method. The results suggest that the config-
Chemistry of Bark 259
H0
OH
H0
H0
H0
H0
OH
OH
Acutissimin A (a furavano-cllagitannin)
Quercul ucutissimu, Q. miyugii, Q. stenophyllu,
Castalagin
Q. tnongolicu var. grossesenutu, Custuneu crenutu
Miricu esculentu [61]
[ 8 5 ] ,Q.petrueu (911, andhogeissus ucurninutu
Quercus petrueu [ 011
var. lunceolutu [ 871
OH
OH OH
Mongolicanin (a procyadino-cllagitannin) l-O-(4-Hydroxybcnzoyl)-,2',5-di-0-
Quereus tnongolicu var. gro.s.serrutu [ 861 galloyl-IJ-D-hamamclofuranosc
Q.petrueu [ 91 1 Hutnutne1i.s virginiunu [ 881
uration of the extender units is almost exclusively 2,3-cis, while the terminal units are
mixed, with 2 , 3 4 7 slightly predominating. The (4+8) interflavonoid bond predominates
over the (4+6) bond by a 4: 1 ratio [ 1051.
Condensed tannins from the barks of a number of angiosperm species have been
characterized. Ohara et al. [80] studied the acetone-water extracts from the bark of Acacia
mearnsii (52.7% of dry bark) and reconfirmed the condensed tannins of this bark to be
mainly 2,3-trans-prorobinetinidinswith a relatively small average DP of 5 (poly dispersity
260 Sakai
3'
5 4
5P C pdttcm Hydroxylation
Propelargonidin 3,4',5,7
Procyanidin 3,3',4',5,7
Prodelphinidin 3,3',4',5,5',7
Proguibourtinidin 3,4',7
Profisetinidin 3,3',4',7
Prorobinetinidin 3.3'.4'.5'.7
1.36),containingsmallamountsofprocyanidinorprodelphinidinchain-extender units.
The angular structure was suggested by the coexistence of angular trimeric proanthocyan-
idins in the same bark extracts. No distinctive difference in molecular size distributions
was observed between the A. nzearnsii bark tannins from China and South Africa [ 1061.
Barks of Quercus species seem to contain both hydrolyzable tannins and condensed tannins
[85,89,91,107,108J. Tannins are supposed to be responsible for the therapeutic effects of
oak (Q. petraeu) bark, which is used against hemorrhoids, chilblains, mouth sores, and
indigestion in Europe. The angular or branched structure, with average DP of 6.1 and a
procyanidin-to-prodelphinidin ratio of 6:4, wasproposedfor the condensedtannins of
Q. perruea bark [107]. The bark of Q. fulcatu, the most important species of red oak in
the forests of the southern United States, is a rich source of quercitrin (quercetin-3-rham-
noside). It contains procyanidin-type condensed tannins made up predominantly of 2,3-cis
extender units and terminated with 2,3-truns catechin units, together with a flow concen-
tration of catechin and three major dimeric procyanidins. The tannins contain only small
amounts of 2,3-trans-procyanidin extender units and only traces of prodelphinidin units
[ 1081.
The condensed tannins of Salix spp. bark have comparatively large DPs for angio-
sperm bark tannins. Those present in S. roridu bark consist of prodelphinidin and pro-
cyanidin types of flavanol units with an average DP of about 8 and with 2,3-truns and
2,3-cis configurations in the extender units and 2,3-truns in the bottom or terminal units.
Prodelphinidin units are more abundant than procyanidin units [IO]. In the case of S. pet-
susu bark, the condensed tannins, with average DP about IO, are composed of both pro-
cyanidinandprodelphinidin types, with the formerdominating.Extender units in the
tannins have largely 2,3-cis and a small proportions of 2,3-rruns configurations [ 1091. The
bark of S. sieboldiann, a shrub common in Japan, contains a homologous series of pro-
cyanidindimersand trimers esterified with 1-hydroxy-6-oxo-2-cyclohexene carboxylic
acid at position 3 of the catechin bottom units [ 1 lo]. Three oligomeric proanthocyanidin
fractions, with average DPs of 4, 5-6, and 6-7, respectively, were found in a blood-red
sap from the slashed bark of Croton lechleri, used by South Americans for the treatment
Chemistry of Bark 26 1
"b, rb, and e denote bark, root bark, and exudate, respectively.
hFound in A. senficosus.
k?
X
m.
Chemistry of Bark 263
icines for centuries from a widevariety of forest plants, including the bark of Alnus rubra,
Tsuga heterophylla, Thuju plicata, Berberis nervosa, Rhamnus purshiana, Oplopanux hor-
ridum, Sambucus racemosu, Pyrus fusca, Prunus spp., Salix spp., and Taxus brevifolia.
Many of the medicinally used plants were rich in procyanidins and associated compounds.
Biological activities of some oligomeric and polymeric proanthocyanidins isolated from
barks were also described in the sections on “Flavonoids and Proanthocyanidins” and on
“Condensed Tannins.”
The antifungal activity of bark components has been known and the activity has
been traditionally utilized in fermentation of Philippinesugarcanewine.Apolyphenol
component obtained from an aqueous extractof the samac (Macharanga grandqolia)bark,
which gave catechin, cyanidin, delphinidin, and sugars upon acid hydrolysis, inhibited the
growth of lactic acid bacteria which caused deterioration of the fermentation mixture [ 1241.
In certain communities of southern Nigeria, the pharmacologically active aqueous extract
of the bark of Sucoglottis gabonensis is commonly used as an additive to palm wine. The
exact biological role of this extract in the beverage maturation process is not clear, but it
exhibits antioxidant properties and at least one constituent, bergenin, is an inhibitor of
yeast alcohol dehydrogenase [ 1251.
It is interesting that hot water extracts from the inner barks of three coniferous species
(Cryptomeria japonica, Chamaecyparis obtusa, and P i n u s d e n s g o r a ) showed almost no
inhibition of the mycelial growth of all the edible mushroom fungi which are wood decay
basidiomycetes, while these extracts exhibited great inhibition of two fungi that are path-
ogenic to the mushroom fungi [ 1261. Acetone extracts of a bark-compost possessed strong
antifungal activity against Fusarium oxysporum, Helminthosporium sigmoideum, Gibber-
ella zeae, etc., fungi of interest to plant pathologists, but the extracts were inactive against
yeast and procaryotic organisms [ 126bl. Mori et al. also observed that all the bark extracts
from 21 conifers belongingtosevenfamilies inhibited the growth of plant pathogenic
fungi more effectively than that of wood decay fungi [ 1271. The feature of the bark being
degraded by wood decay fungi is perhaps of important significance in the mass cycle in
nature, where dead trees are biologically mineralized by organisms including insects, earth-
worms, fungi, and bacteria. On the basis of this property of bark components, biodegrad-
able polyurethane foams were prepared from commercial wattle tannn [ 1281, the barks of
Acacia mearnsii and Ctyptomeria japonica[ 129,1301. The foams produced were gradually
degraded by both white-rot and brown-rot fungi.
However, biocidal activity of condensed tannins against wood decay fungi was in-
tensified by complexing with copper(I1) ions. The best wood-preservative effect was ob-
served with a dual treatment using a sulfited bark extract first, followed by a CuCl, treat-
ment. This method yielded wood with greater resistance to decay than wood treated with
pentachlorophenol [ I3 1 1. Condensed tannin from several tree barks exhibited enzyme-
inhibitory activity against plasmin, thrombin, papain, and trypsin. A tannin from Cedrus
deodara bark showed anti-plant-viral activities against tobacco mosaic virus, potato virus
X, and cucumber green mottle mosaicvirus [ 1321. Antimicrobial properties of tannins, not
only from bark but also from wood, leaves, fruits, and roots of both woody and herbaceous
plants, were reviewed by Scalbert [132b].
Bark extracts influence animals physiologically, too. 3-0-Acetyloleanolic acid, the
predominant triterpene component in Betula nigra outer bark, exhibited significant anti-
feedant activity against the Colorado potato beetle, Leptinotarso decemlinoata, an agri-
culturally important insect pest [40]. Some flavonoids and cerebrosides were isolated from
Quercusdentata,Eucalyptusrubida, and Prunusjamasakura as repellent compounds
against the blue mussel, Mytilus edulis, one of the gregarious fouling organismsthat cause
264 Sakai
serious problems to ships and fishing nets [ 1331. Tannins in the bark extract of Norway
spruce (Picea abies) as a model for debarking waste water were shown to be responsible
for acute and subacute toxicity in carp. Oxidative polymerization of tannins in the extract
into high-molecular-weight polymers abolished the aquatic toxicity [ 1341.
The wastewater from debarking plants is highly toxic to anaerobic microorganisms
also, especially methane-forming bacteria. It was shown that tannin, estimated by adsorp-
tion to insoluble polyvinylpyrrolidone, caused most of the methanogenic toxicity observed
in conifer bark extract [135], and they can be mostly detoxified by alkaline autoxidation
[136], hydrogen peroxide oxidation [ 1371, and enzymatic polymerization with phenol ox-
idases followed by flocculation [ 1381. The detoxified wastewatercan be treated using
conventional anaerobic methods [ 138,1391.
Swan[l401 reviewed health hazardsincluding toxic, allergenic, and carcinogenic
properties associated with extractives from woody plants.
Only recent developments in the technologies for adhesives production from bark tannins
and other recent topics of bark utilization are dealt with here. Laks [ l ] reviewed exten-
sively the utilization of tree barks as a source of adhesives, pharmaceuticals and biocides,
and other utilizations in the first edition.
A. Adhesives
Condensed tannins from the bark of Acacia mearnsii (black wattle) tree have been suc-
cessfully utilized for adhesive purpose for over two decades in South Africa. The wattle
tannin-based adhesives have been progressively displacingsynthetic phenol-formaldehyde
resins, in countries where these tannins are produced, and are consumed by the manufac-
ture of particleboard, plywood, glulam, fingerjointing, and cardboard. Condensed tannins
from other species have been intensively investigated for adhesive applications, but they
could not be exploited as readily as wattle tannins. Pilot production of additives from the
Tsuga hererophylla bark extract in the United States was not cost competitive with pet-
rochemical phenol. Commercial-scale production of adhesives for particleboard from Pinus
rcldiata bark in New Zealand was unfortunately discontinued in the early 1980s. A com-
prehensive description of the situation prior to 1985 was presented in the first edition [l].
The worldwide production of wattle (A. nzearnsii bark) and quebracho (Schinopsis
spp. heartwood) tannins was 150,000 tons per year in the early 1990s, of which not more
than20-30% are available for adhesiveapplication, the rest beingreserved for their
traditional leather market [141]. In China, plantations of the tree A. nzecrrnsii have been
rapidly increasing, and tannin production was expected to increase from 100 tons per year
in 1989 to nearly 2000 tons per year in 1994 [142]. Zhao et al. have studied to develop
production technologies for wattle tannin adhesives which would be usable for exterior-
grade plywood under Chinese factory conditions, which requires a wide range of closed
assembly time from 30 min to 16 h [ 106,142- 1441. The polyflavanoid contents of Chinese
wattle tannins evaluated as Stiasny values were higher and their molecular weight distri-
butions were very similar, compared to those of South African commercial tannin [ 1061.
Adhesives were successfully formulated from Chinese commercial wattle tannin fortified
with phenol-formaldehyde or phenol-urea-formaldehyde resins as cross-linking agents,
instead of paraformaldehyde, whichis relatively expensive in China. Factory trials strongly
Chemistry of Bark 265
indicated that the adhesives would be extremely applicable to the Chinese industry man-
ufacturing exterior-grade plywood.
Recently, Santana et al. reported that the bark of A. mearnsii was liquefied in phenol
in the presence of sulfuric acid catalyst [ 1451. The resulting solution was reacted with
formalin in a basic solution to yield a resol resin that, under the best conditions, performed
similarly to the commercially available phenol-formaldehyde resins in plywood adhesion.
Studies of the adhesive preparation from the bark have been performed with other hard-
wood species, including mangrove trees, Khaya ivorensis and Avicennia alba [1461, and
Moroccan afforestation trees, Eucalyptus astringents, E. Sideroxylon, and Acacia decurrens
[147].
However, commercially feasible technologies for production of adhesives from con-
ifer bark tannins, if developed, would be accepted more widelyin the world, as Coniferous
trees have a very large plantation area, and wood industries in different countries are using
huge amounts of them for timber production and producing large quantities of conifer
bark every year as by-products. Conifer bark tannins will be supplied more stably than
hardwood bark tannins if their utilization is established. Therefore, serious and intensive
work has been done with bark tannins of different coniferous species, including Picea
abies [21,148,149], ktrixleptoides [150], Pinussylvestris [21,148,149], P. halepensis
[151,152], P. pinaster [21,153-1561, P. radiata [14,157-1651, P. caribaea [21], P. elliottii
[21], and the U.S. southern pines [166-1681.
Interest has been increasing in the utilization of pine tannins as materials for adhe-
sives in such countries as Australia, Chile, and the United States. However, difficulties
have been encountered in the utilization of tannin extracts from softwood barks, primarily
due to low extractable yields, excessive viscosity, and much faster reactivity toward for-
maldehyde, as stated in the first edition [ l]. Yazaki et al. reported that the extraction of P.
radiata bark using a four-stage squeeze extraction provided a high extractives yield (ap-
proximately 30%), despite a ratio of 1 bark to 3 solvent used (by mass) [ 1591. The com-
bined extract had viscosities greater than 8000 mPa. S at 40% solid content at 25°C. How-
ever, when part of the extracts was sulfited with sodium metabisulfite under reflux for 2
h and combined with the rest of the extracts, the viscosity of the combined extracts was
suitable for the formulation of adhesives which provided high-quality glue bonds [ 1621.
The gluing properties of the adhesivesderivedfromextracts of differentbarkspecies
appeared to be dependent on their contents of formaldehyde-reactive polyflavanoids as
indicated by their Stiasny values, with a value of 65% being the minimum for producing
a high-quality adhesive by this method [21]. The viscosity modification observed with the
sulfite treatment is based on the findings of Foo et al. that sulfonation of 5J-dihydroxy-
proanthocyanidins involves interflavanoid bond cleavage formingprocyanidin-4-sulfonates
and not formation of polymeric sulfonates as had originally been thought [ 1691. Further-
more, sulfonates (Y to a phloroglucinol ring, namely, procyanidin-4-sulfonates, are good
leaving groups at ambient temperature and pH greater than 8.0. Consequently, under tYP-
ical adhesive formulation conditions, the sulfonic acid groups on tannin derivatives from
conifer barks will be displaced, resulting in water-insoluble polymer [ 1701.
Kreibich and Hemingway have studied ways to develop tannin-based adhesives and
recently reported and reviewed that tannins extracted with sodium sulfite solution from
the U.S. southern pine bark were able to replace about 50% of phenol-resorcinol-for-
maldehyde resin in cold-setting wood adhesives. Bonds in laminates exceeded the require-
ments of the American standards for dry-sheer and vacuum-pressure cold-water soaktests
[ 166,1671. Part of the resorcinol in adhesives made of resorcinol, formaldehyde, and a
I
266 Sakai
styrene-butadiene-vinyl pyridine terpolymer latex for bonding nylon cord to rubber could
be replaced with purified loblolly pine tannin [171].
In Chile a company began in the late 1980s to extract a range of midly sulfited and
nonsulfited tannins from Pinus rudiata and f? insignis mixed bark, to serve the require-
ments of the local leather industry as well as to supply tannins to other applications. Based
on laboratory formulations using commercially produced extracts, Leyser and Pizzi [ 1571
developed a new “honeymoon” fast-set system that is a two-component system in which
one part is a conventional phenol-resorcinol-formaldehyde resin with the pH adjusted to
a value greater than 13 (pot life = indefinite), and the other part is the sulfited pine tannin
extract containing excess paraformaldehyde (pot life 5 h). The result of the fingerjointing
and glulam industrial trial using the system satisfied the relevant international standards
specifications. More recently, mill trials for application of basically the same honeymoon
cold-set adhesive system were conducted for structural end joints in six different mills in
North America. The phenol-resorcinol-formaldehyde/tannin honeymoon system was ca-
pable of producing cold-set, fully exterior-grade end joints at mill production rates [ 1681.
However, yield of the pine tannin, industrially extracted in Chile, was comparatively low
(less than 20%). The phlobaphenesformationand precipitation during sulfite/water ex-
traction of pine tannin from pine bark was minimized by blocking tannin self-condensation
by the addition of smallamounts of astrongnucleophilesuchasphloroglucinol. m-
phenylenediamine, and urea, the latter due to its low cost for industrial application. The
yield of pine tannidurea extracts increased from 19% to 25% in industrial scale, and the
extracts proved to give good thermosetting wood adhesives for panel products [ 1581.
Pinetannin-basedadhesives for exteriorparticleboardhave been obtained by the
reaction of polymeric MD1 (4,4’-diphenylmethane diisocyanate) with mildly sulfited pine
tannin extract [141]. This type of adhesive from wattle tannin was already developed in
the early 1980s [l]. Pinetanningave better results because of a faster reaction of the
phloroglucinolicA-ringwithformaldehyde, resulting in formation of methylolgroups
which in turn easily react with isocyanates to form urethane bonds [141]. The character-
istics of the particleboards obtained industrially by using MDI-fortified pine tannin ad-
hesives satisfied well the requirements of exterior-grade waterproof particleboards. Car-
bohydrates in the bark extract might react also with the MD1 to form urethanes, as pointed
out by Laks [l].
Pizzi et al. reported that autocondensation of polyflavanoid tannins is accelerated by
Lewis acids and cellulose. Based on this observation, particleboards of excellent internal
bond strengths were obtained using a pine bark tannin at pH 10.2 and higher [172]. This
afforded the possibility to prepare interior-grade wood binders, presenting no formaldehyde
emission at all. In this connection, a possibility of the spruce bark reacting with formal-
dehyde during manufacture of particleboards and acting as a formaldehyde scavenger was
pointed out [ 1731.
the former enzyme was the primary agent for the fiberization and the latter plays a sup-
plementary but indispensable role [ 174,1751. A technology was developed for papermaking
from the innerbarkof Cryptomeriajaponica (Japanesecedar),whichhas the largest
plantation area in Japaneseforests [176]. The innerandouterbarks are separated by
treatment with a hot ammonium oxalate solution that dissolves pectic substances. The bast
fiber pulp prepared from alkaline cookingof the separated inner bark is being used suitably
for specialty purposes such as calligraphy, artistic wrapping, paper arts, etc., after blending
with other bast fiber pulps and/or staining in different colors. Acetic acid pulping of the
Pinus pinaster bark after tannin extraction was studied, but properties of the resultant pulp
were not evaluated [ 1771. It is not likely to be feasible to make ordinary paper from fibers
contained in bark.
Enzymatic saccharification of bark polysaccharides has been studied to obtain sugars
and ethanol. Alkali-extracted Pinus pinaster bark was enzymatically hydrolyzed with a
mixture of cellulase and P-glucosidase to produce sugars, but their yield was less than
10% of theoretical. The sugar yield increased to 75% by successive treatments, extraction
with NaOH (15 min), and delignification with acid chlorite (7 h) prior to the enzymatic
saccharification [ 1781. Delignification with hydrogen peroxide andacetic acid had a similar
effect on the sugar yields [179]. These delignification treatments, however, are not eco-
nomically feasible.
In the case of Picea ezoensis bark, NaOH treatment effectively increased the glucose
yield to 40-45% of holocellulose content [ 1801. Glucose in the saccharification mixture
was fermented to accumulate ethanol, and 1 kg of the NaOH-treated bark was expected
to convert to 140 g of ethanol. The barks of three poplar trees (Populus tremuloides, P.
maximowiczii X trichocarpa, and P. trichocarpa X deltoides) were susceptible to dilute
sulfuric acid pretreatment, and relatively high levels of enzymatic digestibility of cellulose
were observed after the pretreatment [ 18 l]. However, the barks from sweetgum (Liquid-
ambar styraciflua)[ l 8 1 ] and Pinus pinaster [ 1821 were unresponsive to acid prehydrolysis
in terms of enzymatic digestibility. Hemicellulosewhichwashydrothermallyextracted
from beech bark at approximately 200°C was saccharified enzymatically with high yields.
However, the tannins had first to be removed at temperatures of 120- 140"C, as they have
an inhibitory effect on xylanases [ 1831.
Oil-absorbent mats were very simply prepared from barks of Cryptomeria japonica
and Chamaecyparisohtusa. Barkswerecrushedinto fibrous fragments of 1-3 cm in
length, treated with water-repellent emulsion, blended with polyester fibers ( 1 cm length)
as binders, molded, and heat-treated at 130°C. The mats, consisting of 90% bark and 10%
polyester, had a density of 0.07 g/cm3 and absorbed fuel oil to more than 3 g/g of mats
[ 1841. The mat may be used to remove mineral and vegetable oils in the effluents from
factories, gas stations, kitchens, etc. Porous spherical tannin resin was prepared from Aca-
cia mearnsii tannin by reaction with formaldehyde in nonpolar media, polybutene, under
stirring at 60°C [ 1851. The spherical resin, with 139 m2/g of surface area and 0.5- 1.O mm
of diameter, adsorbed heavy-metal ions such as Cr(VI), Cd(II), Cu(II), and Fe(J1) [185-
1871. Adsorption of Cu(I1) ion to the resin was shown to be a physical process [186].
Biodegradablepolyurethanefoamswereprepared by the reaction ofdiisocyanate
with poly01 mixtures consisting of commercial wattle tannin or bark, a synthetic polyol,
a siliconesurfactant, catalysts, and wateras a foamingagent[128-1301.Someof the
foams obtained had densities of less than 0.04 gkm' and thermal conductivities as small
as that of commercial polyurethane foams. It is suggested that, based on a model reaction
with catechin, isocyanate groups react preferentially with phenolic hydroxyl groups in the
B-ring of condensed tannins in the barks to form a urethane bond [ 1881.
268 Sakai
REFERENCES
1. P. E. Laks, in Wood and Cellulosic Chernistn (D. N. S. Hon and N. Shiraishi, eds.), Marcel
Dekker, New York, p. 257 (1991), and literature cited therein.
2. M. Trockenbrodt, IAWA Bull., 11: 141-166 (1990).
3. C. M. Chen and J. K. Pan, HolzJorsch., 45: 1 55- I59 ( 199l ) .
4. H. Pereira, Wood Fiber Sci., 20: 82-90 (1988).
5. P. R. Blankenhorm, T. W. Bauersox, C. H. Strauss, G. L. Stimely, L. R. Stover, and M. L.
Dicola, Wood Fiber Sci., 20: 74-81 (1988).
6. G. Vazquez, G. Antorrena, and J. C. Parajo, Wood Sci. Technol., 21: 65-74 (1987).
7. G . J. Leary, K. R. Morgan, and R. H. Newman, Appita, 40: 181-184 (1987).
8. E. Sjostrom, Wood Chenzistv, Academic Press, New York (1981).
9. R. Solar, I. Melcer, and F. Kacik, Cell. Chenz. Technol., 22: 39-52 (1988).
10. S. Ohara and K. Yanagi, Mokuzai Gakkaishi, 41: 406-413 (1995).
11. H. Hafizoglu and M. Reunanen, Holz$orsch., 41: 261-263 (1987).
12. B. Perra, J.-P. Haluk, and M. Metche, HolzJorsch., 47 486-490 (1993).
13. J. G a r p and H. Pereira, HolzJorsch., 51: 225-234 (1997).
14. Y. Yazaki and T. Aung, Holz$orsch., 42: 357-360 (1988).
15. R. Solar, F. Kacik, and I. Melcer, Holz Roll Werkst., 50: 291-294 (1992).
15b. L. J. Proter, in NaturalProducts of Woody Plunrs. I (J. W. Rowe, ed.), Springer-Verlag,
New York, pp. 651 -690 (1989).
16. P. E. Laks, R. W. Hemingway,and A. H. Conner, J. Chenl. Soc., Perkin Truns. I : 1875-
1881 (1987); P. E. Laks and R. W. Hemingway, HolzJorsch., 41: 287-292 (1987).
17. A. Burmester and W. Kieslich, Holz Roh Werkst., 4 4 : 419-422 (1986).
18. T. D. Lomax, K. L. Mackie, R.Meder,M. Croucher,and R. J. Burton, J . Wood Chern.
Technol., 14: 539-561(1994).
19. H. Hafizoglu and B. Holmbom, Ho/z$orsch., 41: 73-77 (1987).
20. N. Dunlop-Jones, J. Huan, and L. H. Allen, J. Pulp Paper Sci., 1 7 560-566 (1991).
21. Y. Yazaki and P. J. Collins, Holz Roh Werkst., 52: 307-3 10 (1994).
22. M. N. Nurulhuda. L. T. Chew, N. M. Y. Mohd, and R. M. A. Abdul, Holz Roh Werkst., 48:
38 1-383 (1990).
23. M. L. Laver and H. H.-L. Fang, Wood Fiber Sci., 18: 553-564 (1986).
24. M. L. Laver and H. L. Fang, J . Agric. Food Chern., 3 7 114- 1 1 6 ( I 989).
25. Q. L. Pu, L. K. Pannell, and X.-D. Ji, Planta Med., 56: 129- 130 ( 1 990).
26. Z . Madar. H. H. Gottlieb, M. Cojocaru, J. Riov, Z. Solel, and A. Sztejnberg, Phytocheniwy.
56: 129-130 (1990).
27. Y. Mimaki, A. Kameyama, Y. Sashida, Y. Miyata.andA. Fujii, Chern.Pharnz. Bull, 43:
893-895 (1995).
28. M. C. Wani, H. L. Taylor, M. E. Wall, P. Coggon, and A. T. McPhail, J. Am. Chenz. Soc.,
Y3: 2325-2327 (1971).
29. S. B. Horwitz, Nurure, 367: 593-594 (1994).
30. K. C. Nicolau et al., Nuture, 367: 630-634 (1994).
31. J. C. Dennis, Correa, and A. E. Greene, J. Org. Chern., 56: 6939 (199 I ) .
32. A. Stierle, G. Strobel, and D. Stierle, Science, 260: 214-2 16 (1993).
33. S. Tachibana, E. Watanabe, Y. Itoh, and T.Oki, Mokuzai Gukknishi, 40: 1254-1258 (1994),
and literature cited therein.
34. N. Vidensek, J . Naturul Prod., 53: 1609- 16 I O ( 1990).
35. G. Chengwei,C.Zhangyu,andC. Yushu, Abstracts2nd Int. S y n p . Chern. Utili:. Tree
Extructives, pp. 188- 189 ( 1995).
36. 1. Ara, B. S. Siddiqui, S. Faizi, and S. Siddiqui, J. Chenz. Soc.. Prrkin Truns. 1, 343-345
( 1989).
37. S. Yaga, K. Kinjo, H. Hayashi, N. Matsuo, F. Abe, and T. Yamauchi, Phytochemistr)!, 32:
141-143 (1993).
Chemistry of Bark 269
111. Y. Cai, F. J. Evans, M. F. Roberts, J. D. Phillipson, M. H. Zenk, and Y.Y. Gleba, Phyto-
chemistry, 30: 2033-2040 (1991).
112. M. H&, M. Heinrich, and H. Rimpler, Phytochemistry, 42: 109-1 19 (1996).
113. S. Achmadi, G. Syahbirin, E. T. Choong, and R. W. Hemingway, Phytochemistry, 35: 217-
219 (1994).
114. H. Kolodziej, Phytochemistry, 28: 3487-3492 (1989).
115. J, J. Karchesy, L. Y. Foo, R. W. Hemingway, D. F. Barofsky, and E. Barofsky, Wood Fiber
Sci., 21: 155-162 (1989).
116. R. Isobe, T. Tanaka, G. Nonaka,and I. Nishioka, Chem.Pharm.Bull., 3 7 1748-1750
( 1 989).
117. D.F. Brofsky, in Chemistry and Significance of Condensed Tannins (R.W. Hemingway and
J. J. Karchesy, eds.), Plenum Press, New York, pp. 175-195 (1989).
118. D. Ferreira and E. V. Brandt, in Chemistry and Significance of Condensed Tannins (R. W.
Hemingway and J. J. Karchesy, eds.), Plenum Press, New York, pp. 153-174 (1989).
119. D. Thompson and A. Pizzi, J. Appl. Polymer Sci., 55: 107-1 12 (1995).
120. S. Ohmura, Medicinal Plants f o r Pharmacognosy in China and Japan (in Japanese), Hi-
rokawa Shoten, Tokyo (1978).
121. T. Kariyone, Japanese and Chinese Pharmacognos.y (in Japanese), Hirokawa Shoten, Tokyo
(1971).
122. E. Malan, E. Swinny, D. Ferreira, and P. Stemberg, Phytochemistry, 41: 1209-121 3 (1996).
123. D. R.Forlines, T. Tavenner, J. C. S. Malan,andJ. J. Karchesy,in Plnnt Polyphenols,
Synthesis. Property and Significance (R. W. Hemingway and P. E. Laks, eds.), Plenum Press,
New York, pp. 767-782 ( 1 994).
124. K. Mura, Y. Iitoi, and W. Tanimura, Hakko Kougaku (in Japanese), 65: 399-404 (1987).
125. Z. S. Okoye and G. E. Neal, Food Chem. Toxicol., 26: 679-689 (1988).
126. M. Samejima and T. Yoshimoto, Mokuzni Gukkaishi, 30: 41 3-416 (1984).
126b. H. Kai, T. Ueda. and M. Sakaguchi, Soil B i d . Biochern., 22: 983-986 (1990).
127. M. Mori, M. Aoyama, S. Doi. A. Kanetoshi, and A. Hayashi, Holz Roh Werkst.. 53: 81-82
( 1995).
128. J.-J. Ge and K. Sakai, Mokuzni Gakkuishi, 39: 801-806 (1993).
129. J.-J. Ge and K. Sakai, Mokuzai Gakkuishi, 42: 87-94 (1996).
130. Y. Nakashima, J.-J. Ge, and K. Sakai, Mokuzai Gakkaishi, 42: 1105-1112 (1996).
131. P. E. Laks. P. A. McKaig, and R. W. Hemingway, Hol;for.sch., 42: 299-306 (1988).
132. J. Zhang, K. Takahashi, Y. Kono. Y. Suzuki, et al., J . Pesticide Sci., 15: 585-591 (1990).
I32b. A. Scalbert, Phytochernistty, 30: 3875-3883 (1991).
133. N. Yamashita, H. Itoh, K. Skata, A. Yagi, H. h a , and K. Ina, Agric. Biol. Chern., 53: 1383-
1385 ( 1 989).
134. J. H. M. Temmink.J.A.Field,J.C. Van Haastrecht, and R. C. M. Merkelbacha. Wnter
RES..23: 34 1-344 ( 1 989).
135. J . A. Field. M. J . H. Lcyendeckcrs, R. S. Alvarez. G. Lettinga, and L. H. A. Habets, Wtrter
Sci. Technol., 20: 2 19-240 ( 1988).
136. J.A. Field, G. Lettinga. andL. H. A. Habets. J. Chern. Techrrol. Biorechnol.. 49: 15-33
( 1990).
137. J. A. Field, R. Sierra-Alvarez, G. Lettinga, and L. H. A. Habets, J . Chern. Techno/.,49: 35-
S3 ( 1990).
138. M. Savolainen and K. Jokinen. in Ligrlocellltlosics, Science. Technolop. De\dqnnent ~ l n d
Use (J. F. Kennedy, G. 0. Phillips. and P. A. Williams, eds.). EllisHorwood. New York.
pp. 393-404 (1992).
139. J.Field, M. J. H. Lcycndeckers. R. Sierra-Alvarez. L. H. A. Habets.and A. G. Lettinga.
Biotechnol. Bioeng.. 37: 247-255 ( 199l ) .
140. E. P. Swan. in Noturd Products o f Woocly Pltrnts, I / (J. W. Rowe, ed.), Springer-Verlag,
New York. pp. 931-952 (1989).
141. A. Pizzi. E. P. von Leyser. J. Valenzuela. and J. G. Clark. Hol;fif~rsc/~..
47: 168- l74 (1993).
272 Sakai
142. L. Zhao, B. Cao, F. Wang, and Y. Yazaki, Holz Roh Werkst., 52: 1 13-1 18 ( 1994).
143. L. Zhao, B. Cao, F. Wang, and Y. Yazaki, Holz Roh Werkst., 53: 117-122 (1995).
l#. L. Zhao, B. Cao, F. Wang, and Y. Yazaki, Holz Roh Werkst., 54: 89-91 (1996).
145. M. A. E. Santana, M. G. D. Baumann, and A. H. Conner, J . Wood Chem. Technol., 16: 1-
19 (1996).
146. E. A. Taiwo and R. A. Ogunbodede, Wood Sci. Technol., 29: 103-108 (1995).
147. M. Fechtal and B. Riedl, Holzforsch., 4 7 349-357 (1993).
148. B. Dix and R. Marutzky, Holz Rolz Werkst., 45: 457-463 (1987).
149. B. Dix and R. Marutzky, Holz Roh Werkst., 46: 19-25 (1988).
150. R. Takano, Research Report of Biomass Conversion Program, No. 29:67-76 (1991) (in
Japanese).
151. A. Grigoriou, E. Voulgaridis, and C. Passialis, Holdorsch. Holzvenuert., 39: 9- I 1 (1987).
152. A. Grigoriou, Holz Roh Werkst., 48: 377-380 (1990).
153. G. Vazquez, G. Antorrena, J. Parajo, and J. L. Francisco, Holz Roh Werkst., 4 7 491-494
( 1989).
154. G. Vazquez, G. Antorrena, J. L. Francisco, and J. Gonzalez, HolzRohWerkst., 50: 253-
256 ( 1992).
155. G. Vazquez, G . Antorrena, J . L. Francisco, M. C. Arias, and J . Gonzalez, H012 Roh Werkst.,
51: 221-224 (1993).
156. G. Vazquez, G. Antorrena, J. Gonzalez, and J. C. Alvarez, HolzRohWerkst., 54: 93-97
( 1996).
157. E. von Leyser and A. Pizzi, Holz Roh Werkst., 49: 25-29 (1990).
158. V. J. Sealy-Fisher and A. Pizzi, Holz Roh Werkst., 50: 212-220 (1992).
159. A. Pizzi and A. Stephanou, J. Appl. Polymer Sei., 51: 2109-2124 (1994).
160. A. Pizzi and A. Stephanou, J. Appl. Pol.ymer Sei., 51: 2125-2130 (1994).
161. Y. Yazaki and P. J. Collins, Holz Roh Werkst., 52: 185- 190 (1994).
162. Y. Yazaki and P. J. Collins, Holz Roh Werkst., 52: 241-243 ( 1994).
163. S. Kim and D. E. Mainwaring, J. Appl. Polymer Sci., 56: 905-913 ( 1995).
164. S. Kim and D. E. Mainwaring, J. Appl. Polymer Sei., 56: 915-924 (1995).
165. S. Kim and D. E. Mainwaring, Holdorsch., 50:42-48 (1996).
166. R. E. Kreibich and R. W. Herningway, Forest Prod. J., 37(2): 43-46 (1987).
167. R. E. Kreibich, in Chemistry and Significance of Condensed Tannins (R.W. Herningway
and J. J. Karchesy, eds.), Plenum Press, New York, pp. 4.57-478 (1989).
168. R. E. Kreibich, R. W. Hemingway, and W. T. Nearn, ForestProd. J . , 43(7/8):45-48
(1993).
169. R. W. Hemingway, in Chemistry und Sign$cance .f Conder~sedTannins (R.W. Herningway
and J. J. Karchesy, eds.), Plenum Press, New York, pp. 265-283 (1989).
170. G. W. McGraw, P. E. Laks, and R. W. Hemingway, J . WoodChem. Technol., 8: 91-109
(1988).
171 K. H. Chung and G. P. Hamed, in Chemistrv and Signifcance of Condetmed Tcnnirzs (R.
W. Hemingway and J. J. Karchesy, eds.), Plenum Press, New York, pp. 479-502 (1989).
172. A. Pizzi, N. Meikleham, and A. Stephanou, J. Appl. Polymer Sci., 55: 929-933 ( I 995).
173. B. Prasetya and E. Roffael, Holz Roh Werkst., 49: 341-344 (1991).
174. H. Tanabe and Y. Kobayashi, Agric. Biol. Chem.. 49: 3595 (1985).
175. H. Tanabe and Y. Kobayashi, Agric. Biol. Chem., 51: 2845-2846 (1987).
176. T. Itoh, Modern Foresh (in Japanese), no. 7 48 (1993).
177. G. Vazquez, G. Antorrena, and J. Gonzalez, Wood Sci. Technol, 28: 403-408 (1994).
178. G. Vazquez, J. Parajo, G. Antorrena, P. Thonart, and M. Paquot, WoodSci.Techrlol., 21:
167-178 (1987).
179. D. V. Penas, M. A. L. Yusty, and J . C. P. Linares, J. Chem. Technol. Biotechrwl., 54: 63-
74 (1992).
180. M. Kuwahara, T. Sawada, Y. Asada, Y. Nakamura, M . Uchikoshi, and M . Komatsu, Mokuzcli
Gakkaishi, 34: 75-81 (1988).
Chemistry of Bark 273
181. R . Torget, M. E. Himmel, and K . Grohmann, Bioresour. Technol., 35: 239-246 (1991).
182. G. Vazquez, G. Antorrena, J. C. Parajo, and X. L. Francisco, Wood Sci. Technol., 22: 219-
225 (1988).
183. E. Walch, A. Zemann, F. Schinner, G. Bonn, and 0. Bobleter, Bioresou,: Technol., 3 9 173-
I77 ( 1992).
184. T. Sakuragawa and T. Miyamoto, Mokuzrti Kogyo, 51: 266-269 (1996) (in Japanese).
18.5. H. Yamaguchi, M. Higuchi. and i. Sakata, J. Appl. Polywzrr Sci., 45: 1455- 1462 ( I 992).
186. H. Yamaguchi, R. Higasida, M. Higuchi, and I. Sakata, J . Appl. Polymer Sci., 45: 1463-
1472 ( 1992).
187. H. Yamaguchi, Y. Iura, M. Higuchi, and I. Sakata, Moklczai Gakkaishi, 37: 815-820 (1991).
188. J.-J. Ge and K. Sakai, Mokuzrri Grrkknishi, 42: 417-426 (1996).
This Page Intentionally Left Blank
Chemical Characterization of Wood and
Its Components
1. INTRODUCTION
Wood analysis comprises the determination of the wood components as well as the iso-
lation, purification, and characterization of the woodconstituents. Wood is chemically
heterogeneous and its components can be divided into two groups: structural components
of high molecular weight (cellulose, polyoses or hemicelluloses, and lignin), which are
the major cell wall components; and nonstructural components of low molecular weight
(extractives and inorganic compounds). The macromolecular components are not uniformly
distributed in wood cells, and their concentration changes from one morphological region
to another. Therefore, knowledge about the distribution of chemical components in the
cell walls is of great importance to understanding the properties of wood. The chemical
composition of wood varies from species to species. In general, hardwoods contain more
hemicellulose than softwoods but less lignin. Figure 1 shows the typical composition of
hardwoods and softwoods [ I ] .
There are different types ofwood analysis. One may consider only the main cell
wall components, holocellulose (cellulose and polyoses), lignin, extractives, and ash; on
the other hand, a very detailed analysis may include functional groups (e.g., acetyl groups),
individual units of the different components (e.g., sugar pattern), different types and fre-
quency of linkages (e.g., p-0-4 linkage in lignin), etc.
Usually, wood is analyzed by the separation of the different components, but there
are serious difficulties in achieving selective isolations. The separation is never complete
and leads to structural changes,predominantly in the lignin. An array of classical, wet
chemicalproceduresandagrowingnumber of instrumentalmethodsare available for
analysis of wood. The wet chemical methods permit the acquisition of data on the gross
composition of wood.and they require the separation ofwood into macroscopicwood
components (e.g.. lignin, holoccllulose, etc.). This is the reason it is always necessary to
report the isolation techniques used. On the otherhand, the instrumentalmethods are
conducive to higher specificity and convenience of wood analysis. Initially, chemical spec-
ificity was achieved in a macroscopic scale; for example, by chromatographic methods the
separation and determination of the individual sugars can be obtained after hydrolysis of
a wood sample.More recently, techniquessuch a s ultraviolet microscopy, electron mi-
275
276 Baeza and Freer
50
40 -
m
gc 30-
W
g
a
20-
10 -
0 -L
croscopy coupled with X-ray analysis, and infrared spectrometry have permitted descrip-
tions of the distribution of chemical constituents in wood and fiber walls.
The methods of wood analysis are more or less standardized. Detailed descriptions
of the analysis of wood are giving in the specialist literature[2-81. The CPPA (Technical
Section, Canadian Pulp and Paper Association, Montreal,PQ, Canada), TAPPI (Technical
Association of the Pulp and Paper Industry, Atlanta, GA), ASTh4 (American Society of
Testing and Materials, Philadelphia, PA) have issued new or revised test methods for the
analysis of pulp and paper materials. There are excellentreviews covering analytical tech-
niques for wood and its components [9-151.
The purpose of this chapter is to inform the reader about some of the methods
available for the chemical analysis of wood. The analysis of the constituents of wood,
according to the scheme shown in Fig. 2, will be described.
The objective of the summative analysis is toaccount for all of a sample. The summative
analysis of wood is based on the isolation and identification of certain groups of wood
components and therefore does not deal with the determination of chemical uniformity of
substances. These groups are mainly cellulose, hemicelluloses, lignin, and extractives, and
often other designations are added, e.g., holocellulose, a-cellulose, or sometimes terms
originated from the method used (such as kraft lignin, Cross and Beavan cellulose).
In evaluating the results of the summative analysis, it is important to take into con-
sideration the methods used. To obtain a complete mass balance of a wood sample, no
constituents may be overlooked or repeated. Nothing should be determined by difference.
Summative analysis will typically account for most of the wood components. Examples
of summative analysis of wood composition have been reported elsewhere [16-211. Sum-
mations may be taken in several ways, and some examples of different types of wood
analyses are given in Table 1.
Valuesof 98-101 are generallyacceptable,butfrequentlyvaluesdeficientor in
excess by about 10% are obtained [ 14,16,19,20]. Possible reasons for failure to achieve a
complete mass balance have been reviewed by Browning [20]. Generally, the summative
analysis is corrected via normalization, giving the same factor of error to all the compo-
nents, independent of the amount in which the different components are present and in
the analytical technique used. The possibility of undetected or unknown compounds in
Chemical Characterizationof Wood 277
EXTRACTION
I I
I I
I
EXTRACTIVES EXTRACTIVE-
FREE WOOD
I
SOLUBLE IN
I
SOLUBLE IN
DELlGNlFlCATlON
A B C D
in the mass balance were extensively discussed. These factors include: the exact contri-
bution of galacturonic acid in the wood; the methanol formed from the demethoxylation
of polysaccharides and lignin; the loss of water in the lignin condensation/dehydration
relations; the wood components included in the polysaccharides and/or lignin moieties that
exist in such small quantities that they are not detected by the analytical methods utilized;
and residual water that remains bonded in the wood after “oven drying” which is released
in the hydrolytic stage and becomes part of the hydrolytic solution. Clearly, the contri-
bution of any of these factors is small. The complex nature of wood seriously complicates
the quantification of these contributions. The improvement of the methods and techniques
of analysis could result in a mass balance close to 100%.
The kind of sampling and sample preparation depends on many factors and on the aim of
the analysis. Thus, the magnitude of the sampling needed for general characterization of
aspecies is quite different than for the evaluation of trees in a specified stand. It is
important to ensure that representative samples are collected which are free from outside
contamination, and properly preserved. No analysis is better than the sample on which is
based. However, for comparison of techniques and methods the only requirement is that
the sample be uniform.
Astandardizedsamplingprocedure is given in TAPPIStandardT-257cm-85 [2].
The procedure given is appropriate for wood in all forms, i.e., logs, chips, or sawdust. A
probability sampling plan and an economic or engineered sampling plan are described. A
detailed discussion of sampling and preparation of samples is given by Browning [22].
Wood for chemical analysis, after air drying, must be milled to achieve complete
penetration by reagents and to ensure uniform reactions. Heating, preparation of very fine
and dusty material, and regrinding coarse material must be avoided. Samples are screened
and normally material passed through a 0.40-mm (40-mesh) sieve and retained on a 60-
mesh sieve. The selected fraction should represent, if not the entire amount of material,
at least 90-95% of the original sample. Theextractiveshould be removedbeforeany
chemicalanalysis,exceptwhere the extraction process and subsequentwashingcould
interfere with the analysis.Aprocedureforfurtherpreparation of wood for chemical
analysis that has been sampled in accordance with TAPPI 257 is provided in TAPPI Test
Method T264 0111-88 121. Neutral solvents, ethanol and benzene, are employed to obtain
extractive-free wood, removing material which is not part of the wood substance or which
may interfere with subsequent analysis. Moisture determination is included. Related meth-
ods are ASTM D1 105 (ANS),“Preparation of Extractive-FreeWood” 131; and CPPA
G.31P. “Preparation of Wood for Chemical Analysis” [4].
It is dangerous to include benzene in the solvent mixture to extract wood. due to its
carcinogenic properties for which it has been long banned. Accordingly, any contact with
the skin or inhalation of benzene vapor must be avoided. Extreme safety precautions must
be taken in carrying out the above procedure. Gloves, good ventilation, and a chemical
fume hood must be used.
A mixture of ethanolandtoluenewasfound to remove the samematerials fro111
wood as ethanol-benzene 1231. However, the mixture of ethanol and toluene does not boil
and reflux at a constant temperature and rate.
Wood samples collected for later analysis must remain moist and cold-stored. Sam-
ples should not be oven-dried to avoid changes in reactivity and lost of volatiles. After
Chemical Characterizationof Wood 279
Extraction
Klason ~ioacido~ysis ................I
Lignin
Extractive-free wood ........................... Lignin
(160
Delignification
IO mg
r .............. 2
.m mmm m mm. m
Klason TFA hydro]ysis
Lignin in Hobcellulose ............................. +
I. Neutralsugar :
holocellulose (80 m@ 20 mg ratio i
r................:
Dissolution by alkali
Water is a natural constituent of all parts of a living tree. In green wood, water is commonly
about 5 0 % of the total weight. When the tree dies or a log is processed into lumber, chips,
etc., the wood loses some of its moisture to the surrounding atmosphere. However, some
water will remain within the structure of the cell even after wood has been manufactured
into lumber, particle, veneer, or fiber product. The amount of residual water depends on
the cxtent of drying and the environmental conditions. The wood-water system is very
important i n many tields of wood technology. The physical and mechanical properties,
resistance to biological deterioration, and dimensional stability of the products are affected
by the amount of water present.
Chemicalanalysis of wood is altnost always performed on air-dried samples, but
results are reported on a moisture-free basis. A moisture determination must therefore be
r ~ for
~ nalmost every sample submitted for analysis.
The amount of water in wood is expressed in two ways:
280 Baeza and Freer
1. The moisture content (MC) is expressed as a percent of the total weight (wood
with water). Thus:
weight with water - OD weight
%MC = x 100
weight with water
The wet weight base is generally used in the pulp and paper industry and
when wood is used as fuel.
2. Themoisturecontent (MC*) is definedas the weightofwaterexpressedas
percentage of the moisture-free or oven-dry (OD) weight of wood. Thus:
weight with water - OD weight
%MC* = x 100
OD weight
Because the denominator is the dry weight, the moisture content calculated in
this way canbeover 100%. Thismethodof calculating moisturecontent is
generally accepted as standard for wood-based materials such as lumber, ply-
wood, particleboard, and fiberboard.
Some of the methods used for moisture determination are described below.
Oven-Dry Method. One of the mostcommonmethods of determining MC is to
weigh about 2 g of the wet sample and oven-dry it at 105 2 3°C until the weight does
not change by more than 0.002 g following a l-h heating period. The details of this oven-
dry method are described in TAPPI Test MethodT264 om-88 [2]. The major disadvantages
of the oven-dry method are that it is a destructive method, it is time-consuming, and some
volatile components other than water can be driven off during drying.
Electrical Moisture Detectors Method. TheMC can be determined by use of
electrical moisture detectors, whichhave the advantage of being relatively simpleand
direct. There are a variety of electrical meters. The resistance moisture meter, used for
lumber, indicates the moisture based on the linear relationship between the logarithm of
electrical resistance and the moisturecontent of woodfrom 7% to the fiber saturation
points [25].Some electric meters are based on the effect that water behaves as a capacitor
when placed in a high-frequency field [26]. However, the dielectric methods have been
described as unsatisfactory for solids such as moist wood. The most common moisture
balance consists essentially of an infrared lamp mounted above the pan of a top-loading
electronic balance. The energy input and drying times are controlled by the operator.
Moisture Determination by Solvent Distillation. Water inwoodmay be deter-
mined by distillation carried out with solvents such as toluene, xylene, and other water-
insoluble solvents. Details of this method by toluene distillation are given in TAPPI Stan-
dard Method T208 om-S9 [2]. The material is boiled in toluene. The resultant water vapor
is condensed, collected in a distilling reservoir consisting of a graduated trap (Stark and
Dean or other type), and measured. This method is specially useful for determination of
moisture content in wood because volatile substances such as turpentine and resins do not
interfere, giving a better measure of true water content than that obtained from oven drying.
Moisture Determination by Karl Fischer Method. Karl Fischer titration is one
of the most sensitive techniques available for the actual measurement of very small quan-
tities of water in different matrices. ASTM Standards E 203 and D 1348 [3] describe the
procedure for determining the moisture content in cellulose and have been applied suc-
cessfully for wood [27]. The Karl Fischer reagent is a solution of iodine, sulfur dioxide,
and pyridine, usually in methanol as asolvent.Methanol may bereplaced by methyl
cellosolve, dioxane, or acetic acid, but pyridine (Py) is essential. It reacts with water as
shown below:
Chemical Characterization
of Wood 281
The general procedure is as follows: the wood is ground and extracted with methanol
to displace the water. The water in methanol is titrated with the Karl Fisherreagent.
Titrations can be performed best by potentiometric endpoint determinations. The results
obtained for this method are not affected by volatile substances other than water in the
sample.
Moisture Determination by Nuclear Magnetic Resonance. Special physical
methods such as infrared, nuclear magnetic resonance, attenuation of p and y radiations
and neutron moderation are also applicable to MC determination i n wood and pulp.
Proton nuclear magnetic resonance ('H NMR) can provide more detailed and quan-
titative information on water in wood than any other method. Information on both mac-
roscopic and microscopiclevels can beobtained. It has been demonstrated that the 'H
NMR signals of solid wood and water are separable, so the MC of wood samples can be
determined by NMR.The MC has been measured in wood samples utilizing wide-line
NMR spectrometers 128,291. Wide-line NMR arethose in which the bandwidth of the
source of the lines is large enough that the fine structure due to chemical environment is
obscured, so only a single peak is associated with each species. These spectra are useful
for quantitative analysis.
The MC in wood relative to a known standard has been measured by pulsed NMR
techniques 1301. These techniques provide a convenient approach to the study of the liquid
adsorbed on or trapped in the solid matrix and the experimental observables include dy-
namics as well as structural information. The relationship between the water content of
wood and spin-spin relaxation time (T,) has been studied [31-331. Menon et al. [34,35]
determined the MC by NMR in samples of western red cedar and Douglas tir. The values
of MC determined by NMR were related to those values determined by the oven-dried
method. Furthermore, on the basis of T2 relaxation times, the water could be separated
into signals for the bound water i n the cell walls and that in the cell lumens. Thus, NMR
images have been obtained for these different water reservoirs. Araujo et a l . [36]carried
out studies of water in wood by using NMR relaxation techniques and NMR relaxation
selective imaging techniques. The distribution of water in sapwood, heartwood, and ju-
venile wood as well as two rehydrated heartwood samples of white spruce were analyzed.
Spectra of T, for white spruce show separate peaks corresponding to the water in different
environments.Thearea under each peak corresponds to the amount of moisture i n a
particular environment, and the T1 value indicates the nature of the environment.The
shape of the Tz spectrum of water in lumen reflects thc radius distribution of water-tilled
cell lumens in wood sample.
The NMR techniques for characterization of water in wood are very useful in the
investigation of a wood drying process.
V. WOOD POLYSACCHARIDES
Wood contains sevcral polysaccharides. of which cellulose and hemicelluloses are the most
abundant. The polysaccharides of wood are built up of a relatively limited number of sugar
residues, mainly D-glucose, D-xylose, D-mannose, D-galactose, L-arabinose, 4-0-methyl-
D-glucuronic acid, D-galacturonic acid, and glucuronic acid. Less common sugar units arc
L-rhamnose, L-galactose. L-fucose. and 0-methylated neutral sugars.
Freer 282 and Baeza
A. Cellulose
Cellulose is the mostabundantorganic material on earth. It is the chief constituent of
wood and cotton. Cellulose is a long-chain polymer of P-D-glucose in the pyranose form,
linked together by 1,4-glycosidic bonds to form cellobioseunits that are the repeating units
in the cellulose chain (Fig. 4).
Though relatively pure cellulose is produced industrially as microcrystalline cellulose
powder from bleached wood pulp by acid hydrolysis, its molecular weight is far less than
that of native cellulose of wood, and it still contains trace amounts of xylose and mannose
[W.
2. Characterization of Cellulose
a. Viscosity, Molecular Weight (Molecular Mass) and Molecular Weight Distribu-
tion. Thecellulosesynthesized in naturehas a certain degree of polydispersitywhich
influences its physical properties. Viscosities andor molecular weight distribution (MWD)
of cellulose are important parameters for all of its end uses. The polymeric properties of
cellulose are sometimes studied in solution, and based on the solution properties, conclu-
sions concerning the average molecular weight, polidespersity, and chain configuration are
drawn.However, its isolation fromwood, extraction, and dissolution often degrade the
polymer, resulting in structural changes, and the separation is never complete [72]. The
length of the molecules forming a cellulose fiber is very important for the internal cohesion
of the fiber, which influences its physical and chemical behavior, such as reactivity and
accessibility L73J. Knowledge of the MWD of cellulose is an essential requirement, par-
ticularly for its processing and industrial utilization.
For molecular studies, the quantitative isolation of pure and undegraded species in
concerned is required. However, in wood cellulose a series of problems caused by ana-
tomical and chemical phenomena make it impossible to isolate undegraded cellulose as a
homopolymeric substance. Problems in isolation of cellulose from wood for the determi-
nation of its molecular properties have been reviewed by van Zyl [74] and by Korner et
al. 1751.
Solvents of Cellulose. To measure the molecularweight,MWD, viscosity, and
other properties of cellulose, it is necessary to dissolve it. Due to the highlyordered
structure of cellulose, resulting from the formation of fibrils and microfibrils via inter- and
intramolecular hydrogen bonding, cellulose is not soluble in common solvents [76J. It has
been found that cellulose could be dissolved in strong acids such as hydrochloric, sulfuric.
and phosphoric acids [ 771, but its solubilization requires high concentration of the mineral
acid,andseveredegradation of the cellulosemolecule and, possibly, substitution will
occur. Different solvent systems for cellulose dissolution have been developed over the
years 178-1071. Cellulose is soluble in no simplesolvents.There are metal-amine (or
ammonia)systems,ammoniumthiocyanate-ammoniumsystems,amineoxidesystems,
and dimethylacetamide-lithiumchloride(DMAC-LiCI)systems.Cellulosedissolves in
these solvents without previous derivatization, but none of these methods is straightforward
and may suffer problems of color, oxidative instability, and degradation of dissolved cel-
lulose. Table 2 showssome of the solvents for cellulose withsome of their relevant
characteristics. Much attention has been given to the investigation of ccllulose derivatives
TABLE 2 Solvents of Cellulose
~~
‘(en): -NH-CH,-CH,-NH-.
Freer 288 and Baeza
(nitro, acetyl, or carbanil), usually as an indirect means to study cellulose. Cellulose de-
rivatives are soluble in common solvents, such as THE acetone, etc. The carboxymethyl
derivative of cellulose has been used in aqueous GPC [ 1 OS]. The conversion of cellulose
into derivatives in some cases is not entirely reliable, and the results obtained must be
considereda reflection of the natureof the derivative itself and onlyas a secondary
description of the original cellulose sample. There are solvents, such as the amine oxides
[96], which are capable of dissolving both cellulose and its derivatives.
The aqueous systems usually dissolve cellulose only after extensive swelling of the
fibers; the diffusion into the cellulose lattice conduces to either decrystallization or the
formation of specific crystalline inclusion complexes. Fast swelling leads to rapid disso-
lution. Dissolution in nonaqueous solvents occurs more readily; the fibers seem to explode
into numerous spindle-like fragments, which are rapidly dissolved.
In general, the metal-amine solvents present good dissolving power, and the relative
stability is dependent on the nature of the metal. The dissolving power is dependent on
the metal and the reacting base concentrations. In these types of solvents the complex-
forming metal ions are partly linked to the cellulose molecules. High concentrations of
both metal and base are required, limiting the use of the resulting solutions essentially to
molecular weight and viscosity determinations. For the determination of molecular weight,
the stability of the solutions of cellulose is the fundamental criterion. Solution in cuoxan
and cuen, which have been considered good solvents for cellulose, are easily autooxidized
and must be used under an atmosphere of nitrogen or hydrogen. Sometimes antioxidants
are added andor the solutions are prepared before using. The oxidative degradation of
cellulose solution in cadoxen is nearly negligible even for the highest-molecular-weight
samples. The rate of degradation in cuen is very much greater than in EWNN and in this
is slightly greater than in cadoxen, but still insignificant. The general characteristics of the
metal-amine solvents are included in Table 2 .
In general, the metal-amine solvents are not convenient for osmotic pressure and
dialysis experiments because the membranes are not resistant to them, nor for light-scat-
tering measurement, primarily because they are very strongly light-absorbing in the region
normally used. However, cadoxen solutions of underivatized cellulose prepared by a mod-
ified procedure have been used for osmotic pressure and dialysis experiments by using
ordinary cellulose membranes and also for light-scattering measurements [891.
Formeasurementofmolecularweight by the osmoticpressure or light-scattering
method, cellulose derivatives (nitrate or acetate) are usually used. Cellulose is solvated in
strongly polar, aprotic solvents, such as N,N-dimethylformamide (DMF) or N,N-dimethyl-
acetamide (DMAC) after introduction of a certain amount of dinitrogen tetroxide (NzO,)
or nitrosyl chloride (NOCI). A highly esterified cellulose nitrite ester is formed and the
cellulose is dissolved in the N,N-dialkylacylamides, which likely act as proton acceptors.
The cellulose nitrite is an unstable compound and unmodified cellulose will be regenerated
easily by contact with protic solvents.
The solubilization of cellulose in nonderivatizing solvent systems, such as DMAC-
LiCI, amine oxide, and liquid ammonidammonium salt systems, occurs only in selected
ranges of composition, and the condition of dissolution becomes critical.
The NH,/sal/H,O ratios that can be used to obtain acceptable cellulose dissolution
are in a very narrow range, and usually outside the conditions normally used for swelling.
N-methylmorpholine N-oxide ("NO), dimethylethanolamine N-oxide (DMEAO),
and a mixture with DMF or DMSO in the presence of different amounts of water have
been used as solvents for cellulose derivatives. The amine oxide/water systems may either
dissolve or only swell cellulose depending on the water concentration. This is of critical
racterization
Chemical of Wood 289
importance: up to a certain critical amount (19% for MMNO and 14% for DMEAO, and
intermediatevalues for the mixtures of bothamineoxides), the liquids are unable to
dissolve cellulose. The dissolving power of the liquid increases by removing more water
from the amine oxide system, the dissolution of cellulose in amine oxide-water systems
occurs only at low water concentrations. Small amounts of water are necessary to lower
the melting point of the amine oxide system and consequently to obtain dissolution of
cellulose at lower temperatures, avoiding its degradation. Apparently, a certain amount of
water is needed to reactivate and reopen the internal areas of cellulose pores that have
been closed or deactivated by drying during the pulping process. To be able to work at
low temperature (near room temperature), it is necessary to dilute the solutions with an
organic solvent (such as DMSO or DMF). The amounts of the organic solvent that may
beadded to cellulosesolutions in MMNO/waterareconsiderablebefore the cellulose
precipitates.
For "NO, the active dissolving part of the molecule is the N - 0 appendage, which
forms up to two H bonds with hydroxylated compounds. MMNO in the presence of 13.5%
of water (corresponding to the monohydrate) dissolves the cellulose easily. When 2 moles
of water (23.5% of water) are present to interact with the N - 0 bond there is no further
driving force for NMMO to interact with the cellulose OH groups. In the case of DMEAO,
a similar rationale has been developed, where the N - 0 group is able to form only one
hydrogenbondwith an external OH due to which it may adoptacyclicconformation
stabilized by the H bond.
In the dissolution of cellulosewithLiCI-DMACsysteman initial activation with
water to induce swelling of the fibers is required. A subsequent solvent exchange proce-
dure, in which the water is replaced with DMAC, must be realized. Complete removal of
water is necessary to achieve thorough penetration of the solvent into the sample and total
dissolution. The LiCl forms a complex with dimethylacetamide, releasing Cl-, which acts
as a base toward the cellulose hydroxyl group hydrogens. A possible mechanism of inter-
action of cellulose with LiCl and DMAC is shown in Fig. 6. Other salts of Li, such as
LiBr, LINO,, etc., do not work. The solutions of cellulose in the DMACLiCI system are
very stable at room temperature for a long time.
Viscosity of Cellulose Solutions. The molecularweight of macromoleculessuch
as cellulose can be determined by absolute and relative methods. The average molecular
weight of cellulose is commonly obtained from the determination of viscosity of a solution
FIGURE 6 Solvated complex of cellulose, LiCI, and DMAC. (From Ref. 97.)
290 Baeza and Freer
= [q]+ k'[q]'C
VSP
7
L
TABLE 3 Constants for the Mark-Houwink Equation [TI = KM" = K'(DP)o as Determined for
Cellulose and Cellulose Derivatives in Several Solvents
system
Solvent K (dL/g) K' (dL/g) a Method Ref.
Cellulose
Cuoxan 6.8 X 10" 0.9 SD
Cuen 9.8 X lo-' 0.9 SD
Cadoxen 3.38 X 10" 0.75 SD
3.85 X 10 0.76 SD, LS
3.15 X 10.' 0.93 GPCNISC
EWNN 6.6 X 10" 1.01 SD
NH3/NH4SCN 6.86 X IO-' 0.95 VISC
MMNO/H?O/DMSO 19.9 X 10" 0.79 VISC
[ 1/1.5 (w/w)]
Type
Number-average
molecular - 2,
tLM, Osmotic
pressure;
reducing
weight M,, = - end group
2,n,
Weight-average
molecular - x, "1 scattering
Light
weight
Z-average molecular
M,v= -
x, ILM,
weight
Viscosity-average Viscosity
molecular weighth
- \ /
nitrate was 1.9 (M,$. = 2.7 X IO', by light scattering) [ 1201. For Novacel K wood pulp the
M,,JM,, was found to be 2.72 with DP,,. = 870 and DP,, = 320, while for Tyrecell N wood
pulp the values were 2.69, DP,,. = 1385, and DP,, = 430 [ 1211.
For MWD determinations, size-exclusion chromatography (SEC), such as GPC and
gel filtration chromatography (GFC), has gained wide acceptance as a preferred method.
SEC is a special form of liquid chromatography; it is an entropically controlled separation
technique that depends on the relative size or hydrodynamic volume of a macromolecule
withrespectto the size andshape of the pores of the packing.GFC is referred to as
aqueousSEC at less than 1 kPa,whilesystemsusinghighpressuresandnonaqueous
solvents are referred to as GPC. With high-performance SEC (HPSEC), the relative MWD
can be obtained. If the column is calibrated with known molecular weight samples and/
or using molecular weight detectors, average molecular weights values can be obtained.
In other words, in conventional SEC the raw data represent the elution volume distribution
of the polymersample by weight,and a transformation of it tomolecularweight is
required.
Cael et al. [ 1221 and Lauriol et al. [ 1231 used a double detection system consisting
of a low-angle laser light-scattering photometer (LALLS) and UV to determine the MWD
of cellulose which had converted into tricarbanilate. SEC-LS detectors have been used to
characterize polysaccharides [ 1241. Triple detector systems LS andviscosity detectors, both
combined with SEC, were used to study carbohydrates [ 1251. SEC-FTIR has been reported
for the characterization of cellulose esters [ 1261. The capability of SEC has been improved
considerably by development of online viscometers and light-scattering detectors which
have greatly extended the usefulness of SEC [ 1241.
The appropriate analytical techniques are generally dependent on the dissolution of
the polymer. Two aspects must be considered in MWD of cellulose:
1. Due to the insolubility ofcellulose in the organicsolventscommonly used as
an eluting phase in GPC, it is necessary to convertcellulose into stable and
soluble derivatives.
2. It is necessary to select standards to obtainameaningful calibration curve so
that data from the GPC chromatograms can be converted into MWD curves to
obtain true average molecular weights.
The calibration procedure has been generally achieved by using the Mark-Houwink
coefficients ( K and a ) derived from both the polymer under study and narrow-distribution
polystyrene standards. Theaveragemolecularweightsand MWD obtainedfromsuch
calibration methods depend on the correctness of K and a for the polymer/solvent pair.
Different efforts have been realized to obtain more realistic calibration curves.
In the initial application of GPC to study cellulose, it was converted into cellulose
nitrate. Some examples are given below.
Segal [ 109,119,127]studied the MWD of nitrated cottonand wood cellulose by
analyzing the chromatograms of the nitrate derivatives in THF solutions, using polystyrene
calibration curves. The values for the average molecular weights were markedly greater
than those measured viscometrically. The authors consider that the extensive solvation of
the cellulose nitrate molecule by THF may account for these results. However, the results
of Meyerhoff and Havanovics [ 1281, by comparing the integral distribution obtained from
the precipitation fractionation of cellulose nitrate and the GPC chromatograms in THF,
gave lower average degrees of polymerization than those obtained by polystyrene calibra-
tion curves applied to cellulose nitrate. Huang and Jenkins [ 1211, using a calibration curve
based on cellulose nitrate, showed that the average molecular weight was in good agree-
aracterization
Chemical of Wood 293
common solvents. Some methods have been devised for the determination of nonderiva-
tized cellulose [98-99,1411. Kennedy et al. [98] reported the MWDs of cellulose samples
dissolved in DMAC-LiCl results. They were determined by SEC using a poly(styrene-
divinylbenzene)GPCcolumnwith a wide fractionation range (10' to IO7 polystyrene
MW).Thesame solventsystemwasused as the mobilephase. The methodinvolves
activation of the cellulose with waterto induce the swelling of the fibers, solvent exchange,
and complexation with LiCI. The amount and the length of the dissolution time required
depends on the cellulose sample. Birch and spruce cellulose samples required less LiCl
than cotton linters. The chromatographic traces of the samples show bimodal distributions
for the softwood and hardwood samples with a peak eluting very close to the exclusion
volume of the column and the other within the fractionation range. Two sharp peaks were
obtained for the birch hardwood sample, whereas for the softwood sample the peaks have
relatively higher dispersities. Timpa [99] appliedthe universal calibration concept to obtain
MWDs for cotton cellulose, corn, and wheat starch flours and avocado cell wall polymers.
Heused the samesolventsystem(DMAC-LiCl) to characterize them by SEC,witha
refractive index and viscometry.
Applications of SEC to the characterization of cellulose and cellulose derivatives
were reviewed [ 142- 1441.
b. OtherChcwacterizationProperties. Variousinstrumentaltechniqueshavebeen
used to characterize cellulose. These include infrared (IR) spectrometry, Raman spectros-
copy, NMR, and X-ray and electron diffraction.
In the solid state a regular system ofthe hydrogen bonds between cellulose molecules
results in an ordered system with crystal-like properties. There is also a noncrystalline or
amorphous portion in cellulose. The degree of crystallinity (crystallinity index), which is
a measure of the crystalline portion in a cellulose sample, depends on the origin of the
cellulose. The crystallinity index can be determined by X-ray diffraction [145,146], and
by IR using ratios of certain absorption bands [ 147,1481. The values of crystallinity index
for wood pulps range from 60% to 70%.
Wood celluloses tend to be less crystalline than other celluloses. therefore less ame-
nable to study by diffraction methods. Also, wood celluloses are highly sensitive to elec-
tron beam damage at doses far lower than those required for usual high-resolution electron
microscopy [149]. On the other hand, Valonia cellulose is highly crystalline and resistant
to radiation. It is a widely accepted standard for natural cellulose. A considerable number
of studies on cellulose structure have used Valonia cellulose [150- 1521.
Variation in the physical structure of cellulose has been observed according to its
sources and developmental stage. This variation, which includes differences in microfibril
crystallographic orientations, degree of polymerization, crystallite size, pattern of glucan-
chain hydrogen bonding, and glucan chain polarity, has made it difficult to determine the
basic crystalline structure of cellulose.
The most common crystalline form (allomorph) of native cellulose is cellulose I,
which is metastable and can be irreversibly converted into another crystalline state. TWO
distinct forms of cellulose I (I, and Io) have been reported by Atalla and VanderHart [ 1531
based on CP-MAS I3C NMR evidences. Morerecently, the electron diffraction experiments
of Sugiyama et al. [ 1541 demonstrated the coexistence of these two phases (I, and Ia) as
regions of different crystalline structure rather than alternative chains along the cell wall.
Cellulose I, and I, differonly in their patterns of hydrogenbonding; their molecular
conformations are identical [155,156]. The I, and I, allomorphs are most easily distin-
guishable via the C4 resonanceof certain solid-state I3C-NMR and by OH stretching
regions of the Raman spectra [155,157]. Electron diffraction patterns are consistent with
aracterization
Chemical of Wood 295
triclinic and monoclinic unit cells in these two forms [156]. The native celluloses can be
classified into two types: algal-bacterial and cotton-ramie-wood types, which are referred
to as celluloses I,, and I,, respectively [ 1581. The cellulose I1 polymorph is formed from
cellulose I by mercerization or by precipitation from solution, and it is the most stable
allomorph known. These two are the most common polymorphs. Other forms (cellulose
111, IV, and the so called cellulose x) have been reported. Transformation of cellulose into
its various lattice modification is summarized by Fengel and Wegener [ 1591. Each cellulose
allomorphcanbe identified by its characteristic X-rayor electron diffraction pattern
[160,161].
Infrared techniques have been used to analyze the crystal structure of cellulose and
to study cellulose reactions. IR in solid-state techniques can be used to distinguish among
cellulosesamplesfromvariousorigins.Theassignmentof the IR absorptionbands of
cellulose is basically derived from that of the glucose unit. Studies of the polarized infrared
spectrum [ 162- 1661, the effect of deuteration [ 1631, and the correlation of the bands or
groups of bands with the bands of chemically related compounds [ 167,1681 have given
information for a complete assignment of the absorption bands. Table 5 shows the relative
intensity, polarization, and assignment of the main absorption bands of celluloses I and I1
[ 1691.
The early work on the vibrational spectra of cellulose was summarized by Blackwell
[170]. The information obtained from conventional IR was limited by lack of resolution
of bands in the spectra.With the introduction of the Fourier-transformspectrometers
(ITIR), new methods for measurement and evaluationaredeveloped. Theincrease in
sensitivity resulting from the combination of FTIR and computer techniques has greatly
enhanced the usefulness of this technique. Improved resolution can be obtained by decon-
voluting the spectra [ 17 1,1721, curve-fitting [ I731 or second derivatization [ 174- 1771.
FTIR spectra of cellulose, wood and other major components are shown in Fig. 7.
Marriman and Mann [ 1781 were the first to point out the difference between spectra
obtained from Valonia and bacterial celluloses, in which a band at 3242 cm” appears,
and those of tunicin, ramie, or cotton, in which this band is absent. They suggested that
the difference was “either due to a different crystal structure in the two cases or to a
difference in degree of perfection of crystals having the same basic molecular arrange-
ment.” Liang and Marchessault [ 1641 also observed an intense band at 3245 cm” exclu-
sively in Valonia and bacterial celluloses. More recently, Michell [l761 studied a series of
native celluloses by using a second-derivativetechnique,supporting the hypothesis of
Marriman and Mann that the native celluloses possess two different crystal structures.
Transformation between phases of cellulose by some treatments has been monitored
by IR. The conversion of I, phase to the most stable I, by an annealing treatment has
been reported by Sugiyama et al. [161]. It was demonstrated by combined infrared and
electron diffraction techniques that some absorption bands in the IR are specific to the I,
(3240 and 750 cm”), while others correspond to I, cellulose allomorph (3270 and 710
cm-’).
Michell [l771 analyzed the changes in structure of a eucalyptus wood cellulose by
treating it with solutions of NaOH ranging from 0 to 20%. Better resolution of the bands
by using second-derivative FTIR spectroscopy can be used in the conversion of native
cellulose into mercerized and into an amorphous form. It can be used to confirm known
changes in band intensities and to discover more subtle changes. The number of bands in
the second-derivative spectra of cellulose sensitive to mercerization is far greater than the
four bands (at 1428, 1111, 990, and 893 cm-’) which were identified earlier by normal
spectra reported by McKenzie and Higgins [ 1791.
296 Baeza and Freer
Frequency Relative
(cm ") intensity
Polarization Assignment
1030 vs
1069 vs CO and CC stretching and CH, rocking
1078 vs
1 l06 vs
1117 vs
1 l61 S
1204 W
I232 W CH2 wagging
1249 W CH and OH deformation
1275 m
1310 m CH deformation
1335 m CH in-plane deformation
1365 m CH deformation
I426 m CH, symmetric bend
1446 m OH in-plane deformation
1630 111
1650 m Adsorbed HzO
1720 m
Ca 2340 vw
Ca 2500 vw
Ca 2700 vw
285 1 W CH2 symmetric stretching
2907 m CH, stretching
2967 W CHZ antisymmetric stretching
Ca 3300 vs OH stretching
Ca 3400 vs OH stretching
By Liang et aLh
663 m
OH out-of-plane bend
-700 m
-740 sd CH2 rocking?
-800 sd Ring breathing (p)
895 W Antisymmetric out-of-phase stretching
985 sd
1000 sd
1015 m CO stretching
1035 S
1058 S
TABLE 5 Continued
Frequency
Relative
(cm-Polarization
') intensity Assignment
1005 S
1020 S CO stretching
1035 S
1060 vs
I078 vs
I l07 5 Antisymmetric in-phase ring stretching
I133 S Antisymmetric bridge oxygen
I200 m OH in-plane bending?
1225 m
I273 W
1277 m CH bending
1315 W CH2 wagging
1335 W OH in-plane bending
I365 m
1375 m CH bending
1416 W CH2 bending
Freer 298 and Baeza
TABLE 5 Continued
Frequency Relative
(cm"') intensity
Polarization Assignment
sh 1440 II
1470 sh I OH in-plane
bending
1635 - - H,O
Adsorbed
2850 W II? CH, symmetric stretching
$1
2874 sh
289 1 S CH stretching
2904 S L
2933 m I? antisymmetric
CH2 stretching
2955
2968
W
W "l
I1
CH stretching
298 1
3175
3305
W
F
S
:I)
I
OH stretching
(intermolecular
hydrogen bonding)
3350 S I
3447
3488
vs
S "l
II
OH stretching (intermolecular hydrogen bonding)
0.90 /- h
a
0
0
a
a
! \ I I
Frequency (cm-')
FIGURE 7 FTIR spectra of E. regnas wood and its major components. (From Ref. 174.)
with the ordered regions (sharper resonances) and those broader on the upfield side of the
crystalline counterparts resonating downfield (associated to disordered regions) allow the
crystallinity to be determined.
The solid-state transformations (I(, + I,) of Valonia cellulose crystal by an alkaline
hydrothermal treatment have been analyzed by comparing the diffraction diagrams before
and after hydrothermal annealing, which shows substantial modifications [161]. This trans-
formation has also been observed by I3C-NMR [194,195]. The spectrum of annealed Va-
Freer 300 and Baeza
lonia cellulose shows a well-resolved multiplicity of two for the signals of the carbon
atoms at C 1, C4, and C6 of the glucose moieties, indicating only two glucose residues
that are magnetically nonequivalent in the unit cell. In the spectra of the original Valonia
samples, at least three signals were observed for these carbons.
The relevant signals in the I3C-NMR spectra that permit distinguishing between I,,
and I, are well resolved in spectra of algal or bacterial celluloses, but poorly resolved in
the spectra of wood or pulp. Studies of combinations of "C-NMR with the analysis of
principal components 11961 and/or improved resolution-enhancement function [ 1971 have
been reported. By using the latter approach. the ratio IJI, is estimated from the ratio of
heights of the peaks at 90.2 and 88.5 ppm assigned to C4 in I,, and I, crystallite interiors,
respectively. Values of 1.8, 0.8,and0.4havebeenreported for softwood timber, ther-
momechanical pulps, and Kraft pulp, respectively [197,198]. The effect of temperature,
time, and alkali concentration in the conversionbetweencrystallineformsofcellulose
during pulping has been reported [1981.
B. Hemicellulose or Polyoses
Hemicelluloses constitute 20-30% of wood. They are found predominantly in the primary
and secondary cell walls. A smaller amount occurs in the middle lamella. They differ from
cellulose by containing various sugar units, with much shorter chains, and by branching
of the chain molecules. They are soluble in alkali, and some hardwood polyoses are even
soluble in water. Hemicelluloses are more reactive than cellulose.
Hemicelluloses are generally classified according to the types of sugar residues pres-
ent. Thus, xylan is a polymer of ~-xylosylresidues and mannan a polymer of D-mannosyl
residues (homoglycans). However, the most frequently found hemicelluloses contain two
to four, and rarely five or six, different sugar types (heteroglycans).
Hemicelluloses in softwoods and hardwoods differ not only in the percentages of
total hemicelluloses but also in the percentages of individual polyoses and their compo-
sition. In softwoods, the major hemicelluloses are partially acetylated galactoglucoman-
nans, and a small amount of arabino-4-0-methyglucuronoxylans.In hardwoods, the pre-
dominant hemicelluloses present are 0-acetyl-4-0-methyglucuronoxylans (acetylation of
the xylose groups is about 70% on C-2 or C-3) with a small proportion of glucomannans.
Partial chemical structures of polyoses are shown in Fig. 8.
COOH
0
OH 1
OH OH OAC I
CH,OH CH20H OH
FIGURE 8 Partial chemical structures of hemicelluloses: (a) 0-acetyl-4-0-methylglucuronoxylanfrom hardwood; (b) 0-acetyl-galactoglucomannan from
softwood.
302 Baeza and Freer
2. Characterization of Hemicelluloses
Once a hemicellulose has been purified, its structure can be determined. The sample is
characterized by determination of constituent sugarsandsugar acids, specific rotation,
linkage and sequences, functional groups, molecular weight, and MWD.
a. Am1y.si.s of Wood Curbohydrates. In the structural analysis of a hemicellulose it
is necessary first to determine the kind and quantity of glycosyl units present. Hemicel-
luloses are commonly hydrolyzed by mineral acids, formic acid [24O], or trifluoroacetic
acid[241,242].Thesubsequentdeterminations of the componentmonosaccharidesare
normally carried out by chromatographic techniques.
Methodologies for analysis of wood sugars have undergone rapid advances in recent
years. Older methods applied to determine carbohydrate composition of wood and pulp
are paper or thin-layer chromatography. These techniques were replaced by GC following
derivatization, which offers the advantage that it can easily be connected to MS. Cation-
exchangeHPLCcoupledwith refractive indexdetectionwasfound tobe simplerand
faster. Recently,anion-exchangeHPLCwithpulsedamperometricdetection(PAD)was
found to have the advantages of greater sensibility and easier sample preparation. Also,
capillary electrophoresis has been shown to be a valuable tool in analyzing carbohydrates.
racterization
Chemical of Wood 305
xylan from unbleached hardwood and softwood were characterized with respect to their
content of mono- and oligosaccharides.
Carbohydrate Determination. TAPPITestMethodT249cm-85[2] or other
equivalent methods (e.g., ASTM D 1915 [3]) are accepted procedures for the analysis of
wood and pulp carbohydrates. The determination of the carbohydrate composition is car-
ried out by gas chromatography of the carbohydrates as alditol acetates [255,256]. The
TAPPI method involves hydrolysis with sulfuric acid of extractive-free samples using a
two-step technique. Inositol is added as internal standard and the solution is neutralized
with Ba(OH),, and then the sugars are reduced to alditols with NaBH,. The alditols are
acetylated with acetic anhydride using sulfuric acid as the catalyst, and the alditol acetates
are precipitated in ice water and extracted with methylene chloride and analyzed by gas
chromatography. The analysis by GCusing the alditol acetates is time-consumingand
involvesnumeroussteps,andeachstep requires skilled care if the procedure is to be
quantitative. The alternative trimethylsilyl or trimethylsilyloxime derivatives are easier to
prepare [257], but the chromatography is complicated by isomerization [258].
Liquid chromatography has been also used for carbohydrate analysis and offers the
advantage of not requiring derivatization to chromatography. Several low-pressure liquid
chromatography systems have been developed for analysis of carbohydrates [259-2621.
HPLC has been widely used for the analysis of carbohydrates. Various methods, including
aminoalkyl bonded silica [263-2651, reverse-phase C,, [266], and ion-exchange columns
[249,252,267-2691, have been reported.
Ligand exchange on calcium and lead cation-exchange columns [267] coupled with
refractive index detection and water as eluent have been widely used for analysis of wood
sugars. The sugar content in pulp mill effluents, TFA-hydrolyzed pulp and wood samples,
by using lead-based columns and refractive index detection, was reported by Paice et al.
[252]. The nature of the sugar-metal ion interaction was reviewed by Angyal [269].
The monosaccharides commonly found in wood and wood pulp hydrolyzates can be
separated by using anion-exchange resins. The separation is based on the weakly acidic
properties of carbohydrates in alkaline conditions. At high pH, neutral and acidic carbo-
hydrates are partially or completelyionizedand thus retained on the column.Column
selectivity is changed using a NaOH gradient. Anion-exchange HPLC (HPAE) with pulsed
amperometric detection (PAD) was found to have the advantages of greater sensitivity and
easier sample preparation [270]. PAD is based on oxidation of the carbohydrates in mul-
tistep potential waveforms (E,) applied to Au electrodes (sensitivity for PDA of carbohy-
drates is larger at Au than Pt electrodes [27 11) in miniature flow-through cells. During the
first potential, E,, the signal from carbohydrates is measured, and the second and third
potentials (E? and E3) correspond to oxidative and reductive cleaning pulses, respectively,
and permit the maintenance of the electrode activity. This technique is applicable to mono-
saccharides, disaccharides, and oligosaccharides.
The analysis of wood and wood pulp hydrolyzates by HPAE-PAD in combination
with acid or enzymatic hydrolysis has been carried out by various authors 1253,270,272-
2781. HPAE-PAD provides rapid and versatile methods for carbohydrate analysis. It is a
sensitive method of high selectivity and specificity. Sample preparation is minimal, without
neutralization or monosaccharidesderivatization;onlygroups that containoxidizable
groups are detected. It is possible to effectively separate and analyze quantitatively trace
amounts of monosaccharides released from hemicellulose hydrolysis in the presence of as
much as 95 times the amount of D-glucose from cellulose hydrolysis. HPAE-PAD chro-
matograms are shown in Fig. 9.
Chemical Characterizationof Wood 307
3 1. Fucosa
Deoxyribose
2. 25
3. Arabinose
25
. Galactose
25
. Glucose
25
1A 8 16. Xylose
25
0 5 10
TIME.rnin
-
15
J
(b)
20
L
1. Arabinose
2. Galactose
3 3. Glucose
4. Xylose
5.Mannose
1 5
I I I I
0 155 10 20
TIME, rnin
FIGURE9 HPAE-PAD chromatograms: (a) common monosaccharides; (b) hydrolyzed wood pulp.
(From Ref. 274.)
The degradation of sugar during the acid hydrolysis and the effects of other param-
eters of sample preparation on the measurement of wood sugars by anion-exchange HPLC
using PAD were investigated 12771. Unhydrolyzed and hydrolyzed sugar standards, and
wood meal (Pinus t m d a ) and wood meal spiked with a standard before hydrolysis were
analyzed. Loss of sugar in hydrolyzed standard ranged from 6.4% for arabinose to lS.796
for mannose. In wood samples to which a standard was added before hydrolysis, the results
were very close to the sum of wood samples and hydrolyzed standard analyzed separately,
except in the case of galactose, which was about 4.7 lower i n the combined sample. These
results indicate that the effect of hydrolysis on sugars is probably the same in wood. Thus,
the other wood components apparently do not protect sugars or enhance their degradation
during the hydrolysis.
Threechrornatographic methods for the quantitativedetermination of wood sugar
were compared for the ease of their operation and accuracy [270]. They were: ( I ) borate
complex anion-exchange chrolllatography-reducing sugars were detected by postcolumn
derivatizntion with Cu-bicinchoninLlte (0.33 mL/min) measuring the absorbance at 546 n m ;
308 Baeza and Freer
Buffer vials
[O - 30 kV1
7 9 i3 1'5 Min
the complete methylation is the absence of a hydroxyl peak near 3500 cm" in the IR
spectrum. Details of the Hakomori procedure are given elsewhere [304,3051.
Modifications to the Hakomori method have been proposed [306,307]. Some poly-
saccharides that show strong resistance to permethylation by the Hakomori procedure may
be completely methylated with a mixture of DMSO and 1,1,3,3-tetramethylurea in a 1 :1
ratio 13071.
If uronic acids are present, the complete permethylation is more difficult and may
yield secondary products. Methylation is facilitated by reducing the carboxyl groups of
the original polysaccharide or the fully methylated polysaccharides to hydroxyl groups. It
has been recommended the following steps: first an initial methylation, forming a uronic
acid methyl ester, then the reduction of this ester with lithium aluminum hydride to a
hydroxyl group, and finally a remethylation [305]. In the case of alkali-labile polysaccha-
rides it is advisable to prereduce with borohydride or borodeuteride to avoid "peeling"
reactions [308].
The methylation techniques have been reviewed by Rauvala et al. 13081.
Periodate Oxidation. Periodate oxidation of polysaccharides is usedin the struc-
tural characterization andmonosaccharidesequencedetermination of polysaccharides
[309]. Carbohydrate residues having glycol groups are oxidized to dialdehydes (Fig. 1 l ) ,
but if the residues contain hydroxyl groups in three adjacent carbons, formic acid can also
be produced. In addition to glycol units, oxidation occurs in a-hydroxyaldehydes, a-hy-
droxyketones, and a-hydroxyhemiacetals. Formaldehyde is produced from the oxidation
of primary hydroxyl groups, and formic acid and aldehyde or carbinol units between two
other hydroxyl-bearing carbons [3 lo].
The oxidation is conducted in a dilute solution of sodium metaperiodate in the dark
at low temperature and at pH 3-3.5. The periodate consumed, formaldehyde and formic
acid produced, and the characterization of the oxidized polymer provide information on
the molecular structure, nature of the end groups, and point of linkages between the con-
stituents. Additional information may be obtained by applying the Smith degradation se-
quence,whichcomprisesaborohydridereductionof the periodateoxidationproducts
followed by a mild acid hydrolysis of the polymer (Fig. 12) 131 l]. Normally, when a sugar
residue of a polysaccharide is cleaved by periodate and reduced, it corresponds to an acetal
which is acid-sensitive. When a sugar unit that survives oxidation is joined to a cleaved
unit, it appears as a glycoside that is relatively stable to acid. Due to the difference of
stability between true acetals and glycosides, the products of periodate oxidation reduction,
and mild hydrolysis are low-molecular-weight glycosides, glycoaldehyde, and polyalcohol.
Analysis by GUMS after conversion to volatile derivatives provides information on the
structure of the parent polysaccharide 131l ] .
'l0dH CH,OH
NalO,
'lot CH,OH
CH0 OHC
OH
CH,OH CH,OH
o j R z '
lo< oO
j Rz
NaBH,
____)
CH0 OHC
CH,OH
HOCH,
CH,OH
1
mild acid HCOH
l
+ CH0
I
+ R,OH
hydrolysis HCOR, CH,OH
I
CH,OH
CH,OH
l
acid HCOH
COH hydrolysis
I + CH0
I + R,OH
I CH,OH
CH,OH
FIGURE 12 Smithdegradation.
Lead tetraacetate also cleaves glycols, similarly to periodates. The reaction is usually
conducted in acetic acid because lead tetraacetate decomposes in water. Methyl sulfoxide
can be also used as solvent [3 121. The formic acid produced in the reaction is further
oxidized to carbon dioxide by lead tetraacetate. Lead acetate has been used to differentiate
between cis and trans hydroxyl groups in glucomannan [313].
Another analytical scheme for identifying periodate oxidation products of polysac-
charides is knownas the Barrydegradationandhasbeendescribed by Lindberget al.
[314]. In the Barry degradation the periodate-oxidized products are reacted with phenyl-
hydrazine in dilute acids, followed by hydrolysis to release the phenylhydrazone-contain-
ing units. Some polysaccharides can be degraded by the Barry method with a sequential
removal of monosaccharides residues. Then the sequence of carbohydrates residues in the
polysaccharide chain can be determined.
Partial Hydrolysis. Partial hydrolysis, by either acids or enzymes, gives a mixture
of monosaccharidesandoligosaccharides.Glycosidicchainlinkagescanbeestablished
through the carbohydratefragments.Rates of monosaccharideproductionalsoprovide
information on the location of single sugar units as branches and the nature of their ring
forms.
The rate constants for the acid hydrolysis of the glycosidic bond in polysaccharides
vary greatly depending on the monosaccharide composition, positions and configuration
of the glycosidic linkages, and the ring form. This different resistance of the glycosidic
linkage can be advantageous. For example, the resistance of the glycosidic linkages of the
uronic acids makes it possible to obtain, by acid hydrolysis, a good yield of aldobiouronic
acids (separated easily) and its glycosidic linkage can be determined after reduction of the
carboxyl group to produce the neutral polysaccharide, which can be analyzed. Uronic acid
Characterization
Chemical of Wood 313
therefore the results are suitable only for comparative purposes and do not represent the
true molecularity.
The elution behavior of 4-O-methylglucuronoxylan,isolated from birch, has been
studied using a Separon HEMA 1000 column and DMSO and DMFA as the mobile phases
[345].The polyelectrolyte effectsweresuppressed by the addition of acetic acid and
lithium bromide (0.03 M). Fractions of the same polymer, characterized by viscometry in
cadoxen, were used for column calibration. Thismethodproved useful to estimate the
MWD of xylans rapidly. HPSEC for the determination of MWD of wood hemicelluloses,
using the same type of column with 0.5 M NaOH as eluent, has been investigated by
Eremeeva and Bykova [238]. Nonexclusioneffects were observed on the Separon S HEMA
1000 column in 0.5 M NaOH. In this method, the analysis of the alkaline extracts from
wood pulps was performed without previousisolation of the hemicelluloses of the solution,
shortening the analysis timeandminimizing the changes in the hemicelluloses due to
sample preparation. Only extraction with basic solutions and filtration are required. Both
a series of dextrans and xylan fractions were used to calibrate the column. It was shown
that the universal calibration between dextran and xylans is valid.
VI. LIGNIN
Different techniques have been reported to characterize lignins. Lignin is the second most
abundant biopolymer. It consists of an aromatic system composedof phenyl propane units.
Figure 13 shows a typical representation of a lignin model [346]. An important charac-
teristic of this natural polymer is the presence of different functionalgroupssuch as
phenolics,methoxyls, aliphatic alcohols, aldehydes, ketones, and ethers [347]. This last
group together with carbon-carbon bonds are present in the formation of the polymeric
bonds. The frequency of thesegroupsand the molecularweight are important in the
characterization of the lignin. The lignin composition depends on different factors. Lignins
from different plants differ appreciably in their structure, and it is possible to distinguish,
for example, softwood lignin (“guaiacyl lignin” with coniferyl alcohol as main polymer
unit) and hardwood lignin (“guaiacyl-syringyl lignin” with coniferyl and sinapyl alcohols
as main copolymer units). Morphological locations from which lignin is isolated are also
important, as shown in Table 6, in which the guaiacyl-syringyl ratios and the distribution
in lignins in white birch, determined by UV-EDXA technique, are given [348]. Another
important factor is the method used to isolate the lignin, due to the appreciable changes
that occur during the isolation procedure.
The chemical structure of lignin affects the reactivity ofwood during industrial
processes, such as pulping and bleaching, and also the final fiber and the possible new
applications of technical lignins. For this reason, the characterization of lignin has been
an important concern and many research groups are working in the development of dif-
ferent analytical techniques to obtain more structural information. A large number of papers
has been the result of this activity.
A. Isolation of Lignin
The analysis may be carried out using lignins isolated by different methods or directly
from wood or pulp, without previous isolation. The method used in the isolation of the
lignin is a very important aspect to be considered from the analytical point of view, and
316 Baeza and Freer
H2COtI
H3CO HC=O
I
HC
ICH2OHI 4 I
HC-O-
l
HC-O-
H2COH
H2:OH
I
I
I
HCOH
H$O 4II
0-CH
I
H2COH
I
HC-0
OCH3
H 3 C 0 p " '
HOH2C"C"C=O
H~;N'\CH
I l
HCOH
0- OCH3
OCH,
OCH3
H,CO
OH OH [U-C1
for that reason it is necessary to specifythe type of lignin.The different lignins from
wood are:
Milled wood lignin (MWL),orBjorkmanlignin, isolated with tolueneafter ball
milling 13491. I t has been postulated that it is a lignin with a chemical structure
close to native lignin.
Brauns native lignin (BNL), isolated by extraction with ethanol, following by pre-
cipitation with ether [350].
Klason lignin. isolated from extracted wood meal after treating with cold concen-
trated sulfuric acid (TAPPI Standard Method T-222 om 8 3 ) 121.
Cellulolytic-enzyme lignin (CEL), isolated from ground wood, using an enzymatic
hydrolysis [ 35 1,3521. This is a slow procedure, but the lignin suffers less struc-
tural alteration than MWL lignin.
Swelled-enzyme lignin (SEL), in which wood nleal is swelled in an organic solvent
before treatment with a cellulase 13531.
Organosolv lignin, extracted with different solvents or as side product in an orga-
nosolv pulping process [354-3581.
Chemical Characterizationof Wood 317
Lignin
Tissue
Morphological Guaiacyl-syringyl concentration
Element region volume (%) ratio (g/&
“Fiberltiber.
hFiber/vessel.
‘Fiberhay.
“Raylray.
Kraft lignin, obtained during the kraft (sulfate) process. This is the most abundant
industrial lignin [359].
Lignosulfonate, obtained in a sulfite pulping procedure [360].
Details of lignin isolation procedureshavebeendescribed by Browning [361]. A
critical review of methods of isolation of lignin has been published by Lai and Sarkanen
[362]. The subject has been also discussed by Lundquist [363], Fengel and Wegener [364],
Brauns [365], and Brauns and Brauns [366].
B. Characterization of Lignin
The chemical characterization of lignins is carried out using:
Degradative procedures, in which the lignin is depolymerized through chemical re-
actions, followed by the identification of the low-molecular-weight degradation
products.
Nondegradativeprocedures, in which the polymer is characterizedwithoutdegra-
dation. Spectroscopicmethods are the mostcommonnondegradative pro-
cedures.
1. Degradative Procedures
In acidolysis, the lignin is refluxed i n a mixtureofdi-
a. Acid~l~~.~i.~-Etlranol~~.~i.~.
oxane-hydrochloric acid (0.2 N) 9:l (v/v). The a- and P-aryl ether linkages are cleavage
due to protonation of the oxygen with the formation of a good leaving group. After com-
318 Baeza and Freer
OH OH OH
FIGURE 14 Monomericphenolsformedduringacidolysis-ethanolysis ( R , = Hor OCH3; Rz =
CH,CH,). (From Ref. 367.)
Ho+cHR, R
,H
,c--- ,~c--
OCH,
FIGURE 15 Thioacidolysisproducts.(From Ref, 37 I .) G ~ t i ~ Ser-ie.s: y l I -(4-hydroxy-3-mcth-
oxyphenyl)- 1,2.3-(tris-thioethyI) propane (R, = SC,H,: R? = H); Syr-ir~gyISeries: 1-(4-hydroxy-3.5-
din1ethoxyphenyl)- I .2.3-(tris-thioelhyl) propane ( R , = SC,H,; R? = OCH,).
Chemical Characterization of Wood 31 9
R 4 bR
R2
Et2 “BF,
R2
(-ROH)
R*
H,CSEt H2COH
L
H SEt
l
HCSEt
l
HCSEt
4R2
R, OH
(5)
EtSH
I
OR
(4)
R2 (-ROH)
OR
(3)
FIGURE 16 Mechanism of lignin thioacidolysis. (From Refs. 371, 373, and 374.)
and (5). This last step is important because it avoids the condensation reaction present in
the acidolysis procedure. When R, is different from R2,50% of each isomer, erythro and
threo, are obtained.
The total and the relative amounts of guaiacyl and syringyl groups give information
about the phenylpropane units involved in arylglycerol-p-aryl ether bonding pattern. Table
7 shows the results reported by Lapierre et al. [375]. These results indicate that lignins of
woody angiosperms contain twice as many units of these types as lignins from woody
gymnosperms.
c. Nitrobenzene Oxidation ( N O ) . In this reaction, whichwasintroduced by Freu-
denberget al. [376], the lignin is treated withnitrobenzene in alkalinemedium ( 2 M
NaOH) at elevated temperature (170- 180°C). The products have been identified by GC-
MS after the residue is silylated (Fig. 17). Under these conditions, softwoods give vanillin
(main product) (17a) and vanillic acid (minor product)(17b). On the other hand, hardwood
lignins, in addition to these two oxidation products, give syringaldehyde (17c), and the
corresponding acid, syringic acid (17d) [377,378].
Reaction Mechanism. It is not clear which is the mechanisminvolved in this
reaction. Chang and Allan [379] proposed a two-electron transfer process with a quino-
nemethide intermediate. Iiyama and Lam [380] also suggested a two-electron mechanism,
but Schultz et al. [381,382] concluded that the oxidation involves a one-electron transfer
through a free-radical process.
320 Baeza and Freer
OH
a) R,=R,=R,=H
b) R,=R,=H, R,=COOH COOH COOH
& c) R,=R,=H,R,=COOH I I
d) R,=R,=COOH.R,=H
e) R,=OCH,,R,=R,=H
R1 9 R,=OCH3, R,=COOH,R,=H
g) R,=OCH,,R,=H.R,=COOH
H3CW O C H ,
OCH, OCH,
OCH,
h) R,=R,=OCH, R,=H
(11
(2)
+0cH3q0cH3
COOH FOOH
H,CO f 1 0 C H 3 COOH
R 0 I
OCH,
OCH,
a) R=H
(4)
b) R=OCH,
COOH
I
I
COOH COOH H,CO Q-QOCH, OCH,
\ a) R=H
OCH, OCH, b) R=OCH,
H,CO W O C H ,
OCH, OCH,
(5)
anol number" as the methanol concentration in units of mg/L produced during 5 min of
chlorination at 25"C, 1% consistency, and an initial chlorine concentration of 3.0 g/L.
Ozonation. Ozone iswell known asa chemicalreagent in the elucidation of
chemical structures. The main characteristic is its ability to react with the aromatic ring,
giving complementary information to that obtained by other oxidative degradation meth-
ods, suchaspermanganateandnitrobenzeneoxidationmethods.Importantinformation
FIGURE 19 Methanol formation. (From Ref. 390.) R = lignin side chain; R, = H, OCH3, or lignin
unit.
Freer 322 and Baeza
about the behavior of the different structural units present in the lignin during the ozonation
has been obtained from the study of lignin model compounds, Sarkanen et al. [394] gave
a list of monomeric and dimeric lignin model compounds that have been ozonated.
Matsumoto et al. E3951 described the products obtained by complete degradation of
the aromatic rings present in the lignin, with the aliphatic carboxylic acids the main prod-
ucts formed (Fig. 20). Several lignins (MWL, klason lignin, dioxane lignin, thiolignin, and
soda lignin) were reduced with sodium borohydride, then treated with ozone, followed by
saponification. Theproductswereanalyzed by "C-NMRandGC.Erythronic, threonic,
glyceric, and glycolic acids were detected. The ratio for the first two acids, which are
derived from erythro and threo isomers of arylglycerol-p-arylether type structures, was
almost 1 :1 for MWL and wood meal. Morerecently, it was reported [396] that when wood
meal is subjected to ozonation the erythronic and threonic acids are predominant, but not
in the case of unbleached kraft pulp. Taneda et al. [397] used this reaction to determine
the relative abundance of the steric structures of lignin side chain. The ratio of erythro
and threo isomers of the p-0-4 structure was determined in model compounds and in the
residue lignin during the kraft process. The results showed that the E R ratio decreased
during cooking.
g. NucleusExchangeReaction(NE). This reaction, developed by Funaoka et al.
[398-4051, consists of the treatment of lignin with boron trifluoride in an excess of phenol,
causing a selective cleavage of C,-C, linkages of phenylpropane and forming methylene
linkages of diphenylmethane type of structural units. The lignin aromatic units that are
displaced by the phenol are extracted with ether from the reaction mixture and estimated
by GC as trimethylsilyl derivatives. From model compounds studies, they found that all
bonds lignin units are cleaved with the exception of diphenyl ether and biphenyl linkages.
Reaction Mechanism. This is three-step reaction: (1) formation of the diphenyl-
methane structure by phenolization at the a-position of the lignin side chain, (2) a nucleus
exchange step that involves displacement of the phenyl nuclei of lignin for phenol, and
(3) demethylation of the methoxyl groups of the released phenols (Fig. 21).
As shown in Fig. 22, in the case of softwood lignins the uncondensed guaiacyl units
are converted first to guaiacol (22a), and this compound is partially 0-demethylated to
give catechol (22b). In addition to these two phenols, the hardwood lignins give 1,3-0-
dimethyl pyrogallol (22c) from the syringyl unit, which by subsequent demethylation gives
1 -0-methylpyrogallol (22d) and pyrogallol (22e).
This method in combination with the nitrobenzene oxidation process has been used
to evaluate the amount of noncondensed, condensed, and diphenylmethane units of guaia-
cy1 and syringyl in residual lignins during the pulping process [406,407]. Chan et al. [408],
more recently, indicated that the nucleus exchange reaction does not allow accurate mea-
surement of aromatic units in lignin. They found erroneously high syringyl contents of
lignins when lignins from eucalyptus woods andkraft pulps were studied. They also found,
from studies of the NE reaction using different models, that diphenyl ether and biphenyl
compoundsgavecathecol as a reaction product.Theyconcluded that the amounts of
noncondensed structures and the ratio of syringyl and guaicyl nuclei in lignin by the NE
reaction and the amounts of diphenyl moieties determined by NE-nitrobenzene oxidation
in modified lignins is likely to give incorrect results.
h.Pyrolysis-GasChromatography-MassSpectrometry(Py-GC-MS). This analyt-
ical techniqueincludespyrolysis (Py) to generate volatile degradationproducts, GC to
separate the degradation fragments, and MS as a detection system. The identification of
the products can also be performed using the normal flame ionization detector (FID), and
Fourier-transform infrared (FTIR). Several pyrolyzers are commercially available, such as
Chemical Characterizationof Wood 323
E
2
0 tL,
C
.-0
m
S
0
W
K
3
PU
324 Baeza and Freer
I
0
0--0--0
0
I
0
I
Chemical Characterizationof Wood 325
OH
JHP-3 model Japan Analytical IndustryCO;Chemical Data System Probe; PYROLA, Pyrol
AB, Lund, Sweden; Fischer, Germany.
This is an effective method for the characterization of polymers in a short time, with
simplepreparation (no special pretreatment), high sensitivity, andusingsmallsamples.
The application includes the analysis of the isolated lignin from the wood or pulp and
also the lignin without previous isolation (total lignin). Some good examples of Py-GC-
MS are the following, reported in the literature.
Obst [409] used this technique to pyrolyze wood and classify lignins as either guaia-
cy1 type or syringyl-guaiacyl type. He was also able to isolate vessel elements
and identified the type of lignin. The pyrograms of the woods and the milled
wood lignins are clearly different, and the products that were identified allow
the distinction between the two types of lignins. Figure 23 shows the pyrograms
of milled wood and the MWL lignin for loblolly pine and white oak. This
technique permits quick and clear separation of guaiacyl and syringyl-guaiacyl
lignins.
Kuroda et al. [4 IO] compared pyrolysis products of sugi wood and the lignins isolated
by different procedures. As an example, the pyrogram of sugi wood is showed
in Fig. 24. The compounds identified are the 4-p-hydroxyphenyl and guaiacyl
types. These authors also studiedthe pyrolysis of milled wood, alcoholbisulfite,
hydrochloric acid, kraft, and klason lignins. In all cases, they observed that the
lignin preparation methods greatly affected the product profiles and the total
yield products. The yields werealsostronglydependent on the number of
condensed units in the lignins.
Analysis of chlorolignin residues was done to obtain information after treatment of
wood chips with sodium chlorite. For this purpose, Pouwels and Boon [411]
identified chlorinated methoxyphenol, guaiacol, and syringol derivatives in the
pyrolyzate of the xylan fraction isolated from beech.
Characterizationofdissolvedorganicsubstanceswas done on anewsprintwhite
water system (thermomechanical pulp, TMP) [412]. The chemical structure of
the dissolved polysaccharides and of the hydrophilic lignin were established.
A comparison between the lignin and sugar composition of pulp, long fibers,
fines, and paper were also done. Previously, Sjostrom and Reunanen[4 131 used
this technique for characterization of water-soluble organic substance isolated
from spruce groundwood pulps.
The syringyl/guaiacyl (S/G) ratios were determined in hardwood kraft pulps by Ta-
naka et al. [414]. The relationship between S/G and cooking time shows similar
behavior to the ratios of syringaldehyde to vanillin (S/V) by nitrobenzene ox-
idation. The authors also analyzed the pulp after bleaching with chlorine.
r
l
h
0;
0
W
t
ti
-
9
-
-I
m
cr
0
i-
L
m
326
puyd 1
lunm UOI
(v
W
a
Chemical Characterizationof Wood 327
C
V
- 0
N
-S
r
U
0
P
Freer328 and Baeza
2. Nondegradative Procedures
a. Ultraviolet Spectroscopy (UV). This is the most simple and classical analytical
method for lignin characterization. The spectrum of a softwood lignin (“guaiacyl lignin”)
shows a maximum at 280 nm (band B), in agreement with the spectrum obtained when
the benzene ring is substituted by hydroxyl or methoxyl groups, shifting the secondary
maximum from 254 nm to 270-280 nm, accompanied by a five- to sixfold increase in
intensity. Due to the presence of the syringyl structure in the hardwood lignins (“guaicyl-
syringyl lignins”), the maximum appears at 274-276 nm (Fig. 25). Two extra bands are
present in typical softwood lignins, a shoulder at 230 nm (band E2) and a sharp peak at
200-210 nm (band E,). The B-band absorptivity values for softwoods are in the range of
18-21 L/g.cm, while for hardwoods the values are much lower (12- 14 L/g.cm). This
band is usually used in qualitative and quantitative UV determination of lignins [416-
4191.
Difference spectra ( A s curves) have been used to help in the interpretation of the
UV spectra. This method has been used to study the different types of chromophores in
softwood and also in the case of more complicated hardwood lignins. However, to obtain
more information about the structure of the lignins it is initially necessary to study the
spectra of model compounds [420,421].
The ionization difference spectrum, Asi, obtained by difference between the absorp-
tivities in alkaline solution (ionized) and in neutral or acid solution (nonionized) (Fig. 25),
represents the absorption of the ionizablephenolicgroups,and the phenolichydroxyl
content can be estimated by comparison with standards. The spectrum obtainedafter treat-
ment with reducing agents such as sodium borohydride resulted in a decrease in the in-
tensity of the band assigned to free phenolic groups conjugated to reducible groups such
as carbonyl, carbon double bonds (300-400 nm) [422,423].
2.1
0.6
0.3
TABLE 8 Band Assignments MWL Lignins of Spruce, Beech, and Bamboo [435]
Maxima"
Range of
maxima"
Bamboo Beech Spruce Peak assignment
"Wavenumbers in cm '
liquors from kraft pulping of Eucalyptus sieberi, obtaining a precise relationship (linear
plot with R' = 0.99) between integrated band intensity (1 118 cm") and kappa number.
Previously, the same author studied the chemical changes in woods during soda and soda
antraquinone processes [439]. Faix et al. [440] evaluated the continuous process control
of pulping by FTIR. The kappa numbers and the yields of pulps obtained in ASAM and
kraft AQ processes were determined.
Freer332 and Baeza
Due to the complexity of wood it is difficult to assign each peak to a single com-
ponent, and interpretation of isolated bands in wood FTIR is misleading [ 174,4411. For
this reason the use of multivariate analysis (MVDA) have been frequently used to correlate
FTIR and near-infrared reflectance spectra data with changes in structure or wood com-
position during chemical or physical processes. Multivariate data analysis is the process
of learning how to combine data from numerous channels to overcome selectivity prob-
lems, gain new insights, and allow automatic detection. It has been shown to be useful
for the evaluation of experimental data, especially for data containing correlated variables.
By usingMVDAtoevaluatespectroscopicdata, it is possible to select useful datato
interpret andmodelhighlycomplex spectra withoverlappingpeaks.Usingthese tech-
niques, several papers have reported rapid methods for estimating the concentrations in
wood or pulp components [442-4451. It is necessary to have a set of samples of known
data, which are used to model the relationship between some quality of the sample and
the absorbance band. This step is called “calibration” and is done by using chemometric
software, which provides methods such as partial least-square regression (PLS), principal-
components analysis (PCA), and principal-componentsregression (PCR)[442,446,447].
After the model is tested by using another set of known data (validation step), it is possible
to predict new, unknown values.
Another lignin behavior that has been studied using FTIR is the photo-induced color
reversion of bleached and unbleached stone groundwood pulp [448]. Changes in the chem-
ical structure were monitored after exposures to UV irradiation. The change in the lignin
content was followed using intensity of the peakat1509 cm”, and the change in the
carbohydrate content followed at 1060 cm”. Also, a study of the yellowing of Eucalyptus
reganns from cold-soda pulp has been done [449]. Michell [450] studied the reaction of
inorganic agents with wood, particularly those that are used to protect the wood surface,
such as chromium trioxide. Due to the reaction with the aromatic rings of the lignins,
changes at 1505 and 1595 cm” were observed.
Near-Infrared Spectrometry (Near-IR). The absorption bands in this region can
be assigned to overtone and combination vibrations, primarily of OH, NH, and CH func-
tionalities. In general, visual interpretations are more difficult because of the nature of the
bands. The spectral bands are often overlapped, and differentiation between similar ma-
terials appears less definitive than in the mid-IR region. In spite of these apparent disad-
vantages, the near-IR region provides useful qualitative information, especially with the
use of computer-assisted data analysis techniques. There are also other advantages, such
as the use of cells of glass or quartz, insensitivity to water, and lower signal-noise ratio
compared with mid-IR. Changes in the bands of near-IR might be used to measure dif-
ferences in wood chemistry such as occur in different tree zones, between woods of dif-
ferent species, between woods grown on different sites, etc.
Near-IR has been utilized to determine the content of lignin and cellulose, and also
toestimatethe fiber orientation, moisturecontent,kappanumber,and the pulp yield.
Recently,Brunneret al. [451]analyzed 90 samples of 12 different species,and it was
possible to distinguish between samples of a given wood species of different origins. For
this purpose, a chemometric software was used. Michell and Schimleck [452] investigated
the origins of the bands in the near-IR spectra of Eucnlypfus glohulus woods. Different
approaches were used, including comparisons of bands in the near-IR with those in the
near-IR of its major components (cellulose, glucuronoxylan, MWL, and hot water extrac-
tives). Figure 27 shows the second-derivative near-IR spectra of E. glohulus wood, and of
some of the different isolated components-i.e., those removed by treatment of the wood
with chlorite and with weak alkali-and also variations in intensity of the major bands
Chemical Characterization
of Wood 333
I
0
0
m
r
0
0
m
r
0
0
t
0
0
z z
0 0
O
m
N
Wavelength ( m )
FIGURE 27 Second-derivative near-IRspectra of E. globulus, cellulose, xylan, lignin, and ex-
tractives. (From Ref. 452.)
in the mid-IR spectra and their connections with near-IR bands and particular chemical
components in wood.Many partial correlationswerefound,showing a highdegree of
intercorrelation between the bands and confirming that the near-IR arose from the com-
bination of several fundamental bands.
Recently [452], the use of diffuse reflectance, near-IR, and multivariate evaluation
have been applied to determine yield, kappa number, lignin, glucose, xylose, and uronic
acid during akraft pulping process. This method has the advantage of being fast. Principal-
components analysis (PCA) has also been used in conjunction with near-IR to discriminate
between woods from pines and eucalyptus, between woods from different eucalyptus spe-
cies, between woods from different provenance, and between woods fromthe same species
of eucalyptus on different sites [453].
Freer334 and Baeza
TABLE 9 Mixture Components, Identifying Adsorptions, and Most Probable Source [454]
source
probableMost
absorptions
Identifying
Components
Kaolin
3695,
3619,
1031,
915
and
938, cm-' Additive
Calcite
1450,
874,
and
710 cm" Additive/environmental
Aragonite
1496
and
855 cm-' Additive/environmental
cm"10 11
CaSO, Processing interaction
Cellulose
3368,
1161,
and
1033
cm" Raw materials
Oxalatesalt1620and1318 cm-' Raw materials (wood)
Defoamer3305,2950,2922,
2866,
2854,
1635, Additive (pulp preparation and/or
making)
paper cm" 1366 and 1462,
Polyethylene 1719,
1339,1257,
1253,
and714
cm"Paper
machine
(wire
material)
Terephthalate
Chemical Characterizationof Wood 335
9.58-9.86 Ar-CH=CH-CH0
CH0 in
7.23-7.90 Ar-H in Ar-COR
6.25-7.23 Ar-H in Ar-R
H-a in Ar-CH=CH-CH0
H-p in Ar-CH=CH-CH0
H-a in Ar-CH=CH-CH,OAc
5.75-6.25 H-a with a-0-Ac in p-0-4 and p-1
H-p in Ar-CH=CH-CH,OAc
5.20-5.75 H-a with a-0-Ac in p-5 (0.09/C9)
H-a with a-0-Ac in p-0-4 and p-l (O.OS/C,)
4.50-5.20 H-p in p-0-4
H-y in Ar-CH=CH-CH,OAc
H-a in p-p
3.95-4.50 H-y in p-0-4, p-5, p-l, and p-p
3.55-3.95 (3.39/C,)Ar-0-CH,,
H-p in 0-5 (O.O9/C,)
H-y in p-p (O.O4/C,)
2.50-3.55 H-p in @ - l , p-p,
others
and
2.20-2.50
H in Ar-OAc except for 5-5 unit
1.50-2.20
H in Aliph-OAc
Ar-OAc
and 5-5in units
1.10-1.50
0.75-1.10
transforms (FT). Some of the advantages are the following: ( I ) it is possible to obtain
information about all the carbon skeleton; (2) there is more resolution over a much wider
chemical shift range and less overlap of signals (the I3C-NMR range is about 200 ppm,
compared with only 12 ppm for 'H-NMR); ( 3 ) spin-spin coupling between carbons almost
does not exist.
The I3C-NMR spectrum of lignin can be divided into three main segments: ( I ) 200-
I65 ppm, carbonyl bonds; (2) 165- 100 ppm, aromatic and olefinic carbons; ( 3 ) 100- I O
ppm, aliphatic carbon atoms. The assignments have been made by using I3C-NMR lignin
model spectra. Table 1 I summarizes the chemical shift data [461] for 'C-NMR of a lignin
from poplar, using DMSO-d, as solvent, and in Fig. 29 appears a typical routine '.?2-NMR
spectrum [462]. Like 'H-NMR and IR, '7C-NMR have been used to distinguish the origin
of lignins. Nimz and co-workers 14631 studied the structural differences using acetylated
lignins.Takingintoaccount that therearethreecharacteristicsignals in thesesamples
(81.1, 75.5, and 63.8 ppm), corresponding to carbons atoms a, p, y in p - 0 - 4 structures,
it is possible to distinguish the aromatic nuclei occurring in lignin guaiacyl (G), syringyl
(S), and p-hydroxylphenylpropane (H).
Several experiments that have employed "C-NMR have been useful in the elucida-
tion of lignin structure, beside simple experiments without irradiation of the protons. One
of these is distortionless enhancement by polarization transfer (DEPT), which simplifies
the spectra. DEPT is a one-dimensional pulse sequence, involving the spin-echo phenom-
enon sequence, which allows separate recording of the NMR signals for the CH,, CH,,
w
w
Q)
C
C
I
C
FIGURE 28 H-NMR of MWL from spruce (400-MHz; solvent DMSO-d,). (From Ref. 459.)
aracterization
Chemical of Wood 337
TABLE l 1 Chemical Shifts and Assignments of "C-NMR Signals for Poplar Lignin Samples
(Solvent DMSO-d,,, T = 323 K ) [46 I ]
Signal 6 (ppmnMS) Assignments
I 171.6 C=O acetyl in xylan
2 168- 167 C=O in benzonic acid
3 163.4 C-4 in H ne
4 152.7- 152.3 C-3/C-5 in S p-0-4 e
5 149.6- 149.2 C-3 in G e
6 147.6-146.9 C-3 in G p-0-4 ne; C-3/C-5 in S p-0-4 ne
7 145.6 C-4 in G p-0-4 ne
8 139.1-138.1 C-l in S p-0-4 and G in p-0-4 e
9 13s- 134 C-4 in S p-0-4 e and ne
10 133- 132 C-l in G p-0-4 ne; C-5/C-5' in 5-5' units
11 131.6-131.3 C-2/C-6 H units
12 129.5 C-a and C-p in vinylic structures
13 I22 C-l in H units
14 119.4-1 19.2 C-6 in G units
IS 115.8-115.6 C-5 in G units, C-3, C-5 in H units
16 111.8-111.1 C-2 in G units
17 107.8-107.3 C-6 in S units with C=O
18 106.8-104.5 C-2/C-6 in S units
19 103.6 C-2/C-6 in S p-p
20 102-101 X-l in xylose units
21 07.6-97.3 X,,: anomerin carbon in xylose units
22 92 X<.: anomerin carbon in xylose units
23-24 87-84 C-p in S and G p-0-4; C-LYin p-p
2s 81 C-dC-p in p-0-4 and a-0-4 units: C-4 in 4-OMeGlu
26 76.9-76.0 X-2 in xylose units linked
27 75.5-75.0 X-4 in xylose units
28 74.7-74.0 X-3 in xylose units
29 72.7 X-2 in xylose units
30 72.2 C-a in p-0-4 units
31 71.8-71.0 C-y in p-p; C-2 in 4-OMeGlu
32 70.1-69.5 C-S in 4-OMeGlu
33 65.7-65.5 C-5 in reducing xylose units
34 63.3 C-y in p-S and p-0-4 with a C=O; X-5 in xylose units
3.5 60.2 C-y in S and G p-0-4
36 56 Aromatic OMe in S and G units
37 53.4-52.7 C-p i n p-p and p-5 units
S = syringyl units; G = gunincyl unlts: H = p-coulnaryl units: c = in etherified struciures: nc = in nonethcrifcd
structures: X , = xylose units: 4-OMcClu = residue of methyl glucuronic acid.
and CH groups. Another clear advantage is a large signal intensity enhance~nent due to
the polarization transfer. This polarization transfer relies on a spin population interchange
fromthehigh-sensitivitynucleus, 'H, to the less sensitivenucleus, "C, to enhancethe
latter. In Fig. 30 is shown a DEPT edited "C-NMR spectrum of birch lignin 14641. This
technique was also used to obtain more information about kraft lignin obtained during the
alkaline delignification of hardwood (poplar) in a flow reactor [461 I andmilled lignin
obtained from softwood ( Pirzus .sylvc..str-is) 14651.
338 Baeza and Freer
-0
' En
. .-c
og
-0
In
-0
-3
2
-3 0
-In
r
-8
N
0
-In
N
Chemical Characterizationof Wood
340 Baeza and Freer
The application of "C-NMR to solid samples gives spectra with weak. broad, and
poorlyresolved signals. Theweak signals havebeenimprovedusing cross-polarization
pulsesequences(CP)and the broadsignalsusingmagic-anglespinning(MAS)[466].
Figure 31 shows "C CP/MAS NMR spectra for different lignins [467].
A "delayed-contact" pulse sequence was used to separate I3CCP/MAS-NMR spectra
into subspectraof kraft pulp components. The method exploits differences in rotating-
frame relaxation time constants for cellulosic and noncellulosic domains within the sample.
Lignin contents were estimatedfor the subspectra, and good agreement was found between
these values and those determined by klason lignin. The author concluded that this tech-
nique is useful in chemical analysis in pulping processes [468].
Different morphological fractions of spruce woods were analyzed usingI3C CP/MAS
NMR spectra [468]. Spectra for the whole wood and five morphologically different frac-
tions (compression, ray cells, cambial, middle-lamella particles, and "fines") were ana-
lyzed. Differences in the relative amounts of cellulose, lignin, hemicelluloses, and protein
were observed. The fractions also present different degrees of cellulose cristallinity; the
"fines" fraction showed the lower value, suggesting that it is derived mostly from the
primary wall.
Two-Dimensional (2D) NMR Methods. The use of 2D-NMR spectroscopy gives
additional information, including connectivity (structural information) and coupling con-
stants (stereochemical information) [470]. A brief review of 2D NMR has been published
by Williams and King [471]. It has been shown that it is possible, by the combined use
of homo- and heteronuclear correlative techniques, such as homonuclear Hartmann-Hahn
C
~ - " - ~ ' ~ " i " " l " " l " - ~ l ' ~ - - i
F1
8 7 6 5 4 3 2
Integral"
Peak F sc DHP MWL Assignment
A 7.5417.34 0.5 -b 0.7 protons
Aromatic
Adjacent to carbonyls
B 7.3416.52 6.6 3.2 5.2 alP:4
C 6.871632 - - 0.4 ?'
D 6.5216.09 9.8 25.3 2.3 dP:3
E 6.1414.62 3.3 5.9 0.5 PI y:3
E' 5.9914.63 3.3d 4.7d 2.6* alp:1
F 5.9215.44 0.5 - 0.3 dP:6
G 5.4513.65 3.3 2.1 1.9 alP:2
H 4.7914.03 2.3 - 1 .o ?
I 4.4313.13 - - 0.5 ?
J 4.20/3.00 1 .o 0.9 - PI y:9
K 3.9811 .81 6.6 2.4 2.6 ylP:5
L 2.521 I .S1 6.6 1.9 2.6 aIulp:5
M 1 S217.34 - - - PI y:4
"Normalized to OCH, = 1000.
hNot observed.
'Unknown.
"Uncertain due to close proximity of E.
'Folded and not integrable.
MWL [477]. Figure 33 shows the HOHAHA spectrum of the side-chain region
of acetylated dioxane-water MWL. The cross-peak in region A, arising from
H,,-H, magnetization transfer in p-0-4 structures, permits the conclusion that
the levels of p-0-4 were approximately equal for each lignin. The correlations
from the other structures were also of similar intensity among the three samples.
However, there are significant variations in region B, which encompasses the
chemical shift ranges of the side-chain proton from p-l structures. The higher
content of p - l units was found in the acid anhydride MWL, being somewhat
less in the acetone-water lignin and significantly higher than that of the di-
oxane-water lignin, suggesting that the p-1 units are not evenly distributed
within the cell wall.
Acetylated residual lignin from unbleached kraft pulp, isolated by cellulase treatment.
wasstudiedusingHOHAHAexperiments in combinationwitha selective
cleavagewithpivaloyl iodide. It wasconcluded that: ( I ) the residual lignin
obtained from beech unbleached kraft pulp still contained p-0-4 and resinol-
typechains; and (2) glycosidicbondsbetween lignin andcarbohydrates are
present in the residual lignin [478].
The use of HOHAHAandHQMCexperimentsshowed that these techniques are
sensitive, rapid, and unambiguous probe for the presence or absence of non-
cyclic benzyl aryl ether (a-0-4) structures in soluble lignin samples. The limits
of detection by 2D NMR techniques is < 0.3 structure of a-0-4 structures per
100 C9 units. Other techniques that have beenused to determine a-0-4 are
the acidolysis and ID NMR techniques, both beingambiguous [479]. In the
59
61
I .
A 63
J _c--
49 4.7 45 49 47 45 49 47 45
45
5.0
J
5.5
-55
*
@ -60
6.5
3
3 P 1 I
B
~-.- . , J
35 34 3.3
3.5 34 33 3.5 3 4 3.3
FIGURE 33 Side-chain region of HOHAHA spectrum of acetylated dioxane/water (Pinus radiafa) and expansion of the regions A and B for acetylated
MWLs: (a) dioxane-water: (b) acetone-water; (c) acetic anhydride. (From Ref. 476.)
344 Baeza and Freer
first case the cleavage of any other ether-linked phenol will give an overesti-
mation of the a - 0 - 4 structures. In the case of I D NMR it is not possible to
do a correct chemical shift assignment.
3'P-NMR Spectroscopy. "H-NMRspectroscopyhasbeen used intensively in lig-
nin chemistry but presents some limitations, such as the limited range of chemical shifts,
poor spectral resolution, and extensive signal overlapping. To overcome these disadvan-
tages, ?IP-NMR spectroscopy has been used. The labile protons of residues of phenols,
alcohols, aldehydes, sugars, and carboxylic acids react with 1,3,2-dioxaphospholanyl chlo-
ride (Fig. 34) [480]. This method presents several advantages: ( 1 ) rapid and quantitative
reaction; (2) a sharp, single "P signal is obtained; (3) phenols, alcohols, and simple car-
boxylic acids give "P signals in separate regions; (4) chemical shift is sensitive to the
chemical environment of the phosphitylated center; (5) it is possible to obtain details about
the stereochemistry i n both lignin modelcompoundsand isolated lignins [481].The
syringyl, guaiacyl, p-hydroxyphenyl-free phenolic groups, primary groups, and secondary
groups can be quantitatively determined by this technique 14821.
The analysis of a set of lignins by this method has been done, andthe results obtained
for the total hydroxyl contents compared with those obtained by using 'H-NMR and wet
methods. The results appear in Table 13, and a typical 31P-NMR is shown in Fig. 35 14831.
Structural changes during the pulping process have been studied by using this tech-
nique. The ortho-quinone contentin mechanical pulp can be evaluated [484].The hydroxyl
content during the kraft process was monitored using "P-NMR and the results were cor-
related with "C-NMR. Also it was possible to establish the stereospecificity of the deg-
radation reaction 14851.In a similar study, dissolved lignin isolated from the three-stage
formic/peroxyformic acid (Milox) pulping process was studied using 31P-NMRand oxi-
dative degradation 14861. Quantitative"P-NMRallowed the determinationof guaiacyl,
syringyl, total condensed free phenolic OH groups, carboxylic acids, and the erythro and
threo forms of the a-OH present in the p-0-4 units. In addition, oxidative degradation
provided information about the fate of specific condensed structures such as biphenyls,
biphenyl ethers, and C5- and C6-substituted guaiacyl units. It was possible to obtain in-
formation on the rate and the topochemistry of the scission reactions of arylglycerol-p-
aryl ethers and the condensation reactions that occur in the lignin. The cleavage of the p-
0 - 4 aryl ether linkages within the lignin was approximated 45% during the first stage and
considerably intensified in the second stage, with a concomitant increase in the total phe-
nolic hydroxyl content. A significant amount ( 1 6%) of fhreo-arylglycerol-~-aryl ethers
seems to remain after the third stage. The guaiacyl phenolic units predominated within
the solubilized lignin after the first stage. This, together with the fact that guaiacyl units
are less reactive than syringyl units towardperoxyacids, indicates that a topochemical
effect is operating during this stage, being a greater accessibility toward the guaicayl-rich
middle lamella lignin. The second stage seems to preferentially cleave the syringyl-rich
ROH + CCp/
\O
'1 PyndinelCDCI,
25 "C
w R-0"P '1 + HCI
\O
(1)
FIGURE 34 Derivatization reaction. (From Ref. 480.) Where R = residues of phenols, alcohols,
aldehydes, sugars, carboxylic acids.
3n,
3.
c)
EL
3
e
m
TABLE 13 Functional Group Distributions Derived by Quantitative "P-NMR Analysis for Lignins" (Figures in Parentheses are Averages Obtained 2
During the International Round Robin Analytical Effort) [483] -.N4
Lignin functional group (mol/C,) z
Alpha-OH in p-0-4 structures
2
Total Total 9
Guaiacyl Syringyl phenolic Primary Total hydroxyl 6
Lignin sample -COOH -OH -OH -OH -OH Eritrho Threo P-0-4-OH content
Steam explosion (aspen) 0.04 0.14 0.27 0.4Zh 0.44 0.24 0.13 0.37 1.27
(0.45) ( 1.26)
Steam explosion 0.00 0.15 0.24 0.4gh 0.36 0.12 0.10 0.22 1.06
(yellow poplar) (0.59) ( I .20)
Ball-milled enzyme 0.00 0.04 0.05 0.15' 0.52 0.39 0.14 0.53 1.20
(cottonwood) (0.18) ( 1.47)
Alcell'" organosolv 0.06 0.25 0.45 0.48 0.32 0.06 0.10 0.16 I .24
(mixed hardwoods) (0.59) (1.20)
Indulin"' AT kraft 0.06 0.2 1 0.00 0.57* 0.4 1 0.06 0.08 0. I4 1.18
(mixed softwoods) (0.67) ( I .23)
w
P
ul
346 Baeza and Freer
PPm
FIGURE 35 Quantitative"P-NMRspectrum of steam explosion lignin produced from aspen.
phosphitylated. (From Ref. 483.)
lignin structures of the secondary wall, but the high syringyl content of stage I1 lignin
may be the reason the guaiacyl units participate more extensively in condensation reac-
tions. Condensationreactions were found to bepredominant in thesecondformic acid
digestion stage, due to the reactivity of the benzylic carbon and the formation of inter-
and intramolecular C,-C, and C,-Cc, carbon-carbon bonds. The results obtained by "P-
NMR have an excellentquantitativecorrelation with thosefromoxidativedegradation
using potassium permanganate and hydrogen peroxide.
JiangandArgyropoulos [487] used "P-NMR to quantified parcr-hydroxyphenyls,
catechols, guaiacyls, and phenols bearing C, and C, substituents, after a Mannich reaction
of softwood kraft lignins and models with piperidine.
19
FNMR Spectroscopy. Barrelle [488-4901 developeda method using "F-NMR
for the quantitative determination of OH groups (hydroxylic, phenolic, and carboxylic).
The fluorobenzoate lignin and the lignin model compound derivatives are obtained using
2- or 4-fluorobenzyl chlorides or the respective anhydrides. This technique permits one
to distinguish between a guaiacyl lignin and a guaiacyl-syringil lignin, to determine the
syryngyVguaiacyl ratio of structural units with phenolic groups, and to approximately
determine the (a-C=O) content.
In Table 14 appear the chemical shifts for lignin model compounds and in Fig. 36
the spectrum of a fluorobenzylatedorganosolvlignin.Barrelleassignedthesepeaks to
ketocompounds,G-compounds, and S-compounds,aftercomparison with the spectrum
obtained from a mixture of eight model compounds 14901.
cl. Molecular Wkight m d Molecular Weight Distribution. Lignin is a polymer with
a wide range of molecular weights and when it is removed from wood, the original value
of the molecular weight is affected. There are a multiplicity of isolation procedures, giving
lignins with different characteristics. MWLs are considered as preparations i n which min-
imal changes have occurred, but these depend on the material and the milling procedure
149 1.4931.
Various techniques are available for the determination of MW and MWD of lignin
samples, as described before for cellulose and hemicelluloses: SEC, viscometry. osmom-
etry, light scnttcring, and ultracelltrifllgation. SEC has been recently and greatly cnriched
TABLE 14 Assignment of Signals in the '"F-NMR Spectra of Model Compounds [490] 3m
z
D
AS A6 rr
a-(C=O) compounds 6" G-S(C=O) a-(C-OR) compounds S a-(C=O)-a-(C-OR)
G
I FBzl-0-G-CHO FBzI-O-G-CH~OH 24.77 0.57 2!
25341 0.27 FBZI-O-G--CH,OG 24.92 0.42 s
2 FBzl-0-S-CHO 25 07 FBzI-O-G-CH~OS 24.93 0.41 s
FBzI-O-S-CH,OH 24.61 0.46 d
}
0
3 FBzI-O-G-CO-CH, 25.25 FBzI-O-G-CHOH-CH1 24.60 0.65 P
0.23
4 FBLI-O-S-CO-CH~ 25.02 FBzl-0-S-CHOH-CH, 24.58 0.44
5 FBLI-O-G-CO-CH~OG 25.28 FBzl-0-G-CHOH-CHZOG 24.89 0.39
i
7 FBz~-O-G-CO-CHOG 25.40 FBz~-O-G-CHOH-CHOG 24.87" 0.53
I I
CH,OH CH,OH 24.90 0.50
8 FBzl-0-S-CO-CHOG 0.20 FBzl-0-S-CHOH-CHOG 24.65" 0.55
I 25.20 I
CH,OH CHzOH 24.62 0.58
G = &OM" s:''fle
348 Baeza and Freer
by the advent of the real-time differential viscometer (DV) and LALLS photometer. How-
ever, due to the characteristics of lignins in solution, mainly their low viscosity, the be-
havior during the determination of molecular weights is quite different from that of pol-
ysaccharides or synthetic polymers 14931.
A general review with 66 referenceson the use of SEC in thedetermination of
MWD in lignin derivatives was given by Himmel et al. [4941. MWD of lignin preparations
by SEC with viscometric detectors and ultracentrifuge sedimentation equilibrium analysis
is illustrated.
A reliable method for estimating MW and MWD requires a suitable solvent, in which
the aggregation (solute-solute interactions), solute-solvent, and solute-packing material
interactions are minimized, in addition to solubilizing lignins over a wide range of mo-
racterization
Chemical of Wood 349
lecular weights. Many lignins are not soluble in suitable organic solvents or water, but
lignin derivatives are soluble in solvents such as THF and DMF. By derivatization, the
adsorptionand association caused by hydroxyl groups are diminished. Acetylated
[492,495-5011, propylated [502], and sylanated [495] derivatives have been used.
SEC permeation chromatograms of MWL have shown multimodal [491,503,504] or
single symmetric distributions [354,492], depending on the experimental conditions. AS a
rule, SEC/THF of acetylated lignins does not show exceptional irregularities. In DMF and
DMF-THF mixes, multimodal elution behavior was observed, which is indicative of as-
sociative phenomena. These can be eliminated by the addition of lithium chloride to the
DMF [503,504]. The effects of the associative/dissociative processes in nonaqueous andl
or aqueous mediaon the MW and MWD of kraft lignins [505-5071 and organosolv lignins
[504] have been discussed.
The early publications dealing with MW and MWD of lignins using GPC employed
soft gel dextran columns [508,509]. More recently, cross-linked gels were employed. Sev-
eral studies have been conducted on GPC on dextran gels columns using DMF, DMSO,
or dioxane-water mixtures as solvents [508-5151 and on agarose gels (Sepharose CL)
[354,5 131. Polystyrene-divinylbenzene copolymer gels columns (e.g.. p-Styragel, Waters
Associates; p-Spherogel, Beckman Instruments; PLgel columns, Polymer Labs) have been
used to conductHPSEC of lignins allowingconvenientrecording of MWD[354,
496,498,501,5171.
Column calibration has been carried out by using dehydrogenation polymers (DHP)
of coniferyl alcohol [509]. Himmel et al. [499] have demonstrated that a series of com-
mercially available standards as well as low-molecular-weight lignins all fit universal cal-
ibration. FaixandBeinhoff[500]discussed SEC column calibration withpolystyrene
standards, lignin fractions,and lignin-like modelcompounds.TheMWdeterminations
calibrated by lignins can deviate from those obtainedby polystyrene calibration depending
on the polydispersity and MW. This can be explained mainly by the fact that lignin mol-
ecules present a spherical shape, being more densely packed in solution than the flexible
chains of polystyrene. Also, as can be expected, due to the chemical structure, a higher
adsorption affinity of lignin to the gel contributes to longer retention times at the Same
nominal MW as the polystyrene. By derivatization, the adsorption and association caused
by hydroxyl groups are diminished. In Table 15, average molecular weights of different
lignins for the underivatized and derivatized samples are shown [495].
compounds
calibration
PS calibration
Model
Sample“ M,,M,, M,, /M,, M,, M,. M,. IM.j
90 120
Retention timdmin
0.8-
G3000SWcolumns,andcombinedHPSECandpyrolysis-gaschromatographic-mass
spectrometric study of lignosulfonates in pulp mill effluents was carried out [523]. The
MWDs show distinct differences between the various lignosulfonate samples and can be
used to characterize structural modifications. Sodium lignosulphonate standard, prepared
under mild conditions,is a relatively polydispersepolymer with a large proportion of
preserved phenylpropaneunits. Lignosulphonates discharged by pulpmills are more mono-
disperse macromolecules, showing lower PD values than the standard, and they also pres-
ent modifications to a greater extent.
Macromolecular characteristics of alkali lignins from western hemlock wood were
reported by Dolk et al. [524]. Wood thin platelets of uniform thickness (0.4 mm), extracted
exhaustively with amixture of ethanol-benzene (1:2 v/v), were delignified with 1.0 N
aqueous sodium hydroxide solution, and the dissolved lignins were fractionated into acid-
insoluble (AIL) and acid-soluble lignin (ASL) fractions. For the AIL, M,, and M,? were
determined by VPO and LALLS, respectively, and for ASL, the two MWs were estimated
by SEC. For AIL, the M,, values are relatively small, but they show a steady increase as
the delignification process advances; values of 874 and 1804 were obtained after 10 and
420 min reaction time. With M,c values a more striking increase occurs (2,318 and 20,685
after 10 and 420 min reaction time), giving rise to a significant increase of the polydis-
persity of the lignins as the delignification proceeds. The values of M,, and M,, for the
total ASL material using the SEC pattern with column calibration were around 400 and
630, respectively. The SEC pattern for ASL shows a number of peaks, andthe values were
assigned to the modes of these peaks (230, 440, 700, and 1020) suggest the presence of
monomeric, dimeric, trimeric, and some larger fragments of lignin.
The condensation of a lignin has been determined by analyzing the molecular weight
distribution of the lignin thioacidolysis products [525].The MWD can be used to give a
measure of reactive to unreactive or condensed bonds in the lignin. A higher proportion
of unreactive linkages in the lignin conduces to higher-molecular-weight fractions in the
thioacidolysis products. The MWD was analyzed by HPSEC on polystyrene columns with
THF as eluent. There was no relationship between MW of the thioacidolysis products and
the MW of the starting lignin.
The LALLS methodology and the correction procedure for optical effects (fluores-
cence, absorption, and optical anisotropy) of lignin solution for the determination of MW
of kraft lignin havebeendescribed by DongandFricke[526].Based on the absolute
molecular
weight
characterized
with
LALLS, the Kuhn-Mark-Howuwink-Sakutara
(KMHS) equation was developed, providing the KMHS constants for the kraft lignin. In
Table 16, KMHS parameters of kraft lignins in DMF and 0.5 N NaOH are shown [527].
Based on these results, it was considered that the lignin molecules in solution are approx-
imately spherical and only slightly solvated.
VII. EXTRACTIVES
The extraneous components of wood are substances which are not considered as essential
structural parts of the cell wall or middle lamella. Unlike cellulose, hemicelluloses, and
lignin, the extraneous components are nonpolymeric (except pectins and condensed tan-
nins) andmay be separated from the insoluble cell wall materials by their solubility in
water or organic solvents. They cover a wide range of chemical compounds even though
they generally represent only a small part of the wood. Because most of the extraneous
compounds are commonly isolated from wood by solvent extraction, they are called ex-
tractives. Strictly speaking, the two terms are not synonymous. However, in most cases
the distinction between extractives and extraneous materials is academic.
Knowledge of the composition and amount of extractives in wood is of great interest.
Many differences in the properties of woods are determined by the composition of the
extractives. Many woods contain extractives which aretoxic to fungi, bacteria, and termites
[528,529]. Other extractives can add color and odor to wood, accent the grain pattern, and
enhance strength properties [530]. On the other hand, extractives can cause some negative
or undesirable properties. For example, the presence of extractives results in corrosion of
metals in contactwithwood[531], inhibition of setting of concrete, glue, and finishes
[532], contribution to color reversion in pulps, pitch problems during papermaking [533],
etc.Furthermore,theextractiveshave industrial importance.Forexample, tall oil and
turpentine have been used traditionally in cosmetics, paints, and varnishes, and have been
proposed as a source for diesel fuel, and energy for steam and electric power generation
[5341.
The composition and the amount of the extractives are dependent on the wood spe-
cies, within and among trees, tree age, and the environmental conditions under which they
grew. Details of the composition of the extractives are given in the literature [531,535-
5381. The extractives are sometimes characterized into chemical classes that have a direct
influence on the pulping process, namely, saponifiables andunsaponifiables [539-5421.
Saponifiables (fatty acids, resin acids, some steryl esters, and glycerides) are considered
to becompounds that formsolublesoapsunder alkaline conditions.Unsaponifiables
(waxes, some steryl esters, diterpene alcohol and aldehydes, sterols, triterpene alcohols,
and fatty alcohols) do not form soaps and have a tendency to deposit and cause pitch
problems [543-5451.
The techniques of analysis of extractives involve the isolation of components (ex-
traction, distillation of volatiles, chemicalorchromatographicseparation)and analysis
(GC, LC, G C M S , NMR, IR, etc.). A large variety of analytical techniques are used in the
analysis of extractives, but the methods vary depending, among other things, on the type
of information required from the analysis.
No single solvent is capable of removing all the substances considered as extractives,
and no single sequence is applicable to all woods. Different schemes of separation and
sequences can be found in the literature [546,547]. For example, according to the scheme
of Kurth [548,549], successive steam distillation and extractions with ether, ethanol, and
water remove different types of extractives. The general scheme outlined by Kurth is of
great value as a general guide. Figure 38 gives an overview of the groups of extractives
from an analytical standpoint,withexamples of subgroupsandindividualcompounds
[SO].
Many other extraction sequences have been employed [g], the selection depending
on the safety, reproducibility, and the desired extractives of the material being examined.
In the isolation of extractives, normally halogenated compounds (mutagenic compounds)
Baeza and Freer
I
racterization
Chemical of Wood 355
and aromatic compounds, especially benzene which is carcinogenic, are used as solvents.
Laboratories are advised to avoid the use of these solvents. Acetone has been found to be
asuitablesolventfor the analysis of extractives in wood, pulp, andpapersamples
[5,542,55 1-5541.
To reduce the hazards associated with the use of large amounts of potentially harmful
organic solvents, together with costs and environmental dangers of wastedisposaland
emission of the solvent into the environment during sample concentration, some alternative
methods of extraction havebeen applied. Areview of some of the modem analytical
methods that can be used in the analysis of extractives from wood and pulp is available
[555].Two broad aspects are discussed, sample isolation procedures and analytical pro-
cedures. Furthermore, to reduce or eliminate the use of toxic organic solvents, the new
proceduresdiscussed are simpler than the traditional methods,whichincludemultistep
procedures. Among the techniques that have been used for extraction of wood and wood
products are Soxtec extraction, gas-phase extraction, which includes headspace sampling
and supercritical fluid extraction (SFE), and sorbent extraction, being solid-phase extrac-
tion (SPE), the most commonly used sorbent extraction technique [556].
The Soxtec method is based on Soxhlet and Goldfisch extractions.It consists of three
steps: (1) boiling-initial extraction, in which the sample is completely immersed in the
boiling solvent; (2) rinsing-condensed solvent washes last traces of soluble matter from
the sample; and (3) solvent recovery-solvent is evaporated, condensed, and collected. A
scheme of the operation and features of Soxtec extraction is shown in Fig. 39. By using
Soxtec extraction, the extraction canbeperformedfasterand the solventvolumes are
about 3.5 times less than the traditional solvent extractions using Soxhlet extractors[555-
5571. However, the Soxtec values tend to be lower than those of Soxhlet [558,559], due
probably to an inefficient washing during the rinsing stage. This problem can be overcome
by performing a second extraction of the sample with fresh solvent [558].
Solid-phase extraction (SPE) methods also require only small quantities of organic
solventand are rapid[553,560,561]. The sorbent is packed in disposablecolumnsor
cartridges. SPE is especially suited for sample preparation of diverse compounds such as
extractives in deposit and wood resin in pulp.
Supercritical fluid extraction (SFE) [563] has recently been used in the separation of
extractives in wood and woodproducts[552,562].Because supercritical fluids possess
bothgaslikemass transfer and liquidlike solvatingcharacteristics, SFE is an attractive
solvent-free sample preparation technique. It is rapid and simple, but it requires heavy
equipment for on-site field analysis. The SFE extraction method has Some distinct advan-
tages over others: thermally unstable compounds are undamaged, extraction times can be
short, and nontoxic solvents can be used [555].
A. VolatileMaterials
The methods for determination of volatile materials (collectively called essential oils or
volatile oils) are based on steam distillation. These components include cyclic hydrocar-
bons(terpenesand terpenoids), aliphatic hydrocarbons,phenols,alcohols, ethers, alde-
hydes, and lactones [564]. In general, the amount of essential oils in hardwoods is neg-
ligible, but they are present in considerable amounts in softwoods. The essential oils of
pines are called turpentine, which consists primarily of monoterpenes.
The total amount of volatile oils can be found by loss of weight of the wood sample
after steam distillation [565]. Thismethod is not suitable when the amount of volatile
356 Baeza and Freer
aracterization
Chemical of Wood 357
materials is small. The volatile oils can also be determined after recovering by conden-
sation techniques or the adsorption of vapor on charcoal or other suitable materials
The water-insoluble volatile oils can be determined by using distillation equipment
with a suitably calibrated trap. The wood is distilled until no more oil comes over, and
the volume is read directly.
The amount of volatile oils can also be determined by extraction of a team distillate
with ethyl ether. Fresh wood meal is steam-distilled, normally for 2 h, although the time
can be extended 3-4 h to improve the recovery of high-boiling-point compounds. The
distillate is extracted with ether. The organic solution is dried and the solvent is removed
to yield the volatile oils. Details of the procedure are given elsewhere [567,568].
Distillation may be conducted with a caustic solution in order to prevent degradation
and isomerization [569,570]. The basic solution may also contain ethylene glycol [569].
To minimize bumping during distillation, ground wood is placed in a cheesecloth bag that
is placed on a supporting metal screen in a resin kettle [571].
The yield of volatile oilshasbeendetermined by gaschromatographyusing an
internal standard 15721. A known amount of tetradecane as an internal standard was added
to the ground wood sample prior to distillation. Yield was determined by GC recording
the proportion of tetradecane to terpene components in a portion of the distillate. With
this method the turpentine yields are about 5 % greater than those obtained by the volu-
metric procedure.
The characterization of the volatile oils is commonly determined by GC and GC-
MS [570,573,574].
Extraction of volatile compoundswith supercritical carbondioxideandhot-water
distillation was conducted for coniferous woods, and the extracts were analyzed by GC
and GC-MS 15621. The yields by SFE for 30 min were greater than those by hot-water
distillation for 8 h. For example, by SFE of western red cedar and Douglas fir, the yields
were0.61% and 0.95%(0.d.b.)(at 300 kgf/cm' and 40°C), respectively, while by hot-
water distillation the yields were 0.12% and 0.07% (o.d.b.), i.e., 5.6 and 13.6 times more,
respectively. The yields by SFE are dependent on the conditions and time of extraction.
For example, the yields of extractives for 30 min were about 80% of those for 90 min.
The effect of pressure on the yield is given in Fig. 40. The components of SFE extracts
and essential oils by hot-water extraction were similar for some species, but they were
quite different for western red cedar and Douglas fir woods.
Headspace volatiles emitted from extracts of SFE using carbon dioxide and essential
oils by hot-water distillation from seven species of woods were collected and analyzed by
GC. They were compared with those of woods [575]. The composition of the three head-
space(volatilesfromwood,SFEextracts,and essential oils)weredifferentfromeach
other, and in general, a-pinene and/or P-pinene were the main compounds of each head-
space volatiles.
Another method for monitoring volatiles involved FT-IR measurements. Emission of
terpenesfromchip piles wasdetermined[576].Themethodusesapolymer film wind
tunnel and measures a-pinene, P-pinene, and y-3-carene on a semicontinuous basis with
a detection limit of 1 mg/m3. The total hydrocarbonswasmonitored simultaneoLlsly by
GC-FID.
4.0 A
\
3.0 \
n
S
W
in
9
Q) 2.0
F
Hinoki
1.o sugl
Hinokiasunaro
Alaska cedar
W.redcedar
I I I
300 200 100
Pressure (kgf/cm2)
FIGURE 40 Effect of extraction pressures on yields (40°C.30 min). (From Ref. 562.)
diterpenoids and occur naturally in conifers. The major acid is often dehydroabietic acid
(DHAA), although which acid is predominant depends on tissue type, age, and species.
Several of the resin acids have two conjugated double bonds, and it is these which appear
to be least resistant to chemical degradation. In softwood species such as Pinus radiara,
free resin acids may comprise up to halfof the extractable organic compounds present
[577], although the amount present varies with the age of the trees and the conditions
under which they grow [578].
Anumber of standardprocedureshavebeenpublishedwhich are very similar in
principle and procedure [579]. These include the TAPPI Standard, Method T-204 om-SS
[2] and ASTM D 1107, D 1108, and D 1794 [3].
The general procedure is as follows. A sample of air-dried wood sawdust is weighed
in an extraction thimble and placed in a Soxhlet extraction apparatus. The extraction is
carried out with a suitable solvent for 4-8 h, having solvent siphons over at least 6 times
per hour. The flask is removed from the apparatus and the solvent is partially evaporated
to reduce the volume to 20-25 mL. The extract is transferred to a tared weighting disk
by washing it with small amounts of fresh solvent. The solvent is evaporated to near-
dryness. The dish and contents are dried in an oven for 1 h at 105°C (1 15°C for ethanol-
toluene), cooled in a desiccator, and weighed. The sample can be also dried to a constant
weight in a vacuum oven at 60°C.
aracterization
Chemical of Wood 359
I ETHANOL- BENZENE
EXTRACTIVES l
Carefully concentrate to dryness
Extract with ether
a Neutrals
I Add water-acldified
Extract with ether
-
I I 1
I
+
Glycerol
Unsaponifiables
Extract withdil.Aq. NaOH
Selective esterification
preparation; (2) solvent extraction with acetone to determine the amount of extractives
(freeze-dried samples were extracted with acetone in a Soxtec extraction apparatus, then
the remaining soluble extract in the sample and thimble walls were rinsed with solvent,
and the extractdriedusinganitrogenstream);(3) fractionation of the extractivesinto
weak acids (pK,, > 5), strong acids (pK,, < 5 ) , and neutrals by ion-exchange chromatography
using a DEAE-Sephadex ion-exchange procedure developed by Zinkel [560,561]; (4) de-
rivatization and GC analysis using short columns to separate the components in the ex-
tractives (methylated); and (5) identification and quantitation of the extractives components
by using a spreadsheet program. Figure 42 showsa chromatogram for a fresh wood sample.
The peaks were assigned by using individual standards ranging from fatty acids to tri-
glycerides. The use of a short column permits the elution of high-boiling-point fractions
and fatty acid-methylatedesters in a reasonabletime(30min)andwithmuch better
resolution than is obtained with packed columns.
Acetone extracts of white spruce and trembling aspen woods and a series of kraft
pulps, were analyzed by GC [544]. Chemical changes in wood resin during pulping per-
formed over a range of initial effective alkali (EA) from 11 .O to 44.9 g/L, and the impli-
cations of the results in deresination and pitch control were discussed. The ratio of resin
acids and wood fats (fatty acids and glycerides) to resin (sterol and steryl esters) is of
great importance in the deresination of the pulps. For a complete dissolution the ratios
werefound to be 2.6: 1 and 1.4: 1 for spruce and aspen pulps, respectively. Below this
value, deresination is not complete and some insoluble components are not washed out of
the pulp. Consequently, it can be predicted that poorer deresination will result in the aspen
sample than in the spruce. More frequent occurrence of pitch problemsfrom the kraft
pulping of aspen due to relatively higher amounts of neutral and unsaponifiable materials
can be generally observed. With EA values in the range 25.0-40.0 g/L for spruce and
30.0-44.9 g/L for aspen, over 70% of the total acetone-extractable was removed during
pulping, while at EA values below25.0 g/L and 30.0 g/L for spruce and aspen,respectively,
the percentage of extractsincreased rapidly. This increasemight be due to incomplete
removal of the lignin degradation products during cooking, which are later extracted with
the other acetone-soluble compounds. Under the normal cooking conditions (EA concen-
tration of -40 g& forspruceand -30 g/L for aspen), the glycerides are practically
completely saponified into soaps. The concentrations of fatty acids, resin acids, and steryl
esters in spruce, and fatty acids and steryl ester in aspen, are reduced compared to the
original values in the woods. The steryl esters content is decreased considerably by their
hydrolysis to sterols. At lower initial EA concentration, the total amount of extractives
remaining in the pulps increases rapidly. The glycerides and steryl ester increase in both
species. The content of fatty and resin acids in spruce, and fatty acids in aspen, increases
towards its original value in the wood.
An analysis of the acetone extractives of fresh trembling aspen (Populus tremuloides
Michx.) wood was reported by Dunlop-Jones et al. [542]. Freeze-dried samples were ex-
tracted for 18 h with acetone. The dried acetone was dissolved in a diethyl ether-meth-
anol-water mixture (89: 10:1 ) and fractionated into weak acids, strong acids, and neutrals,
using a DEAE-Sephadex column. The neutral fraction was saponified and fractionated into
saponifiables (acids) and neutral (unsaponifiable) using an ion-exchange column. The free
acids and combined saponified fatty acids were analyzed as to their methyl esters by GC.
Theunsaponifiableswere silylated with a mixture of N,O-bis(trimethylsily1)-trifluoro-
acetamide(BSTFA)andtrimethylchlorosilane (TMCS), andthenanalyzed by GC. The
neutral componentswere identified by comparisonwithknownpurecompounds or by
mass spectrometry. The ratio of saponifiables to unsaponifiables found for the fresh aspen
05
o.201
fD
3.
c)
m
c
U
v) 0.151
r
0 0
> 0
P
W'
cn
z
0
a
cn
W
E
I /-
I 8
I I
I I
I I
I I
I r 1 8
0 5 10 15 20 25 30
was about 2: 1, and this low value is a reason for its tendency to give pitch problems.
Other workers [580,581] have obtained higher ratios, but the amount of seasoning of the
woods analyzed has not been discussed. Seasoned wood chips exhibit fewer pitch prob-
lems, since during the storage, extractives undergo volatilization, enzymatic hydrolysis,
and air oxidation, but long storage times promote microbial deterioration [582,583]. The
storage effect is greater in chips than in roundwood.
SPE was used to separate and quantify lipid classes in acetone extracts of wood and
pulp [553]. A method for rapidly separating acetone wood or pulp extractives into five
different classes was described. The acetone extracts were absorbed onto an aminopropyl-
phase column and the recovery and quantification of different classes were carried out by
eluting with a sequence of solvents. The samples were further analyzedby GC and HPLC.
The method was applicable to both hardwoods and softwoods. Separation of the different
lipid classes was highly reproducible. The five classes are (1) fatty acids (FA) and/or resin
acids (RA), (2) steryl esters (SE)/waxes (W), (3) triglycerides (TG), (4) monoglycerides
(MG), and (5) sterols (S)/diglycerides (DG)/fatty alcohols (Fal) (Fig. 43). The separation
takes about 2 h per sample, with a recovery rate of 95-99%. This method is presented as
an alternative to the traditional one used in the wood and paper industries, in which acidic
compounds are separated from neutral by ion exchange. Neutral lipids are hydrolyzed and
finally analyzed by GC [565,566].
Analysis of oriental beech (Fugus orientalis) wood fatty acids by supercritical ace-
tone extraction (30 min, 240°C, 6.0-6.5 Kpa) was carried out and the results were com-
pared with Soxhlet extraction [552]. While the yield of the Soxhlet extract was 2.5470,
the yield of the SFE extract was 9.55% (dry wood basis). The fatty acids present in the
extracts were separated by chemical and chromatographic methods and analyzed by GC-
MS. Among the fatty acids, from both Soxhlet and SFE extracts, linoleic acid was the
major constituent, followed by linolenic and palmitic acids. Palmitic acid appears to be
the main saturated fatty acid. The proportion of the dienoic fatty acid is lower in the SFE
extracts due to the high temperature used in the extraction.
The main resin and fatty acids (RAFA) of Pinus elliottii Engelm. were characterized
before and after pulping (cooking liquor, methano1:water 80:20) [584]. The RAFA were
saponifiedand/ormethylatedandcharacterized by GC andGC-MS.Determination of
RAFA isolated from wood, pulping liquor, and pulp were carried out without fractionation
of the extracts before the GCanalysis.The total resin acidcontentdoes not undergo
quantitative changes after pulping, but instead qualitative modifications occur as a result
of eitherisomerizationand/oroxidation reactions, leadingtoformation of abietic and
dehydroabietic acids as the preferred end products. Selective removal of the extractives
from the black liquor was obtained by using a liquid-liquid extraction with diethyl ether
without affecting the solvent composition of the black liquor.
A rapid spectrophotometric procedure for the determination of total resin and fatty
acids (RAFA) in pulp and paper matrices was developed [585], which can be applied to
wood chips, whitewaters and effluents. The method involves the following steps: ( 1 ) ex-
traction of the fatty and resin acids from the matrix, (2) complex the free fatty and resin
acids with copper(I1) ions to yield blue complexes, (3) extraction of the complexes with
a solvent, and (4) measuring of the absorbance at 680 nm. The calibration curves were
obtained using oleic acid as the standard. Compounds containing resin and fatty acids,
such as glycerides, and metals salts, do not react with cupric ion, but the bound RAFAs
can be determined by the difference between the total RFA value obtained before and after
the hydrolysis of the extractives. This method is simple and can be used as a rapid pro-
cedure to estimate the amount of RAFAs in extractives, yielding results which agree with
Chemical Characterizationof Wood 363
Extract
(TG, SE, FA, RA
DG, S , MG, W, Fal)
4 CHC:,:? + Elh; j
Column 1
R
A- R
A-
FA
DG
MG
S
Fa1
H W
SE
TG
tj, H Fa1
RA
FA
(Save)
Fraction A
Dry down
Hexane
+
Fraction B
Dry down
Ethyl acetate
Fraction C
Dry down
Ethyl acetate
Fraction D
(Save)
PE & 6!: 0
Fraction E
(Save)
TG
Fraction F
(Save)
Fraction G
(Save)
MG
those obtained by chromatographic procedures. This method has a higher limit of detection
and requires much larger volumesofsamplesthanthose used in chromatographic
procedures.
Wood extractives released during pulping or as by-products of the pulping process
have been reviewed by SitholC [lo].
There is abundant literature on the analysis of by-products of the kraft pulping pro-
cess. Proton NMR has been used to determine DHAA and total resin acids in rosin [586].
Asimple, rapid, andaccuratemethod for the quantitativedetermination of DHAA in
commercially disproportionated rosin acids was developed and tested [587]. The method
entails converting the acid into its methyl ester derivative before analysis by capillary GC
with a FID detector, using methyl stereate as the internal standard. The method can also
be applied to the quantification of DHAA in rosin or other rosin derivatives.
The analysis of pulp mill process water, sediments, and fishes is very important from
an environmental point of view. Investigations have been centered fundamentally on the
study of organochloride compounds (which originate in the bleaching process with ele-
mentalchlorineorchlorinedioxide),butsomemethodshavebeendeveloped for the
determination of wood extractives. Resin acids are toxins to fish at very low concentrations
( 1 -2 mg/L) [588,589].
An analytical procedure for the rapid determination of fatty acids, resin acids, and
triterpenoic components in pulp mill process water was published by Backa et al. [590].
The extractives were isolated from alkaline aqueous samples using reversed-phase meth-
ods. The highest yields of extraction were achieved when an octadecyl phase, C,,, was
used (C,-C,, reversed-phase chain lengths were tested). The cations of the adsorbed acid
salts are exchanged in situ for quaternary ammonium ions. The acids are methylated by
thermal decomposition of the quaternary salts in the injector of the gas chromatograph.
The analysis scheme for determining extractives is: sample application + reversed-phase
adsorption + rinsing with alkali -+ ion pairing -+ elution + pyrolytic methylation and
gas chromatographic separation. A chromatogram of extractives isolated from a kraft black
liquor is shown in Fig. 44.
Resin acids in effluents, river waters, and sediments from a paper mill from Australia
were determined by HPLC and GC (FID and MS) [591]. The resin acids in effluent were
extracted by passage through a C,, cartridge at pH 9 and determined by HPLC and GC
as their 7-methoxycoumarin-4-yl and 7-acetoxycoumarin-4-yl esters. The sediments were
extracted with acetone, the extracts were dried, dissolved in water at pH 11 (KOH), and
loaded onto a C,, column prepared as for the water extraction. The resin acid, eluted with
acetone, was derivatized with diazomethane and analyzed by GC. The results confirmed
the presence of resin acids derivedfrom the paper mills. The major resin acids in the
effluent, water, and sediment samples were dehydroabietic, palustric, abietic, and pimaric
acids. Smaller amounts of isopimaric, neoabietic, and sanderacopimaric acids were also
found. The study concluded that dehydroabietic acid could be used as a tracer for organic
matter derived from the paper mill.
Wood extractives often cause problems in pulping and papermaking. Not only are
they responsible for the formation of pitch deposits in the process system or in the pulp,
they may affect the quality of the product.Successfulpapermaking often requires that
compounds which cause pitch problems be removed from the system or otherwise neu-
tralized. Rapid, accurate, andsensitivemethods of analysis are neededto classify and
quantify the resin components. Knowledge of the extractives of the wood is important for
the implications for deresination and pitch control. Methods for separating and identifying
components of wood pitch have been reported [543-545,592-5951. They can be deter-
Chemical Characterizationof Wood 365
0 10
l 20
Time, min
FIGURE 44 A gas chromatogram of extractives (Kraft black liquor). (From Ref. 590.)
C. PhenolicExtractives
Phenolic compounds possess free phenolic functional groups in their structure. Extractable
phenolic compounds from wood, bark, and foliage range in complexity from simple phen-
olics (for example, vanillin) to polymeric condensed tannins. In some woods the amount
is small. It is usually high in barks and foliage.
Thephenolicextractivesaresoluble in the morepolarsolventssuchasacetone,
alcohol, or water, and they are also soluble in aqueoussodiumhydroxideandsodium
carbonate solutions. Some of the compounds of low molecular weight appear among the
volatile components.
The phenolic compounds constitute a heterogeneous class of compounds, which may
be divided into the following groups [598].
366 Baeza and Freer
The phenolic extractives are normally isolated from dried, ground, preextractedwood
with petroleum ether [599,600] or benzene [601] to remove resinous materials. They are
also extractable withethanol[602-6051,acetone[606],acetone-water[599,600,602-
605,6071, water [608], and dilute caustic solutions. SFE with acetone, THF, dioxane, and
toluene have also been used [609]. Resolution of the complex mixture of phenolic sub-
stances obtained by total extractives or fractions obtained after selective separations re-
quires the application of different techniques, such as selective solubility, chromatography,
ionexchange,countercurrentliquid-liquidextractions,andformation of derivatives or
reaction products [610]. Paper chromatography and thin-layer chromatography (TLC) have
been widely used to monitor separations. After isolation of the phenolic compounds, they
are characterized by various analytical techniques: GC [600,609,61 l], MS [5991, UV, IR,
and NMR [612].
D. Water-Soluble Components
The water-soluble substances are those which are dissolved by cold or hot water. They
include water-soluble carbohydrates, some organic acids, many of the phenolic materials,
and some inorganic constituents. Some of these are also soluble in organic solvents, and
the amount obtained in the water fractions depends on the previous solvent extractions.
Subsequent separations are necessary before identification. The isolation and characteri-
zation of the different fractions are given by Browning [613].
Extractable carbohydrates of wood have been analyzed by different very well estab-
lished techniques, such as "C-NMR, HPLC, and GC. These were used to characterize the
extractable carbohydrates for Norway spruce trees growing in different SO,-polluted sites
[614]. Samples representingtypical growth periods were selected. Each of them was milled
to a particle size of 0.25 mm, with cooling to avoid thermal decomposition, dried to a
moisture content of 13% relative to the dry weight, and stored at -20°C. Samples were
extracted with cold water, 1 g in 20 mL and 1 g in 100 mL for HPLC and NMR analysis,
respectively. For GC analysis, the samples were treated with TFA (10 mg of wood with
500 p L of 2M TFA) for 4 h at 100°C. The hydrolyzates were dried, washed twice with
methanol, reduced overnight with NaBH4. After neutralization, drying and washing, the
samples were neutralyzed, acetylated, and analyzed for GC using 2-deoxy-glucose as an
internal standard. All the samples contained carbohydrates composed of glucose, mannose,
galactose, xylose, and arabinose, and small amounts of rhamose. A higher content of xylose
andarabinosewasfound in the heartwoodregionthan in sapwoodofeach tree. The
average content of both glucose and fructose was found to be about 3% higher in the
haracterization
Chemical of Wood 367
highly stressed trees, and the content of free D-fructose was found to be higher than that
of free D-glucose in trees from the less damaged forest site.
The inorganic part of wood is analyzed as ash by incineration of the organic matter. The
inorganicmaterialscomprisefrom0.1% to 0.5% ofoven-dryweight of wood in the
temperature zones and up to 3-4% in tropical woods. There are numerous reports on the
ash content of various woods [615]. The main components found in wood ash are potas-
sium, calcium, magnesium, sodium, iron, silica, sulfate, phosphate, chloride, and carbon-
ate. Trace levels of many other elements have been detected. The amounts and types of
inorganic components depend on the soil in which the tree has grown, the fertilizers used,
air pollution, and on preservatives applied to lumber. The components are not distributed
uniformly through the tissue.
A. Determination of Ash
The determination of ash is always accomplished by incineration to remove the organic
matter. Some loss of volatile components, such as alkali metal chlorides and ammonia
salts, may occur. Procedures for ash determination in wood are given by Browning [616],
TAPPI Test Methods T 211 om-93 [2], ASTM D 1102 [3], and CPPA G. 10 [4]. All of
these utilize different temperatures. The value specified in the procedure for ash deter-
mination in wood, pulp, paper, and paperboard given by TAPPI Test Method T 2 11 om
93 [2] is 525°C. However, the user must specify the temperature used in order to present
accurate results and desired information. Combustion at 900°C is useful when an under-
standing of the noncellulosic materials present in the sample is required (TAPPI T 41 3
t21).
More reproducible and somewhat higher values are obtained from the sulfate ash
determination. The inorganic salts are converted to nonvolatile sulfates by adding sulfuric
acid before the ignition is completed. The general procedure is as follows: wood is heated
at a low temperature until most of the volatile materials are removed and a carbonaceous
residue remains. Some drops of 50% sulfuric acid are added, and the crucible is heated
until excess sulfuric acid is fumed off, and the ignition is completed at 700-800°C [616].
B. Determination of Elements
In studies related to the growth, metabolism, and feeding of trees, it is important to know
the concentration of a number of elements in various parts of the trees.
Analysis for trace and major elements in solid samples generally requires decom-
position of the organic matter followed by dissolution to give a solution for subsequent
analytical determination. The decomposition may be achieved by various classical or re-
cently developed procedures. each of which has particular advantages and disadvantages
[617]. The analysis of adigestedsamplegenerally is carried out by atomicabsorption
spectrometry (AA), inductive coupled plasma emission spectrometry (ICP-AES), electro-
chemical, colorimetric. gravimetric methods. and more recently, ion chromatography [61 S].
There are twogeneralprocedures for destroyingorganic matter: dryashingandwet
digestion.
368 Baeza and Freer
Dry ashing is simple, but some elements are lost in the process. The ash from wood
is used directly for analysis by arc emissionspectrography,or it is dissolved in dilute
hydrochloric acid or nitric acid for the analysis, generally by ICP-AES or AA. An alter-
native to the muffle furnace is the low-temperature plasma. Ash content of loblolly pine
was found to be higher when it was determined by the plasma method (0.335%) than that
obtained by the muffle procedure (0.250%) [6191.
Wet digestion is carried out by using mixtures of acids, including nitric and perchloric
acid [620], as well as perchloric, nitric, and sulfuric acids [621]. Hydrogen peroxide has
beenincluded in the digestionmixturetoavoid the useofperchloricacid[622-625].
Bomb digestion with hydrogen peroxide provides a relatively rapid means of decompo-
sition that assures a complete recovery of elements in wood samples [626]. Microwave
ovenshavebeen used for aciddigestion of manytypesof solids [627].Digestion by
microwave has proved to be a rapid and reliable method for plant tissues [628,629], which
may be applicable to wood samples. The digestion procedure depends on the matrix and
the elements to be determined. Samples are digested with either nitric acid, hydrochloric
acid, fluorhydric acid, or hydrogen peroxide in closed Teflon PFA vessels in a microwave
oven.
Organic matter can also be destroyed by combustion in a Schoniger flask or Parr
bomb [630]. This technique is time-consuming, and only small samples can be burned.
Loss of volatile elements is prevented because these devices are closed. Chlorine may be
lost during ashing and wet digestions. Therefore, samples containing organically bound
Cl are usually burned in a Schoniger flask or oxygen bomb [631]. The chloride formed is
usually analyzed by potentiometric titration with silver nitrate.
Nitrogen content is usually determined by Kjeldahl digestion, and sulfur in solution
as sulfate is analyzed gravimetrically by precipitation with barium, or by ion chromatog-
raphy [6 1 81.
Analysis of tree leaves, bark, and wood by sequential ICP-AES for Ca, Mg, K, Na,
P, Mn, Fe, Al, B, Cu, and Zn was carried out [6321. Samples were shredded and air-dried
before being powdered and dried at 105°C. A 5-g sample of this material was ashed at
500°C. The residue was extracted with 5 mL of I : 1 hydrochloric acid, filtered and diluted
with water up to SO mL. Leaf and bark extracts were diluted, while wood samples did not
need additional dilution. The solutions were analyzed byICP. The choice of the line to
be used for the analytical measurement was essentially determined by the sensitivity of
the line, or lack thereof, and by the presence of spectral interference.
After decomposition of hydrogen peroxide, which is used to treat wood, and elec-
trothermal vaporization, analysis with ICP-AES of wood samples of red spruce and sugar
maple were carried out [626]. Dried wood samples were decomposed in a bomb made of
Teflon with 50% hydrogen peroxide and heated in an oven at 125°C for 4 h. The element
concentrations were obtained sequentially by electrothermal vaporization ICP-AES using
S-pL sample aliquots.
Due to the high volatility of mercury, the commonlyuseddigestionmethods are
susceptible to mercury loss, requiring special wet techniques. The loss of mercury was
virtually eliminated by using aqua regia in the digestion of pulp and paperboard samples
[633]. This technique is also useful for wood. The measurement of mercury was done by
cold vapor atomic absorption. Neutron activation analysis (NAA) is extremely sensitive
and accurate and there are no requirements for the destruction of the organic matter 16341.
A discussion of the use of NAA in the analysis ofwood samples is provided by
Meyer and Langwig [63S], and data for different species are available [636,6371. For this
aracterization
Chemical of Wood 369
REFERENCES
1. E. Sjiistriirn. Wood C/wruisfryF/rrldrrr,rc.rlrtr/.sm d Applicnfior1.s. 2nd ed., Academic Press. New
York. p. 249 ( I 993).
2. TAPPI 7?sf Mc.fhod.s, Technical Association of the Pulp and Paper Industry, Atlanta, GA.
3. ASTM Sffmf[rrd.s,American Society for Testing and Materials. Philadelphia, PA.
4. CPPA S/mckrrd.s. Canadian Pulp and Paper Association, Montreal. Quebec, Canada.
5. Appiftc Stant/trrd.s, Australian Pulp and Paper Industry Technical Association.
6. I S 0 Sfcrrrhrds, International Organization for Standardization.
7. DIN-Norrrwrl, Deutsches lnstitut fur Normung.
8. B. L. Browning. M e t h o d s of Wood Chcw~isrry.Interscience, New York (1967).
9. D. B. Easty and N. S. Thompson. Fiber Sci. Technol. Ser., l / : 49 (1991).
IO. B. B. Sithole. Arltrl. Cllcwl.. 67: 87R (1995).
1 1. D. Fengel and G. Wegener. Wood Chc~r,ri.sfr;y,U//rur.strrcc.trcrc,h'e~rc/ior~.s, Walter de GrLlyter,
Berlin. chap. 3 (1989).
12. S. Y. Lin and C. W. Dence (cds.). M r f h o t l s i r l Ligrlirl C'/wrrli.stry, Springer-Verlag, New York
( 1992).
Freer 370 and Baeza
13.H.H.Nimz,in Proc. 8th Int. Symp. on Wood and Pulping Chentistry, Helsinki, Vol. 1, p. 1
(1995).
14. W. E. Moore and D. B. Johnson, Procedures f o r the Chemical Analysis of Wood and Wood
Products, US. Forest Products Lab. (1967).
15. N.Morohoshi,in D. N.-S.HonandN.Shiraishi(eds.), Woodand Cellulosic Chemistry,
Marcel Dekker, New York, chap. 8 (1991).
16. J. A.Smelstorius, Holdorsch. 28: 67 (1974).
17. D. Fengeland D. Grosser, Holz Roh-Werkstofl 33: 32(1975).
18.G.Wegener, Das Papier; 28: 478(1974).
19. C. M. Stewart. J. F. Melvin, S. H. Tham, and E. Zerdoner, Cellulose Chem. Techrzol. 7: 371
(1973).
20. B. L. Browning, Methods of Wood Chemistry, Interscience,NewYork,p.12(1967).
21. W. E. Kaarand D. L.Brink, J. WoodChem.Technol., 11: 479(1991).
22. B. L. Browning, Methods of Wood Chernistr?,, Interscience, New York, p. 45 (1967).
23. M. Goetzler, TAPPI, 65: 149(1982).
24.T.Ona,T.Sonoda,M.Shibata,andK.Fukazawa, TAPPI J., 78: 121 (1995).
25. R.T.Lin, ForestProd. J., 1 7 54 (1967).
26. R. T. Lin, ForestProd. J., 17: 61 (1967).
27.G.LohseandH.H.Dietrichs, Holz Roh-Werkstofl 30: 468(1972).
28.A.J.Nanassy, Wood Sci., S: 187(1973).
29.A. J. Nanassy, Wood Sci., 9: 104(1976).
30. A. R. Sharp, M. T. Riggin, R. Kaiser, and M. H. Schneider, Wood Fiber Sci., 10: 74 (1978).
3 1. E. Hsi, R. Hossfeld, and R. G. Bryant, J . Colloid Interface Sci., 62: 389 ( l 977).
32. E. Brosio, F. Conti, C. Lintas,and S. Sykora, J. FoodTechno/., 13: 107(1978).
33. J. E. CarlesandA.M.Scallan, J. Appl. Polytner Sci., 17: 1855(1973).
34. R. S. Menon, A. L. MacKay, J. R.T.Hailey,M.Bloom,A. E. Burgess,and J. S. Swanson,
J. Appl. Polymer Sci., 33: 1 141 (1987).
35. R. S. Menon, A. L. MacKay, S. G. Flibotte, and J. R. T. Hailey, J . Mag. Res., 82: 205 ( 1 989).
36. C. D. Araujo,A.L.MacKay,J.R.T.Hailey,K. P. Whittall, and H. Le, Wood Sci. Technol.,
26: 101 (1992).
37. A. Isogai,A.Ishizu, J. Nakano, S. Eda,andK.Kato. Carbohydr. Res., 138: 99 (1985).
38.G.J.Ritterand E. F. Kurth, lnd. Eng. Chetn.,25: 1250(1933).
39. T. E. Timelland E. C. Jahn, Svensk Papperstid., 54: 831(1951).
40.T. E. Timell, Pulp Paper Mag. Can., 60: T26(1959).
41.G. W. Holmesand E. F. Kurth, TAPPI, 42: 837(1959).
42. E. F. KurthandA. A. Swelim, TAPPI, 46: 591(1963).
43. Useful Method 249, in TAPPI Useful Methods, Technical Association of the Pulp and Paper
Industry, Atlanta, CA.
44.B.L.Browning, The Chemistry of Wood, Interscience,NewYork,p.389 (1967).
45.G.Jayme, Cellulose Chern., 20: 43(1942).
46.G.JaymeandG.Schwab, Papierfabr.. 40: 147 (1942).
47. L. E. Wise, Ind. Eng. Clwm., Anal. Ed., 17: 63 ( 1945).
48. E. L. Lovell, Ind. Eng. Chenl., 3 7 1034 (1945).
49. L. E. Wise,M.Murphy,andA.A.D'Addieco, P q w r Trarlt. J.. 122: 35 (1946).
5 0 . N. S. Thompsonand 0. A.Kaustinen, TAPPI, 47: 157 (1964).
51. N. S. Thompsonand 0. A.Kaustinen, TAPPI, 53: 1502 (1970).
52. B. L. Browning, The Clwrnistry of Wood, Interscience,NewYork.p.396 ( 1967).
53. D. Fengel, H. Uqar,and G. Wegener, Dtrs Papier. 33: 233 (1979).
54. B. L. Browning, Methor1.s qf Wood Chtvnistty, Vol.2.Wiley-Interscience.NewYork.p.394
( 1967).
5 5 . A. Poljak. Angen: Chcw~..AGO: 45 (1948).
56. J. S. AlbrechtandG.A.Nicholls. P q m i Puu. 58: 49 (1976).
57. H. Haas. W. Schoch,and U. Str6le, D(/s Ptrpier; Y: 469 (1955).
Chemical Characterizationof Wood 371
148. L. Ferrlis and P. Pa&, Cellulose Chem. Technol., 11: 633 (1977).
149. J. Sugiyama, H. Harada,and H. Saiki, 1nt. J. B i d . Mucromol., 9: 122 (1987).
150. A. Bourret,H. Chanzy,and R. Lazaro, Biopolpers, 11: 893(1972).
151. K. H. Gardnerand J. Blackwell, Biopolymers, 13: 1975 (1974).
152. W. Claffeyand J. Blackwell, Biopolymers, 15: 1903(1976).
153. R. H. Atalla and D. L. VanderHart. Science, 223: 283(1984).
154. J. Sugiyama, R.Vuong, and H. Chanzy, Macromolecules, 24:4168 (I991).
155. J. H. Wiley and R. H. Atalla, in R. H. Atalla (ed.), Strucfure of Cellulose, ACS Symp. Ser.
340, American Chemical Society, Washington, DC, p. 151 (1987).
156. J. Sugiyarna. J. Persson, and H. Chanzy, Macromolecules, 24:2461(1991).
157. D. L. VanderHart and R. H. Atalla, Mucrornolecules, 17: 1465 (1984).
158. F. Horii. A. Hirai, and R. Kitarnaru, Mrrcrotnolecules, 20:2117 (1987).
159. D. Fengel andG. Wegener, Wood Chemistry, Ultrastructure Reactions, Walter deGruyter,
Berlin, p. 87 (1989).
160. A. Sarko, TAPPI. 61:59(1978).
161. J. Sugiyarna, T. Okano, H. Yamamoto, and F. Horii, Macromolecdes, 23: 3196 (1990).
162. S. E. Darrnon and K. M. Rudall, Disc. Faraday Soc., 9: 251 (1950).
163. M.Tsuboi, J. Polymer Sci., 25: 159 (1957).
164. C. Y. Liang and R. H. Marchessault, J . Polymer Sci., 37: 385 (1959).
165. C. Y. Liang and R. H. Marchessault, J. Polytner Sci., 39: 269 (1959).
166. R. H. Marchessault and C. Y. Liang, J. Polymer Sci., 43: 71 (1960).
167. H. G. Higgins, C. M. Stewart. and K. J. Harrington, J. Polymer Sci., 51: 59 (1961).
168. K. H. Bassett, C. Y. Liang, and R. H. Marchessault, J . Polymer Sci. A, 1 : 1687 (1963).
169. D. N.-S. Hon and N. Shiraishi (eds.), Wood u r d Cellulosic Chernistty, MarcelDekker.New
York, p. 341 (1991).
170. J. Blackwell, in J. C. Arthur (ed.), Cellulose Chemistry and Technology, American Chemical
Society. Symp. Ser. 48, Washington, DC. p. 206 (1977).
171. D. Fengel and M. Ludwing, D m Papier, 45: 45 (1991).
172. D. Fengel, Dus Papier, 45: 97 (1991).
173. L. G. Braznik, R. G. Zhbankov, T. A. Tsetsokho,and A. M.Shishko, Zh. Prikl. Spektrosk,
53: 633 ( 1991).
174. A. J. Michell, Appitcr. 41: 375 (1988).
175. A. J. Michell, Curlmhydr. Res., 173: 185 (1988).
176. A. J. Michell, Curbohpdr. Res., 197 53(1990).
177. A. J. Michell, Cellulose Chern. Techno/., 27: 3(1993).
178. H. J . Marrinanand J . Mann, J . Polymer Sci. 21: 301 (1956).
179. A. W. McKenzieand H. G. Higgins, Svensk Puppersridn., 61: 893 (1958).
180. D. Fengel, Holzjiirsch., 46: 283(1992).
181. D.Fengel, Ho1;forsch.. 47: 103(1993).
182. D. V. Whittaker. L. A. S. Parolis, and H. Parolis, Curbohydr. Res., 253: 247 (1994).
183. G. Grbnberg, U. Nilsson. K. Block, and G. Magnusson, Carbohydr. Res., 257: 35 ( 1994).
184. J . Sano, N.Ikushima, K. Takada, and T. Shoji, Cnrbohydr. Res., 261: 133 (1994).
185. T. J . Rutherford, C. Jones, D. B. Davies, and A. C. Elliott. C~trbohylt:Res., 265: 79 (1994).
186. M. Matulov6, R. Toffanin, L. Navarini, R. Gilli, S. Paoletti, and A. Ceslro, Carho/Zydt:Res..
265: 167 ( 1994).
187. S. M. A. Homlbeck, P. A. Petillo, and L. E. Lerner, Riocherni.stc\; 33: 1426 (1994).
188. M. Bruix, J . JimCnez-Barbero,and P. Cronet, Curld~ycir.Res., 273: 157 (1995).
189. R. P. Gorshkova, V.V. Isakov, E. N. Kalmykova,and Y. S. Ovodov. Curbohydr: Res., 268:
249 (1995).
190. A. J . DeBruyn, J . Curbohydr: Chem., 14: 135 (1995).
191. L. Poppc and H. van Halbeek. N u t u r d Sfr’uct. Bid., 1: 215 (1994).
192. B.Adarns and L. E. Lerner, M q n . Reson. Chern., 32:608 (1994).
193. J . M. Hackney. R. H. Atalla, and D. L. VanderHart, I n t . J . B i d . Mucrotnol., 16: 215 (1994).
Freer 374 and Baeza
243. M. F. Chaplin and J. F. Kennedy (eds.) Carbohydrate Analysis, a Practical Approach. IRL
Press, Oxford-Washington, DC (1986).
244. Y. C. Lee, Anal. Biochern, 189: l5 1 (1990).
245. G. W. Monier-Williams, J . Chem. Soc., 119: 803 (1929).
246. J. F. Saeman, W. E. Moore, R. L. Mitchell, and M. A. Millett, TAPPI, 37: 336 (1954).
247. G. Jayme and H. Knolle, Z. Anal. Chem., 178: 84 ( 1960).
248. J. E. Jeffery, E. V. Partlow, and W. J. Polglase, Anal. Chem., 32: 1774 (1960).
249. M. L. Laver, D. F. Root, F. Shafizadeh, and J. C. Lowe, TAPPI, 50: 618 (1967).
250. R. C. Pettersen, V. H. Schwandt, and M. J. Effland, J . Chrornatogr. Sci., 22: 478 (1984).
251. D. Fengel and G. Wegener, in R. D. Brown, Jr. and L. Jurasek (eds.), Hydrolysis of Cellulose:
Mechanisms of Enzymaticand Acid Catalysis, Adv. Chem. Ser. 181, American Chemical
Society, Washington, DC, p. 145 (1979).
252. M. G. Paice, L. Jurasek, and M. Desrochers, TAPPI, 65: 103 (1982).
253. M. Tenkanen, T. Hausalo, M. Siika-aho, J. Buchert, and L. Viikari, in Proc. 8th Int. Svmp.
on Wood and Pulping Chemistty, Helsinki, Vol. 2, p. l89 (1995).
254. A. Rydlund and 0. Dahlman, in Proc.8th Int. Svrnp. on WoodandPulping Chemistry,
Helsinki, Vol. 2, p. 159 (1995).
255. E. P. Crowell and B. B. Burnett, Anul. Chenz., 39: 121 (1967).
256. L. G. Borchardt and C. V. Piper, TAPPI, 53: 257 (1970).
257. Pierre Handbook and Catalogue, 1979-1980, pp. 173 and 182, Methods 2 and 21.
258. K. M. Brobst and C. E. Lott, Jr., C e r e d Chem., 43: 35 (1966).
259. L.-[. Larsson and 0. Samuelson, Svensk Papperstid., 70: S71 (1967).
260. M. Sinner, M. H. Simatupang, and H. H. Dietrichs, J . Wood Sci. Technol., 9: 307 (1975).
261. M. Sinner and J. Puls, J. ChromrrtogI:, 156: 197 (1978).
262. M. R . Ladisch, A. L. Huebner, and G. T. Tsao, J . Chromutogr., 147: 185 (1978).
263. A. D. Jones, I. W. Bums, S. G. Sellings, and J. A. Cox, J. Chrornatogc, 144: 169 (1977).
264. R. Schwarzenbach, J . Chromatogr., 1 1 7 206 (1976).
265. R. T. Yang, L. P. Milligan, and G. W. Mathison, J. Chromntogr., 209: 316 (1981).
266. A. Heyraud and M. Rinaudo, Liy. Chromatogr., 3: 72 I (1 980).
267. H. D. Scobell, K. M. Brobst, and E. M. Steele, Cereal Chem., 54: 905 ( I 977).
268. F. E. Wentz, A. D. Marcy, and M. J. Gray, J. Chromatogr. Sci., 20: 349 (1982).
269. S. J. Angyal. Chenl. Soc. Rev.. 9: 415 (1980).
270. R. C. Pettersen and V. H. Schwandt, J . Wood Chem. Technol., 11: 495 (1991).
271. G. G. Neuburger and D. C. Johnson. A n d . Chem., 59: 203 (1987).
272. T. Hausalo, Proc. 8th M . S w p . Wood trnd Pulping Chemistry, Helsinki, Vol. 3, p. I3 1 ( 1995).
273. D. A. Mead. L. A. Larew, W. R. Lacourse, and D. C. Johnson, in P. Jandik and R. M. Cassidy
(eds.), Advunces i r l I o n Chror,latogrtrphy. Vol. 1. p. 13 (1989).
274. W. T. Edwards. C. A. Pohl, and R. Rubin. TAPPI J., 70: 138 (1987).
275. M. L. Lavcr and K. P. Wilson, TAPPl J., 76: IS5 (1993).
276. M. Suzuki, R. Sakamoto, and T. Aoyagi, TAPPI J.. 78: 174 (1995).
277. J . J. Worrall and K. M. Anderson. J. Woocl Chem. Technol., 13: 429 (1993).
278. L. Van Nifterik. J. Xu, J. L. Laurent. and J. Mathieu. J. Chrornutogf:,640: 335 (1993).
279. J . Puls, T. D. Glawischnig.A.Herrmann,A.Borchmann, andB. Saake. in Proc. 8 t h I n t .
Syrnp. on Wood and Pulpir~~g Chcwistry, Helsinki. Vol. I , p. S03 ( 1 995).
280. C. Huher. E. Grill, P. Oefner, and 0. Bobleter, Fre.senirr.s’J . A n d . Chcrn., 348: 82.5 ( 1994).
281. B . L.Browning. TAPPI. 32: l19 (1949).
282. M. Bylund and A. Donetzhuher. S ~ v n s kPuppr,v~id.,71: SOS (1968).
283. Z. Dische. J. Biol. Chcvn.. 167: 1 89 ( 1947).
284. R. W. Scott, W. E. Moore, M. J. Effland. and M. A. Millett. Antrl. R i o c ~ h c ~ n r21:
. . 68 ( 1967).
285. T. Bittcr and H. M. Muir. A/?tr/.BiochenI., 4: 330 ( 1962).
2x6. R. W. Scott, Anrrl. Chcwr.. S/: 936 (1979).
2x7. T. M. Jones and P. Albcrsheinl. Pltrnt Physiol., 3Y: 926 (1972).
288. S. Johnson and 0. Samuelson. Anul. Chirn. Actu. 36: 1 ( 1966).
I
375. C. Lapierre, B. Pollet, and C. Rolando, in Proc. 8th h t . Symp. on Wood Pulping Chemistry,
Helsinki, Vol. 2, p. 131 (1995).
376. K. Freudenberg, A. Janson, E. Knopf, and A. Haag, Chem. Ber.. 69: 1415 (1936).
377. F. D. Chan, K. Nguyen, and A. F. A. Wallis, J. Wood Chern. Techno/., 15: 329 (1995).
378. K. Y. Wu and D. L. Brink, J. Agric. Food Chern., 25: 692 ( 1 977).
379. H. M. Chang and G. G. Allan, in K. V. Sarkanen and C. H. Ludwig (eds.), Lignin, Occurrerzce,
Formation, Structure crnd Kecrction, Wiley-Interscience, New York, p. 433 ( 1 97 l).
380. K. liyama and T. B. T. Lam, J. Sri. Food Agric., 51: 481 (1990).
381. T. P. Schultz, T. H. Fisher, and S. M. Dershem, J. Org. Chenz., 52: 279 (1987).
382. S. M. Dershem, T. H. Fisher, S . Johnson, and T. P. Schultz, Ho/@rsch., 42: 163 (1988).
383. S. Larson and G. E. Miksche, Acta Chem. Scrmd., 25: 647 (1971).
384. W. G. Glasser and N. Morohoshi, TAPPI, 62: 101 (1979).
385. N. Mohroshi and W. G. Glasser, Wood Sci. Techno/., 13: 165 (1979).
386. E. Adler, I. Falkehag, and B. Smith, Acta Clwrn. Scand., 16: 529 (1962).
387. E. Adler, S. Hernestam, and I. WalldCn, Svensk Pupperstid., 61: 641 (1958).
388. Y.-Z. Lai, in S. Lin and C. W. Dence (eds.), Methods in Lignin Chemistry, Springer-Verlag,
Heidelberg, p. 423 ( 1 992).
389. Y.-Z. Lai and X.-P. Guo, Wood S c i Techno/., 25: 467 (1991).
390. Y.-Z. Lai, X.-P. Guo, and W. Situ, J. Wood Chem. Technol.. I O : 365 (1990).
391. Y.-Z. Lai and M. Funaoka, Holiforsch., 4 7 333 (1993).
392. Y.-Z. Lai and M. Funaoka, J. Wood Chenz. Technol., 13: 43 (1993).
393. Y, Ni, G. J. Kubes, and A. R. P. van Heiningen, J. Pu/p Puper Sci., 16: 583 (1990).
394. K. V. Sarkanen, A. Islam, and C. D. Anderson, in S. Lin and C. W. Dence (eds.). Methods
in Lignin Chemistrv, Springer-Verlag, Heidelberg, p. 387 ( 1 992).
395. Y. Matsumoto, A. Ishizu, and J. Nakano, Holiforsch., 40: 81 (1986).
396. Y. MatsumotoandG.Meshitsuka, in Pro(.. 8th I n t . SynqJ. on Wood Pulping Chenzistty,
Helsinki, Vol. 1, p. 557 (1995).
397. H. Taneda, N. Habu, and J. Nakano. Ho/;for.sch.. 43: 187 (1989).
398. M. Funaoka, I. Abe. and V. L. Chiang, in S. Lin and C. W. Dence (eds.), Methods i n Lignin
Chcmi.st~?t,Springer-Verlag, Heidelberg, p. 369 ( 1992).
399. M. Funaoka and I. Abe, Mokuxri Grrkktrishi, 24: 256 ( 1978).
400. M. Funaoka and 1. Abe, M o k u x i Gakkrrishi, 28: 529 ( 1 982).
40 I . M. Funaoka and I. Abe, Mokuzrri Gnkknishi, 28: 621 ( I 982).
402. M. Funaoka and 1. Abe. Mokuzui Grtkknishi, 28: 718 (1982).
403. M. Funaoka and 1. Abe. Mokumi G a k k a i s h i , 3 / : 67 1 (1985).
404. M. Funaoka and I. Abe, Mokuzc~iGakknishi, 32: 47 (1986).
405. M. Funaoka, M. Shibata, and 1. Abe, Ho/&rsc/z., 44: 357 (1990).
406. V. L. Chiang and M. Funaoka, Ho/;for.sch., 44: 309 (1990).
407. V. L. Chiang and M. Funaoka, Ho/;fi)rsch.,44: 147 (1990).
408. F. D. Chaw K. L. Nguyen. and A. F. A. Wallis. J. Wood Chcwr. 7 k ~ h n o l . .1 5 : 473 ( 1995).
409. J. R. Obst, J . W(~ ot/Chem. Techno/..S: 377 (1983).
410. K. Kuroda. A. Yamaguchi, and K. Sakai, Mokuztri Gakkrrishi, 40: 987 (1994).
411. A. D. Pouwels and J. J. Boon, J . Wood Chem. T c . r ~ h n o / . . 7 : 197 ( 1987).
412. M. Klecn and K. Lindstriim. Nord. Pu//J P q w r K o s . J.. 2: 1 I 1 (1994).
413. J. Sjostrom and M. Reunanen. J . A n d . AI?/>/.Pyro/ysis, 17: 305 ( 1990).
414. J. Tanaka. A. lzumi. H. Ohi. K. Kurodn. and A. Yamaguchi. in Proc.. 8th I n / . Symp. on Wr~ocl
Pulping Chc~rnistr:v.Helsinki. Vol. 3, p. 143 (1995).
415. D. N.-S. Hon and N. Shiraishi (eds.). Wood trnd Cdlrrlosic Chc~rrri.st~:v, Marcel Dekkcr. New
York. p. 317 (1991).
416. D. B. Johnson. W. E. Moore, and L. C. Zank, TAPPI, 44: 793 (1961).
417. A. G. Schiining and G. Johansson. S ~ ~ c w sPupperstid.,
k 68:607 ( 1965).
41 8. J. Marton. 7APPl. 50: 335 (1967).
419. P. A. Ahlgrun and D. A. I . Giiring. Ctrn. .l. Chc)trr..49: 1272 ( l97 I ).
Chemical Characterizationof Wood 379
466. G. J. Leary and R. H. Newman, in C. Dence and S. Lin (eds.), Lignin Cl?ernistry, Springer-
Verlag, Berlin, p. 147 (1992).
467. J. F. Haw and G. E. Maciel, Anal. Clre~n.,56: I323 (1984).
468. G . J. Leary, K. R. Morgan, and R. H. Newman, Appita, 40: 181 (1987).
469. G. J. Leary, K. R. Morgan, and R. H. Newman. Hol;forsch., 40: 221 ( 1986).
470. R. M. Ede, G. Brunow, L. K. Simola, and J. Lemmetyinen. Holdorscl?.,44:95 (1990).
47 1. K. R. Williams and R. W. King, Anal. Cl?em.,67: AI 25 (1990).
472. A. Bax and D. G. Davis, J. Magn. Reson., 65: 355 (1985).
473. M. F. Summers, L. Marzilli, and A. Bax, J. Am. Chern. Soc.., 108: 4285 (1986).
474. R. M. Ede and Y. Kipellinen, Res. Chern. Interrned., 21: 313 (1996).
475. R. M. Ede, J. Ralph. K. M. Torr, and B. S. W. Dawson, Ho/;forsch., 50: 161 (1996).
476. I. Kilpellinen, J. Sipill, G. Brunow, K. Lundquist, and R. M. Ede. J. Agric. Food Clwnl.,
42: 2970 ( I 994).
477. R. M. Ede and J. Ralph, in Proc. 8th I n t . Synlp. 011 Wood crnd Prrlpirlg Cl?c~nri.stry. Helsinki,
Vol. 3, p. 97 (1995).
478. N. Fukagawa, G. Meshitsuka, and A. Ishizu, J. Wood Ckern. Tecllr~ol.,12: 425 (1992).
479. R. M. Edeand 1. Kilpellinen, in Proc. 8th h1t. S y n p . o n Wood rrrzd Pulpirlg Cl?ernistry,
Helsinki. Vol. I , p. 487 (1995).
480. D. S. Argyropoulos, H. I. Bolker, C. Heitner. and Y. Archipov, Hol$orsck, 47: 5 0 (1993).
481. D. S. Argyropoulos, H. 1. Bolker, C. Heitner, and Y. Archipov, J . Wood Chenl. fi~chrlol.,13:
1 x7 ( 1993).
482. D. S. Argyropoulos, J. Wood cI?L'1?1. E~h~lol., 14: 45 (1994).
483. D. S. Argyropoulos, J . Wood Clrern. Techrwl., 14: 65 (1994).
484. D. S. Argyropoulos and C. Heitner, Hol2;forsch., 48. Suppl.: I12 (1994).
485. P. Malkavaara and R. A l h , in Proc. 8/11 h i t . Syrnp. on Wood c r n d Prrlping Clwrnistty, Helsinki.
Vol. 3, p. 103 (1995).
486. D. S. Argyropoulos, Y. J. Sun, M. Mazur. B. Hortling, and K. Poppius-Levlin. Nord. Pulp
Puper Res. J., 10: 68 ( 1995).
487. Z.-H. Jiang and D. S. Argyropoulos, in Proc. 8th I11t. Symp. on Wood m d Pulping Cherni.st~y,
Helsinki. Vol. I , p. 5 1 1 (1995).
488. M. Barrelle, J . Wood Cl~rrn.Tecknol., 12: 413 (1992).
489. M. Barrelle, Hol&rsc/1.. 47: 261 (1993).
490. M.Barrellc, J . Wood Chenl. Tech~lol..15: 179 (1995).
491. G . Wegener and D. Fengel, Wood Sci. Tecllr~ol.,/ I : 133 (1977).
492. 0. Faix, W. Lange, and E. C. Salud. Hol;for.sc.l1., 35: 3 (1981).
493. D. A. I. Giiring, in K. V. Sarkanen and C. H. Ludwig (eds.), Lignins, Occ~rrrcr~cc~, Formr/ioII,
Structur-e c r n d Rerrctiorls, Wiley-Intersciencc, New York. p. 695 ( 1 97 1 ).
494. M. E. Himnlel. J. Mlynar. and S. Sarkancn, Clzmnltrtogr: Sci. SL'I:, 6 9 353 (1995).
495, J. Pellinen and M. Salkinoja-Salonen, J . C h r o ~ ~ ~ r 328: ~to~ 299
~ ~( :1985).
.
496. 0 . Faix. W. Langc. and 0. Beinhoff, Hol;forsch.. 34: 174 (1980).
497. W. Langc, W. Schwcers. and 0. Beinhoff, H o l ~ ~ r . s c h35: . , 119 (1981).
498. H. L. Chum, D. K. Johnson, M. P. Tucker, and M. E. Himmel. Hol;fi)r.sch., 41: 97 (1987).
499 M. E. Himmel, K. Tatsumoto, K. K. Oh, K. Grohmann. D. K. Johnson, and H. L. Chum. in
W. G. Glasser and S. Sarkanen (eds.), Lig11i11 Properties ( r n d Mtrtc.rirr1.s. ACS Symp. Ser. 397,
chap. 6, p. X2 ( 1989).
500 0. Faix and 0. Beinhoff, Hol$)r:sch.. 46: 355 ( 1992).
50I W. G. Glasser. V. DavC. and C. E. Frazier, J. Wood Che111. k h r ? o l . . 13: 545 (1993).
502. E. J . Siochi. T. C. Ward. M. A. Haney, and B . Mahn. M~lcro~ilolecrr/r.s. 23: 1420 (1990).
503. S. Sarkanen. D. C. Tellcr. J . Hall. and J . L. McCarthy. Mrrc.r~)111olec./r/(~.s. I d : 426 (1981).
504. W. J. Connors. S. Sarkanen, and J. L. McCarthy, Hol;/i)r.sch.. 34: X0 (1980).
505. S. Rudatin. Y. L. Scn. and D. L. Woerncr, in W. G. Glasser and S. arka an en (cds.). Li,qlfill
PropvticJs m d Mtrterids, ACS Sytnp. Scr. 397. chap. 11. p. 144 ( 1989).
506. S. Dutta. T. M. Carver, Jr.. and S. Sarkancn. i n W. G. Glasser and S. Sarkanen (eds.). 12ig~tifl
P I - O ~ C ~ It - r~ d; CM .NS~ c I . ~ u / sACS
, Symp. Ser. 397. chap. 12, p. I 55 ( 1989).
aracterization
Chemical
Wood of 381
548. E. F. Kurth, Ind. Eng. Chem., Anal. Ed., 11: 203 (1939).
549. B. L. Browning, The Chemistry of Wood, Interscience, New York, p. 119 (1967).
550. D. Fengel and G. Wegener, WoodChemistry,UltrastructureReactions, Walter de Gruyter,
Berlin, p. 33 (1989).
551. D. W. Reeve and A. B. McKague, J. Pulp Paper Sci., l 7 5115 (1991).
552. A. Demirbag, Wood Sci. Technol., 25: 365 (1991).
553. T. Chen, B. Breuil, S. Carrikre, and J. V. Hatton, TAPPI J.. 7 7 235 (1994).
554. B. B. SitholC, J. L. Sullivan, and L. H. Allen, Hol7,forsch.. 46: 409 (1992).
555. B. B. SitholC, Appita, 45: 260 (1992).
556. Z. Zhang, M. J. Yang, and J. Pawliszyn, Anal. Chem., 66: 844A (1994).
557. H. Halvarson and 0. Hult, TAPPI J.. 66: 105 (1983).
558. B. B. Sitlholt, P. Vollstaedt, and L. H. Allen, TAPPI J., 74: 187 (1991).
559. E. J. Nurthen, B. V. McCleary, P. L. Milthorpe, and J. W. Whitworth, Anal. Chem., 58: 448
(1986).
560. D. F. Zinkel, TAPPI, 58: 109 (1975).
561. D. F. Zinkel, J . Wood Chem. Technol., 3: 131 (1983).
562. F. Terauchi, T. Ohira, M. Yatagai, T. Ohgama, H. Aoki, and T. Suzuki, Mokuzai Gakkaishi,
39: 1421 (1993).
563. F. Bright and M. E. P. M. McNally (eds.), Supercriticul Fluid Technology Theoretical and
Applied Approaches in Analytical Chemistry, ACS Symp. Ser. 488 (1992).
564. B. L. Browning, The Chernist~):,, of Wood, Interscience, New York, p. 91 (1967).
565. A. W. Schorger, Ind. Eng. Chem., 9: 556 (1917).
566. M. Myer, B. Hausmann, N. Zelmann, and K. Kratzl, Proc. TAPPI Research and Development
Division CO$, Technical Association cf the Pulp undPcrper Industry, Atlanta, GA, p. 155
(1984).
567. A. Sato and E. von Rudloff, Can J. Chem., 42: 635 (1964).
568. D. F. Zinkel, in M. H. Esser (ed.), Proc. 6th Annual Lightwood Research C o n j , Southeast
Forest Experiment Satation, Asheville, NC, p. 147 (1979).
569. J. Drew, J. Russell, and H. W. Bajak, Sulfate Turpentine Recovery, Pulp Chemical Association,
New York (1971).
570. J. Drew and G. D. Pylant, Jr., TAPPI, 49: 430 (1966).
571. D. F. Zinkel and C.R. McKibben, in M. H. Esser (ed.), Proc. 5th Annual Lightwood Research
C o n j , Southeast Forest Experiment Satation, Asheville, NC, p. 133 (1978).
572. S. V. Kossuth and J . W. Munson, TAPPI, 64: 174 (1981).
573. E. Zavarin, Y. Wong, and D. F. Zinkel, in M. H. Esser (ed.), Proc 5th AnnuolLightw.ood
Resetrrch Cotrf.. Southeast Forest Experiment Satation, Asheville, NC, p. 19 (1978).
574. R. P. Adanis. M. Granat, L. R. Hogge, and E. Rudloff, J . Chromotogt: Sci.. 17: 75 (1979).
575. F. Terauchi, T. Ohira, M. Yatagai. T. Ohgama, H. Aoki, and T. Susuki, Mokuzai Gakkcrishi,
39: 1431 (1993).
576. H. Axelsson. C. Bostrom, D. Cooper, and U. Svedberg, Norrl. Pulp Papt,r RPS.J., 7: 155
(1 992).
577. R. C. McDonald and L. J . Porter. N.Z. J . Sci., 12: 352 (1969).
578. J. M. Uprichard and J. A. Lloyd, N.Z. J. Forest Sci.. IO: 55 I (1980).
579. I . D. Suckling and R. M. Ede, Appitcc. 43: 77 (1990).
580. B. L. Browning and L. 0. Bublitz. TAPPI, 36: 418 (1953).
581. H. Wienhaus. Chcwr. Trchnol. (Berlin), 5 : 25 (1953).
582. J. P. Casey (cd.), P d p und P q w r . 3rd ed., Wiley. New York, Vol. I , pp. 126. 424. 487 (1980).
583. A. M. Mimms, M. J . Kocurek. J. A. Pyatte, and E. E. Wright (eds.), Krqf? Pulping, A Corn-
pilnrion of N o r c ~ s TAPPI
. Press. Atlanta, CA. p. 2 1 ( I 993).
584. A. Quinde and L. Paszner, Ho/;fi)r.sch. 46: 5 I3 ( 1992).
585. B. B. SitholC. TAPPI J.. 76: 123 (1993).
586. M. Yang. Frw.ui Hlrtr.une, 21: I148 (1993).
587. M. J . Brites. A. Guerreiro, B. Gigante. and M. J . Marcelo-Curto, J. Chrotncrto~I:.641: 199
(1993).
Chemical Characterizationof Wood 383
Nobuya Minemura
Hokkaido Forest Products Research Institute, Hokkaido, Japan
1. INTRODUCTION
Wood is an excellent material to absorb and reflect light. This physical interaction produces
wood whose color may range from almost white, as in the sapwood of many species, to
almost black, as in the heartwood of black ebony. The color characteristics depend on the
chemical components of wood that interact with light. Hence, the reaction of wood com-
ponents to light, heat, and chemicals will change the color of wood. Extensive studies and
observations have shown that most, if not all, wood species of commercial importance,
and in particular those used for furniture, paneling, and decks, are prone to discolor with
age. Discoloration occurs both indoors and outdoors. Manyfactors and elements participate
in the discoloration of wood. In this chapter, the major factors playing a role in discol-
oration, as well as the methods of removing and avoiding discoloration, are discussed.
The color of wood varies with wood species. In this section, a general concept of color,
the coloration of wood, and characteristics of the color of wood species are discussed.
A. Mechanisms of Coloration
Isaac Newton, the English physicist/mathematician, said, “Rays are not colored.” Color
is recognized only when a rayof light enters the eye and is absorbed in the retina by
light-sensitive receptor cells called cones and rods.
Visible light, which produces the visual sense for human eyes, is part of an electro-
magnetic wave. Its wavelength ranges from 380 to 780 nm, as shown in Fig. 1. Ultraviolet
(UV) light is at the lower end and infrared (IR) light at the upper end. Visible sensitivity
varies with wavelength. In a dark place, a wavelength of 500 nm can be seen. In a bright
place, the wavelength must be increased to 550 nm before the human eye can distinguish
it. UV light does not reach the retina because it is absorbed into the cornea or crystalline
lens; IR light reaches the retina but is not registered.
385
Minemura
386 and Hon
Sho-ave
X ray
Vacuum UV
c,
"
uv
Visible
IR
radar
Wave
for
-
Wave U e l e v i s l o n
t",
I 1
400 5 00 600 700 (nm)
P ' Db ' B I G ' Y ' 0 I R
P: Purple Db. Deep blue 0 : Blue G . Green
Y' Yellow 0 : Orange ,VR R: Red
When visible light strikes an object, if all the light is reflected, we recognize the
object's color as white. In contrast, when all the light is absorbed, we recognize the color
as black. Most materials absorb certain wavelengths and reflect the rest. The reflected light
is recognized as acolor, which is dependent onthe composition and amount of the reflected
light. For example, the reflection of wavelengths longer than 590 nm produces an orange
color.
Absorption of light by a material excites its electrons. Generally, electrons are in the
lowest energy state or ground state. If adequate energy is absorbed by the electron from
outside, the electron will transit toahigherenergystate, or excited state. Light is an
aggregate of photons that have energy, so depending on its wavelength, it can provide the
energy necessary for electron excitation and transition.
Betweentheenergyofaphoton ( E , kcal/mol) and thewavelength (A, nm), the
following equation can be derived:
2.86 X 10'
E=
A
Figure 2 shows the relationship between energy and wavelength. The shorter the wave-
length, the higher the energy will be
The electron of an unsaturated bond (e.g., )C=C(, )C=O, )C=NH, -N=N-)
can transfer easily to an excited state with a small amount of energy. In molecules con-
taining many unsaturated groupings that are all conjugated, the molecular orbitals con-
taining the electrons in the system will extend over these groups. The resulting high degree
of delocalization of the electrons means that the energy required for a transition decreases.
For example, one unit of )C=C( absorbs light at 190 nm, but @-carotene, in which I 1
units overlap, absorbs the light at 520 nm to give red. An atomic group having rr electron,
such as an unsaturated bond, is called a chromophore. An atomic group having isolated
electron pairs, such as " O H , " C O O H , and "OR, is called anauxochrome.Auxo-
Color and Discoloration 387
(kcall
160'
150 ~
140 -
130.
120 -
110-
100-
W
F go-
E, 80-
c
5 70-
6 60.
50.
40.
30
200 400 600 800
Wavelength
chromes assist the action of chromophores by intensifying the coloration or enabling the
absorption of light having a longer wavelength.
Minemura
388 and Hon
L a b
Q White birch 78.1 2.0 16.3
@ Japanese larch 63.0 11.3 21.5
@ Mizunara 48.8 8.9 16. 2
@ Black walnut 36.2 5.7 8.6
0
400 500 600 700
(nm)
Wavelength
FIGURE 3 Spectral reflectance curves and numerical color values for some woods.
Color and Discoloration 389
L a b
Ebony
Before 2.6
19.2 2.2
After 19.9 2.0 2.6
Rosewood
Before 7.7
37.0 7.9
After 44.0 9.4 8.9
70 -
..Q
.. .. .Q
68-i : .. ..
.. ..
0 180 360
,. ..
E
-U . . .. ..
(deg )
Angle
.-c 6 6 - ; : . .
.. .. a
. .
0 0
'2 1
60
0 180 360 0 180 360
(deg) (de91
Angle
L a b
L a h
~~ ~
30
0
OD 35 o 25
0 o o
0
€jb
0
0
30 20
0
0 0
0
25 15
b 0
€P oo 0 0 0 0
B, 8 0
0
Ip 20 0 10 0 0
0
8
0
15 5
I W
0 0
W
O 10 0
0 5 10 15 20
a
FIGURE 5 Distribution sphere of color of various woods.
Allof the woods distribution are in positive sphere of chromaticness indices a and b.
Numerical values of lightness range from 20 to 85 [5].
In Figs. 6 and 7, the correlation of three color factors by the Munsell specification
system is shown [55,58]. When reddish extent becomes large, Munsell value decreases.
When yellowish extent becomes large, Munsell value increases. Saturation is in proportion
to Munsell value for typical tropical woods.
AA
2
1 2 3 4 5 6 7 8
Saturation
FIGURE 7 Relationship between Munsell value and saturation of typical tropical woods in Asia.
Japanese-style houses and are widely used as ceiling board, wall paneling, and posts. Red
birch is popular for furniture and interior doors. Rosewood and Chinese quince are widely
used as decorative materials in house construction, as well as for carving, instruments,
high-quality art objects, and so on.
c.Yellowish-BrownWood. Yellow and orange colors convey warm, homelike feel-
ings. Teak and keyaki are widely used as building materials and for furniture. Tsuge is
used for stamps and combs. Keyaki, which has a peculiar pattern, is the most precious
wood.
d.Light Brown Wood. Mizunara is a light-colored, expensive material. Mizunara
and nire are used for furniture as well as window frames, building construction, high-
quality goods, and so on. Trend in color furniture has been considered an indication of
the economic climate. For example, if light-colored woods such as mizunara are popular,
it implies economic recovery. In contrast, if dark colors such as deep red are popular, it
suggests an economic recession.
e.BlackWood. Blackconveys the impressionoforderandcalm.Ebony is used
for carvings, furniture, decorative materials in home construction, Buddhist family altars,
and so on.
5 Contrast of Color in Sapwood and Heartwood. The color of heartwood and sap-
wood differs in manywoods.Thisdifference is often used decoratively. Forexample,
decorative pillars of ebony display the brown color of heartwood and the white-yellow
color of sapwood; decorative pillars of hinoki utilize the natural contrast of red and white.
Such a contrast also is emphasized in fancy goods, souvenirs, cufflinks, and so on.
Because wood is porous, water-soluble substances or salts are often deposited in its voids
during the course of growth or after logging. Such deposits can change the wood’s color.
Except in the caseofdecay by microorganisms,discoloration does not meanan
accompanying loss of wood strength and is usually limited to the surface layer. Because
wood color is an important factor that strongly affects value, discoloration is a serious
problem from the viewpoint of commercial worth. Table 4 summarizes the characteristics
of discoloration according to the factors that influence the change in color.
After logging Addition of source of stain Biological source Propagation of microorganism Blue stain
from the outside
Chemical source Bonding of metal ion Iron stain
Bonding of acid Reddish discoloring of zelkova
Bonding of alkali Adhesion of cement
Physical source Heating Sticker mark
Irradiation of light Discoloration by sunlight
Immanence of source of Metal ion Blackish discoloring of sugi
stain Enzyme Red discoloration of alder
Resin Exudation of resin
In shade Imperfect pruning Brown stripes
Deposition of substance Existence of specks
396 Hon and Minemura
induced discoloration is calculated based on the total difference of the lightness and chro-
maticity indices before and after light irradiation. As lightness generally changes more
than chromaticity indices, lightness and the quantity of photo-induced discoloration have
a strong mutual relationship. Whiteness is also calculated from the sum of the lightness
and chromaticity indices. As the lightness of wood is higher than chromaticity indices,
whitenessandlightnesscorrelate in mostcases. In the case of asmallchange in the
chromaticity indices, the declining rate of whiteness and the quantity of the photo-induced
Color and Discoloration 399
Fading only
Rosewood 15.9 - 30.3
Indian rosewood 12.7 -50.8
Fading-darkening
Teijsmanniodendron 8.2 -9.7
Nire, Ulmus davidiann davidiana 5.2 -1.1
var. japonica
26 -
0
24 - 0
22 -
.
A 0
20 - 8 A
0
18 - 0 0
0
0
% 16-
0 0
14 -
0 0
00
t A
L 12 - B
0 0 A
5 10
-o o a A
L
ta
2
V
8 - 0
0 d o
0
0
6 - 0 AAA A
4 -
02 0
A A
A'
2 -
0 " " " " " '
0 2 L 6 8 1810
161412
a
ChangeInduced by exposure '
26 -
24 -
b 22-
/
20 - <l’ Numerals in the figure show exposure
!ime(hr) to carbon arc light.
l8 L I
2 4 6 0 10 12 14
a
FIGURE 9 Change of chromaticity indices ( a and h ) of a todomatsu during exposure to a carbon
arc lamp for 100 h.
discoloration also have a mutual correlation. In the 100 wood species described earlier,
10% haveahighervalue of whitenessafter 100 h of light exposureascompared to
unirradiated wood.
Photo-Induced Discoloration of Sapwood and Heartwood. Comparedwith
heartwood, sapwood has a pale color. The photo-induced discoloration of woods that have
a distinct contrast in color betweenheartwoodandsapwoodshows that the sapwood
usually keeps darkening, even if the heartwood changes from darkening to fading in color.
The quantity of the photo-induced discoloration, therefore, is higher in sapwood than in
heartwood. The discoloration pattern is similar to that illustrated in Fig. 9. It changes in
the direction of red with high saturation.
Photo-Induced Discoloration WhenLeft Indoors for a Long Time. Timber of
50 wood species were left indoors for 1800 days and their photo-induced discoloration
determined [59]. As shown in Fig. 10, the color of the timbers changed in various direc-
tions within 780 days, but after 980 days all the woods changed only in the yellowish
direction. In Fig. 11 the change of photo-induced discoloration of skinanoki, as a function
of the change of irradiated wavelength, is shown. Photo-induced discoloration becomes
large when the wood is irradiated with light in the near-UV region. When the wavelength
changes from short to long. hue changes are as follows: yellow + red + purple + blue.
b. WavelengthsParticipatinginPhoto-InducedDiscolorution. When 75 kinds of
commercial wood were exposed to light, 62% of the woods discolored with UV light and
28% of the woods discolored with visible light [6]. Figure 12 shows the amount of photo-
induced discoloration when the sapwood and heartwood of karamatsu were covered with
various glass filters andthenexposed to light [7]. In Fig. 12, a positive valuemeans
darkening and a negative value means fading. It is clear that the heartwood of karamatsu
discolorsstronglywith UV light and slightly with visible light. In the photo-induced
discoloration of sapwood, light over 390 nm gives rise to lightening, and light under 390
nm brings about darkening. The wavelength range for lightening (or bleaching) is consid-
eredtobebetween 390 and 580 nm. This range for lightening is also discernible in a
photo-irradiated newspaper 181.
Color and Discoloration 401
b*
b' I
35 -
30 -
25 -
20 -
15-
-980th day
lot
310 nm
I I I I I I
a'
-5 t
u t I I I
FIGURE 12 Discoloration of a Japanese larch covered with various light filters when exposed to
xenon light.
Rosewood shows typical fading with light irradiation. The influence of various wave-
lengths on the fading is shown in Fig. 13 [ 5 ] .The fading becomes stronger when UV light
with shorter wavelengths is used. The previously described range for lightening does not
fortify fading in this case.
Figures14and 15 show the rates of increaseanddecrease in the reflectance of
irradiated karamatsu which was covered with various filters during exposure, against unex-
posed wood [7]. When a filter was not used or when only UV light was used, heartwood
and sapwood changed toward redin color. The change in reflectance shows the largest
decrease at 410 nm, likely due to the formation of a quinonoid structure [ 91. Judging from
the irradiation wavelengths and the pattern of the reflectance curves of karamatsu heart-
wood, it appears that wavelengths between 300 and 390 nm cause discoloration to red,
390 to 580 nm cause discoloration to yellow, and over 580 nm results in scarcely any
change. Wavelengths of 390 to 590 nm also cause lightening of heartwood.
In the sapwood of karamatsu, light of 300 and 390 nm cause yellowing, 390 to 580
nm cause lightening, and over 580 nm cause no discoloration. That hue and saturation
differ with wavelengths probably means that the reaction occurring in wood during photo-
induced discoloration isnot simple. Details of the photochemistry and photoxidation re-
lated to discoloration can be found in Chapter 11.
c. Wood Cor?lpound.s Relutecl toPhoto-Induced Discolorcrtion. Most woodcom-
pounds related to photo-induced discoloration are high-molecular-weightcompounds
which are insoluble in solvent [56],although a few low-molecular-weight compounds are
known. In western hemlock sapwood, five lignans. one neolignan, and three minor con-
Color and Discoloration 403
-5
C
=
.c
D
-10
-b ----"""_ - -0
0
0 -15
-0
~~
0 25 50 75 100
(W
Exposuretimetocarbonarclight
Transmittedwavelength , x , Infrared only , 0 ; Over 660 run , A ; Over 630 nm.
0 , Over 500 nm , v , Over 530 nm , B Over 480 nm. A , Over 430 nm
I .
, Over 370 nm , A , Over 360 nm , o ;Over 320 nm, 0 , Ultravidet only,
v . No fillerwasused.
stituents have been isolated as the causing materials for photo-induced discoloration [60].
Almost all ofthemhave a quaicylringstructureandanoxygenationstructureat the
neighboring a-position on the aromatic ring. Concerning the photo-induced discoloration
of western red cedar, the participation of the causative compounds are plicatic acid, pli-
catinaphtol, and plicatinaphtalene in a rough ratio of 5:2:2 1611. In rosewood, 4-methoxy-
dalbergione has been isolated as the causative material [57].
d. Restoration of Sound Color to Discolorated Wood. One method for restoring
sound color is todecomposethesurfacechromophoricstructuresofthesystem with
bleaching chemicals such as hydrogen peroxide or sodium chlorite. Another method is to
sand the surface with sandpaper or a plane. Surface treatment is effective because discol-
E
c o
EY -l0
-20
0,
Q
.-c
-30
g -40
C
2 -50
V 500 600 700
Wavelength (nm)
FIGURE 14 Change in percent reflectance o f a karamatsu heartwood that was covercd with light
filtersandthenexposed to a xenonlight.Transmittedwavelength: 0 : Ultraviolet only (300-415
n m ) , 0 : Over 430 nm. 0 : Over S80 nm. A: N o lilter was used. Light source: Xenon light.
404 Hon and Minernura
("IO)
-
2C 30
2 20 -
-
G 10
?!
+
c 0 -
a,
Y -10 -
Q
a
c
-20 -
._
0 Mizunara
0 0 Sawagwurnl
5 -40 0 Kri
V A Bund
-50
600
400 500 700
(nm)
Wavelength
FIGURE 15 Change in percent reflectance of a karamatsusapwood that was covered with light
filters and then exposed to a xenon light.
oration is usually limited to the surface layer. For a faded surface, painting with dye or
pigment is also effective.
e. ControllingPhoto-InducedDiscolorution. Although the photo-induced discol-
oration of wood can add to an impression of dignity and age, for most applications dis-
coloration is regarded as an unfavorable reaction. Various mechanisms have been proposed
with respect to photo-induced discoloration. On the basis of the accepted mechanisms, one
or acombination of thefollowingmethods may beadopted to prevent photo-induced
discoloration:
1. Cutting off UV light
2. Modifying the light-absorbingstructures
3. Destroyingthestructuresparticipating in discoloration
4. Eliminating oxygen or capturing the singlet oxygen
5. Scavengingfreeradicals
6. Extractingtheprecursors of discoloration
The method selected must not damage the original wood color. It must be easy to carry
out, low in cost, and safe.
Cutting off UV Light. One way to prevent photo-induced discoloration is to coat
the wood surface with asubstance that absorbs UV light, i.e., to use a UV absorber.
Normally, commercially available UV absorbers are colorless or pale yellow and absorb
light below 400 nm. Some UV absorbers change their structure by absorbing light and
releasing the absorbed energy as heat to regain its original structure without degradation.
About 35% of 70 commercial wood speciesdiscolored by light canbecontrolled by
coating with UV absorbers [6].
The general treatment is to coat the wood surface with organic solvents or paint in
which the UV absorber dissolves. Coating with paint is ideal because a film of UV absorber
forms on the wood surface, preventing degradative UV light from reaching the surface.
However, if the film is peeled off, the protection disappears. A better method is to use a
UV absorber with functional groups that will react with wood. For example, coating with
2-hydroxy-4-(2,3-epoxypropoxy) benzophenone. accompanied by heating under pressure,
Color and 405
reducing chemicals is somewhat effective for initial control: Na,SO,, NaHSO,, Na,S20,,
( N H M O , , (NH,),S2O3, ascorbic acid, etc. For some tropical woods, control is achieved
in several ways [65]. Sodium chlorite treatment gives the best result for white meranti,
while acetylation is best for selangan batu and sepetir. Hydrogen peroxide treatment and
methylation give the best result for yellow meranti and nyatoh, respectively. Coating of
chitosan onto Douglas fir wood leads to a decrease of photo-induced discoloration [66].
Destroyingthe Structures Participating in Discoloration. Destroying the func-
tional groups and precursors of discoloration is also a method of preventing discoloration.
Acetylationcombined with oxidativebleaching and treatment with NaBH, areknown
[2O,2 1 1.
Wood coated with polyethylene glycol (PEG) becomes white with light irradiation,
as shown in Table I O [22]. Figure 17 shows the behavior of photo-induced discoloration
of manggashinoro coated with PEG. As the amount of coating increases, the values of n
and b diminish,implyingdiscoloration in the direction of achromatic color. Figure 17
shows the increases of lightness and whiteness, and Fig. 18 shows the effects of discol-
oration for various woods. This treatment works well on light-colored wood such as mang-
gashinoro, ezomatsu, and Douglas fir, but has negative effects on dark-colored wood such
as walnut and rosewood, because it turns such woods white. Regardless of the size of the
molecule, PEG has controlling effects on discoloration. The coated wood, however, has a
tendency to be sticky and look wet because of the hygroscopic nature of PEG, especially
if a large amount of low-molecular-weight PEG is used. Although wood bleached with an
oxidativebleachingagentsuch as hydrogen peroxideorsodiumchloritebecomesdark
Color and Discoloration 407
0 25 50 75 100
Exposure
time
to carbon arc light (hr)
when irradiated with light, a coating of PEG on bleached wood has a good controlling
effect on discoloration [23]. When newsprint made of softwood is coated with PEG and
irradiated with light, it gives lower levels of photo-induced discoloration [24], lower al-
kaline extract, and lower initial radical formation than the untreated form. As shown in
Fig. 19, PEG-coated filter papers impregnated with various phenolic substances gives very
low quantities of photo-induced discoloration [25].
For whitening of wood coated with PEG, the mechanism was elucidated 1551. When
the terminal hydroxyl groups of PEG are methylated or methacrylated, the effect is the
same as with unmodified PEG [ 5 ] . When saturated hydrocarbon is replaced by hydroxyl
groups, the same controlling effect is recognized. This suggests that the hydroxyl group
plays an important role. Polypropyleneglycol (di-ol type) exhibitsasimilarcontrolling
effect. Pulp which is saturated with alcohol is bleached when it is irradiated with near-
ultraviolet in the presence of oxygen [67]. This is presumably due to the formation of a-
hydroxyhydroperoxides from alcohol. The structure of polyoxymethylene is similar to that
of PEG. It is well known that, with light irradiation, this compound splits the main chain
to generate free radicals -CH,O. and .CH20- [26]. Based on the above-described facts,
the following mechanisms are considered. When wood coated with PEG is irradiated with
light, radicals are generated by two different routes. One route is by abstraction of hydro-
gen in the ethylene oxide chain by the excited a-carbonyl group. The other is photolysis
of PEG. The radicals being generated react with oxygen in air to give peroxy radicals.
These reactions are summarized below:
@ I I
C=O + h v + C=O”
I l
408 Hon and Minemura
0-0.
I
-CH?-CH-O- + 0, + -CH2-CH-0-
@ -CH2-CH2-O-CH2-CH2-O- + h u 4 -CH2-CH2"O*
+ .CH2-CH2-O-
.CH2-CH,-O- + 0 2 + .00-CH2-CH2-O-
These peroxy radicals will destroy the coloring structure or its precursor. PEG does not
absorb sunlight, but it does associate with phenolic compounds. Therefore, in wood coated
with PEG, there is the possibility that a complex is formed between PEG and penols or
Color and Discoloration 409
0 25 50 75 100
(hr)
Exposuretimeto carbon arc light
FIGURE 17 Photo-induced discoloration of woods coated with polyethylcne glycol after exposure
to a carbon arc lamp for 100 h. Coat weight of PEG (gln?): 0 ; 0, 0 ; 6 , A: I 1, X ; 22. Numerals in
the figure mean exposure time.
lignin that absorbs light. Use of a peroxide such as benzoperoxide in combination with
PEG shows a higher whitening effect in some cases [25].In the use of wood coated with
PEG, additional coating with paint is necessary. The adhesivestrength of polyurethane
film on wood coated with PEG is high even after light irradiation. It is assumed that a
urethane bond is generated due to the reaction of PEG with isocyanate. Instead of paint,
a coating of wax may be used 1271.
PEG is also effective for whitening of yellowed paper [ 5 5 ] .
Eliminating Oxygen or Capturing the Singlet Oxygen. If light penetrates a sur-
face isolated from oxygen, discoloration does not occur. Wood-plastic complexes show
lower rates of discoloration when exposed to light. However, since void space in a wood
cell is filled with plastic, the diffused reflection of light vanishes and the following phe-
nomena occur: appearance of wet color, decline in lightness, and a rise in saturation level.
A singletoxygenquencher traps the excitedenergy of IO2 that actsasacatalyst
duringphotoreaction.p-Carotene retards the light decomposition of lignin model sub-
stances [28].Nickel complex and 1,4-diazabicyclo[2,2,2]octanearealso well known as
singlet oxygen quenchers. Many quenchers are colored substances and do not regenerate
after quenching. Hence, applications of single oxygen quenchers to wood are limited.
Scavenging Free Radicals. A coatingcompound that has activehydrogens such
as phenolic derivatives and phenolic amines can be used to captureradicals.However,
since it is not reclaimed after use and often crystallizes when used in a thick coat, its use
in wood is limited.
Extracting Solvents. Extraction can remove discoloration that has been caused by
solvent-soluble components. This has been reported for pencil cedar [29]and rengas (30)
extracted with methanol. Incomplete extraction can cause an accumulation of the discol-
ored components on the surface, due to migration from the inner wood when the solvent
evaporates.
410 Hon and Minemura
Color difference A €
0 5 10 15 20 25
Manggasinoro
Hinoki
Melapi
Larch
Douglas fir
Redcedar
Ezomatsu
Hiba
Todomatsu
Shiurizakura
Yamaguwa
Kihada
Walnut
Rose wood
Coat weight
of PEG (g/ma) :
from the top , 0, 6 , 11, 22
FIGURE 18 photo-induced discoloration of woods coated with polyethylene glycol after exposLIre
t o a carbon arc lamp for 100 h.
2. Discoloration by Iron
During the woodworking process, a black stain often appears on the surface of veneer 01'
lumber. This stain usually is caused by a chemical reaction between iron ions and wood
components. Generally, iron stain is seen i n heartwood rather than sapwood. Such iron
stains account for about 70% of the discoloration problems i n the wood industry. Causes
of iron stain, methods of stain removal, and staining prcvention methods are discussed in
this section.
Color and Discoloration 41 1
Color difference AE
0 5 10
1 . , . 1 , ' " ' 1
In transporting the yeneer from the slicer to the dryer, iron stains often occur when
the veneer touches the metal fittings of the joints of the carrying belt. The veneer dryer
usually is made of iron. Therefore, the wet veneer often is stained when it touches the
dryer.
If the slicer process is used to produce veneer, the same phenomena are observed as
when a rotary lathe is used. The thin veneer made from flitch is about 0.2 mm thick. This
veneer is dried by hanging it at room temperature. Black marks often occur when sup-
porting iron hooks are used.
ExamplesinPlywoodManufacturing. Plywood is made by pressing with a hot
press. The plate of the press is made of iron, so an iron ion is produced when the moisture
in the glue layer evaporates during pressing and touches the hot plate. When this iron ion
transfers with water onto the surface of the plywood, an iron stain is produced.
If the vessel used in mixing gluing components (adhesive, water. filler, and hardener)
is made of iron, the surface coated with glue can be discolored.
Examples During Sawing. When green lumber is bound with steel belts for trans-
portation, the wood may become black where it touches the belt.
Prior to transporting, greentimber is treated withamold-proofingsolution. This
water-soluble, antimold agent is usually prepared by diluting the chemicals with water and
placing them in an iron vat. In this case, iron dissolves into the solution. When the lumber
is immersed into this solution and dried. the surface of the lumber often becomes black.
Round poles are made by shaving long green logs withan exclusive shaving ma-
chine. Small guide rollers are fitted into the machine to prevent the logs from bending.
Because the rollers are made of iron, linear stains often are formed where the wood touches
the rollers. This usually occurs at high temperatures and in heartwood.
Examples of Other Woodworking Processes. In the manufacture of carved wood,
wood is cooked with hot water and fixed with a molding flask. A black stain occurs when
the wood contacts bolts fastening the flask.
When the lumber is dried, stains occur when the wood makes contact with carrying
metal fittings that contain iron or with condensed water from the ceiling in the drying
room.
Water-soluble adhesives such as polyvinyl acetate or urea resin are used in manu-
facturing laminated wood. If the instruments for adhesion (the vessel for preparing glue,
the spreading apparatus, the pressing tool, etc.) contain iron, the wood surface where the
glue is applied often becomes black. When furniture is painted, sealing material is applied
to the wood surface prior to painting. Red sealing material often contains iron oxide, so
surfacescoatedwith it turn black. Red putty is oftenused to fill cracksorcavities in
plywood.When the surface of the plywood is laminatedwithadecorative veneer, the
veneer on the filling often becomes black.
Occurrenceof Stain inFinishedWoodProducts. Woods used asexterior wall
panels or in fences form black stains around nails because iron from the nails dissolves
into rain water and reacts with the wood. A similar phenomenon is observed in wooden
entry doors equipped with iron fittings. In wood flooring, iron plates are fitted vertically
under the flooring and are inlayed into green concrete. The water in concrete dissolves
the iron and penetrates into the flooring to form iron stains. In many cases, the stain does
not reach the surface because of the thickness of the flooring.
Wood plates often are used as chopping boards. Black marks can occur when a wet
kitchen knife is left on a board.
EvaluationofIron Stain. Iron stains are very similar to stainscaused by black
mold. In order to remove the stain, rapid judgment of whether or not the stain has been
Color 413
caused by iron is required. Iron stain usually occurs in heartwood. The stain is flat, not
swollen like a mold stain. Small iron pieces may be seen in the center of the stain. When
the stain is coated with 5% oxalic acid solution for several minutes, if it fades, the stain
has been caused by iron.
b. Chemical Factors Influencing the Occurrence of Iron Stain
Concentration of Iron Solution. Oak can be stained with 0.0001% of iron [31].
Thirty-three wood species of decorative veneers widely used in furniture or interiors can
be stained with 0.00005% of iron (Table 1 1 ) [32]. Sixty percent of the tested wood was
stained at one-tenth of this concentration, whereas hinoki and kiri were stained at one-
fiftieth of this concentration.
Tannin Content. Iron ions are widelyused in the qualitative analyses of various
phenolic substances because such ions react easily with them to produce coloring sub-
stances. In wood there are many phenolic substances, such as lignin and tannin. Gallotan-
nin, catechin tannin [33], and gallic acid [34] are recognized as sources of iron stains. The
relationship betweenthetannincontentand iron stain produced in 16woodspecies is
summarized in Table 12 [35]. Woods that contain high levels of tannin show a significant
decline in lightness. Generally, the degree of the stain is higher in heartwood that contains
a large amount of tannin.
pH. The pH values of woods vary with species, as shown in Table 12. Woods that
have low pH values seem to stain easily. The relationship between the pHof wood and
staining with ferric chloride solution has been examined in about 55 species of Japanese
woods and 54 species of tropical woods [36]. In Japanese wood, the amount of staining
generally decreases as the pH value increases, but in tropical woods this tendency is not
clearly recognized. The relationship between the pH of the iron solution and the amount
of staining is shown in Fig. 20. Itis clear that the staining is greatest at a pHof 4 and
lowest at a pH of 7 [37].
Moisture Content. Nailingdriedlumberdoes not cause a change in color, but
nailing greenlumbercancause a black stain around the nail. The lower limit of the
moisture content of wood that causes an iron stain is at the fiber saturation point [33].
Table 13 shows the relationship between humidity and the occurrence of iron stains [37].
This experiment used veneers prepared at various humidities. Iron powder was sprayed
on the veneersand left undervarioushumidity conditions. At 100% relative humidity
(RH), all woods stained, but below 95% RH staining occurred only in Douglas fir. The
equilibrium moisture content at 100% RH is the fiber saturation point, which is 22-35%
at toom temperature.
Concentration (%)
TABLE 12 Relationship Among Tannin Content, pH, and Decreasing Rate of Lightness of
Woods Treated with 0.1% Ferric Chloride Solution
Decreasing rate of
Tannin lightness (%)
pH
content of
Species (%l wood 5 min 4 days
Softwood
Douglas fir (heartwood) 0.3 3.75 50.9 86.4
Sugi (heartwood) 0.3 6.05 38.4 57.3
(sapwood) 0.1 5.40 25.5 53.4
Lawson cypress (heartwood) 0.2 4.35 28.2 63.9
(sapwood) 0. l 5.OO 43.5 65.7
Akamatsu (heartwood) 0.1 4.55 20.4 69.9
Hinoki (heartwood) 0.1 5.30 23.1 59.2
Hardwood
Sawagurumi (heartwood) 2.1 4.20 66.2 77.0
Mizunara (heartwood) 5.6 4.65 68.0 79.9
(sapwood) 1.2 5.10 70.0 76.0
Painted maple (heartwood) 0.6 4.75 58.2 73.7
Black walnut (heartwood) 2.0 4.70 51.6 58.8
Buna (heartwood) 0.4 5.60 40.4 77.0
Kiri (heartwood) 0.6 4.80 42.4 54.7
Japanese red birch (heartwood) 0.3 4.60 32.1 50.1
Manggashinoro (heartwood) 0.2 3.95 20.8 47.8
Magnolia (heartwood) 0.4 5.05 21.2 39.7
(sapwood) 0.2 5.60 14.7 38.4
Swamp ash (heartwood) 0.2 5.40 26.0 34.6
Teak (heartwood) 0.4 5.00 4.9 16.8
Test specimens were soaked in 0.1% ferric chloride solution for S rnin or 4 days.
Oxygen. The amount o n staining is less in a nitrogen atmosphere than in air. When
a test specimenwastakenout of anitrogenatmosphereand leftin air, the degree of
staining became the same as that of the specimen left in air from the beginning 1371. This
indicates that oxygen is necessary to accelerate staining.
c. Physiccrl Fcrctors I~lfluerrcit~g theOccurrence of Irotl Stnirl
Time. Because iron staining is a chemical reaction, it is influenced by temperature
and time. The timenecessary for iron staining to occur is generally 3 min at ordinary
temperature and 1 min at high temperatures [37]. However, the time is affected by such
factors as the method of contact and the wood species.
Table 14 shows the time needed for iron staining to occur for two methods of contact
in 33 species widely used as decorative wood [32].On contact with iron solution, all the
tested woodsshowedstainswithin 45 S, and 40% of the tested woodshoweda stain
immediately. Iron powder takes two to three times longer to stain than does iron in so-
lution. Theorderoftimesneededfor staining in the tested wood species is the same
regardless of the method of contact. Woods that show strong stains seem to stain quickly.
The relationship between contact time with iron and quantity of stain is shown in Figs.
21 and 22. At the initial stage of contact, the quantity of the stain increases in proportion
to the contact time, but it does not change after a certain amount of time has passed [37].
Color and Discoloration 415
0.1% Ferricchloridesolution
v
2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0
1 solution
chloride
0.001% Ferrlc
-
.P
.c
0
40
v)
rd
S 0-
-lo.
2:O 3:O 4:O 5:O 6:O 710 810 910 10.0
PH
FIGURE 20 Relationship between pH of ferric chloride solutions and quantity of stain of some
woods immersed in the solution for 2 days.
Temperature. The time for staining to occur after wood comes into contact with
iron diminishes as the temperature increases. When the temperature increases, the times
needed for staining with iron powder that are used in Table 14 are shown in Table 15
[32]. At temperatures between 85 and 95"C, all tested woods quickly showed stain. When
an iron solution was used, the degree of staining increased in proportion to temperature
(Fig. 23). The degree of staining showed the same properties when wet veneer was put
into contact with an iron plate of various temperatures, as shown in Fig. 24 [37].
Light. Because iron stains also occur in dark places, light probably does not par-
ticipate in the stain reaction [37].
d. Removal of Iron Stain. There are two ways to remove iron stains: decoloring by
means of a chemical reaction and sanding with a planer or sander. Removal by sanding
is effective when the stain is limited to the surface and is small in size. However, iron
stains in woodworking occur mainly in wood with high moisture content and, therefore,
are often large and deep. Such stains must be removed by chemicals.
Removal with Chemicals. It is well known that a coating of oxalic acid solution
is effective in removing iron stains. Oxalic acid has a higher decoloring ability than sulfuric
acid or hydrochloric acid [32]. In decoloring iron stains in mizunara, phosphorous acid,
hypophosphorous acid, and phosphoric acid are also effective, as shown in Fig. 25 [38].
The order of decoloring ability of these chemicals is oxalic acid > hypophosphorous acid
TABLE 13 Relationship Between Moisture Content of Wood and Occurrence of Iron Stain
-
After iron powder was scattered on the test specimens, they were conditioned under various humidities for 20 days.
f : Iron stain occurred.
-: Iron stain did not occur. I
0
3
m
S
P
E
3
8
3
s
Color 7
TABLE 14 Time Required for the Iron Stain to Occur in Woods Treated with Iron
Test specimens for scattering iron powder were previously Immersed in water for S min.
FIGURE 21 Relationship between immersion time in 0.01 % ferric chloride solution and quantity
of stain of some woods. (For key, see Fig. 20.)
418 Hon and Minemura
0 5 20 LO 60 120
(min)
Contact time
FIGURE 22 Relationship between contact time of some wet woods onto iron plates and quantity
of stain. (For key, see Fig. 20.)
> phosphorous acid > pyrophosphoric acid = orthophosphoric acid. Chelate chemicals such
as a disodium salt of ethylenediaminetetraacetic acid (EDTA-2NA) can be used on light
stains. This chemical has a lower decoloring ability than acid, but adding acid to this
chemical increases its decoloring ability [32].
When woods decolored with the acid described in Fig. 25 are left under sunlight,
only the wood decolored with oxalic acid becomes dark again, as shown in Fig. 26 [38].
Ultraviolet light has a significant influence on this restaining, but visible light also causes
black staining, as shown in Fig. 27. Irradiation under nitrogen atmosphere yields only a
small amount of restaining, as shown in Fig. 28. It is clear that oxygen has a great influence
on restaining [38].
Controlling the Restaining of Woods Treated with Oxalic Acid. A small amount
of oxalic acid can decolorize iron stains and given them almost the same color as the
sound wood. However, the wood has a tendency to restain. Washing the treated wood with
fresh water controls the restaining. Because oxalic acid is an acidic chemical, incomplete
washing leaves a red color on the surface involved. After washing, drying is required.
When treated wood is washed with water, the oxalic acid occasionally elutes more quickly
°C Species
Douglas fir 30-35 3-8 0-3
Sawagurumi 10-15 0-3 0-3
Buna 0-3 15-20 0-3
Japanese red birch 25-30 3-8 0-3
Zelkova 40-45 3-8 0-3
Painted maple 0-3 10-15 0-3
Iron powder was scattered on the test specimen.
Color and Discoloration 419
than the iron ions and the surface becomes black again. If a stain must be removed from
the surface of fabricated furniture or thick lumber, washing cannot be used; therefore,
coating with a chemical that has the ability to prevent restaining and red discoloration is
needed. Chemicals that have this ability are dihydrogen phosphate, hypophosphite, hydro-
gen phosphite, and hexametaphosphate [38].
With respect to price, operator safety, degree of discoloration, etc., sodium dihydro-
gen phosphate (NaH2P0,) is the best chemical to use. This chemical is used in an aqueous
420 Hon and Minemura
20
Phosphorous acid
Hypophosphorous acid
8
2 % Unstained
5 5 P S o d i u m fluoride
"@Benzenesulfonic acid
50 L
" 0 1 2 3 4 5 6 7
a
solution. Thick coating causes crystallization on the surface after drying; therefore, the
coat weight of the chemical must be below 10 &m'. Wood decolorized with oxalic acid
together with NaH'PO, displays the same behavior of photo-induced discoloration as does
sound wood, as shown in Fig. 29. The chemical solution can be applied with a spreader,
sponge, or brush, but a brush with iron wire should not be used. An iron-free vessel made
with plastic or glass must be used to dissolve the chemical.
Painting Performance of Wood Decolorized with Oxalic Acid and
NaH,PO,. Both oxalic acid and NaH'PO, remain on the wood surface after decoloring.
Becausedecolorized wood is usually used in furnitureorinteriorwood,thesurfaceis
often painted. We might think that the existence of a chemical layer between the wood
surface and the paint could cause such problems as weakening of the bonding strength of
the paint film or discoloration of the paint. However, painting with a commercial poly-
urethane indicates no such problems, as shown in Fig. 30 and Table 16 [38].
Mechanism of Occurrence and Prevention of Iron Stain. Water, iron, and phe-
nolic substances are necessary for staining to occur. When iron dissolves in water, iron
ions are produced. Since wood is an acidic substance, it accelerates ionization. Iron ions
react with the hydroxylgroups in phenolicsubstances in wood and formdeepblack
substances (iron stain). This reaction is accelerated by oxygen.
0 1 2 3 4 5
Coat weight (g/ma)
""VI h
0 10 25 50 100
(hr)
Exposure time to carbonarclight
When oxalic acid is applied to an iron stain, it removes iron fromthe stain and
makes ferrous oxalate. This product is superior to the bonding strength between the phe-
nolic hydroxyl group and iron ions. Ferrous oxalate is pale yellow, but this color does not
show in wood when it is present in very small quantities. It is thought that ferrous oxalate
has low photostability and decomposes easily when it absorbs UV light. If a phenolic
substance is present, the regenerated iron ion can react with it and form a black substance.
However, if phosphoric ions are present, the iron ions might react with them before the
phenolic substances and form ferrous phosphate. Ferrous phosphate is very stable under
light. These results are summarized as follows:
Iron - water
iron ion >-
phenolic
substances
phenolic
and
wood
black
complex
substances
of iron ion
-
(PI)
NaH,PO, ferrous
phosphate
oxalic acid ferrous
oxalate
PI "----+ substances
phenolic
phenolic
substance hv
oxalic acid
NaH,PO,is a weak acidic substance that keeps the woodsurface in weak acidic
condition after treatment with oxalic acid. It is possible to use phosphoric acid instead of
oxalic acid. In such a case, however, the surface acidity is high. A weak acidic substance
such as NaH,PO, must be used simultaneously to reduce the surface acidity. For some
woodswith light stains, it is possibletoremove iron stains with NaH,PO, alone.The
bonding strengths of iron ions and phenolic substances probably differ according to the
kinds of phenolic substances involved.
422 Hon and Minernura
Color difference AE
0 10 20
Inthe air
'/////////////////////////=I
In N, gas
b
FIGURE 28 Quantity of photo-induced discoloration of iron-stained mizunara decolorized with
oxalic acid when exposed to sunlight with an intensity of 400 mW min/cm2 in the UV region.
20 -
b 15-
-
e" -e
e--.-.
Unstained
Oxalicacld
Oxalicacld l-
Sodiumdihydro-
genphosphate
(Numerals In thefigure
10 - 4 mean exposure time.)
L loo l
5 10 15
a
FIGURE 29 Photo-induced discoloration of iron-stained zelkova decolorized with chemicals when
exposed to carbon arc light for 100 h.
Color and Discoloration 423
0 Unstained
A Oxalic
acid
0 Oxalic acid + Sodium
dhydrogen-
phosphate
0 25 50 75 100
(hr)
Exposure time to carbon arc iight
FIGURE 30 Quantity of photo-induceddiscoloration of iron-stainedmizunarathatwas decolor-
ized with chemicals, painted with polyurethane, and exposed for 100 h.
such as NaH,PO, should be added. If EDTA-2Na is used, it should be added in the amount
of 0.5 g per gram of iron ion, as shown in Fig. 31 [32]. If water is dark brown in color
due to the contamination of iron ions, the addition of aluminum potassium sulfate or a
high-molecular-weight coagulant can remove the iron ions.
If there is a possibility that the knife can break when a rotary lathe is used, the sliced
veneer should be immersed in an EDTA-2Na solution. If condensed water drops from the
surface of cold machinesin winter, the machines should be heated before use or the surface
should be coated with paint. In the manufacture of laminated veneer lumber with a side-
driving system, gears made of stainless steel should be used.
If veneer touches metal fittings fastened to the carrying belt when green veneer is
carried to the dryer, the fittings should be coated or covered with vinyl tape. When a dryer
is used, the wire netting on which the green veneer is placed should be made with stainless
steel and the temperature of the hot wind should be kept above 140°C. High-temperature
air brings on rapid vaporization of water from the veneer and does not produce a stain.
High temperature also helps control deterioration of the exhaust pipe, because the mois-
ture-containing acid component of the wood does not condense on the pipe.
Prevention During Plywood Manufacturing. Because adhesives used in plywood
production are water-soluble, it is desirable to useenameledironware or vessels made
with stainless steel or plastic in preparing the glue. If a press is used in the gluing process,
analuminum plate or duralumin plate withgoodheatconductivityshould be inserted
between the plywood and hot plate to prevent staining.
Prevention During Working of Lumber. To prevent stains where the surface of
green lumber touches steel belts during packing, a piece of wood or cardboard should be
inserted between the belt and the lumber. When packing a small amount of lumber, plastic
tape should be used. If an iron vessel is used for mold-proofing wood with water-soluble
chemicals, a vessel made from a thin plate of polyvinylchloride or stainless steel should
be placed inside the iron vessel. A wooden vessel covered with thick polyethylene sheets
can be used instead of an iron vessel. During the process of laminating wood, both the
vessel used to prepare glue and the instrument for coating should be iron-free
VI
20
x EDTA.4Na
Q,
No added
0
Q,
0
C X
ul
fd
5-10 ' ' ' L
8 -0.01 0.3 0.5 1.0 1.5 0.01 0.3 0.5 1.0 1.5
Concentration ("'0) Concentration ( %)
FIGURE 31 Relationship between concentration of EDTA solution and quantity of stain. Wood
specimens were immersed in 1000 cm' of 0.01% ferric chloride solutionwith 10 cm7 of EDTA
solution of various concentrations.
Color and 425
3. Discoloration by Acid
Acid stain is caused by acid interacting with wood. Acidic substances are not used very
often in woodworking processes. Therefore, acid stains do not occur very frequently.
a.Occurrence of AcidStuin in WoodworkingProcesses. Aminoalkyd resin is
widely used as an abrasive-resistant and inexpensive paint for wood coating. This paint is
mixed with two liquors immediately before use. One of the liquors is a hardener such as
paratoluenesulfonic acid. Wood paintedwithexcesshardenerand left in sunlight often
turns red. Zelkova is usually used as a thin veneer because it is expensive but has a fine
grain. It is stained easily with iron, and the surface of the flitch becomes black. Therefore,
the flitch is often dipped into a solution of oxalic acid prior to slicing. When sliced veneer
is glued to the base wood, the edge that contacts the solution of oxalic acid often turns
red. Urea formaldehyde resin is moderately waterproof and is used widely as an inexpen-
sive wood adhesive.Ammoniumchloride is added to it immediatelybefore use. This
chemical reacts slowly with the formaldehyde in the resin and produces hydrochloric acid.
On rare occasions, a plywood surface with this resin on it may turn red.
b. Factors Affecting the Occurrence of Acid Stain
pH. The acids involved in acid staining during woodworking are hydrochloric acid
and oxalic acid. The extent of stains that occur when several woods are immersed into
acid solutions with various pH values and then left indoors is shown in Fig. 32 [39]. In
the pH range from 5 to 2, all wood species show weak stains that are not recognizable to
the naked eye. However, at pH below 1.5, strong red or reddish-purple stains occur.
Light. Figure 33 shows the change in color of four kinds of wood after immersion
into a solution of oxalic acid with a pHof I , whichwere left to dry in a dark place,
indoors or under the light of a mercury lamp [39]. Under the mercury lamp, a maximum
color change was obtained after 5 min of exposure. With indoor exposure, 5 days were
required to obtain the same extent of staining. When stored in a dark place, a slight change
of color occurred. However, when the immersed specimens were exposed to indoor light,
the colorchanged rapidly to reach the sameextent of staining as the indoor-exposed
specimen. From this, it is evident that discoloration caused by acid stain is accelerated by
UV light.
Oxygen. When wood is immersed in a solution with a pH of 1 and then irradiated
with a mercury lamp under nitrogen atmosphere, it shows the same extent of staining as
wood irradiated in air [39]. This clearly indicates that oxygen does not participate in the
stain’s development.
Wood Extractives. Table 17 shows the relationship betweentannincontent and
stain [39]. Softwood shows high staining, which might be due to the existence of con-
densed tannin. Catechol tannin causes acid stains. Because bubinga and koa contain leu-
coanthocianin, which easily turns red with acid, this substance might be the cause of the
stains in both woods. When woods that show a significant stain are treated with acid after
426 Hon and Minemura
15[ o\ 0 ;Sugi
0 ;Akamatsu
A ; Manggasinoro
hot-water extraction, they do not discolor. This indicates that the source of the acid stain
is not lignin but a phenolic extractive.
c. Removal of Acid Stain. If the stain is limited to the very top surface, it can be
removed by planing or by sanding with sandpaper. For deeper stains, destruction with a
bleaching agent or neutralization with alkali is effective to a certain degree. The use of
sodium chlorite is desirable because it acts under acid conditions. Other bleaching agents
such as hydrogen peroxide or sodium hypochlorite can also be used. Sodium bicarbonate
or calcium carbonate can be used for neutralization.
d. Prevention of Acid Stain. If paints or adhesives that harden with acid are used,
the amount of hardener added should be kept to a minimum. To produce adhesion with
Indirectsunlight
Mercury
lamp
(thinline:darkplace)
lot
-
. ----x-------
0 10 20 30 0 5 10 15
Exposure
time (day) (hr) time
Exposure
FIGURE 33 Relationship between exposure time and color difference after soaking in oxalic acid
solution at pH 1 . 0 : Sugi; 0 : Akamatsu; A: Manggasinoro; X: Buna.
Color and Discoloration 427
heat in plywood production, penetration of acid into the veneer should be prevented by
raising the viscosity of the glue, diminishing the coating amount of the glue, lowering the
moisture content of the veneer, and lowering the pressure and temperature. When oxalic
acid is used to remove iron stains, sufficient washingwithwater or the addition of
NaH,P04 is required.
4. Discoloration by Alkali
Alkali stain is the discoloration caused by the reaction of alkali chemicals with wood. This
stain is observedmoreoftenduring the useofwoodproductsthan in woodworking
processes.
a.Examples of Occurrence of theStain. Freshconcrete is strongly alkali. When
wood contacts it in the presence of water, an alkali stain often occurs. Flooring is often
bonded on concrete. If water overflows on the floor and reaches the concrete layer, the
water in concrete becomes alkaline and penetrates into the flooring to form a brownish
alkali stain. When an excess amount of alkaline water penetrates, even the surface of the
flooring becomes discolored.
A plate made with calcium silicate is used often as a flame-retardant board. On the
plate, a decorative veneer with fine grain is often laminated and used in the interior field.
The base plate is inorganic and alkali. When glue is coated on the base plate, the alkaline
substances in the plate dissolve and react with the veneer to form a brownish stain.
b. Factors Affecting the Occurrence of the Stain
pH. Figure 34 shows the extentofalkalinestains in fourkinds of woodwhen
immersed in solutions with various pH levels of calcium oxide or sodium hydroxide [40].
428 Hon and Minemura
$ 25
30
0
t 0 ,Sawagurumi X
0; Buna /
I
X
(L,
l
"20
C
Q,
Z
.- 15
U
-0b 10
V
5
FIGURE 34 Relationship between pH of an alkaline solution and quantity of stain when immersed
in the solution for 5 min and then dried in air in a dark place.
Under pH 11.4, little staining occurs, but beyond this pH rapid discoloration is observed.
The color differs according to pH. For example, the color of sugi is reddish-brown up to
pH 12.5, but beyond this pH it becomes bluish.
Light. The effect of light on stains is shown in Fig. 35 [41]. Woods immersed in
a solution of pH 12 (in Fig. 34) were left in a dark place, indoors, and under a mercury
lamp. The wood left in a dark place retained its original color at immersion, whereas the
wood left indoors faded in color and the wood under a mercury lamp showed stronger
fading. These results show that light is not required for staining to occur. The opposite is
true with acid stains.
Oxygen. Whenwoodimmersed in an alkali solution isleftin a nitrogenatmo-
sphere, the extent of discoloration is less than in air. When wood in a nitrogen atmosphere
is taken out and left in the air, it discolors to the same extent as wood left in the air from
2ot
d" e"--,
Indirectsunlight
(thin
line : dark
place)
Mercurylamp
-0
the beginning. This means that the production of an alkali stain requires oxygen and the
colored substance is formed by oxidative polymerization [40].
Wood Extractives. Table 18 shows the relationship betweentannincontentand
amount of staining. In the 16 wood species tested, woods with high tannin content had
more of a tendency to stain [41]. It has been found that if sufficient extraction with hot
water is completedbefore the wood is immersed in analkalinesolution, the stain is
scarcely recognized. Only a small yellowish-ocher stain occurs above pH 13. From these
results, it can be surmised that the alkali stain is due mostly to the water-soluble phenolic
components. Lignin also participates in discoloration under alkali conditions.
c. Removal ofAlkali Stain. Most alkali stains occurring in a short period of time
can be removed with bleaching agents. The stain on the surface of concrete blocks with
alkalinecementdescribedearliercanbe easily removedwithhypochloritesolution, as
shown in Table 19 [42]. Concreteblocks are highly alkali-resistant, but notvery acid-
resistant. Because the surface is rough and has a lot of voids, it is very difficult if not
impossible to dissolve the colored substances of the stain with an alkaline solution. Hy-
pochlorite is an alkaline bleaching agent;its solution can be used to decompose the colored
substances without damaging the block.
Stains on the top layers of wood surfaces can be removed by planing or sanding.
Stains at a certain depth that cannotbecompletelyremoved by these methods can be
bleached with a bleaching agent or coating with dilute acid.
d. Prevention qf AlkaliStain. Plywood that does not discolorwith the alkali of
cement or plywood coated with alkali-resistant paint can be used to frame concrete for
hardening.
When a decorative veneer is laminated on the alkaline inorganic plates, the use of
an adhesive film is desirable. When a water-soluble adhesive is used, the following meth-
ods should be considered: increasing glue viscosity, diminishing water, lowering the mois-
ture content of veneer, decreasing pressing time and temperature, coating the plate with
an alkali-sealing paint, and so on.
5. Discoloration by Microorganisms
Approximately half of wood components are carbohydrates, and in green wood they con-
tain a moderate volume of water. When green wood is left under certain conditions, mi-
croorganisms propagate on the wood. This often is accompanied by the discoloration or
lowering of wood strength.
The microorganisms that cause discoloration are bacteria, mold, and basidiomycetes.
Mold discolors the surface of wood but does not diminish its strength. Basidiomycetes
cause a decline in strength. Among the basidiomycetes are brown-rot fungi and white-rot
fungi. The former mainly decompose cellulose and hemicellulose, whereas the latter also
decompose lignin. Bacteria may occur in stored wood when it is immersed in or sprayed
with water.
Discoloration is due to the pigments of the microorganism or coloring compounds
produced by the reaction of the woodcomponentswith the secretions of the micro-
organism.
a. ExcImples of Occurrence ?f Stains in Woodworking Processes or Wood Products.
When fresh green lumber is stacked on a warm day under high humidity, many colonies
of mold with various colors can grow on the surface of the wood overnight. When sliced
veneer is transported without drying, discoloration caused by fungi and bacteria can occur
on the surface. When logs are piled outdoors for a long period of time, brownish discol-
Minemura
430 and Hon
Color Tannin
Stain content
grade (%) ( m
oration may occur in the cross sections of both sapwood and heartwood. This stain is
caused by basidiomycetes. A blue stain is observed only in sapwood and does not bring
about a decline in strength.
Ilomba wood often discolors after felling to give a reddish-brown color. This wood
is normallyallsapwoodandcontains substrates which are suitablefor the growthof
bacterial [68].Some bacterials propagated on the lumber produce ammonia as a metabolite
and form colored substances by the reaction of the components with ammonia [43]. For
wood components which are responsible for the discoloration, (+)-catechin and (-)-epi-
0: Excellent.
A: Common.
X: Poor.
Color 431
catechin are confirmed [69]. In the brown-stained region of hemlock, dark-pigmented fungi
are predominant.Thesefungiinducebrown discoloration in the sapwood.Browning is
accompanied by an increase in pH from 5 to 7, a decrease in total soluble phenols, and
oxidation of phenols such as catechin. Intensive discoloration occurs at pH 7, and oxygen
is indispensable for the development of the discoloration [70-721. Brownish discoloration
in beech wood is caused by bacterials which produce ammonia to give a pH 7.3 [73].
Yellow discoloration of oak heartwood is caused by mold fungus. It is assumed that
metabolic compounds of the fungus react with hydrolyzable tannins and give yellow sub-
stances [74].
From the blue-stain fungi, the dark coloring pigments have been isolated. They are
classed with the group of melanins and are associated with carbohydrates and proteina-
ceous components [75].
Concerning pink stains of angiosperm and gymnosperm woods caused by fungi, a
red pigment has been isolated and identified as 5,8-dihydroxy-2,7-dimethoxy- 1,4-naphtha-
lenedione [76].
Some injurious insects penetrate tropical woods to the inside. Such insects often
carrymicroorganisms.Forexample, in aplacewhere Limnoria lives, there is awhite
corpse of Ambrosia beetle. When the insect moves in the tangential direction in wood, the
discoloration is dappled in the radial section and striped in the tangential section. Larvae
of the sugi bark borer feed on the wood of living sugi trees and induce discoloration [77].
In the discoloredsapwoodand in the reaction zone of the soundsapwood-discolored
sapwood boundary, potassium and magnesium begin to accumulate within one year, and
calcium within two years. Discolored sapwood has a greater cation-exchange capacity.
b. FactorsAffectingtheOccurrence cf Stains. Thefollowinggrowth factors are
essential for the propagation of microorganisms: water, air (oxygen), moderate warmth,
and nutrients. Wood itself is a nutrient. Generalgrowingconditionsare3-40°C, 90%
relative humidity, and 20- 150% wood moisture content.
c.Removal of Stains. Stainscaused by moldcan beremoved by planingor by
coating with a bleaching agent. Because stains causedby basidiomycetes often occur deep
in wood, they cannot be removed completely. In such cases, it is effective to immerse the
wood in a bleaching agent. For example, the brownish stain on a shina log that is caused
by invasion of basidiomycetes can be removed by immersing the log in a dilute solution
of sodium hypochlorite for several hours [44].
For removing fungal stain of ponderosa pine sapwood, 2% hydrogen peroxide so-
lution with sodium hydroxide and sodium silicate as a buffer give a good result [78].
Comparedwithchemical fungicides, biological control is generallybenign to the
environment.Concerning the biological control of sapstain fungi,metabolitesobtained
from two fungi were examined on stained pine veneer disks and it was found that they
remove sapstain and kill existing fungal growth [79].
d. PreventingStains
Addition of GrowthInhibitors. To preventstains, it is effective to adhere pre-
servatives or antimold agents to wood. This can be done by coating, spraying, immersing,
and pressure impregnation. The chemical should be selected on the basis of low toxicity
and slight color. Organic compounds containing tin or iodine are soluble in organic sol-
vents, and solutions of these materials have good permeability to wood. For prevention of
mold growth during drying, pretreatment of green lumber with propionic acid is recom-
mended [74]. As a preventing chemical for brown stain i n hemi fir, a quaternary ammo-
nium compound, didecyldimethylammonium chloride, is effective [SO]. The preservative
Minemura
432 and Hon
6. Discoloration by Enzymes
Variousenzymes in wood participate in manymetabolismsystems.Someenzymes are
active even after logging. In sawing or veneering, when fresh green wood makes contact
with oxygen, these enzymes often take part in discoloring the surface of the wood.
U . Exurnples of Stains Occurring in Woodworking Processes. Alder generates red-
dish-orange discoloration immediately after felling. This discoloration is caused by inter-
action of catechol oxidase and hirsutoside, which is a xyloside of diarylheptanoid con-
taining two catecholic nuclei [81]. Shina is widely used for plywood production. When
veneer sliced with a rotary slicer is left without drying for several hours, the surface often
becomes yellow. When sliced walnut veneer is allowed to stand in the same manner, it
becomes black. The fresh green lumber of todomatsu become yellow.
Kiri wood is widely used for furniture in Japan. This wood changes color to dark
brown when sawn immediately after felling. This discoloration might be caused by cata-
lytic oxidation with oxidase. Caffeic acid sugar esters have beenisolated as the compounds
responsible for discoloration [82,83]. Peroxidase consists of heat-labile and heat-resistant
enzymes. The activity of the latter enzyme occupies about 12% of the total [84].
Concerning brown stain in sapwood of Douglas fir, its enzymatic extract showed
two pH optima for activity (pH 5.5 and 8.0) and highest activity at 35°C. It showed also
highest activity for (-)-epicatechin and dihydroquercetin [85].
When beech wood chip is stored outdoors, it changes color to deep brown after a
few days. As beechwoodcontains sufficient amounts of activeperoxidaseandmalate
dehydrogenase,phenoxy radicals are formed in the lignin andsubsequently new chor-
mophores are formed [86].
b. FuctorsAffectingDiscolorution by Enzymes. Moistureandhumidity signifi-
cantly affect discoloration. For discoloration to occur, the surrounding humidity must be
about 100%. Temperaturealsoinfluences discoloration, with discoloration occurring
slowly below 20°C. Phenolic substances in wood might oxidate to colored substances by
means of enzymes in the wood and oxygen in the air.
c. Stuin Removul. Yellow stains can often be removed by bleaching with hydrogen
peroxideor extraction with hot water. Because the sapwood of todomatsu is usedfor
chopsticks, extraction with hot water is recommended.
Color 433
Hydrochloric acid A e e
Sulfuric acid 0 0
Phosphoric acid A 0 a
Hypophosphorous acid A e
Nitric acid A 0 e
Boric acid X X A
Formic acid X X A
Acetic acid X X X
Oxalic acid A 0 e
Ascorbic acid X A 0
Benzensulfonic acid X 0
Semicarbazide hydrochlorlde X A A
Sodium bisulfite 0 0 0
Sodium sulfite A A 0
Sodium hypophosphite X X X
Sodium nitrite X X X
Fornlaldehyde X X 0
Urea X X X
Thiourea X X 0
EDTA. disodium salt A 0 0
434 Hon and Minemura
Drying condition
ying Wet-bulb Dry-bulb Color
temperature
temperature ti me
("C) ("C) (h) L (1 h
of colored substances from a phenolic compound oxidized with air and the formation of
dark materials from hydrolysis of hemicellulose have been considered the causes of dis-
coloration. When the material causing discoloration is water-soluble, these materials rise
to the surface and accumulate there, discoloring the wood [49]. It is assumed that many
substances that change color with heat also discolor with enzyme action.
Brown discoloration oftenoccurs in Europeanoakwoodduring kiln drying.This
discoloration occurs at a wood moisture content between 30% and 60%, kiln temperature
above 25”C, and relative humidity of about 70% [89,90]. In brown stain of oak, colored
polyphenolic polymers and complex esters, hexahydroxydiphenolesters, are found in larger
amounts than in nonstained wood [91]. This discoloration is presumably due to oxidative
coupling of compounds related to gallic acid. Discoloration during kiln drying may be the
result of hydrolysis and oxidative transformation of ellagitannins 1921.
Concerning sticker stain in sugar maple, chemical analysis of stained sapwood has
been conducted [93]. The amount of acetone-water-soluble materials in the stained part
was less than in the clear part. This suggests that phenolic extractives accumulated under
the sticker stain during drying and then oxidized to insoluble polyphenolic compounds.
Scopoletin was isolated from both the stained and unstained parts as the phenolic com-
pound produced during drying.
As anatomical characteristics of brown stain after kiln drying in hemlock, the stain
exists in sapwood, particularly in the earlywood, and is recognized mainly in longitudinal
tracheids [94]. From a study of the correlation between loss of brightness in mechanical
pulp and storage time of western hemlock chip, it is suggested that d-catechin polymerizes
oxidatively to give an insoluble polymer and to cause the brightness loss [95].
The heartwoodcolor of sugi is classified into three types:normalreddish-brown
color type, black color type, and color-changeable type. The latter occurs when the wood
is left at room temperature; the surface color changes from reddish brown to black in 30
min after sawing. This phenomenon is observed in wood grown on tree farms. This type
of stain cannot be controlled by oxalic acid; therefore, it is not an iron stain. The char-
acteristics of this phenomenon are as follows [96,97]: The stain occurs either in the dark
or under light. Atmospheric oxygen is necessary. In nitrogen, discoloration does not occur.
When the color-changeable sugi is extracted with water preliminarily, it gives no more
black color. The water extract contains a potassium hydrogen carbonate(0.4% w/w), which
keeps the wood weakly alkaline. When normal reddish-brown sugi wood is immersed into
awater solution of KHC03, it changescolor to black. So, this inorganicchemical is
recognized as one of the causative materials of the black discoloration. The characteristics
of black sugi which is alreadyblack when standing in the forest, before logging, have
been examined. The black sugi has high moisture content and alkalinity, and contains K’,
Na’ , and HCO, [%l. The black substances are presumably a polymer of water-soluble
norlignans such as segurin-C and plicatinaphthol [99].
Heartwood of murasakitagayasan changes rapidly i n color from brownish yellow to
dark purple after sawing. For this discoloration, oxygen is indispensable. Light and water
acceleratc the reaction [ 1001. As the substancewhich contributes to the discoloration,
7,3’,4’-triacctoxy-6’-n~cthoxyisoHav-3-enehas been isolated [ 1 0 1 1. This discoloration
might be caused by the autooxidation of this compound to give the quinonoid structure.
h. Rrr~~o\w/ of ~ i . s c ~ o / o r t r r i o rWhether
~. sawed lumber has been artitically seasoned
or not. its surface is rough. When the lumber is used as an interior wood o r for furniture,
its surfacc is usually planed. B ~ C ~ Lheat I S Cdiscoloration is mostly limited to the surface,
a sound surface should reappcar after planing to ;l thickness of 2 mm.
436 Hon and Minemura
W
Q
Color and Discoloration 437
L n b
REFERENCES
Minemura
440 and Hon
Yuan-Zong Lai
SUNY College of Environmental Science and Forestry, Syracuse, New York
1. INTRODUCTION
Wood and other lignocellulosic materials are labile to a wide variety of chemical changes.
These transformations, depending on the conditions of reaction environment, may vary
from an undesirable discoloration (Chapter 9) to a selective breakdown of the major cell
wallcomponents [ 1-31. In the chemical utilization of these lignocellulosic substrates,
lignin usually plays a negative role, and must be modified, partially degraded, or totally
removed, depending on the end uses of the final products. The commercial pulping and
bleaching operations 141 generally are very nonselective, being accompanied by a signif-
icant degradation of the polysaccharide components. For example, the yield of lignin-free
softwood pulp for the most widely used kraft process is only about 44%, as compared to
a theoretical 67% for pine [ 5 ] .Thus, a great technical challenge for the paper industry is
how to improve the delignification selectivity or carbohydrate stabilization.
The fundamental chemistry of the degradation of isolated polysaccharide and lignin
samples as well as related model compounds is now reasonably well understood [l-31.
The detailed kinetics of reactions involving wood components in situ, however, are still
not fully clarified, and are complicated by their heterogeneous nature across the cell wall
[6,7], the possible role of lignin-carbohydrate complex (LCC) or linkages [ 1 -3,8.9], and
the pore structure of the cell wall matrix. Theoretically, accessibility is a significant factor
affecting the degradation behavior of wood polymers in situ, and its significance varies
with the nature of chemical environments.
This revised chapter largely retains the original format [21 for easy reference, and
discusses the chemistry and controlling factors in the degradation of cellulose, hemicel-
lulose, and lignin under acidic, alkaline, and oxidative conditions.
11. REACTIVESITES
A. Polysaccharides
The major functional units of wood polysaccharides are reducing end groups, glycosidic
linkages, and hydroxyl groups. The reactivity of these units, however. varies considerably
among the cellulosc and hemicellulosecomponents,contributing largely to their differ-
ences in supramolecular and chemical structures.
443
444 Lai
1. ReducingEndgroup
Allnatut-al polysaccharide molecules contain a reducing end group which, being hemi-
acetal i n nature, is partially converted to an open-chain aldehyde function i n solution. This
functional group can he reduced and oxidized to a n alditol and aldonic acid moiety, re-
spectively. Also, the anomeric hydroxyl group (at the C1 position). being the nlost acidic
[ I O ] . can be selectively etherified [ I I].
Reduction with sodium borohydride is often used for quantitative estimation of the
reducing end-group contenl [13_]. The reported contents of g l ~ ~ c o smannose.
c, and xylose
end groups in wood 1 1 31 are gcnerally consistcnt with the molecular tnasscs established
for cellulose. glucomannan, and xylan.
Regarding the accessibility of reducing end groups. it was reported by Gentile et a l .
[ 141 for a libroushydrocellulosesample based on the assumption that an amorphous
cellulose was totally accessible. The latter sample was prepared by regenerating ;I cellulose
solution i n a dimethyl sulfoxide (DMSO)-parafortnaldehyde (PF) solventsystem [ 121.
Rcducing end groups were clctertnined by reduction with tritiated sodium borohydride i n
dilute alkalis. Approximately 12%. of the reducing end groups i n the fibrous cellulose were
shown to be inaccessible t o the borohydride treatment. Interestingly, a large difference i n
hydroxyl accessibility between ;I native and ;I regenerated cellulosesample ( S 1 versus
99%) was previously indicated by the deuteration method [ IS]. Since the penetration of
reagent into thc crystallites would be negligible under the mild conditions used (an 0.25
M borohydride solution a t ambient temperature). it appears that the concentration of re-
ducing end gt-oups is significantly higher i n the amorphouscomponent than in thc
crystallites.
Reducing end groups i n alkalis undergo readily a series of the so-called Lobt-y de
Bruyn-Alberda van Ekenstein transfonnations [ 1 S ) . and play a dominant rolc i n the a l -
kaline degradation of polysaccharides.
2. Glycosidic Linkages
The glycosidic bonds. being acetal i n nature, are hydrolyzable under acidic, alkaline. and
oxidative conditions. Acid hydrolysis proceeds very readily and forms the basis of :I sac-
charification proccss, whereas the alkaline cleavage reaction requires more drastic condi-
tions. This hydrolytic reaction i n general weakens the mechmical properties of wood and
fibers.
3. Hydroxyl Groups
The intcrunits of ccllulosc and hemicelluloses contain one primary hydroxyl group for
each :mydro-hcxose u n i t and two secondaryhydroxylsfor each anhydro-hcxosc and
-pentose u n i t . Thew hydroxyl g r o ~ p saresusceptible to oxidation, and the resulting a l -
dehyde or keto g r o ~ ~may
p initiate further degradation reactions. such ;IS dehydration and
cleavage o f glycosidic linkages. Among the three hydroxyl groups, the ?-OH group is the
most acidic 1 1 1,16- 191. and this has been gencrolly attributed t o an activating effect o f
the nnomeric center. Hearne et a l . 1201 observed t h a t methyl P-D-ribopylanoside was more
acidic t h m methyl P-D-xylopyranoside, and they differ only i n the C3 conformation. Thus.
the acidity of the ?-OH group is also likely influenced by other factors such ;IS the hy-
drogen bonding system.
The hydroxyl reactivities in a heterogeneous system are further affected by the X-
cessibility factor. and thc reported data on cellulose have been shown t o vary considerably
with the type and the conditions of reactions used. Inaccessibility may arise from either a
Chemical Degradation 445
B. Lignin
Lignin occurring in plants is well known for its variability or heterogeneity in terms of
both morphological distribution and chemical characteristics. Significant variations have
been observed between juvenile and mature woodlignins; among the normal, compression,
and tension wood lignins; and for lignins in different morphological regions. Our present
understanding of lignin structure has been obtained largely from analysis of milled wood
lignin (MWL) preparations [9], which are usually obtained at less than 50% yield. The
millingprocessused in MWLpreparation is knowntoinducesomechemicalchanges,
notably an increase in the phenolic hydroxyl group content [21,22]. Additionally, MWL
has been shown to originate mainly from the secondary wall lignin [8,23-251. Thus, the
extent to which MWL may represent the lignin in situ requires further evaluation.
Although the approximate contents of major lignin linkages are now reasonably well
understood, the chemical structure of lignin, unlike that of cellulose or the hemicelluloses,
cannot be defined precisely.
Sincecarbon-carbonlinkages are generallyvery resistant tochemical attack, the
degradation or fragmentation of lignin is limited largely to cleavages of ether units at the
a - and P-positions. The nature of these hydrolyzable units and other functional groups
having a significant impact on the reactivity of lignin is outlined below.
1. HydrolyzableEtherLinkages
The hydrolyzable ether units in lignin are the P-aryl, a-aryl, and a-alkyl ether linkages
(Fig. I ) . As summarized in Table I , the P-aryl ether based on phenyl propane (C,) units
constitutes approximately 50% and 60% of spruce and birch MWL, respectively [21], and
is present as two isomers. Proton and I3C NMR analysis indicates that spruce lignin con-
tains about equal proportions of the erythro and threo forms, whereas the erythro form
dominates in birch lignin [26,27].
Spruce MWL was reported to contain 6-9% of acid-labile (presumably noncyclic
a-aryl) ether units by a mildly acidic hydrolysis reaction [28], but an appreciably lower
value (<3%) by a 2-D NMR analysis [29]. The latter method gave a higher value ( 5 % )
for a birch MWL. Also, Lai and Guo [30] obtained a higher value for aspen (6%) than
for spruce (4%) wood lignin when measured in situ by a selective acid hydrolysis. On the
other hand, spruce MWL was shown to have a higher content of the cyclic a-aryl ether
(as in p-5 units) than birch MWL (9-12% versus 6%) [31]. Recently, Brunow et al. [32]
suggested the possible presence of dibenzodioxocin unit [(4) in Fig. l ] in lignin.
Additionally, lignin may contain some a-ethers linked to carbohydrate [ 1 -3,8,9,33].
Thesebenzyl alkyl ethers, like a-arylethers, are susceptible to acid hydrolysis, but at
considerably lower rates.
2. PhenolicHydroxylGroups
The phenolichydroxylgroup is one of the mostimportant functionalities affecting the
physical and chemical properties of lignin polymers. It plays a prominent role in com-
mercial delignification processes by virtue of its ability to promote alkali-catalyzed cleav-
ages of interunitary ether linkages and the oxidative degradation of lignin [ 1 -3,8,34,35]
as well as lignin modification reactions [36]. Reported content of this functional group in
446 Lai
OR
OCH3 Q- OR
OCH3
1 2
R,= aryl
OCH3
I H3C0
CH -0 0 0
CHzOH
CH3
OR OCH3
OH
3 4
Percentage of
intermonomeric
linkages
Type o f linkage Structure (Fig. Birch
I) Spruce
4. Uncondensed Units
The uncondensed units of lignin, in general, may be defined as those units with positions
at C2, C3, C5, and C6 being free orsubstituted only by a methoxyl group. Reported values
for softwood lignin varied slightly with the analytic method used: 50-60% by oxidation
with potassium nitrosodisulfonate (Fremy’s salt) 1401; 50-55% by ‘H NMR [41,42]; and
45-57% by nucleusexchange reactions [43].Thus,softwood lignin containsapproxi-
mately an equal proportion of uncondensed and condensed units.
Hardwood lignins containingsyringyl units, as expected,have a highcontent of
uncondensed units, and a value of 83%wasreported for sweetgum lignin by nucleus
exchange reactions 1431.
5. Unsaturated Groups
Lignin contains some unsaturated groups, mainlyas coniferyl alcohol and coniferaldehyde
end groups. Carbonyl groups may also occur as a-keto or nonconjugated units. The latter
type is probably associated mostly with a detached side chain such as a glyceraldehyde
group.
6. Ester Groups
Ligninfrom certain species may contain a significant amount of estergroups [60,61].
Brauns lignin preparation from aspen was shown to contain more than 7% of p-hydroxy-
benzoic ester groups [44]. Also, grass lignins contain significant amounts of p-coumaric
acid and ferulic acid moieties, which were reportedly esterified mainly at the y-hydroxyl
group (80%) with the remainder at the @-position [45-471.
Additionally, the wester functionwould also include the possible lignin linkages
connected to the 4-0-methyl-glucuronic acid unit of xylan molecule, as summarized by
Lai [2]. Obst estimated [481 that approximately IO-20% of lignin-carbohydrate linkages
in aspen lignin were labile to mild alkali treatments (0.1 M NaOH at 20°C for 90 min),
and these were assumed to be present as an a-ester type.
I
448 Lai
7. Methoxyl Groups
Methoxygroupcontent often serves as an indication of the approximateproportion of
different phenylpropane units in the lignin. Methoxyl groups are relatively resistant to both
acidic andalkaline hydrolysis. They are, however,hydrolyzablewithconcentratedhy-
driodic acid [491, hydrosulfide and methyl mercapitide ions [SO], or sulfite ions [SI].
8. Accessibility
Lignin appears to be amorphous, occurring in plant tissues and in isolated samples, and,
like celluloseandhemicelluloses,ithas a hightendency to formhydrogenbondings.
Michell [52]concluded, from infrared analysis of MWL samples and related lignin model
compounds, that all detectable hydroxyl groups in lignin were involved in hydrogen bond
formation. Both a- and phenolic hydroxyl groups appeared to be involved preferentially
in intramolecular hydrogen bonding. Similarly, crystal p-0-4 lignin model dimers were
shown to contain a variety of intra- and intermolecular hydrogen bonds 1531.
The morphology or fine structure of lignin, unlike that of wood polysaccharides, has
not been studied comprehensively [ 1-3,7,8].
C. Lignin-CarbohydrateComplexes
The concept of lignin-carbohydrate complexes (LCC) has been derived largely from the
difficulty encountered in separating the carbohydrate impurity from a lignin preparation,
or vice versa.
The accumulated evidence strongly supports the existence of chemical linkages be-
tween the lignin and polysaccharides components in the cell wall matrix [ l -3,7-9,21,33].
Three major possible types of lignin-polysaccharide linkages are the benzyl ester, benzyl
ether, and glycosidic bonds (Fig. 2 ) . The reactivity of these units varies considerably with
their chemical structures and the reaction environments. For example, in alkali media, the
ester type is hydrolyzed readily, whereas the benzyl ether ofan etherified unit is fairly
stable even under alkaline pulping conditions.
A. Polysaccharides
The acidic degradation of polysaccharides is affected by their physicalstructures, the
conformation of sugar constituents, and the nature of an acidic medium.
1. Fundamental Aspects
(1. Acid Hydrolysis. Figure 3 illustrates one of the generally accepted mechanisms
for the acid hydrolysis of glycosidic linkages [2,66-691. It involves an initial protonation
of the glycosidic oxygen followed by decomposition of the resulting conjugate acid (9).
The rate-controlling step is likely in the formation of the carbonium-oxonium ion (lo),
which may exist in a half-chain conformation (11).
Table 2 summarizes the relative hydrolysis rates [70] and activation energy [68,71]
for a series of methyl pyranosides pertinent to wood polysaccharides. In general, the p-
anomers react faster than the corresponding a-forms, with an exception of the L-arabi-
Chemical Degradation 451
H0
CH20H
OH
Ho&F
CH20H
=
H0
8 11
-.+l l+.+ l1
CH2OH CH20H
H0Q&?!!+ slow H@
0 @ - I s H20 + products
Glucose
other
OH
9 10
noside. Also, the glycosidic linkages of nonglucose units are generally more reactive than
the glucoside. For the p-series, the relative hydrolysis rate increases in the order glucoside
( l ) , mannoside ( 3 ) , galactoside and xyloside (4.8), and rhamnoside (IO).
It is evident from Table 2 that the presence of a carboxylic group at the C5 of the
methyl glucoside (glucuronoside) reduced the hydrolysis rate by about 50-70%. A larger
reduction (97%) wasobservedbetween the twoglycosidiclinkages of cellobioseand
cellobiouronic acid [72]. Thereduced reactivity of these glucuronosides had beenattributed
to both inductive 173-761 and conformational [71,78-80] effects. It is of interest that the
presence of a C6 aldehyde group increased the rate by a factor of 20 to 70 [81]. Also, the
two glucosidic linkages of cellotriose behaved differently [82]. The linkage at the nonre-
ducing end was about 50% more reactive than the one at the reducing end, and it was
comparable to that of cellobiose.
Additionally, the reactivity of glycosidiclinkages is profoundlyinfluenced by the
ring size of sugar units. The aldofuranosides, because of their more strained structures,
TABLE 2 Relative Rate and Activation Energy for Acid Hydrolysis of Methyl Pyranosides with
0.5 M Acid at 60-90°C
Relative rate'' AEh (kcalhol)
~~
are hydrolyzed much faster than the corresponding pyranosides [83], by a factor of 100
between the methyl a-D-galactof~lranosideand methyl a-D-galactopyranoside [ 10l ] . The
energy of activation for the hydrolysis reaction is very similar among the neutral methyl
pyranosides (32-35 kcal/mol), and is lower for the glucuronosides (Table 2).
6 . Dehydrution m c l Condensation Reuctiorls. Although the immediatehydrolysis
products of polysaccharides,mono-andoligosaccharides, are reasonablystableunder
mildly acidic conditions, they can undergo further dehydration, fragmentation, and con-
densation reactions to yield a variety of nonsugar products including furan [841 and phe-
nolic [85-871 compounds.
Furan Compounds. A good yield of 2-furaldehyde (14) (88% of the theoretical)
is generally obtained by distillation of a xylose solution (12) in concentrated hydrochloric
acid (12% HCI) (Fig. 4). This reaction constitutes the classical method for determination
of the pentosan content [88], and can be carried out in dilute acid (around 0.1 M H2S04)
at elevated temperatures (100-280°C) as well [89]. Glucuronic acid, upon heating with
acid, also forms 2-furaldehyde, but at a much lower yield than from xylose.
Hexoses (13) are more stable than pentoses under acidic conditions, and the major
dehydrationproductsare 5-(hydroxymethyl)-2-furaldehyde (HMF) (15), levulinic acid
(16), and polymeric materials. The yield of HMF from glucose is generally low (10-20%),
and can be increased by raising the pH of the reaction mixture. A high yield of HMF
(45%) was obtained by heating glucose in a pyridine-phosphoric acid system at 200-
238°C for 20 min [90]. HMF polymerizes readily underacidicconditionsandcanbe
further degraded to yield levulinic acid (16) and formic acid.
Phenolic Compounds. Popoff and Theander [85-871 isolated a variety of phenolic
compounds (Fig. 5 ) from mildly acidic treatments of monosaccharides (pH 4 at 96°C for
48h).Compound (19) was the majorphenolicproductproducedfrom the reaction of
hexoses (at 3.4% yield). The yield of phenolic products is generally higher from pentoses
and hexuronic acids than from hexoses. Phenolic compounds were also formed from the
acidic degradation of erythrose and dihydroxyacetone, indicating the complexity of the
chemical transformations involved.
2. Cellulose
a. Overall Process. The heterogeneous degradation of cellulose is characterized by
two distinct phases, an initial rapid reaction followed by a slow process [9 1-94], and may
be monitored either by weight loss equivalent to the formation of soluble materials (Fig.
6) [95], or by DP changes in the cellulose residue (Fig.7) [91]. The initial phase associated
with a small cellulose dissolution (7%) and a sharpDP reductionhasbeengenerally
attributed to degradation in the easily accessible region. This observation was often used
to indicate the percent amorphous component of a cellulose sample. The results, however,
varied significantly withhydrolytic conditions, e.g., acidconcentrationandtemperature
~911.
The slow degradation stage was characterized by relatively small DP changes, es-
pecially after reaching the so-called leveling-off DP (LODP). Nelson and Tripp [93] ob-
served that LODP was not appreciably affected by hydrolytic conditions, and is charac-
teristics of a cellulose structure [92-941. Both the initial and slow degradation processes
are i n agreement with pseudo-first-order kinetics. The initial hydrolysis rate of amorphous
components was generally one to two orders of magnitude greater than that of crystallites
[92,951.
Chemical Degradation 453
12 14
CH2OH products
Condensation I
CH2
@ € € O H Y 1 I
acHo
CH2
H0 - 3H20 l 0
c=
OH HOHzC HC02H + I
13 0 CH3
15 l0
17 R = H, OH, CH3, 19
CO2H or COCH3
OH
0
OH
18 R = CH3 or C H 0 20
o a 16 24 32 40 48
Time, h
FIGURE 6 Rate of weight loss in the acidic degradation of a-cellulose in 2 M and 4 M HCI at
90°C. (From Ref. 95.)
> 1500
a
n
1000 L
500
0 L
0 60 120 180 240 300
Reaction lime, h
FIGURE 7 Rate of DP reduction in the acidic degradation of cellulose in 1 M HCI at 50°C. (Data
from Ref. 91.)
Mercerized Hydrocellulose
Sanlple DP,,. k . 10' min- ' Dl',,. k . 10' min"
effect was observed for linen and a-cellulose samples, which showed an approximate 30%
reduction. It is evident that the rayon sample had the highest reactivity.
Based on kinetic analysis, it was concluded that the end-attack model proposed by
Sharples [98,99] can be applied only to a cellulose I1 structure and not to a cellulose I
crystallite. Also, a slightly lower activation energy was observed for cellulose I samples
(40.7 kcal/mol) than for cellulose I1 samples (43.0 kcal/mol). Thus, the conformation of
cellulose appears to be an important factor affecting its reactivity, and possibly the hydro-
lytic mechanism as well.
c. Reaction Conditions. The degradation pattern of cellulose depends considerably
on the hydrolytic conditions, especially those factors that affect the swelling of crystallites.
Acid Concentration. The hydrolysis rate of cellulose generally increases with in-
creasing acid concentration (Fig. 6). The hydrolysis of cellulose in dilute acid at elevated
temperatures often resulted in glucose yields not exceeding 60% [105]. A complete cel-
lulose hydrolysis for quantitative analysis requires an initial treatmentwithstrong acid
(72% H2SOJ)to effect dissolution, followed by boiling in dilute acid (3-4%) [106]. When
usingconcentrated acids, e.g., 51% HISO, [107], 12% HCl [108,109], or 70% H,PO,
[ 1lo], the hydrolysis process was considerably enhanced, approaching that of a homoge-
neous reaction.
Acidconcentrationalsoappearstohaveaninfluenceon the apparent activation
energy of the hydrolytic process. As summarized in Table 4, hydrolysis conducted in dilute
acid at elevated temperatures ( > l 50°C) was associated witha rather high activation energy.
A similar tendency was noted previously by Nelson [92]. Most of the valuesobtained
below 100°C or using concentrated acids were close tothat of model glucoside, cellotriose,
or methyl P-D-glucopyranoside.
Solvents. The acid-catalyzed degradation of cellulose depends considerably on the
nature of the solvent [ 102,1131. The addition of ethanol, propanol, or methyl ethyl ketone
accelerates the degradation process, including the formation of nonglucoseproducts,
whereas dimethyl sulfoxide (DMSO) has a negative effect. The solvent effect has been
explained in terms of it affecting the hydronium ion reactivity [ 1131, or a r e h a t i o n of
structural stress in the cellulose [ 1021. Thus, nonswelling media may help preserve the
structural stress of cellulose and enhance its hydrolytic degradation.
In addition, cellulose can be effectively depolymerized in ethylene glycol at high
temperatures(200-240°C) [ 1141 whileminimizing the oxidativedegradation reaction.
Also, the depolymerized residues were shown to have a high accessibility toward cellulase
in enzyme-catalyzed hydrolysis.
3. Hemicelluloses
Hemicelluloses are amorphous materials and also contain a variety of nonglucose units
[6,115- 1171. The nonglucose units, because of their different ring structures and hydroxyl
configurations, generally have higher reactivity than the glucose residue, and often can be
selectively removed from lignocellulosic substrates.
Consistent with the behavior of simple glycosides (Table 2), the relative-hydrolysis
rate of p-( 1 +4)-linked polysaccharides in a homogeneous system [ 118,1191 increased in
the order cellulose (1) < mannan (2-2.5) < xylan (3.5-4) < galactan (4-5). A heteroge-
neous hydrolysis of these polysaccharides displayed even greater variations in reactivity,
following an order cellulose (1) < mannan (60) < xylan (60-80) < galactan (300). This
further demonstrates the important role of accessibility in acidic degradation reactions.
Acetyl groups presents in hardwood xylans and softwood galactoglucomannans are
also hydrolyzable by acid, especially at elevated temperatures. The acetic acid released
contributes significantly to the acidity of reaction media.
a. Xylan. The presence of uronic acid groups has a profound impact on the hydro-
lysis of xylans, as it reduces appreciably the hydrolysis rate of glycosidic linkages (Table
2). Thus, high yields of aldobiuronic acid dimers were generally obtained upon partial
hydrolysis of xylans 16.1171. Also, the initial hydrolysis rate of various hardwood xylans
(with water at 170°C) was closely related to their uronic acid contents [120]. A higher
stability of softwood xylans compared to that of hardwood xylans in sulfite pulping may
be partly attributed to their higher uronic acid group contents [6].
The acetyl groups in hardwoodxylanswerefoundto exhibit remarkable stability
under the relatively drastic conditions of acid sulfite cooking [ 1211. They were reportedly,
in steamtreatmentsof birch wood,more stable than the 4-0-methyl-glucuronosyl unit
11221. On the other hand, the arabinofuranosyl linkage in softwood xylans is very labile,
and can be selectively hydrolyzed in dilute H2S0, (0.05 M at 97°C for 3 h) [ 1231 or oxalic
acid (0.01 M at 100°C for 1.5 h) 11241.
b. Glucornannan. The a-D-( 1+6)-galactosidiclinkage in galactoglucomannans is
very labile, andcanbe selectively hydrolyzed in dilute oxalic acid (0.05 M) at100°C
[125]. Its high reactivity 1125,1261, however, cannot be satisfactorily explained in terms
of the behavior of simple glycosides (Table 2). Interestingly, the alkali-induced deacety-
lation of glucomannans increased its resistance toward acid hydrolysis, as evidenced in
acid sulfite cook [ 127,1281. This was presumably caused by the deacetylated glucomannan
being adsorbed onto the cellulose or partially crystallized.
B. Lignin
1. General Aspects
Figure 8 illustrates the general types of acidic degradation reactions for a lignin model
trimer (21b) containing both a- and @-etherunits [21,34,129,130]. The reaction is initiated
by protonation of the benzyl oxygen, followedby a-ether elimination of the corresponding
Chemical Degradation 457
GI-CH -C-CH,
CH2OH bH 8
CHzOH c4
I G,-C-CH .CH,
7 2 6 OH
OCH, OR
OR
21 a: RI = H 25 26ketones
Hibberts
b: R,= a r y l
OR
30
@H2$
OCH, B H,CO
OH OH
31
T
t
+ OCH,
I CH0
OR
27 28a !ab
OCH)
phenol or alcohol to give the benzylic carbonium ion intermediate (23), which may un-
dergo three major competing processes.
Pathway A leads to the formation of the C&-enol ether (24), which readily under-
goes acid-catalyzed hydrolysis to give the a-P-keto1 (25), and then Hibbert's ketones (26).
Pathway B involves a carbon-carbon bond cleavage between the P- and y-positions via
458 Lai
reverse Prins reaction to give formaldehyde and the C,C,-enol ether (27), which is de-
graded slowly to homovanillin (28a). Pathway C consists of an intermolecular electrophilic
addition of benzylic carbonium ion to another aromatic unit. giving mainly the a-6-di-
phenylmethane (DPM) unit (30) and some a-5 condensed structure. Additionally, formal-
dehyde generated from reaction B may condense with two aromatic units to form another
DPM-type condensed structure (31). These condensation reactions have been well estab-
lished in lignin model reactions [ 13I , 1321.
Table 5 illustrates the relative hydrolysis rate of lignin model compound reactions
(with 0.2 M HCI in 9: 1 dioxane-H20 solution at 50°C) reported by Johansson and Miksche
[ 1331.It is evident that both the a - and P-aryl ether hydrolyses were enhanced by the
presence of a phenolic hydroxyl group. Also, the a-aryl ether was much more reactive
than the P-aryl ether, roughly by factors of 25 and 65 for the phenolic and etherified units,
respectively. Reactivity of the a-ether units varies with their chemical nature [ 1361. Leary
and Sawtell [l341 showed that a p-hydroxybenzyl a-aryl ether was about 400 times more
reactive than a p-hydroxybenzyl alkyl ether.
2. a-ArylEther
Model compound studies, especially by Meshgini and Sarkanen [ 1361, indicate that the
overall a-aryl ether hydrolysis (pathway A, Fig. 8) was significantly affected by the nature
of the benzyl and a-ether groups.
As shown in Table 6, benzyl units (ring A) of the syringyl type, as compared to that
of a guaiacyl or p-methoybenzyl moiety, reduced the hydrolysis rate, andhad a higher
activation energy. The reaction was also retarded by the presence of a @-aryl ether unit.
On the other hand, a syringyl moiety on the a-ether unit (R, group) had a positive effect.
The activation energyvariedfrom 19 to 28 k c a h o l , and appears to be related to the
stability of the ether linkage.
Also, solvents play anoticeable role in the a-aryl ether hydrolysis 134,1361. The
rate generally increases with increasing solvent polarity or a decrease in the proportion of
organic solvent (dioxane or ethanol) in aqueous systems.
Linkages
Structure
(Fig. I) Relative rate
~~ ~ ~
P - A r y l ether
Nonphenolic (la), R = CH, I
Phenolic (la), R = H 12
a-Aryl ether
Nonphenolic (lb), R = CH, 6.5
Phenolic (lb), R = H 305
a-Alkyl
Phenolic” G-O-(CH2)3“G 0.8
Compound
(39)rate Relative Formula" A E (kcal/mol)
H,CO@-CH,-0-@CH, 1 24.5
H,CO@-CH,-0-@CH, 30 21.7
H,CO@-CH,-0-@CH, 30 23.7
H,co@-cH~-o-@ 20 21.3
H,co@-cH~-o-@ 33 18.9
H,CO@-CH,-0-@ 3 21.9
H,CO@-CH-0-@CH, 12 22.8
I
O-@CH2
h H,CO@-CH-O-@CH, 0.4 28.2
.'G
= guaincyl; @ = syringyl: @ = p-hydroxybenzyl.
Sortrce: Ref. 136.
3. /?-Aryl Ether
The reactivity of P-aryl ether linkages, like that of a-aryl ethers, is substantially enhanced
by a phenolic hydroxyl group [133,137-1411, and is influenced by their structures and
the reaction conditions.
a. Reaction Mechanism. The major degradation pathway of P-ether units shown in
Fig. 8 is generally accepted as proceeding through an ionic mechanism under acid-cata-
lyzed conditions [ 142- 1461.
Under typical acidolysis conditions (0.2 M HCl in 9: 1 dioxane-water, 4 h at lOO"C),
the ether cleavage reaction (pathway A) predominates, yielding Hibbert's ketones (26).
These reactions have been used extensively in lignin analysis [21,144]. However, different
mechanisms appear to be involved for hydrolytic reactions conducted in the absence of
an acid catalyst. The phenolic P-aryl model compounds (32) and (33), when treated with
water at 100- 130°C [ 138,1391 or with SO% aqueous dioxane at 180°C [ 147,1481 gave a
variety of transformation products, including dimers of the P-S (35), P-l, and P-P types,
and other condensation products (Fig. 9).
Formation of these coupling products was explainedin terms of a radical mechanism
by Sano et al. [147,148]. It should be noted that the etherified P-aryl ether dimer of (32)
was unreactive in aqueous solution at temperatures below 130°C. Also, the phenolic P-
aryl ether dimer of the syringyl type, unlike the guaiacyl dimer, was reactive even under
steam treatment conditions [140], and gave complicated products in aqueous dioxane at
180°C [ 1481.
b. Solvent. Solventshave a significant influence on the overall degradationof P-
aryl ethers. The hydrolysis reaction conducted in an aqueous solution was enhanced by
the addition of dioxane, and unlike the a-aryl ether hydrolysis [ 1361, was retarded in the
presence of ethanol [ 1371. These organic solvents were shown to favor the ChC2enol ether
formation (reaction B in Fig. S), especially at elevated temperature [l451 as illustrated in
Fig. 10. Thus, the reduced ether hydrolysis in ethanol solution can be partially explained
460 Lai
OCH3 RI
Q
OH
OCH3
H20
130°C. 6 h
.
OH
OCH3
c
OH
Condensation
products
(12.5%)
32 34 35
(2 1.5%) RI= CH20H (61.2%) R1= CH20H (16%)
RI= CH =CH-C&OH RI= CH=CH-C&OH (12.5%)
(6.9%)
OH
CH3 Q OOi C H 3 1 G- CH
=CH -CH20H
(2.6%)
33 38
G - p (p-5)G- CH=CH-CH20H (4.5%)
H2O- Dioxane
1 80°C, 20 min (1.4%)
+
v
CHzOH
J
Condensation products (65.9%)
QOCH3
0 0
36 37 39
W",
FIGURE 9 Hydrolysis products of phenolic guaiacylglycerol P-aryl ether in water at 130°C for 6
h (from Ref. 160), and in 50% aqueous dioxane at 180°C for 20 min (from Ref. 147).
in terms of increased formation of C&-enol ethers which, as noted earlier, are relatively
resistant to acid hydrolysis.
Hoo et al. [ 1451 studied the kinetics of acidic hydrolysis of P-aryl ether dimers in
50%aqueousethanolcontaining0.2 M HCl, and obtainedasimilaractivationenergy
(-36 kcal/mol) for both the phenolic and etherified types. This value is substantially higher
than that of the a-aryl ethers (Table 6) [ 1361.
Chemical Degradation 461
1.200r 8
I
I
l
m 17OoC
i
0 14OT} E ~ HHP: f
I
1709c}Dioxane:H20 l
0.900 O 140°c I
0 17OoC, Pure H 2 0 /
(0.002M HCI) l
/
/
4. Carbon-Carbon Linkages
Acidic cleavages of the carbon-carbon linkages in lignin are limited mainly to the bond
between the p- and y-carbon atoms, as indicated in reaction B (Fig. 8) for a p-0-4dimer
with a simultaneous release of formaldehyde. Similar reactions may also occur with a p-
1 (40) or p-5 (42) type units (Fig. 11) [ 129,1521. These reactions do not contribute sig-
nificantly to the formation of low-molecular-weight lignin products.Formaldehyde re-
leased, however, may participate in the lignin condensation reactions.
Under acidolysis conditions [ 1.521, the formaldehyde yield from lignin model dimers
decreased in the order p-1 (15%) > p - 5 (9%) > p-0-4 (3%). The main product from a p-
5 dimer (dihydrodehydrodiconiferyl alcohol) (42) was a phenyl courmarone (43b) (75%).
The latter product has a characteristic UV absorption, and is often used in quantitative
estimation of the p-S units [ 153,1541.
5. Lignin-Carbohydrate Complex
Among the three possible types of lignin-carbohydrate (L-C) bonds (Fig. 2), the ester
(5) and glycoside types (7) are probably more labile to acid hydrolysis than the ether type
(6). Model compound reactions [ 1351 indicate that the benzyl ethers of vanillyl methyl
462 Lai
HCHO
H 0 OCH3 - OCH3
OCH3 OCH3 OCH3
OCH3
42 43a 43b
FIGURE 11 Formation of formaldehyde from the acidic degradation of p-1 and p-5 lignin model
dimers. (From Ref. 144.)
ether (44a) and methyl 4-0-veratryl a-D-glucopyranoside (44b) were substantially more
stable than the glycosidic linkage of methyl a-D-glucopyranoside (44d) (Table 7). The
reactivity of benzyl ethers was also significantly enhanced by the presence of a phenolic
hydroxyl group [ 1331.
Judging from the behavior of model compounds (Tables 5 and 7), LCC of the a-
ether type, if present in lignin, can only be hydrolyzed slowly [135]. Also, the etherified
units are more stable to acid hydrolysis than the glycosidic linkages.
6. Condensation Reactions
Lignins are known to undergo condensation reactions even under mildly acidic treatments
[9]. This is attributed mainly to the high reactivity of the benzyl hydroxyl groups. Three
major types of lignin condensation processes are possible based on the lignin model com-
pound reactions.
a. Phenolation. This type of intermolecular condensations occurs between the ben-
zyl carbon and another aromatic nucleus, mainly at the 6-position (30) (80%) (Fig. 8) plus
some at the 5-position [131,155-1601. The condensation reaction varies with the nature
of the phenyl units and reaction conditions. Yasuda and Ota [ 1601 reported that syringyl
nuclei condensed more readily than guaiacyl nuclei upon reaction with vanillyl alcohol in
5% sulfuric acid at 100°C. The formation of benzyl chloride was observed upon treatment
of p-0-4 dimers in hydrochloric acid, and this would reduce condensation at the benzyl
position.
b. Formaldehyde Addition. The condensation of aromatic units with formaldehyde
results in the formationofmethylenecross-links (31) withpossiblysome1,3-dioxane
derivatives [159]. Acidic hydroxymethylation occurs largely at the C, or C, position of
aromatic nuclei, which may be phenolic or etherified. The initial hydroxymethylation for
syringyl units was faster than for guaiacyl units, and was promoted by the presence of a
phenolic hydroxyl group, whereas the subsequent cross-linking reaction was facilitated by
an increase in acid concentration and reaction temperature [162].
c. Intramolecular Type. Yasuda et al. [ 1631 identified a phenylcoumaran-type con-
densation product (46) in the acidic treatment (5% H,SO, at 100°C) of a p-ether dimer
Chemical Degradation 463
C H30
B 1
(3343
OH
CH20H
D 20
OCH3
OH
7. Lignin In Situ
The overall degradation of lignin, like model compound reactions, depends considerably
on the acidic environment. In aqueous media, lignin condensation reactions dominate and
lead to the formation of acid-insolubleresidues.This principle servesas the basis for
quantitative determination of lignin content in plant materials [ 1641. Lignin condensation
45 R=HorCH3 46
reactions, however, can be minimized by using mildly acidic conditions in the presence
of organic solvents, or nucleophiles.
a. Atyl-Ether Cleavages. Lai and Guo [ 130,1651 determined the acid-catalyzed hy-
drolysis of aryl ether linkages in wood lignin, which was evaluated in terms of the phenolic
hydroxyl group generated. As indicated in Fig. 13, temperature had a significant influence
on the aryl ether hydrolysis reaction. The low-temperature reaction (<65"C) displays two
distinct phases, notably in the case of spruce wood lignin. The rapid phase was likely
attributable to the hydrolysis of very reactive noncyclic a-aryl ether or possibly the di-
benzodioxocinlinkages (4) (Fig. 1) [32], whereas the slowphasewas likely associated
with the P-aryl ether hydrolysis. Accordingly, the highly reactive a-aryl ether units (pre-
sumably the noncyclic type) were determined to be 4% and 6% of C,, units for spruce and
aspen lignin, respectively [30]. It is also evident that the aspen lignin contained a high
proportion of P-aryl ether units with hydrolysis rates substantially higher than those of
the spruce lignin.
Nimz [ 166,1671 subjected wood to water percolation at 100°C for several weeks.
Approximately 20% and 40% of the spruce and beech wood lignin, respectively, became
soluble, and contained a variety of lignin oligomers. Sakakibara [ 1681 obtained similar
results upon lignin hydrolysis in 50% aqueous dioxane at 180°C. These soluble products
were assumed to arise mainly from the cleavage of a-aryl ether units.
b. Condensation. The extent of lignin condensation depends on the acid used, and
is oftenmeasuredbyareduction in the yield of simplealdehydes upon nitrobenzene
oxidation [8,169]. It is generally higher in sulfuric acid than in hydrochloric acid solution
[9,170,17l], probably because of the chloride ions being able to form a stable benzyl
chloride adduct [ 1611. Similarly, nucleophilic species, such as bisulfite and thioglycolic
acid, are known to reduce lignin condensation reactions in dilute acid by being able to
undergo sulfonation or formation of thioglycolic acid derivatives [ 1461. The sulfonation
reaction [SI] is essentially the chemistry of delignification in commercial sulfite pulping
[ 1721. Also, the extent of lignin condensation reaction varies with the nature of aromatic
units, beinghigher for the guaiacyl than for the syringyl units, asobservedduring the
initial hydrolysis of sweetgum wood in concentrated hydrochloric acid [173j.
C. Applications
The chemical analysis and utilization of lignocellulosic components require a quantitative
separation or a selective degradation of individual components, which, however,is difficult
to achieve, especially for lignin [ 1741. even under an ideal laboratory condition. Several
analytical and practical processes have been developed based on controlled degradation
of the individual cell wall components [2].
1. Acidolysis
Under typical acidolysis conditions (with 0.2 M hydrogen chloride in a 9: 1 dioxane-water
mixture at 100°C for 4 h) [21,144], lignin was depolymerized through a- and P-aryl ether
cleavages to givesolubleproducts.The yieldof monomericanddimericproductswas
substantially higher from birch (30%) than from spruce ( 1 7%) lignin. The aryl ether hy-
drolysis reaction can be further enhanced by solvolysis in a dioxane-ethanethiol solution
containing boron trifluoride ethereate (thioacidolysis) [ 1751. These low-molecular-weight
products provide valuable information about lignin structure.
Chemical Degradation 465
t-
60 0 2 4 6 8 IO 12
Reaction T m e . h
85-c
466 Lai
2. Acid-Sulfite Pulping
Conventional calcium-based acid sulfite pulping [4] uses an acidic liquor (pH = 1.5-2)
containing 6% total and 1% combined SO,, and is conducted in the range 120-140°C for
5-20 h. The mechanism of delignification [34,51,172] is attributed mainly to the hydro-
lysis of benzyl ether linkages (Table 5 and Fig. 8) and the sulfonation reactions. Wood
polysaccharides, especially hemicelluloses, are very susceptible to dissolution and hydro-
lytic degradation reactions.
Table 8 illustrates typical compositionof sulfite pulpsfromspruceandbirchas
compared to that of the original wood [5]. It is evident that the hemicellulose loss was
very substantial, notably in the case of birchxylan (83%). A large proportion of the
dissolvedpolysaccharideswas in the formofmonosaccharides [ 1761. The acid sulfite
process, because of the problem of species limitation, chemical recovery, and weaker sheet
properties, had been largely replaced by the kraft process.
3. Prehydrolysis
Hemicellulosesfrombiomasscanbe preferentially removed by hydrolysis in water (at
170°C) [177,178], in dilute acid (0.1 M HCl at 120°C) [120], or in concentrated acid (20-
30% HCI at 40°C) [ 1791. These prehydrolysis conditions resulted in little dissolution of
the cellulose and lignin components, and may be used in connection with kraft pulping to
produce dissolving pulps [4] or with wood saccharification. To preserve pulp quality for
the production of dissolving pulp and other applications [ 180- 1821, the prehydrolysis is
normally carried out in water at 170°C for about 2 h withthe organic acid released (mainly
acetic acid) acting as catalysts.
On the other hand, prehydrolysis for wood saccharification emphasizes the quality
of the hydrolyzates. In terms of a xylose production, prehydrolysis with dilute acid (0.4%
H2S0,) gave better results than steam hydrolysis alone [ 1771. The xylan removal from
hardwood generally displayed an initial rapid phase followed by a slow reaction [ 120,1831.
Thus, xylan contains fractions of different reactivity. Reported activation energy for the
initial-phase reaction varied from 28.2 kcal/mol based on wood solubilization in HCl (0.5-
6.0 M) [ 1841 to 22.8 kcal/mol based on a xylose formation in 30% HCl solution [ 1791.
4. Autohydrolysis
Autohydrolysis is basically a steam hydrolysis process conducted at elevated temperatures
(175-220°C) [ 1851. The bulk of hemicelluloses become solubilized, while lignin can be
extracted from the residue by organic solvent or alkali. After a short hydrolysis of aspen
at 215°C (for 4 min) or at 195°C (for 25 min), about 90% of the lignin became soluble
in a 90% aqueous dioxane solution. The lignin dissolution resulted largely from cleavages
of the a-aryl ether plus some P-ether units [ 186- 1881 and had an activation energy (29.3
kcal/mol), similar to that of acid hydrolysis for an etherified syringyl a-aryl ether model
dimer (28.2 kcal/mol) 11361.
At a given temperature, the extractability of lignin as a function of time went through
a maximum [ 1851, indicating that lignin condensation is a controlling factor [187-191].
Phenolic compounds, especially 2-naphtho1, were reported to be effective in preventing
lignin repolymerization [ 19 1,1921.Thus, hardwood under proper autohydrolysis treatment
followed by solvent extraction can be separated roughly into its three major components.
The autohydrolysis process is generally unsatisfactory for softwood with respect to
delignification. Also, the pulp produced is not attractive for papermaking, because of its
poor sheet properties caused by the hydrolytic degradation of cellulose [ 1931. The latter
degradation, however, can be reduced slightly by continuous removal of the steam con-
densate during autohydrolysis [ 1941.
6. Organosolv Delignification
The principle of acid-catalyzed organosolv pulping is basically a controlled steam hydro-
lysis i n the presence of organic solvents 1207-21 I ]. Among various pulping processes
reported, the Alcell process, usingaqueousethanolunderautocatalyzcdconditions, is
currently the only one in a pilot-scale operation [212,213].
In recent years, solvent-assistcd alkaline delignification (organosolv pulping) has also
received much attention, notably in sodacooking with an acidic methanolpretreatment
(Organocell), and acombined alkaline sulfite-anthraquinone(AQ)-methanol (ASAM)
system [ 2131.
In acid-catalyzed organosolv processes. the partially degraded lignin and hemicel-
lulose components are solubilized simultaneously, and thc delignification selectivity varies
with the nature of the solvent and catalyst used ;IS well ;IS their concentrations. Although
ethanol. methanol, and sulfuric acid are most commonly used i n laboratory studies, delig-
468 Lai
nitication with organic acids, e.g.. acetic 12 141 and formic [ 149- 1 S l ] acids. also has re-
ceived much attention.
Available data [30,165,2 15-2 171 suggest that the a- and p-ether units cleavages arc
very important to acidicdelipnilication. Also. the hydrolysis of L-C linkages may bc
involved 12071. because isolated organosolvlignins generally have a low carbohydrate
contcnt.
Lai and Mun [ 2 16,7 171 recently reported the significance of methanol and aryl-ether
hydrolysis i n the acidic delignitication of aspen wood. As indicated i n Fig. 14, the acidic
system without methanol resulted i n preferential retnoval of the carbohydrate components.
For the two mcthanolic systems, the autocatalyzed process had ;I slightly higher deligni-
fication selectivity than the acid-catalyzed one.
Figure 1 S illustrates the dependence of acidic delignification on aryl-ether clcavagc
reactions as indicated by the l’ormation of phenolic hydroxyl groups. A nearly linear re-
1 0 3
L
c
0
’=
a
.-V
60
20
0
I
Chemical Degradation 469
lationship was observed initially, and extended to approximately 30% and 60% delignifi-
cation for the watersystemand the twomethanolsystems, respectively. The reduced
impact of aryl-ether hydrolysis on delignification observed in the water system was likely
caused by lignin condensation reactions 12151.
7. WoodSaccharification
The hydrolysis of cellulose in the wood for glucose production (saccharification) is usually
performed on prehydrolyzed residues with low hemicellulose content. Cellulose, because
of its physical structure and crystalline nature, is relatively resistant to acid hydrolysis.
This physical inaccessibility presents a formidable task for the commercial production of
glucose from ligno-cellulosic substrates. The glucose yield varies considerably with hy-
drolytic conditions. With dilute acid (0.1 - 1.6% HISO,) at hightemperatures(around
200"C), the glucose yield was usually less than 60% [ 105,220,22 I], because of the for-
mation of nonsugar products. In general, the net glucose yield favored a short reaction
duration at high temperatures; e.g., a 54% yield was obtained with 0.4% acid at 260°C
for 27 S [2201.
With concentrated acid at low temperatures (around 40"C), cellulose is extensively
swelled and eventually dissolved, and the hydrolysis proceeds rapidly to give glucose in
nearly quantitative yield [ 1061. The use of hydrochloric acid up to 16 M in the 20-50°C
range was also reported [log]. The reactivity of the hydrolysis lignin residue depends on
the type of acid used. Lignin obtained from hydrolysis in hydrochloric acid, as compared
to sulfuric acid, was shown to be less condensed [ 170,17 l], and may be more attractive
for by-product applications.
Additionally, there are nonconventional methods that use organic acids or solvents.
The useof trifluoroacetic acid[222],hydrogen fluoride [223,224],amixture of acetic
acid-acetic anhydride-dimethylformamide-sulfuric acid [225], or aqueous acetone with
acid catalyst [2261 has been explored.
The organosolv delignification of hardwood was reported to have an activation en-
ergy of 19.2 kcalhnol(forblackcottonwood in aqueousmethanolcontaining 0.05 M
H2S0,) [218], or 16.2- 19.8 kcalhol (for birch in 60%aqueousethanol)[219].These
values are slightly lower than those for the hydrolysis of a-aryl ethers dimers (18-28
kcal/mol) (Table 6) [ 1361, and substantially lower compared to that of P-aryl ether com-
pounds (36 kcalhol) [ 1371. This variation may suggest that diffusion is a significant factor
in the acidic organosolv delignification process.
The primary hydrolytic degradation of wood components in alkaline media, like that under
acidic conditions, involves the cleavage of hydrolyzable linkages in lignin ( a - and /?-aryl
ethers) and of glycosidic bonds in polysaccharides. However, the alkaline and acidic deg-
radations proceed through distinctly different mechanisms. The major loss of polysaccha-
rides from the alkaline degradation process is caused by endwise depolymerization reac-
tions (peeling),leading to the formation of carboxylic acid derivatives.Thealkaline
degradation of lignin plays a dominant role in the utilization of lignocellulosic components,
and constitutes the fundamental chemistry of the alkaline pulping or kraft process, whereas
the alkaline degradation reactions of polysaccharides cause undesirable losses in pulp yield
and necessitate the use of an excess of alkali to neutralize the acidic degradation products.
470 Lai
A. Polysaccharides
The chemistry of the alkaline degradation of polysaccharides and related model compounds
has been extensively studied and reviewed [ 1-4,130,2271.
H2COH
R 0e o - - RO-
OH
OH
47
t 51
OH H0 0
49 50
48
I
R&H=0
52
OH
0 OH
54
[:gyH 56
-
1 l
H2COH 1 I
C02H
R 04 7 c 0 2 H I
R0 CHOH
OH I
OH CH3
53 55 57
or alkalinity below the mercerizing strength (<2 M NaOH). An average of 68 and 40 was
observed for a native and mercerized cellulose, respectively.
As anticipated, the peeling process wasconsiderably more extensive for soluble
polysaccharides.e.g.,amylose [2291 orglucolnannans 12391, because of the lack of it
physical stopping process. It was observed that mlylose could be totally degraded by a
peeling reaction alone in dilute alkali treatments at low temperatures (2291.
Saccharide Composition. The peeling reaction of wood polysaccharides is af-
fectcd by their- composition, a s reflected i n the behavior of disaccharides in alkalis
1240.241 1. Thc disaccharide reactivity increases in the order mannobiose < ccllobiose <
472 Lai
CH3
I
R c=o
&OH
OH
OH
17 R=H 58
R CH3
l
QOH
OH
OH
18 R= CH3, CH0 59
CH2-0
CH2OH CH2OH
e CH20H
o -A " B K 0 - 0 - R k
OH N B " OH ___) OH -+
-OR H
H H H 0 OH
OH 0-
474 Lai
Table 9 for a variety of methyl pyranosides pertinent to wood polysaccharides. The re-
actions (conducted in 2.5 M NaOH at 170°C) were reported mainly by Janson and Lind-
berg [270]. All the relative hydrolysis rates were based on methyl a-D-glucopyranoside
being unity.
It is evident that the P-glycosides of D-glucose, D-galactose, and D-xylose residues
were substantially more reactive than the respective a-anomers, whereas the reverse was
true for D-mannoside and L-arabinoside. These variations, however, are consistent with
the contention that glycosides with the aglyconic and the C2 hydroxyl groups being in a
trans position react faster than the corresponding cis anomer [270]. Also, high reactivity
was observed for a glucuronoside under alkaline conditions [271].
In wood polysaccharides, the bulk of glycosidic linkages are present in a P-pyran-
oside form except for the galactose, arabinose, and glucuronic acid residues, which are in
a-form. Also, arabinose units, unlike others, are present as a furanosidic structure. Thus,
in a homogeneous system, the relative reactivity of various glycosidic linkages present in
the wood would be expected to increase in the order of saccharide moieties: galactose (1)
< mannose ( 1 . 1 ) < glucose (2.5) < xylose (5.8) < arabinose (32) < glucuronic acid (280).
This order ofreactivity is distinctly different from that of acid hydrolysis (Table 2), notably
for the high stability of the galactose side groups and the high reactivity of the glucuron-
osides [271] under alkaline conditions.
Alkali Concentration. Hydroxyl-ionconcentrationwasshown to significantly in-
fluence the hydrolysis rate for a variety of simple glycosides [260,265,267,272,273]. The
alkaline cleavage reaction as illustrated in Fig. 18 [260,272] can be described kinetically
i n terms of proceeding via anionic intermediates according to Eqs. ( I ) and (2):
K
GlcOR + OH = H,O + GlcOR
K1
GlcOR- + products
k,K[OH-]
=
1 + Kl0H-I
k,h*
where GlcOR is a glycoside, GlcOR- is an anionic intermediate such as (61) (Fig. 18), K
Relative rate
Methyl pyranosides of P-Anomer
a-Anomcr
Chemical Degradation 475
3. Cellulose
The alkaline degradation of cellulose is often encountered in alkaline pulping as well as
in hot-alkali refinement of dissolving pulps.
a. OverallProcess. Figure 19 illustrates the overall degradationprocess of cellu-
lose in alkali at elevated temperatures (150- 180°C) [237,274,275]. The existing reducing
end groups will rapidly initiate the endwise degradation (peeling) ( k , ) .This primary peeling
process, however, is generally negligible for cellulosebecause of its high DP (around
10,000). The alkaline hydrolysis of glycosidic linkages (k,,) generates new reducing end
groups which give rise to a similar peeling process (k?), resulting in a loss of about 65
glucose units [274,275].
Both the peeling and glycosidic cleavage reactions of cellulose, like a homogeneous
reaction of simple glycosides [272] and amylose [229], have been shown to conform with
pseudo-first-order kinetics [235,274,275]. The peeling process was relatively rapid in the
150-190°C range and roughly by a factor 10' times faster than the alkaline hydrolysis
reaction, according to Lai and Sarkanen's estimation [275]. Although the alkaline hydro-
lysis of cellulose is a relatively slow reaction resulting in small weight loss, its impact on
pulp viscosity in alkaline pulping is appreciable. The activation energy for the alkaline
I
GG-G-G-G-G-G-G-G,y
k , ( Rapid )
+ R,
1 k h ( Slow )
l
"G-G-Gs
k2 ( Rapid )
+ X,
G : Anhydroglucose
units 2 n:Acidicdegradationproducts
Gr: Reducing
end
groups CS: Stable
end
groups
hydrolysis of cotton cellulose (36 kcal/mol) [275] was similar to those reported for simple
glycosides 1260,2761. A substantially higher activation energy (43 kcal/mol) has been re-
ported for the cleavage of wood cellulose in kraft pulping of sprucewood as measured by
a reduction in pulp viscosity [277].
h. Physical Structure. The overall alkalinedegradation of cellulose is profoundly
influenced by its morphological structure, which has a significant impact on the physical
accessibility and possibly the reaction mechanism as well.
As noted earlier, the supramolecular structure of cellulose exerts a dominating influ-
ence on termination of the peeling process, as the degrading ends may be stabilized when
they reach regions inaccessible to alkali. At low temperature, the majority of degrading
chains terminate with a normal reducing end group at the crystalline-amorphous transition
region. The observed, approximately constant, degradable chain lengths in native cotton
cellulose, as Lai and Sarkanen suggested 12751, must be attributed to the submicroscopic
structure of microfibrils or more specifically to the average length of accessible segments
of the cellulose molecules. For mercerized cellulose,the degradable chain length was lower
ascompared to native cellulose (40 versus 68), and this may beascribedtoashorter
accessible segment of the mercerized cellulose molecules.
After alkaline degradation, the residue from mercerized cellulose had a higher con-
tent of stable acidic end groups than that of native cellulose 1237,2381. Thus, it appears
that the fringe area of ordered regions in the cellulose I1 type are inaccessible to the peeling
reaction but conversionto stable acidicendgroups may occur, while in corresponding
areas of the native cellulose both reactions are impeded. Also, both alkaline peeling and
chemical stopping reactions occurred more rapidly in the amorphous region of a regen-
erated hydrocellulose than in the disordered regions of a fibrous hydrocellulose 1141.
In glycosidic cleavage reactions conducted in the range 150- 190"C, the hydrolysis
constant for mercerized cellulose was approximately 70% higher than that of native cel-
lulose [275].This suggests that the number of accessible glycosidic linkages in mercerized
cellulose is higher by a factor of 1.7 than in native cellulose. In addition, Gentile et al.
1141 detected the alkaline hydrolysis reaction for amorphous hydrocellulose under rela-
tively mild conditions ( 1 M NaOH, 60-80°C). They suggested that the disordered regions
associated with cellulose I and I1 polymorphs had different degrees of structural order and
reactivity. Thus, molecular conformation, along with molecular mobility and accessibility,
appears to influence the alkaline susceptibility of glycosidic linkages.
4. Hemicelluloses
Hemicelluloses in alkali are susceptible to both physical changes and chemical reactions
includingswelling,dissolution, saponification, reprecipitation. peeling, and glycosidic
cleavage reactions [4].Alkali-induceddeacctylationandhydrolysis of the uronicacid
group of xylan proceed readily under alkaline pulping conditions, and contribute signifi-
cantly to its redeposition or adsorption onto the fibers. Consistent with the simple glycoside
reaction (Table 9), the galactose side chain in galactoglucomannans is fairly resistant to
alkaline hydrolysis. Glucomannans were generallyless stable than wood xylansin alk,A I'me
pulping 141, although this trend was not reflected in the reactivity of model glycosides
(Table 9).
U . Xyltrrls. Thepeelingprocess of xylans is basically similar to thatof cellulose
(Fig. 16), but it can be retarded chemically by their unique structures. Extended treatment
of a rye-four arabinoxylan with dilute sodium hydroxide at room temperatures resulted
in only a 29% reduction i n tnolecular mass 12781. The peeled xylose units were converted
Chemical Degradation 477
to qdo-isosaccharinic acid plus several other acids. Chemical stopping was also shown to
involve the formation of xylo-metasaccharinic acid end group. As anticipated, the poly-
saccharide after reductionwithsodiumborohydridewascompletely stable to alkaline
peeling.
The moderate alkaline stability of xylans can be partly attributed to their unique end-
group arrangements, and the presence of 4-0-methyl-glucuronic acid groups. Johansson
and Samuelson [279,280] showed that the peeling of xylan molecules was retarded when
a galactouronic end group was substituted at the C2 position with a rhamnose unit, or a
xylose end group containing a 4-0-methyl-glucuronic acid unit at the C2 position. The
former rhamnose unit, stable in dilute alkali at low temperatures, was labile to elimination
at moderate temperature (95°C).
Similarly, the retardation effect of 4-0-methylglucuronic acid groups was appreciably
reduced at elevated temperatures, because these groups are unstable under alkaline pulping
conditions [238,281-2851 and maybe degraded according to the scheme shown in Fig.
20 [279,282,283]. It involves a series of @-elimination reactions, typical for uronic acid-
containing carbohydrates [286].
The key step is a base-catalyzed demethoxylation of (64) to give the hexene-uronic
intermediate (65), which after another p-elimination releases the anticipated acidic moiety
(67). The 4-deoxyhex-4-enuronic acid group (65) was recently reported to be present in
pine kraft pulps[287].Thisproposedmechanism is consistentwiththedegradation of
glucuronic acid groups in xylan as a reaction dependent on alkali concentration [283].
Softwood xylans, in addition to having a higher content of 4-0-methyl-glucuronic
acid groups than hardwood xylans, contain some arabinose units attached to the C3 po-
sition of xylose residues. The latter substitution pattern also induces the chemical stopping
reactionand certainly contributestohigher alkali stability of softwoodxylans [4,284J.
Similarly, the beneficial effects of arabinose groups are probably diminished at elevated
pulping
temperatures,
because
they are extensivelyremoved
from the xylans
[239,284,288J.
Regarding the alkaline hydrolysis of xylans, the relative reactivity of glycosidic link-
ages, based on methyl glycosides reactions (Table 9), decreases in the order glucuronoside
(280) < arabinofuranoside (32) < xyloside (5.8). The high reactivity ofglucuronosidic
linkages was attributed largely to the presence of a carboxylic function, which is capable
of initiating a series of p-elimination and transformation reactions similar to those shown
in Fig. 20.
470 Lai
B. Lignins
The alkaline degradation of lignins, like that under acidic conditions, involves mainly the
cleavage of a- and P-aryl ether linkages, which also may be accompanied by condensation
reactions. The nature of these chemical reactions, derived mostly from lignin model com-
pound reactions, hasbeenextensivelyreviewed, [ I -3,21,34,35.146], but it is still not
entirely clarified in the degradation of wood lignin, especially with regard to condensation
processes.
1. General Aspects
Figure 21 illustrates a general scheme for the alkaline degradation of hydrolyzable ether
units, which are present in both phenolic (68) and etherified (75) types. The reaction of
phenolic units is initiated by a phenoxide ion (69) to yield a quinonemethide intermediate
(QM) (70) with elimination of the P-ether unit as R,OH.
The QM may participate in several reactions depending on the alkali environment.
In soda cook, it undergoes mainly a carbon-carbon bond cleavage of the P - y linkage to
yield formaldehyde and C&-enol ether (74) (reaction C). The latter enol ether may un-
dergo further degradation slowly [294,29S]. Thus. soda cooking of phenolic P-aryl ether
units resulted in only limited ether cleavages (reaction A) 1296,3021. In kraft liquor, the
QM reacts readily with hydrosulfide ion and the resulting adduct (71) then undergoes an
intramolecular displacement process leading to the P-ether cleavage (reaction B).
On the other hand, alkaline cleavage of the etherified P-aryl ether unit (75) proceeds
directly through an intramolecular displacement mechanism (reaction D).
Under kraft cooking conditions, the P-aryl ether cleavage of phenolic type could be
12 to SO times faster t h a n that ofan etherified type, depending on the hydroxide- and
sultide-ion concentrations (3031. Lignin condensation reactions may include the formation
a-S-diphenylmethane (DPM) (79) and a-carbohydrate ether linkage (81), derived possibly
from the quinonerncthide (70) (2 1,34.304-308] or the epoxide (77) (309.3101 intermediate.
Formaldehyde release from reaction C may participate in the formation of a 5.5’-DPM
u n i t (80) [3OXl, while coniferyl alcoholproducedfrom reaction B may involve in the
formation o f a P-y-linked condensed unit (82) (306,3071.
The overall degr;ldation of ether units. asrevealed by lignin model reactions, is
profoundly influenced by their chemical structures and reaction conditions.
479
Chemical Degradation
CH2OH
I
CH
CH3
CH
-G
____)
OCH,
OH
71 72 73
t SH- HCHO
OCH,
o,,
0
70
OH
74
OCH,
Condensation Products
)$4
Go-CH CH-OG
II
+ H,CO OCH,
OH 0-
79 80
OCH,
CHzOH
t
OCH, om,
OCH, OH
75 81
lI
H
' OCH, CHzOH
6
CHzOH CH2OH
&H
o:hH
Q -
kH- O b CH3 +XI
$ -
6HO-
OCH,
D
-G OCH,
OH-
OCH, om,
76 77 78
2. a-Ether
The alkaline cleavage of phenolic wether structures (68) is generallyaccepted as pro-
ceeding through the formation of quinonemethide intermediates (70),and is significantly
affected by the nature of the ether group. This cleavage reaction occurs quite readily for
an a-aryl ether unit under mild alkali treatment ( I M NaOH at 25°C) [31 I], whereas the
wether linkage of a lignin-carbohydrate model was shown to be stable under the same
conditions [ 1351. Concerning the reactivity of a-aryl ethers of etherified units, they were
generally shown to be stable in alkali even at elevated temperatures ( 2 M NaOH at 170°C)
[312], although some a-aryl ether models were recently reported to be hydrolyzed slowly
[3 131.
3. P-Aryl Ether
The alkaline cleavage of P-aryl ether linkages generally proceeds through an intramolec-
ular displacement (by a neighboring-group participation) process, with distinctly different
mechanisms for the phenolic and the etherified types. The reactivity varies appreciably,
with the P-etherconformationbeinghigher for the erythrothan for the threoisomer
1314,3 151. The nature of the phenyl units also has a noticeable effect on the degradation
process, being especially facilitated by the syringyl nuclei [316-3181.
a. PhenolicUnit
Mechanism. The dominant reaction of phenolic P-ether models in soda liquor, as
noted earlier, is the formation of the C6C,-enol ether (74) (Table 10) resulting from elim-
ination of the y-carbinol group as formaldehyde. Theether cleavage, being a minor process
(12-33%), is generallythought to involve an ionic mechanism[296,300,305] via qui-
TABLE 10 Effect of Alkali, Sodium Sulfide, and AQ on the Alkaline Degradation Products of
Phenolic Guaiacylglycerol-P-Guaiacyl Ether (33)
CH20
A OCH3
OCH3
0 OH
33 36 74
Q0 OCH3
CH:
I 7H20H
CH
I
C HI
79 83 84
nonemethide intermediates (reaction B) or the epoxide pathway (reaction C ) (Fig. 22). The
possible involvement of a radical mechanism also has been suggested for the syringyl P-
ether units [3 16-3 181. The observation that the cleavage reaction was facilitated in a high-
alkali concentration (80% cleavages in 4 M NaOH) [296],andfor the erythroisomer
[314], however, suggest that the epoxide pathway is a viable reaction.
Structural Effects. The reactivity of phenolic P-ether units in soda liquor is con-
siderably affected by their chemical characteristics, and is higher for the erythro than for
the threo isomer [316]. Also, the syringyl P-ether model dimer displayed distinctly dif-
ferent behavior than the guaiacyl dimer, notably an unusually high ether cleavage (Table
11) [3 171, whereas the enol ether formation is largely suppressed. As illustrated in Fig. 23
for the degradation products of a syringyl dimer (85) (in 1 M NaOH at 140°C for 180
min), the yield of enol ether was quite low, being 5% [316] as compared to approximate
70% for a guaiacyldimer. The syringylP-ethercleavage reaction wasreported to be
relatively unaffected by the hydroxyl-ion concentration in the range 0.1- 1 .O M [318].
Additives. The cleavage of phenolic P-ether linkages was greatly facilitated by a
variety of additives, notably sodium sulfide [298-300,3 19-3221 and anthraquinone (AQ)
derivatives[322,323-3251. The enhancedcleavage is usually explained in terms of a
nucleophilic addition to the QM intermediate followed by an intramolecular fragmentation
of the adducts (71)(Fig. 21) and (91) (Fig. 24) to cleave the P-ether group.
In the case of AQ, an alternative mechanism based on electron transfer reactions also
has been suggested [325,326]. Other types of additives effective in enhancing the P-ether
482 Lai
TABLE 11 Influence of Syringyl Units on Relative P-Ether Cleavages in Soda (0.1 M NaOH)
and Kraft (0.1 M NaOH + 0.015 M Na,S) at 130°C Estimated at the Half-Life Reaction
G G-l 25 84
S G-I 52 92
S S-I 80 90
G GI1 14 90
S GI1 6 74
S95SI1 80
S<JUKY:Ref. 317.
$$
HO-&H HO-&H HO-CH
OCH3
O H OCH3
OH OH
OCH3
CH20H
CI H - O O C H 3
&:cCH3 - IMNaoH
140T
180 min
H3C0
OH
L
85 86
R H3CO$CH_
R = CH3 (75%) OH
R=H
R =CH0 (5%)
R =COCg
87 88 89
influenced by the side-chain structure (Table 12). The presence of a y-methyl group re-
sulted in a slightly lower activation energy than that of a glycol-type unit (26-28 versus
31 kcal/mol) in a soda cook. This difference, however, was not evident in the kraft cook,
and a distinctly higher value (33 kcal/mol) was reported for the threo isomer.
b. Nonphenolic Units. Alkaline cleavage of nonphenolicP-etherstructuresgener-
ally proceeds through an intramolecular displacement process and yields the epoxide in-
termediate (77)(Fig. 2 I ) [312,315,3411. This reaction is facilitated by an ionized hydroxyl
group at the a- (76) or the y-position.Consistent with the stereochemicalrequirement,
the erythro isomer was more reactive than the corresponding threo unit by a factor of 4
[315]. Also, the cleavage reaction was not affected by the addition of sodium sulfide [339]
or AQ [342], and was appreciably influenced by alkali concentration, the nature of the
solvent, and structural modification.
Alkali Concentration. As a base-catalyzedreaction via anionicintermediates
[260], the cleavage of nonphenolic P-ethers (75) increases initially with the hydroxyl-ion
concentration and levels off to a constant value at a certain alkali concentration. Limited
data [315,339] indicate that the cleavage reaction is directly proportional to the hydroxyl-
ion concentration within the range of alkali used (0.1- 1 M). An estimation based on these
data using Eq. (3) indicates that a linear relationship between the reaction rate and alkali
concentration extends to about 4 M alkali.
404 Lai
+ -
p&o
H0
0
36 90a 91
1
Carbohydrate
Oxidation
&
l
0 +
- QOC",
0-
yH20H
CH
1
l
CH
G 00
O C H j
90 73
FIGURE 24 A proposed anthraquinone catalyzed alkaline cleavage of phenolic P-aryl ether link-
ages. (From Ref. 34.)
TABLE 12 Activation Energy for Alkaline Cleavages of Phenolic and Etherified Lignin P-Aryl
Ether Model Dimers
a R, = R2 = H 0.5 - 31.3
0.5 0.08 28.9
b R, = H, R? = CH,
Erythro 0.1 - 28.0
0.2 0. l 28.7
Threo 0.1 - 26.1
0.2 0.1 33.3
C R , = CH,, RZ= H 0.5 33.5
d R, = CH3, R? = H 1.o 28.7
e R , = R2
CH,, = CH20H 1 .o 31.6
OCH3
OR,
92
Chemical Degradation 485
Solvent and Alkali. Limited data on the P-ether model reaction indicate that the
effect of organic solvents on the alkaline cleavage reaction varies with the side-chain unit
[343]. In the case of veratrylglycerol-P-guaiacylether (75), the cleavage rate was slightly
reduced by using dioxane as a co-solvent, but was unaffected by methyl cellulose. Inter-
estingly, the cleavage reaction was enhanced in a monoethanolamine (MEA)-sodium hy-
droxide solution 13421.
For veratryglycol-P-guaiacol ether, the ether cleavage was suppressedby the addition
of methyl cellosolve andespecially dioxane 13431. Also, the reaction was greatly facilitated
in a DMSO-potassium tert-butoxide system[338];and64%cleavagewasobtained at
75°C in 30 min.
Structural Effects. The hydrolysis of nonphenolicP-etherlinkages (75) canbe
greatly improved by modification of the a- or y-hydroxyl group [328,339,344-3471 (Fig.
25). The presence of an a-carbonyl or a p-tolysulfone group was reported to give the
highest enhancement, and the alkaline hydrolysis was observed even under ambient con-
ditions [345]. A possible mechanism, as indicated in Eq. (4) (Fig. 25),may involve a series
of @-elimination, a hydrosulfide-ion addition to the enone intermediate (94), and an intra-
molecular fragmentation of the keto thiol intermediate (95) to cleave the P-ether linkage
[344]. Similarly, the P-ether bond cleavage was facilitated by an a-sulfonate [345] (98),
an a-2,4-xylenol (102) 13451, or a y-p-toluenesulfinyl 13471 group. The chemical effect of
a 2,4-xylenol group (102) on the P-ether cleavage may be accounted for by it being able
to initiate an intramolecular displacement process [Eq. (6)] [327,328].
93 94 95 96 97
I
R
98 99 100 101
H CH20H
102
5. Condensation
The nature of alkaline condensation reactions [34,307,358-3611 has been extensively stud-
ied using lignin model compounds. There are four major types: conjugate addition, for-
maldehyde addition, epoxide addition, and aldol reaction. One of the key factors in these
reactions is the formation of carbanions, which has been reported for several aliphatic
carboxylic acids and simple phenols at 170°C [362].
a. ConjugateAddition. Theadditionofnucleophilesderivedfromphenolicor
enolic units to a QM (quinonemethide) or an extended QM (from coniferyl alcohol) in-
termediate leading to the formation of stable carbon-carbon or ether linkages has been
welldemonstrated[21,34,304-3081.Thetypesofcondensed units formedinclude a-5
DPM (79), a-carbohydrate ether (81) [363], and p-y linkages from dimerization of con-
iferyl alcohol (82) [306,307] (Fig. 21). Also, the formationof a - l (114) and a-a (116)
units (Fig. 27) was reported after alkaline treatment of a ðer dimer (112) (in 0.5 N
NaOH at 140°C) [305].
The participation of an enediol intermediate arising from the reducing end groups of
carbohydrates has also been suggested [300]. These reactions clearly indicate that a wide
variety of carbanions may be generated from lignin- and carbohydrate-derived products
and are involved in the condensation process. The conjugate-type condensation was re-
ported to be suppressed in the presence of additives, e.g., AHQ [364,365], and could be
inhibited by steric factors in a two-phase system [366].
b. Formaldehyde Addition. The alkali-induced addition of formaldehyde to simple
phenols occurs mainly at the C5 position via a Lederer-Manasse reaction, leading to the
formationof5,5’-DPM (80) (Fig. 21).Thiscondensation reaction is wellestablished
[308,367,368], and may involve formaldehyde generated in situ from the reaction of lignin
and related dimers [353,358,369,370]. Also, the formaldehyde-type condensation was re-
tarded in the presence of sulfide or AQ [371].
Chemical Degradation 487
+
0
0 0
X
8X
+
v)
z
8
F 0
zQ
0
lo
U -U -U
488 Lai
OH
OH
115 116
G= +OH
OCH,
FIGURE 27 An example of the formation of a-l and a-a' condensed units. (From Ref. 305.)
-OH,A
v
- QOCH3
OCH, -
OCH,
( + isomers)
120 122
123 124
C. AlkalineDelignification
The kraft process, using mainly sodium hydroxide (-0.8 M) plus sodium sulfide (-0.2
M) at elevatedtemperatures ( 160- 180°C) for 1-2h, is the mostwidelyused in the
production of chemical pulps for papermaking and cellulose applications [4,227,358]. Un-
fortunately, suchalkalineconditionsrequired for delignification are alsoconductiveto
polysaccharide degradation.
1. Overall Process
The kraft delignification process is known to consist of three distinct stages: the initial,
bulk, and residual phases [373-377). They differ considerably in delignification selectivity,
as illustrated in Fig. 30 for a relative change of polysaccharide and lignin content during
a kraft cooking of spruce and beech wood [378]. The residual phase, associated with the
removal of the last 10% of the lignin, is accompanied by severe cellulose degradation. In
e Beech
P OSpruce
40 I
l , l , l , l , l , L
0 5 IO 15 20 2 5 30
Residual Ltgnin (%of Wood)
FIGURE 30 Relative changes of the polysaccharide and lignin content showing theinitial (ID),
bulk (BD), and residual (RD) phases of kraft delignification process. (From Ref. 378.)
490 Lai
commercial practice, pulping is usually terminated at the transition between the bulk and
residual phases to maintain fiber quality.
Table 13 illustrates the composition of typical kraft pulps from pine and birch as
compared to the original wood [ 5 ] . The yield of lignin-free pulp is quite low, e.g., 44%
compared to a theoretical 67% for pine. The loss in pulp yield is largely a consequence
of the alkaline degradation of polysaccharides (see Section IV.A), being higher for glu-
comannan (70-80%) than for xylan (40-50%). The complex mixture of nonvolatile hy-
droxyacids plus formic and acetic acid produced consumed the major portion of alkali
charge (around 63%) in kraft pulping [4]. Improved methods for enhanced delignification
and carbohydrate stabilization are still a technical challenge for the pulp and paperindustry.
a. Initial Phase. The initial delignification occursduring the heating-upperiodto
about 15OoC, and is generally lower for softwood (l8-24%) [288,292,380,381] than for
hardwood (25-30%) [373,380,382]. The carbohydrate loss, however, is severe, amounting
to about 50% and 60% of the total yield loss encountered in the kraft pulping of softwood
and hardwood. respectively.
Delignification. In softwood, the initial delignification (ID) was further shownto
consist of two distinguishable subphases: rapid (ID,) and slow (ID,) [292,380]. In the ID,
subphase, the delignification reaction was affected by the chipmoisturecontent[375].
However, the reaction was independent of chemical concentrations within the ranges 0.2-
1.5 M for NaOH and 0.1-0.4 M for NazS 13751, and has a low activation energy ( I O - 12
kcal/mol)[292,375];whereas the reaction in the ID, subphasedependedon the alkali
concentration [380] and has a slightly higher activation energy (17.5 kcal/mol) [292]. In
the case of hardwood(aspen), the IDz subphasewas not discernible, and the ID was
practically unaffected by increasing alkali concentrationabove the 0.1 M level 13801.
Although the ID reaction is generally thought of as being a diffusion-controlled process
[375], the proposedcleavages of some reactive a-arylether bonds[292,303,359]was
supported by a recent report [379].
Carbohydrates Loss. The pulp yield loss encountered in the IDstage is largely
limited to the hemicellulose components, and is noticeably higher and faster for hardwood
and for softwood [380]. Also, in the case of softwood, two subphases were discernible for
carbohydrate loss in the ID stage [292,380]. These variations may be attributed to the loss
of xylan and glucomannans, caused mainly by the physical dissolution and the peeling
process, respectively. The activation energy for the latter reactions is likely in the range
21.2-24.5 kcal/mol, as observed for amylose [229] and isolated glucomannan 12911.
Component Birch
Source: Ref. 5 .
Chemical Degradation 491
b. Bulk Phase. The bulk delignification (BD), which occurs at temperatures above
150°C andremovesan additional 70% of the lignin, involvesmainly the cleavage of
etherified P-aryl ether units [303,359]. It was recently shown [379] to be closely related
to the aryl ether cleavage reactions. The delignification process follows pseudo-first-order
kinetics [373,376,383-3861, and was shown to have activation energies in the range 32-
35 kcal/mol [383-3861, similarto that of lignin modeldimers [315,339]. In addition,
preferential cleavage of the more reactive erythro isomer (Section IV.B.3.b) was observed
in the early BD phase [387,388]. This is consistent with the reactivity of different isomeric
P-aryl ether dimers [3 151.
Thecarbohydrate loss during the bulkphase is causedmainly by the secondary
peeling resulting from the alkaline cleavage of glycoside linkages including cellulose (Sec-
tion 1V.A).
c. Residual Phase. The residual phase, which begins at about 90% delignification,
is characterized by a slow delignification coupled with a rapid carbohydrate-degradation
reaction (Fig. 30). The chemical nature of residual lignin causing its resistance to degra-
dation is still not entirely clear, although it is known to contain few intact P - 0 - 4 ether
structures [389], a highproportion of enol ether units [389-3911, and a highphenolic
hydroxyl group content [392].
The residual delignification was reportedly influenced appreciably only by the hy-
droxide ion concentration and pulping temperature, and had a higher activation energy
than the bulk phase (30 versus 35 kcal/mol) [393]. Also, it was relatively unaffected by
the sulfide concentration [376,393] and the aryl ether cleavage reaction [379]. These ob-
servations are consistent with a contention that the cleavage of carbon-carbon linkages is
important to the residual-phase delignification [359].
2. Aryl-Ether Cleavages
The significance of aryl-ether cleavages in alkaline delignification has been well recog-
nized [303,359], and is clearly illustrated in a gradual increase of phenolic hydroxyl groups
in both dissolved [394,395] and residual pulp [379,394] lignin. Also, the dissolved lignin
has a considerablyhighercontent of this functionalgroupthanthe residual lignin
[394,396-3981.
LaiandKuo [379] recently reported that the impact of aryl ethercleavages on
alkaline delignification, like delignification selectivity (Fig. 30), displayed three discernible
phases as shown in Fig. 31 for a kraft cook of Norway spruce. As indicated, initial delig-
nification was associated with the hydrolysis of few very reactive aryl ether units, whereas
the bulk delignification extending to an approximate 90-95% level was correlated directly
to the aryl ether cleavage reaction. On the other hand, the ether hydrolysis has relatively
little impact on the residual delignification.
U)
.-(
W
Q
80
20
0
:-
0
/
100
Phenolic hydroxyl
I
200
I I I
300
c
+
150°C
160°C
170°C
g r o u p , mmo1/100 g lignin
400
I
FIGURE 31 Influence of temperature in kraft pulping of Norway spruce wood on percent delig-
nification as related to the phenolic hydroxyl group content of pulp residual lignin. (From Ref. 379.)
A. Polysaccharides
The oxidation of polysaccharides, besides a glycol cleavage, may be discussed in terms
of reactions at the reducing end group, at the hydroxyl group, and at the anomeric (acetal)
position, and as illustrated in Fig. 33. The tendency of these oxidations depends consid-
erably on the nature of the oxidants and the hydroxyl configuration.
1. Oxidation Types
a.ReducingEndGroups. The reducingendgroups of polysaccharides are very
reactive and can be readily oxidized to a gluconic acid moiety by a variety of agents [ 161,
including bromine and hypohalites. In alkaline solutions, oxygen [414-4181 and AQ (an-
thraquinone) [419-4211 degrade the cellulose end groups to a mixture of gluconic (129),
mannonic, arabinonic, and erythronic acid moieties. These aldonic end groups are gener-
ally resistant to the peeling reaction, and thus reduce the yield loss in kraft pulping [335]
and in O2 bleaching [422].
h. Glycol-Cleavage Oxidation. Oxidants such as periodate and lead tetraacetate are
effective in the cleavage of a,@-diol units, giving the dialdehyde derivative (131). This
oxidation isvery specific and has been used extensively in both structural analysis and
modification of polysaccharides [3,6,423]. A cis diol unit is generally more reactive than
a trans diol. Thus, the cleavage of a cis diol in the mannose residues proceeds faster than
that of a trans diol in the glucose or xylose residues.
c. Hydroxyl Oxidation. The selectivity of hydroxyl oxidation, as shown for simple
glycosides, varies significantly with the type of oxidants and substrates.
Oxidant. Platinum black was shown having a high selectivity in oxidation of the
6-OHgroup of methyl a- and P-D-glucopyranosides to the correspondinguronic acid
494 Lai
CH,OH CH20H
A R 2 K C 0 2 H
R2 OH OH
128 129
131
c _R2i o OH
132
R 1 "e"'
R2
133
OH
CH20H
- Further
Degration
@"l __I
R2
1
OH
130
135
l
CH20H
E
- R2QR'
137
1
CH,OH
CO,H
OH
128 129
derivatives, which were obtained at 87% and 68% yields, respectively [424]. A slightly
lower preference was observed upon oxidation with nitrogen dioxide, which also gave the
2-keto, 3-keto, and 4-keto derivatives [425]. Reaction with chromic acid [426] or Fentons's
reagent (Fe" plus H20,) resulted in rather nonspecific oxidations [427].
On the other hand, the initial oxidation of methyl P-D-glycoside in oxygen-alkali
[428-4311 or peroxide-alkali [429] solutions occurred mainly at the 2-OH or 3-OH group.
The resulting ketoglucosides could be further oxidized to the 2,3-dicarbonyl intermediate
(133, leading to the formation of methyl 2-carboxy-P-D-pentofuranoside derivative (136).
Different Glycosides. For 0,-alkali oxidation (in 0.5 MNaOH at 120°C for30
h), the reactivity of methyl glycosides (as percent oxidation) increased in the order p-
xyloside (12%) < a-glucoside (18%) < p-glucoside (20%) < a-mannoside (28%) [431].
Also, 1,5-anhydroribitol was more reactive than 1,5-anhydroxylitol [432], whereas higher
reactivity was observed for methyl P-D-ribopyranoside than for methyl p-D-xylopyrano-
side [20]. Thus, glycosides containing an axial hydroxyl group seem to promote the oxi-
dative degradation process.
Also, the reactivity of methyl glycosides toward an aqueous bromine oxidation [433]
maybe related to their hydroxylconformation.Asindicated in Table 14, methyl a-D-
glucopyranoside yielded mainly the 2-keto and 4-keto derivatives, whereas oxidation of
the corresponding@-glucosidewas less specific and resulted in additional formation of
the 3-keto compound. The a-anomerof manno- and galactopyranoside, containing anaxial
hydroxyl group at the C2 and C4 positions, respectively, gave mainly the corresponding
2-keto and 4-keto products. The p-anomer of these two glycosides also gave a noticeable
amount of the 3-keto derivative. Steric factors were assumed to be a contributing factor
in determining the specificity of the hydroxyl oxidation.
cl. Anomer-ic Site. Anotheroxidationtype is direct attack ontheanomericcarbon
(Cl), resulting in glycosidic cleavages and the formation of a gluconic acid moiety. These
oxidative degradations occurred in chlorination [434,435], and were reported to be a dom-
inant mechanism in ozonation [436-4401.
2. Cellulose
Controlled oxidation of cellulose provides a means to prepare useful cellulose derivatives
[423,441], but it is generally difficult to achieve good selectivity. Most of the cellulose
Galactoside cy - - 31 34
P - IO IC) 36
496 Lai
degradations encountered during pulp bleaching [412,413] are very undesirable, resulting
in the loss of fiber strength.
a. Controlled Oxidation. A classical example is in the aging of viscose to obtain a
proper pulp viscosity for making cellulose xanthate and other products [4,442]. An alkali-
0, system is normally used to induce the oxidative cleavage of glycosidic linkages. Other
controlled oxidations are related to the preparation of oxycellulose.
Glycol Cleavage. Oxidation with periodic acid selectively converts cellulose to the
2,3-dialdehyde derivatives (131)(reaction B of Fig. 33), which are useful as intermediates
for preparing nitrogen-containing products [443].
The initial periodate oxidation of cellulose, like other chemical reactions, is largely
limited to the readily accessible component, and has been used to indicate the accessibility
of cellulose substrates [444,445].
Hydroxyl Oxidation. Amongvariousoxidants,nitrogendioxidewas relatively
more selective in oxidation of the 6-OH group to yield an oxycellulose containing 87.5%
of the glucuronic acid moiety and about 6% of the 2-keto and 3-keto groups [446]. Oxi-
dation with DMSO-acetic anhydride (AC,O) gave a good yield of the C6-aldehyde (47-
62%) and a significant amount of the C2- or C3-keto groups [447]. However, oxidation
with DMSO-AC20 in a DMSO-PF solvent system gave selective oxidation at the 3-OH
group [448]. This was suggested as being attributed to a reversible formation of hydroxy-
methyl and polyoxymethyleneol groups at the C2 and C6 positions.
b. UncontrolledOxidation. Oxidativedepolymerizationofcelluloseandhemicel-
lulase is the most undesirable side reaction encountered in pulp bleaching [412,413], no-
tably with alkali-oxygen, peroxide, and ozone. It involves mainly an oxidative cleavage
of glycosidic linkages resulting in a reduction of pulp viscosity, which may lead to a loss
of fiber strength. The cleavage reaction may result from direct oxidation (reaction F, Fig.
3 3 ) , e.g., with ozone, or may be promoted by a ketol group present in the polymer. Ketols
formed at the C2, C3, or C6 position are known to readily undergo an alkali-induced p-
alkoxy elimination reaction [422,449-45 I].
Figure 34 illustrates one of the most common depolymerization reactions facilitated
by a C2-keto group. Such a process, like a peeling reaction, can proceed under mild alkali
conditions. For 0, and peroxide bleaching conducted in an alkali medium, the ketol for-
mation and the p-elimination process can occur in a successive manner. For bleaching in
an acidicmedium, the keto group formed also initiates a similar p-elimination process
during a subsequent alkaline stage.
-Re
@oeoR
CH2OH CH2OH
CH2OH @OR
.OH + OH
R
OH 0 OH OH
138 128 139
Chemical Degradation 497
3. Hemicellulose
The oxidative degradation of hemicelluloses is expected to be similar to that of cellulose
in reaction type, but it will be more extensive because of the relatively high accessibility.
The oxidation of birch xylans with hypochlorite resulted in the formation of a variety
acids including aldonic acid end groups, uronic acid, formic acid, and glycolic acids 14601.
The degradationreactions of corn cob [ 4611 and birth [462] xylans in oxygen-alkali
treatments were shown to be very similar to that of cellulose.
B. Lignin
Theoxidative degradation of lignin involvesextensivecleavages of the side-chain and
phenyl units, and has been discussed extensively by Gierer [35,360,361], notably on the
development of a general concept. This section is limited to a brief outline of the major
types of lignin oxidation and 0, delignification. These reactions, unlike delignification i n
498 Lai
pulping, are characterized by the cleavages of carbon-carbon linkages and the formation
of acidic groups from ring degradations.
1. ReactionTypes
Most of the bleaching systems are not selective and can achieve a wide range of reactions,
which are outlined in Figs. 35 and 36.
U . Substitution
Reactions. The substitution reactions commonlyobservedwith
chlorine and nitric acid C4631 include electrophilic substitution at anunsubstituted ring
position (141) and at the C1 position with a side-chaindisplacement (142). The latter
displacement reaction depends on the nature of the benzyl unit, and is possible with a
benzyl alcohol unit.
b. Side-ChainCleavages. Theoxidativecleavageof lignin side-chain units is
largely confined to the a-p linkage of ring-conjugated structures containing an a-carbonyl
(143) or an ethylenic (145) group. These functional units may be generated during the
pulping stage, e.g., stilbene types, or may be formed in situ during the bleaching process.
Both unsaturated units are very reactive under alkali-0, or peroxide bleaching conditions,
whereas the ethylenic units are also rapidly degraded in ozonation.
c. Ring Degradation. The oxidation of aromatic nuclei to dicarboxylic acids (Fig.
36) represents one of the most distinct features of bleachingactions, and has been observed
in nearly all the common oxidants, including 03,peracetic acid, CIO?, Cl,, and alkaline-
0, and-peroxide. The ring openingbetween the C3-C4 bondto yield muconic acid
5CH0
RI RI +
0 CHOH
OR
A
X=CPNO?
OCH,
* xQ
CHOH
OH
+
X
QOCH3
OH
140 141 142
I
H-C-R
143 144
I
l O=FR
CR
Chemical Degradation 499
I 1
OH OH 0
147 148 149
I 1
___)
B
C CO,H
150 151
v
R P
HO
OH
154 152 153
derivatives (151) may proceed directly (reaction B), or may be preceded by a demethox-
ylation (148) or an ortho-quinoneformation (149). Additionally, ring degradationwith
peracetic acid may be initiated by a hydroxylation (152) followed by oxidation to maleic
derivatives (154) 1464,4651. This pathway is probably also applicable to an acidic hydro-
gen peroxide system.
2. 0, Delignification
In the current trend of producingbleachablechemical pulps, 0, bleachinghasbecome
nearly a standard process following kraft pulping, but it can only readily remove approx-
imately 50% of the pulp residual lignin without causing excessive degradation of the tiber
[466]. Despite extensive efforts, the nature of the residual lignins, especially those resistant
to 0, delignitication, is still not well established.
Lignin model compound reactions reveal that phenolic and enolic structures are the
major reactive sites under 0, bleaching conditions [35,360,3611, andtheir chemical nature
has a considerableinfluence on the reactivity 14671. The biphenyl unit is generally accepted
as being relatively resistant to degradation [468].
Recent reports [469,470] indicate that the phenolic lignin DPM (diphenylmethane)
model dimers are quite reactive in alkali-0, solution, resulting in the formation of mo-
nomeric and oxidative coupling products, as illustrated in Fig. 37 for an a-5 dimer (155)
(4691. However, a preliminary analysis on kraft pulp lignins s e e m to suggest thatthey
contain some DPM-like structures that are resistant to degradation in 0, bleaching 14711.
These stable “diphenylmethune” units were suggested to be probably o f an etherified type.
500 Lai
OH NaOH
OCH, OCH,
_,, HOH,C
OH OH
155 156 157
i i
VI. TOPOCHEMISTRYOFDELlGNlFlCATlON
The topochemistry of delignification refers to the relative rates of lignin being removed
from various morphological regions. It has an appreciable impact on the overall pulping
and bleaching efficiency. If the lignin component in the middle lamella (ML) and cell
corner (CC) can be selectivelyremoved, this will greatly facilitate the fiber separation
process, and thus improve the pulping efficiency. Also, the morphological distribution of
residual lignin will have a significant bearing on the bleachability and papermaking prop-
erties of pulp fibers.
This subject has been studied by various groups, among which Goring and co-work-
ers deserve special mention [472]. They reported, in kraft cooking of spruce wood [473],
that the initial delignification occurred preferentially i n the secondary wall (SW); and at
about 50% delignification, lignin in the ML and CC regions began to dissolve faster than
the SW lignin, leaving most of the residual lignin in the SW (Fig. 38). A smaller topo-
chemical effect was observed in sulfite cooking, whereas none was found in chlorite treat-
ments [473].
The bulk of reported data[472,475-4781 generally support a contention that the
nature and extent of topochemical effects vary significantly among different pulping sys-
tems. They are, however, not entirely consistent even for a similar process obtained by
different authors. For example, Giidda [479] observed an opposite linding to that of Procter
et a l . [473] in kraft cooks. The initial kraft delignification of pine wood was shown to be
faster in the ML than in the SW region.
The reported contrary results, although they were obtained with different species
(pine versus spruce), may be largely attributed to variation in the experimental conditions
used. This contention is supported by an interesting finding of Westermark and Samuelson
Chemical Degradation 501
-
-
-
-
I I I I I I I I I I I I -
0 20406080 0 2 0 4 0 6 0 8 0 020406080w3o
DELlGNlFlCATlON OF W O D
[480]. They observed that the ML tissue was delignified more rapidly than the whole wood
by using a highliquor-to-wood ratio (40:l), whereas the oppositewasfound in a low
liquor-to-wood ratio cook (4: 1).
Similarly, as discussed by Chen et al., the reported topochemistry of acid sulfite and
organosolv delignification [477,478] was not entirely consistent. Most reports indicate that
the ML lignin was chemically attacked by the liquor at the early stage of a cook. Also,
the initial removal of the ML lignin was either faster or comparable to that of the SW
lignin. In contrast, based on studies of isolated tissue, the SW lignin was clearly more
responsive to acidic degradation than the ML lignin. These seemingly conflicting findings
can only be reconciled as the topochemical delignification being the result of an interplay
among the morphological, physical, and chemical properties of the cell wall matrix. This
subject merits further investigation.
REFERENCES
I. D. Fengel and G. Wegener, Wood-Chemistry,Ultrastrrrcture, Rerrctions. Walter de Gruyter,
Berlin (1984).
2. Y.-Z. Lai, in WoodandCellulosic Chenzistry (D. N.-S.Hon and N. Shiraishi, eds.), Marcel
Dekker, New York, p. 455 (1990).
3. E. Sjostrom, Wood Chenzistry-Furzdcrrnentnls rrtzd Applicatiom, 2nded.,AcademicPress,
New York, London ( 1993).
4. S. A. Rydholm, Pulping Processes, Wiley-Interscience, New York (1965).
S. E. Sjostrom, Tappi, 60(9): 15 1 ( 1 977).
6. T. E. Timell. A h . Cmhhydr: Chern., 20:409(1965).
7. D. A. I. Goring, in Lignin: Properties and Muterinls (W. G. Glasser and S. Sarkanen, eds.),
ACS Symp. Ser. No. 397, American Chemical Society, Washington, DC, p. 2 (1989).
8. C.-L. Chen, in Wood Structure and Cornpositions (M. Lewin and I. S. Goldstein, eds.), Marcel
Dekker, New York, p. 183 ( 1991).
9. Y.-Z. Lai and K. V. Sarkanen, in Lignin (K. V. SarkanenandC.Ludwig,eds.), Wiley-
Interscience, New York, p. 165 ( I97 1 ).
IO. J. A. Rendleman, Jr., A h . Chern. & K , No. l 1 7 5 1 (1973).
I 1. A. H. Haines. A h . Carboh>vlt: Chenz., 33: I 1 (1976).
12. L. R.Schroeder, V. M. Gentile,andR. H. Atalla, J. Wood Chenz. Techno/.. 6(1):1 (1986).
13. 0. Samuelson, Proc. l Y K l h r . S w p . WoodPulping Cham.. SPCI.Stockholm, vol. 2, p. 78
(1981).
502 Lai
14. V. M. Gentile, L. R. Schroeder, and R. H. Atalla, in The Structure r$Ce//u/ose (R. H. Atalla,
ed.),ACSSymp. Ser. No. 340, AmericanChemical Society,Washington, DC, p. 272
(1987).
15. J. C. Speck, Jr., A h . Carbohyd,: Chem., 1 3 5 3 (1958).
16. J. W. Green, in The Curbohydrutes (W. Pigrnan and D. Horton, eds.), Academic Press, New
York, vol. IA, p. 1101 (1980).
17. J. M. Sugihara, Adv. Carbohydr. Chem., 8:1 (1953).
18. R. W. Lenz, J. Am. Chern. Soc., 82: I82 (1960).
19. J. A. Rendleman, A h . Carbohyd,: Chem., 21:209 (1966).
20. D. 0. Hearne, N. S. Thompson, and L. R. Schroeder, J . Wood Chern. T e c h ~ ~ o11(3):307
/.,
(1991).
21. E. Adler, Wood Sei. Techno/., / l :169 (1977).
22. Y.-Z. Lai, in Methods in Lignin Chemistry (S. Y. Lin and C. W. Dence, eds.),Springer-Verlag,
Berlin, Heidelberg, p. 423 (1992).
23. N. Terashima, K. Fukushima, and T. Imai, Ho/<fi,r.sch.,46:271 (1992).
24. A. Maurer and D. Fengel, Ho/i$or.srh..46417 (1992).
25. A. Maurer and D. Fengel, Hol&rsch., 46:471 (1992).
26. H. H. Nimz, U. Tschirner, and M. Stahle, J . Wood Chem. Techlo/., 4(3):265 (1984).
27. M. Hauteville, K. Lundquist, and S. von Unge, Actcc Chern. S c m d . , 840:31 (1986).
28. E. Adler, G. E. Miksche, and B. Johansson, Ho/i$or.sch., 22:171 (1968).
29. K. Lundquist, Actcc Chem. Scurld., B35:497 (1981).
30. Y.-Z. Lai and X.-P. Guo, Ho/;fi)rsch.,46:3I 1 (1992).
31. M. Erickson, S. Larsson, and G. E. Miksche. Actcc Chuu. Scnrrd.. 27:903 (1973).
32. G. Brunow, J. Sipila, K. Syrjanen, P. Karhunen, H. SetiilY, and P. Rummakko, Pro(.. X t h
S j w p . Wood Pulping Chemistry, Helsinki, p. 33 ( 1995).
33. T. Koshijima, T. Watanabe, and F. Yaku. in Lignin: Propertic~sn r l d Matc.rin/s (W. G. Glasser
and S. Sarkanen. cds.). ACS Symp. Ser. No. 397. American Chemical Society, Washington,
DC, p. I I (1989).
34. J. Gierer, Wood Sci. Techol.. 19289 (1985).
35. J. Gierer, Wood Sci. Techrwl.. 20: I ( 1986).
36. S. Y. Lin, in Progress i n H i o r ~ ~ c r s Conwr.sion
.s (D. A. Tillman and E. C. Jahn, eds.), Academic
Press. New York, vol. 4. p. 3 1 ( I 983).
37. D. R. Robert, M. Bardet. G. Gellerstedt, and E. L. Lindtors. J . Wood Chcwr. 7 k h o l . . 4:239
(1984).
38. H.-M. Chang. E. B. Cowling. W. Brown. E. Adler. and G. M. Mikshe. Ho/;/?)r.sc.h.,29:153
( 1975).
39. Y.-Z. Lai and X.-P. Guo. Wood Sci. TL'(.IIIIO/., 25:467 ( 199 I ).
40. E. Adler and K. Lundquist, Acta Chcwr. SUM/., /5:223 (1961).
41. C. H. Ludwig, B. J. Nist. and J. L. McCarthy, J . Afjl. Chc,rn. Soc.. XO: I l96 ( 1964).
42. N. Morohoshi and A. Sakakibara, Mokuxri Gctkktrishi. 17:393 ( 197 I ).
43. M. Funaoka, l. Abe. and V. L. Chiang. in M e t h o d s i r l L i p t i r l C / ~ e ~ r i . ~(S. y Lin and c. W.
t ) .Y.
Dencc. eds.). Spring-Verlag. Berlin, Heidelberg. p. 369 ( 1992).
44. D. C. C. Smith. J. Chert/. Soc.. 2347 (1955).
45. Y. Nakamtlra and T. Higuchi. Ho/;fi)r.sc.h..30(6):187 (1976).
46. Y. Nakamura and T. Higuchi. Cc//.Chcwl. 7 k h t 1 o l . . 12: 199 ( 1978).
47. Y. Nak:unura and T. Higuchi, CO//.Chcw. 7 i d f l l o / . . /2:209 ( 1978).
48. J. R. Obst. Ecppi, 634): l09 ( 1982).
49. C.-L. Chen. i n Mc,/hot/,s irl Li<yllirr Chcvtristry (S. Y. Linand C. W. Dence, etls.). Springer-
Verlng. Berlin. Heidelberg. p. 465 (1992).
SO. T. Enkvist, J. Turenen. and T. Ashorn, 7hppi, 45: 118 ( 1962).
51. G. Gcllerstcclt. S I Y . I ~Ptrppc,rstit/rl.,
,~ 7Y:537 ( 1976).
52. A. J. Mitchell. CO//.Chef)/.7 i ~ r . h I r o l . . 16:87 (19x2).
53. R. Stomberg and K. Lundquist. Nordic. P I / / / )P q w r K c s . ./.. 9:37 ( 1974).
Chemical Degradation 503
102. B. Philipp, V. Jacopian, F. Loth, W. Hirst, and G. Schulz, in Hydrolysis of Cellulose: Mech-
a n i s m of Enzyrnutic ccnd Acid Cuta1ysi.s (R. D. Brown and L. Jurasek, e&.), Adv. Chem.
Ser. No. 18 I , American Chemical Society, Washington, DC, p. 127 (1979).
103. A. Sharpies, in Cellulose ancl Cellulose Derivutives (N. M. Bikales andL. Segal, e&.) Wiley-
Interscience, New York, part V, p. 991 (1971).
104. M. A. Millett and M. J. Effland, in H.ydrolysis of Cellulose (R. D. Brown and L. Jurasek,
eds.), Adv. Chem. Ser. No. 181, American Chemical Society, Washington, DC, p. 7 1 (1979).
105. J. F. Saeman, Ind. Eng. Chenl., 37( 1):43 (1945).
106. C. J. Biermann, Adv. Crrrbohydr. Clzem. Biochenz., 46:25 1 ( I 988).
107. K. Freudenberg and C. Blomqvist, Be,: Drsch. Chem., B68:2070 (1935).
108. I. S. Goldstein, H. Pereira, J. L. Pittman, B. A. Strouse, and F. P. Scaringelli, Biotech. Bioeng.
Synzp., 13:17 (1983).
109. E Bayat-Makooi and I. S. Goldstein, in Cellulose and Its Derivatives (J. F. Kennedy, G. 0.
Phillips, D. J. Wedlock, and P. A. Williams, eds.), Ellis Horwood, Chichester, U.K., p. 135
( 1 985).
1 IO. K.Garves, Holzforsch., 4 7 149 (1993).
111. N. Bhandari, D. G. Macdonald, and N. N. Bakhshi, Biotechnol. Bioeng., 26:320 (1984).
112. D.H. Foster and A. B. Wardrop, Austrczl. J. Sci. Res., A4:412 (1951).
113. K. Garves, Cell. Chem. Technol., 18:3 (1984).
114. J. Bouchard, G. Garnier, P. Vidal, E. Chornet, and R. P. Overend, Wood Sci. Techno[., 24:
1S9 ( 1990).
1 15. K. C. B. Wilkie, Adv. Carbohydr. Chern. Biochem., 36:215 (1979).
116. K. Shimizu, in Wood and Cellulosic Chemist~y(D. N.-S. Hon and N. Shiraishi, eds.), Marcel
Dekker, New York, p. 177 (1990).
117. T. E. Timell, Wood Sci. Technol., /:45 (1967).
118. J. N. BeMiller, Adv. Carbohyd. Chetn., 22:25 (1967).
119. J. N. BeMiller, in Starch Chernistr~and Technology (R. L. Whistler and E. F. Paschall, eds.),
Academic Press. New York, vol. I , p. 495 ( 1 965).
120. E. Springer and L. L. Zoh, Tnppi, 51(5):214 (1968).
121. G. E. Annergren and I. Croon, Svensk Pnpperstidn., 64:618 (1961).
122. H. E. Korte, W. Offermann, and J. PUIS,Hol&r.sch., 45:419 (1991).
123. S. K. Banerjee and T. E. Timell, Tappi 43: 10 (1960).
124. K. Lundquist, R. Simonson, and K. Tingsvik, Svensk Papperstidn.. 86:R44 (1983).
125. J. K. Hamilton, E. V. Partlow, and N. S. Thompson, J. Am. Chern. Soc., 82:451 (1960).
126. T. E. Timell. Tuppi, 45:734 (1962).
127. G. E. Annergren, I. Croon, B. F. Enstrom, and S. A. Rydholm, Svensk Papperstidn., 64386
(1961).
128. I. Croon, Svensk Pupperstidn., 66: 1 (1963).
129. K. Lundquist, Appl. Poljvner. Sci., 28: 1393 ( 1976).
130. Y.-Z. Lai, in Chelniccrl Modifccrtion of Ligr~ocell~rlo.sic~
Mcrtericr1.s (D. N.-S. Hon, ed.), Marcel
Dekker, New York, p. 35 (1996).
131. J. M. Harkin. Aclv. Chem. Set:. 59:65 (1966).
132. T. Ito, N. Terashima, and S. Yasuda, Mokuzai Gakktrishi, 27(6):484 (1981).
133. B. Johansson and G. E. Miksche. Actu Chern. Scarzd.. 26:289 (1972).
134. G. J. Leary and D. A. Sawtell, Hol<fi)r.sch.,38( 1):53 (1984).
135. B. KoSikov6, D. Joniak, and L. Kosikovi, Holzforsch., 33( I ) : I l (1979).
136. M. Meshgini and K. V. Sarkanen, Holdorsch.,43(4):239 (1989).
137. K. V. Sarkanen and L. H. Hoo, J . Wood Chern. Technol., I( 1): 1 1 ( 198 1).
138. H. Nimz, Angetc: Chern.. 5:843 (1966).
139. H. Nimz. U. Tschirner, and M. Roth, Proc. I n t . S y n p . Wood Pulping Chemistry, Tsukuba.
Japan, vol. I , p. 90 ( 1 983).
140. U.Tschirner, UntersuchungenzumLigninabbau in Wassrig-organischenLiisungsmitteh
Genehmigte Dissertation, Universitiit Karloruhe (1984).
Chemical Degradation 505
226. L. Paszner, A. A. Quinde.and M. Meshgini, Int. Symp. Wood Pul/in<q Clzern., Vancouver,
Canada, p. 235 (1985).
227. D.Clayton, D. Easty, D. Einspahr, W. Lonsky, E. Malcolm, T. McDonough.L.Schroeder,
and N. Thompson, in Pulp and Paper Manujiacture. Vol. 5, Alkaline Pulping (T. M. Grace
and E. W. Malcolm, eds.), The Joint Textbook Committee of the Paper Industry, Atlanta, p.
3 (1989).
228. W. Pigman and E. F. L. J. Anet, in The Crcrbohydrate Chenlistty and Biocherni.str~v(W. Pigman
and D. Horton, eds.), Academic Press, New York, vol. IA, p. 165 (1972).
229. Y.-Z. Lai and K. V. Sarkanen, J. Polymer Sci., Port C, 28:15(1969).
230. G.MachellandG. N. Richards, Tappi, 4/:12 (1958).
23 1. 0. Sarnuelson and A. Wennerblorn, Svensk Pappersticfn., 57:827 ( I 954).
232. M. H. Johanssonand 0. Samuelson, Cczrbohydt: Res., 34:33(1974).
233. I. ForsskHhl, T. Popoff, and 0. Theander, Carbohydr. Res., 48:13 (1976).
234. R. L. Colbran and G. F. Davidson, J . TextileInst., 52:T73 (1961).
235. D. W. Haas. B. F. Hrutfiord, and K. V. Sarkanen, J . App/. Polyrner Sci., //:S87 (1967).
236. K. Christofferson and 0. Samuelson, Svensk Papperstidn., 63(20):729(1960).
237. K. Christofferson and 0. Samuelson, Svensk Pqqwrstidn., 65( 13):517 (1962).
238. U. Albertsson and 0. Sarnuelson, Svensk Papperstidn., 6324): 1001 (1962).
239. H. Meier, Svensk Papperstirh.. 65( l6):589(1962).
240. B. Lindberg, 0. Theander, and J.-E. Uddegard, Svensk Pq~perstidn.,69( 10):360 (1966).
241. R. Malinenand E. Sjostriim, Paper; Puu, 565395 (1974).
242. Y.-Z. Lai, Carlmhydr: Res., 28: I 54 ( 1 973).
243. Y.-Z. Lai and D. E. Ontto, J . Appl. Polymer Sci.. 233219 (1979).
244. V. L. Chiang and K. V. Sarkanen, J . Wood Chetn. Techno/., 5(2):203 ( 1966).
245. D. W. Clayton and L. M. Marraccini, Svensk Pqqwrstidn., 69(9):31I (1966).
246. A. R. Procter, Cell. Clwrn. Technol., 4:269(1970).
247. V. L. Chiang and K. V. Sarkanen, J . Wood Ckem. Technol., 4( 1 ): 1 (1984).
248. F. L. A. Arbin, L. R. Schroeder, N. S. Thompson, and E. W. Malcolm, Cc4l. Chetn. Techno/..
15:523 (1981).
249. 0. Sarnuelson and L. Stolpe. 7hppi, 52(9):1709 ( 1969).
250. H. Kolrnodin and 0. Samuelson, Svensk Pcpper.stidn., 73(4):93 (1970).
25 I . T. Vuorinen and E. Sjostrom, J. Wood Chern. Technol., 2(2): I29 ( 1982).
252. L.Liiwendahl and 0. Samuelson, Actcc Chern. Scand. B3353 1 ( 1 979).
253. L. Lowendahland 0. Samuelson, 7kppi 61(2):19(1978).
254. U. Carlson and 0. Sarnuelson, Svemk Pqqx~r.~tidn., NO( I7):549 ( 1979).
255. K. Ruohoand E.Sjiistrom. %ppi. 6(7):87(1978).
256. T. Vuorinen. Corbolrytlr. Res.. IIh:61 (1983).
257. T. Vuorinen. Cccrhohydc Res., 14/:307 (1985).
258. P. Ahlgren, A. Ishizu, I. Szabo, and 0. Theander, S w n s k I ’ l / / ~ / ) ~ t . . ~ / ; ~ 7/(9):355
ltl., ( 1968).
259. B. Alfredsson and 0. Samuelson, S ~ w ~ P~cpprstidn.,
sk 72( 1):361 (1969).
260. Y.-Z. Lai. Pro(.. 1081 Int. Sy~np.Wood Pulping Cl~c~tn., SPCI, Stockholm, vol. 2. p. 26 (1981).
261. W. G.Overend, in The Ccr~hol~ytlrtc~~~.s-CChc~r~~i.stry trntl Bioc~herni,s/~~v (W. Pigman andD.
Horton, eds.). Academic Press, New York. vol. I A , p. 279 (1972).
262. C. E. Ballou, Ctrrbohydr: Chert/.. Y:SY ( 1954).
26.3. R. C. Gasman and D. C. Johnson, J . Org. Chern.. -?I: I830 ( 1966).
264. C. M. McCloskcy and G. H. Colcman. J . Orq. Chern., I O : 184 ( 1985).
265. Y.-Z. Lai and D. E. Ontto. Ctrrbohytlt: K t ~ s . .75:sI (1979).
266. G. H. Coleman, M ~ t h Crrrholtyrlt:
. Clwrn.. 2:397 ( 1963).
267. R. E. Brandon. L. R. Schroedcr, and D. C. Johnson. ACS Sytnp. Set:. /0:125 (1975).
268. D. A. Blythe and L. R. Schroeder, J . Wood Clw~tt.7 k h 1 1 o / . , 5(3):3 l3 ( 1985).
269. D. R. Dimmcl and R. M. Knylor. Proc. I n t . Syrrrp. Wood f / / / p i n g Cher~r..Helsinki. vol. I , p.
309 (lY95).
270. J . Janson and B. Lindbcrg. A t ~ t rClwrn. S c m d . . 14(9):205I (1960).
508 Lai
367. J. Marton, T. Marton, S. I. Falkehag, and E. Adler, Adv. Chem. Se!:, 59:125 (1966).
368. K. H. Ekman, E&, 48(7):398(1965).
369. N.Terashima, H. Araki, and N.Suganuma, MokuzaiGakkaishi, 23(7):343 (1977).
370. H. Araki, Y.Tomimura, and N.Terashima, Mokuzai Gakkaishi, 26(2):102 (1980).
37 I . T. J. Fullerton, J . Wood C h e m Technol., 7(4):441 (1987).
372. J. Gierer and W. Wannstrom, Holdorsch., 38(4): 181 (1984).
373. P. J. Kleppe, Tuppi, 53(1):35 (1970).
374. P. Axeglrd, S. Norden, and A.Teder, Svensk Papperstidn., 8/(4):97 (1978).
375. L. Olm and G. Tistad, Svensk Papperstidn., 82(15):458(1979).
376. A.TederandL. Olm. Puperi PULI,63(4a):315(1981).
377. P. Axegard and J.-E.Wiken, Svensk Pappersticln., %:R178 (1983).
378. V. MaSura, Cell. Chem. Technol., /R421 (1984).
379. Y.-2. Lai and C.-H. Kuo, Ahstr: 40th Anniverscuy Cor$ of the Japnrz Woocl Research SocieV,
Tokyo, p. 392 ( 1995).
380. Y.-2. Lai, A. R. Czerkies, and I. L. Shiau, Appl. Polymer S.ynzp., 37:943 (1983).
38 I . J . R. Obst. Toppi, 68(2):100 ( 1985).
382. R. Aurell, SIvnsk Prrpperstidn.. 67(2):43(1964).
383. H. Wilder and E. J. Daleski, Jr., Tappi, 48(5):293(1965).
384. T. N. Kleinert, fifppi,49(2):53 (1966).
385. S. Lemon and A. Teder, Svensk P q p x s f i d n . ,7h( 11):407 (1973).
386. G. Wilson and A. R.Procter. Pulp Paper Mrrg. C m . , 7/:T483 (1970).
387. Y. Matsunloto, A. Ishizu and J. Nakano, Hol$or.sc/t., 4O(Suppl.):81 (1986).
388. Z.-H. Jiang and D. S. Argyropoulos, J. Pulp PuperSci.. 20(7):J183 (1994).
389. G. Gellerstedt and E. L.Lindfors, Sverzsk Papperstih., 87(9):R6I (1984).
390. D. R. Robert,M. Bardet. G. Gellerstedt,andE.-L. Lindfors. J . Wood Clzeru. Techno/., 4 ( 3 ) :
239 (1984).
391. A. M.Croucher,J. A. Lloyd, K. Nishi. M. F. Pasco,and I . D. Suckling, Pro(,. Int. Synrp.
Wood Pulping Chen~..Helsinki, vol. I , p. 201 ( 1995).
392. Y.2.Lai. S.-P. Mun, S.-G. Luo, H.-T. Chen, M. Ghazy, H. Xu. and J. E. Jiang. Ho/;for.sch.,
49319 (199s).
393. C. T. Lindgren and M. E. Lindstrom, J . Pulp Puper Sci., 22(8):J290 (1966).
394. G. Gellerstedt and E.-L. Lindfors, S V C I I SPnppersfidn.,
~ 87( 15):R 1 15 ( 1984).
395. E. Evstgneycv. H. Maiyorova. and A. Platonov, 7kppi J.. 76(5):l77 (1992).
396. T. Yamasaki. S. Hosoya.C.-L.Chen, J. S. Gratzl, andH.-M.Chang, Proc. 1981 Int. S J w p
Wood Pulping Chew.. SPCI. Stockholm, vol. 2, p. 34 (I981).
397. G. Gellerstedt and E.-L. Lindfors, Pro(.. h r . Pulp Blenching Cor$. Stockholm. vol. I . p. 78
(1991).
398. G. Gellerstcdt. K. Gustafsson. and R. A. Northey, Nordic Pulp Ptrper Res. J.. 3(2):87(1988).
399. G. Gellerstedt and E.-L. Lindfors, Ho1;fimch.. 38(3):15I (1984).
400. G. Gellerstedt and K. Gustafsson. J. Wood C / w m Trchrzol.. 7( 1):65 (1987).
401. V. L. Chiang and M. Funaoka. Hd;forsch., 42(6):385(1988).
402. V. L. Chiangand M. Funaoka. H ( J / ~ ) I x .44(2):147
~.. (1990).
403. F. D. Chan. K. L. Nguyen, and A. F. A. Wallis. J . Wood Che/u. 7 k h r w l . . 15(3):329 ( 1995).
404. P. M. Froass. A. J. Ragauskas. T. J. McDonough, and J. E. Jiang, Pro(.. h i . P d p Rlmchirfg
CO$. Washington.DC.vol. 1. p.163 (1996).
405. B . Wortling. E. Turunen. and J. Sundquist, Nordic Pulp P q w r Res. J.. 3 : 144 (1992).
406. T. Iversen and S. Wiinnstriim, Ho/;forsch, 40: 19 (1986).
407. A. Isogai. A. Ishizu. and J. Naknno, .I. Wood Chenf. 7 k h 1 ? 0 / . . 7:463 ( I 987).
408. U. Westermark and K. Gustafsson. Ho1;forsc.h.. 48(Suppl.):146 (1994).
409. 0. Karlsson and U. Westermark. Proc.. Eppi Pltlpir~gCon/:.San Diego. v o l . I , p . I ( 1994).
410. J. Sundquist and T. Rantanen, Ptrperi Puu. 65( I I ):733 (1983).
41 I . J. Sundquist. T. Rantanen. and L. Hovi, P a p 4 P ~ I 66(3):253
~I. (1984).
412. R. P. Singh (cd.), The Bleccching o f P d p , 3rd ed.. Tappi. Atlanta (1979).
Chemical 511
413. C. W. Dence and D. W. Reeve (eds.), Pulp Bleaching: Principles n t d Prccctice, Tappi Press,
Atlanta (1996).
414. 0. Samuelson and L. Thode. Tappi, 52( 1 ):99 (1969).
415. 0. Samuelson and L. Stolpe, Tappi, 52(9):1709 (1969).
416. R. Malinen and E. Sjostrorn, Pap. Puu, 54:45 1 ( 1 973).
417. R. Malinen and E. Sjostrom, Pap. Puu, 55547 (1973).
41 8. T. Vuorinen and E. Sjostrom, J . Wood Chern. Technol., 2 : 129 ( I 982).
419. L. Lowendahl and 0. Sarnuelson, Acta Chern. Scand., B33:531 ( I 979).
420. U. Carlson and 0. Samuelson, Svensk Pqyxrstidn., 82:48 (1979).
421. T. Vuorinen, J . Wood Chen~.Technol., 13(1):97 (1993).
422. J. S. Gratzl, Tappi Oxygen Delig~~i'ccctior~Symp. Notes, Tappi Press, Atlanta, p. 1 (1990).
423. M. Yalpani (ed.), Polysaccharides: Synthesis, ModiJcarions, and Stnccture/Proj?erties Relcr-
tions, Elsevier, Amsterdam (1988).
424. S. A. Barker, E. J. Bourne, and M. Stacey, Chem. Ind. (London), 970 (1951).
425. A. Assarsson and 0. Theander, Acta Chern. Scar~d..1 8 5 5 3 (1964).
426. A. Assarsson and 0. Theander, Acta Chern. Scand., 183727 (1964).
427. A. N. de Belder, B. Lindberg, and 0. Theander, Acta C h e n ~Scanrl., 17: 1012 (1963).
428. B. Ericsson, B. Lindgren, 0. Theander, and G. Petersson, Carbohyd~:Res.. 23:323 (1972).
429. H. Kolrnodin, Carbohydr. Res., 34:227 (1974).
430. B . Ericsson and R. Malinen, Cell. Chern. Technol., 8:327 (1974).
431. R. M. Malinen and E. Sjostrom, Cell. Chenz. Technol., 9:231 (1975).
432. E. C. Millard, L. R. Schroeder, and N. S. Thompson, Carhohydr: Res., 56:259 ( 1 977).
433. 0. Larm, E. Scholander, and 0. Theander, Carbohydr: Res., 49:69 (1976).
434. P. S . Fredricks, B. 0. Lindgren, and 0. Theander, Svensk Papperstidn., 74597 (I971 ).
435. J. Gierer, Holiforsch., 44(6):395 (1990).
436. A. A. Katai and C. Schuerch, J . Polymer Sei. Part A - l , 4:2683 (1966).
437. P. Angibeaud, J. Defaye, and A. Cadelle, in Cellulose and Its Deri1wtive.s (J. F. Kennedy, G.
0. Phillips, D. J. Wedlock, and P. A. Williams,eds.),Ellis Horwood, Chichester, U.K., p.
161 (1985).
438. G. Y. Pan, C.-L. Chen, H.-M. Chang, and J. S. Gratzl, Proc. b ~ tS. y n p . Wood Pulpiug Chern.,
SPCI Stockholm, vol. 2, p. 132 (1981).
439. G. Y. Pan, C.-L. Chen, J. S. Gratzl, and H.-M. Chang, Res. Chern. Intenned.. 21:205 (1995).
440. T. Kishimoto and F. Nakatsubo, H o l ~ o r s c h .50:372
, (1996).
441. T. P. Nevell, in Cellulose Chetnistly und Its Applications (T. P. Nevell and S. H. Zeronian,
eds.), Ellis Horwood, Chichester, U.K., p. 243 (1985).
442. E. E. Treiber, in Cellulose Chemistry and I t s Applications (T. P. Nevell and S. H. Zeronian,
eds.), Ellis Horwood, Chichester, U.K., p. 455 (1985).
443. E. Maekawa and T. Koshijima, in Cellulose: Strueturd and Functionul Aspects (J. F. Ken-
nedy, G. 0. Phillips,and P. A.Williams,eds.), Ellis Horwood,Chichester, U.K., p. 337
(1989).
444. E. R. Cousins, A. L. Bullock, C. H. Mack. and S. Rowland, Textile Res. J., 34:953 (1964).
445. R. Jeffries, J. G. Roberts, and R. N. Robinson, Textile Res. J., 38:234 (1968).
446. T. J. Painter, Carbohydr. Res., 55:95 (1977).
447. S. L. Snyder, T. L. Vigo, and C. M. Welch, Carhohydr: Res., 34:91 (1974).
448. C. Bosso, J. Defaye, A. Cadelle, C. C. Wong, and C. Pederson, J . Chern. Soc., Perkin Trms.
I : 1579 (1982).
449. 0. Samuelson, Svensk Pcrpperstidn., 83(8):205 (1980).
450. 0.Theander, in Chemistry of Delignijication with Oxygen, Ozone, and Peroxides ( J . S. Gratzl,
J. Nakano, and R. P. Singh, eds.), Uni, Tokyo, p. 43 (1980).
45 1, E. Sjostrom, Paperi Puu, 63(6-7):438 ( 1 98 1).
452. I. Norstedt and 0. Samuelson, Svensk PLrpperstidn., 68565 (1965).
453. R. Malinen, Paperi Puu, 57(4a):193 (1975).
454. E. Sjostrom, Paperi Puu, 63:438 (1981).
512 Lai
455. E. Sjostrom, in Chernistry of Deligni’cation with Oxygen, Ozone, and Peroxides (J. S. Gratzl,
J. Nakano, and R. P. Singh, eds.), Uni, Tokyo, p. 61 (1980).
456. A. Robert and A. Viallet, ATIP Re\]., 25:237 (1971).
457. A. Robert and C. De Choudens, ATIP Rev., 36(6):332 (1982).
458. P. J.-M. Lauriol, P. Froment, F. Pla, and A. Robert, Hol7forsch., 41:215 ( 1 987).
459. D. R. Dimmel, Proc. 1nt. Pulp Bleaching Confi, Vancouver, p. 107 (1994).
460. S.-l. Anderson and 0. Samuelson, Svensk Pupperstidn., 81:79 (1978).
461. H. Kolmodin and 0. Samuelson, Svensk Papperstidn., 74:301 (1971).
462. H. Kolmodin and 0. Samuelson, Svensk Papperstidn., 76:71 (1973).
463. C. W. Dence. in Lignins (K. V. Sarkanen and C. H. Ludwig, eds.), Wiley-Interscience, New
York, p. 373 (1971).
464. Y.-Z. Lai and K. V. Sarkanen, Tappi, 5 / ( 10):449 (1968).
465. D. C.Johnson, in Chemistry of Deligni’cation withOxygen, Ozone and Perosides (J. S.
Gratzl, J . Nakano, and R. P. Singh, eds.), Uni, Tokyo, p. 21 7 (1980).
466. T. J. McDonough. in Pulp Blecrching: Principles and Practice (C. W. Dence and D. W. Reeve,
eds.), Tappi Press, Atlanta, p. 213 (1996).
467. E. Johansson and S. Ljunggren, Pmc. Int. S ~ w p Wood
. Pulping Chern., Beijing. p. 180 (1993).
468. G. Gellerstedt, K. Gustafsson, and E.-L. Lindfors, Nordic Pulp Paper Res. J., 3:14 (1986).
469. H. Xu, S. Omori, and Y.-Z. Lai, Hol$or.sch., 49323 (1995).
470. H.Xu and Y.-Z. Lai, Proc. Tappi Pulping Cor$, Nashville, TN, p. 197 (1996).
471. Y.-Z. Lai, M. Funaoka, and H.-T. Chen., Hol+r.sch., 48(4):355 (1994).
472. D. A. 1. Goring, Appita, 38(1):3 1 (1985).
473. A. R. Procter, W. Q. Yean, and D. A. I. Goring, Pulp Paper Mug. C m . , 68(9):T445 (1967).
474. J. R. Wood, P. A. Ahrgren, and D. A. I. Goring, Svensk Pupperstidn.,75(1):15 (1972).
475. T. E. Timell, Compression Wood in Gymnosperons, Springer-Verlag, New York, vol. I , p.
433 ( 1986).
476. L. Paszner and N. C. Behera, Hol$orsch., 43:159 (1989).
477. H.-T. Chen, Y.-Z. Lai, and Y.-C. Ku, Q. J. Chin. Forest., 26( I ):93 ( 1 993).
478. H.-T. Chen, Y.-C. Ku, and Y.-Z. Lai, Q. J. Chin. Forest.. 27(3):105 (1994).
479. L. Gadda, Pciperi Puu. 63:793 ( 1981 ).
480. U. Westermark and B . Samuelson. Holzjiwsch., 4O(Suppl.): I39 (1986).
11
Weathering and Photochemistry of Wood
1. INTRODUCTION
513
514 Hon
tions, which determine the usefulness of wood products. Moreover, the deleterious effects
of acid rain on wood buildings and wood-based products have come to the fore in the
past three decades and are the subject of continuing investigation. There is an evident
need for enhancing the resistance of wood to weathering and photodegradation, increasing
its durability as a substrate and extending the service life of wood [4]. In responseto
technological demands. the field of weathering, photochemistry, and photophysics of wood
has, during the last three decades, moved from a largely empirical body of accumulated
practical knowledge to an increasingly sophisticated science employing the most advanced
techniques of physics and chemistry. As a result, our understanding of weathering and
photodegradation is not insignificant. For a comprehensive view of wood photodegrada-
tion, a number of specialty articles are available [5-71. However, for the purposes of this
chapter, it is instructive to consider the general features of weathering, photodegradation,
and photooxidation of wood and its chemical components. The degradative effects of air
pollutants and acid rain on the wood surface quality are also considered.
Wood exhibits beautiful color because its chemical constituents eitherreflect, scatter, trans-
mit, or absorb light. Unfortunately, because of its light-absorption properties, a small part
of this absorbed energy may well be specific enough to trigger undesired photophysical
and photochemicalprocesses.Suchphotochemicalprocessescaneventuallychange the
chemical, physical, optical, and mechanical properties of wood surfaces. In fact, the quan-
tum energies associated with light at the short ultraviolet end of the sunlight spectrum are
more than sufficient to break many of the chemical bonds present in wood constituents,
namely, cellulose, hemicelluloses, and lignin (Fig. l ) . For example, carbon-carbon, car-
bon-oxygen, and carbon-hydrogen bonds that form the basic backbone skeleton of ho-
locellulose and lignin might be expected to degrade with light of quantum energy at or
120
110 -
h
.
32 100-
-E
90 -
Y
80-
W
C
W
70 -
z
CS 60:
above the value associated with the bond strength. Hence, the absorption of light is a topic
of theoretical and practical interest to wood scientists and technologists.
Wood is a composite material containing intrinsic and extrinsic color that possesses
functional groups or systems, i.e., chromophores, which will absorb light. The chromo-
phoric groups or systems may control the course of photoreactions by acting as donors or
acceptors in energy-transfer processes. Chain impurities of wood polymers may play a
special role a s weak links or energy sinks and thereby control the act of chain scission in
photoreaction of wood.
An inspection of the ultraviolet (UV) absorption curves of wood, cellulose, hemi-
celluloses, and lignin (Fig. 2 ) reveals that cellulose absorbs light stronglybetween200
and 300 nm, with a tail of absorption extending to 400 nm. Lignin and polyphenols, i.e.,
extractives, absorb light strongly below 200 nm and have a strong peak at 280 nm, with
absorption down through the visible region. The combination of the U V absorption curves
of cellulose and lignin makes up the absorption curve of wood. According to Norrstrom
[S], lignin contributes 80-95%, the carbohydrates 5-20%, and the extractives about 2%
of the absorption coefficient. Pure cellulose is not a good light absorber. The absorption
by cellulose may be due to the presence of a carbonyl group that is accidentally introduced
into cellulose molecules during its isolation and purification. In addition to the carbonyl
group, it has been suggested that the absorption chromophores in cellulose also can be
contributed by acetal 19-1 I] or ketonic carbonyl [ 121 groups at the C l position of the
nonreducing glucose unit. Because of the structural similarity, the light-absorption char-
acteristics of hemicellulose should resemble those of cellulose. Unlike cellulose, lignin is
a good light absorber. The absorption occurs at chromophoric structural elements within
the molecular network of lignin. Hon and Glasser 1131 have classified the potential chro-
mophoric system as follows:
1.0
0
250 300 350 400 450 480
Wavelength km)
3. Leucochromophoric systems:
methylenequinones,
phenanthraenequinones,
phenylnaphthalenediones, bimethylenequainones, etc.
4. Intermediates:free radicals
5. Complexes:chelatestructureswithmetalions
Because of lignin’s predominant light-absorption properties, it absorbs more light,
resulting in more degradation than cellulose. Moreover, because of lignin’s phenolic struc-
tures, the photon energy absorbed by cellulose is likely to delocalize and transfer to lignin.
Hence, the presence of lignin will protect cellulose from photodegradation to some extent
[ 141, although lignin itself is consequently discolored and degraded. The protective effect
of lignin also has been observed by Kleinert [ 151. He explained that this phenomenon is
due to the high absorption and the strong capability of autoxidation of lignin.
Nonetheless, all of the components in wood are capable of absorbing enough ultra-
violet and visible light to trigger photochemical reactions, leading ultimately to changes
in esthetic. physical, chemical, and mechanical properties.
Because of the wide range of chromophoric groups or systems distributed at wood
surface components, light cannot easily penetrate into wood. Furthermore, the absorption
of light at the wood surface triggers photochemical reactions that result in rapid discol-
oration that forms a shielding layer to avoid further light penetration. Essentially, discol-
oration of wood by light is a superficial phenomenon. The dark-brown surface layers of
ponderosa pine and redwood that are affected by light extend only 0.5-2.5 mm into the
wood [ 16,171. As photooxidation progresses, most woods change to a grayish color, but
only to a depth of about 0.10-0.25 mm. Visible (400-750 nm) light, as measured spec-
trophotometrically, can penetrate into wood asfar as 2540 p m [ 171. The gray wood surface
layer was reported to be 125 p m thick; beneath the gray layer was a brown layer from
508 to 2540 p m thick. Moreimportant,thesecolorchanges are the consequences of
photochemical reactions.
The use of UV light-transmission techniques to measure free-radical chain reactions
that are always involved in the penetration of light through radial and tangential sections
of different woods as a functionof thickness revealed that UV light cannot penetrate deeper
than 75 pm. Visible light, on the other hand, can penetrate up to 200 p m into wood [ 181.
Although visible light can penetrate deeper in wood, its energy (400-700 nm) is insuffi-
cient to cleave chemical bonds in any of the wood constituents because the energy is less
than 70 k c a h o l (see Fig. l). Consequently, the brown color formed beneath a depth of
508-2540 p m could not be caused by visible light. It must be due to a sequential free-
radical chain reaction. It has been suggested that the aromatic moieties of wood compo-
nents at wood surfaces initially absorb UV light, and that an energy-transfer process from
molecule to molecule dissipates the excess energy to create new free radicals [ 181. The
use of electron spin resonance (ESR) techniques to monitor free radicals generated un-
derneath different layers ofwood confirmed that primaryfree radicals weregenerated
underneath 80 p m of exposed wood [ 181. This fact also had been substantiated by studying
the changes in infrared (IR) absorption of carbonyl groups and lignin of wood exposed to
UV light underneath different layers of wood [191.
The energy-transfer processes betweenelectronically excited groupsat the outerlayer
of the wood surface and another group underneath the wood surface thereby account for
the photoinduced discoloration of wood underneath the surface, which absorbs practically
no UV light. Furthermore, free radicals generated by light are high in energy and tend to
undergo chain reactions to stabilize their parent radicals. Consequently, new free radicals
formed in this way may migrate deeper into wood and cause discoloration reactions.
Weathering and Photochemistryof Wood 517
In
c
-
0)
E: 150
-m
c
.-
m
v,
.-
% W
-2
c
m
30
1 l
240 60
180 120
IrradiationTime(min)
FIGURE 3 ESR signal intensity (recorded at 77 K) of wood as a function of irradiation time and
storage time at ambient temperature.
518 Hon
for promoting free-radical formation. When woodis exposed to a weaker light such as
fluorescent light at ambient conditions,the free radicals generated are active toward oxygen
and other active gases such as sulfur dioxide and nitric oxide. They will be discussed in
Section 1II.E.
FIGURE 4 Free-radical formation in cellulose irradiated with ultraviolet light of h > 340 nm.
Weathering and Photochemistryof Wood 519
\.m*
.;eyoHm
H0
H&OH OH
\
FIGURE 5 Free-radical formation in cellulose irradiated with ultraviolet light of A > 280 nm.
CHSOH j .CH0 t H,
FIGURE 6 Free-radical formation in cellulose irradiated with ultraviolet light of A > 254 nm.
520 Hon
radicals in vacuo have an affinity for recombination and hydrogen abstraction to stabilize
themselves in the presence of oxygen, and they are transformed rapidly into hydroperoxide
radicals to build up hydroperoxide. This rapid oxygenation reaction is further accelerated
when excited oxygen is presented [32]. The involvement of singlet oxygen in oxidation
of cellulose has been a question of considerable debate [33,34], although several scientists
have suggested that singlet oxygen is involved [35-371.
Although cellulose is not sensitive to ultraviolet light of wavelengths longer than
340 nm, the presence of metal ions [37], particularly ferric ions [38], dyes [39-451, and
many sensitizers [46-491, promotes free-radical formation even when cellulose is exposed
to light longerthan 340 nm [37,38]. In addition to wavelengthsand sensitizers, other
factors that have significant effectson free-radical formationanddegradation rate are
morphology [50] and humidity and wetness [51].
I l
HC-OH
L
$0CH3 +OCH3
I I
HC-OH
0
-6 0
OCH,
\
HC-OH
60
OCH,
FIGURE7 Formation of o - and p-quinonoid structures during ultraviolet light irradiation of lignin.
troscopy coupled with Fourier transform infrared spectrophotometry (DRIFT) can be used
to detect hydroperoxide without any sample preparation or damaging the oxidized surfaces
[55].The appearance of a doublet absorption peak at 3550 cm", due to the hydroperoxide,
can be detected from the tangential section of a southern yellow pine, irradiated with lights
of A > 223 and h > 300 nm. More hydroperoxide is detectable from the specimens irra-
i=O
CEO
-0
-0
QOCH,
-0
FIGURE 8 Norristype I photoreactionin lignin.
522 Hon
Energytransferredhere
here
-0
"c.I 4 -
0
"0
FIGURE 9 Dissociation of the @-aryl etherlinkage in lignin by an energy-transferprocess via
excited a-carbonyl group.
diated with the latter light source. Competitive reactions between hydroperoxide formation
and decay revealed that at the initial 90 days of irradiation with the light of A > 300 nm,
the rate of formation exceeded the rate of degradation. When the wood is irradiated with
light of A > 223 nm, most of the hydroperoxide is generated and converted simultaneously
into carbonyl group. This chemical conversion is also observed from the specimen irra-
diated above 65°C.
ESR can also be used to monitor the formation and decay of hydroperoxide radical.
A typical ESR spectrum of peroxy radicals in photoirradiated wood is shown in Fig. 10.
The peroxy radicals seek to complete their unsatisfied valences, which may be done by
abstracting a proton from a nearby molecule to form a hydroperoxide. The hydroperoxide
is relatively unstable toward heat and light, and is usually transformed into a new chro-
mophoricgroupsuchasacarbonyl or carboxylicgroup.Thehydroperoxideimpurities
FIGURE 10 A typical ESR spectrum of peroxy radicals in photo-irradiated southern yellow pine.
ochemistry
andWeathering of Wood 523
hv
Initiation: RH”+
R. + H* (1)
R. + ROi product
non-radical (5)
R. + R - ”% non-radical
product (6)
d[o’l - k,[R.][O2]
”
dt
Using the usual steady-state assumptions, the rate of chain initiation can be illustrated
as follows:
+
R, = k,l[ROO*]2 2krl2[R*][ROO.] kf2[R.]’ + (8)
"
dt
b. Formation Mechanisms. When wood is irradiated with ultraviolet light, free rad-
icals are generated at the surfaces due to the dehydrogenation, dehydroxylation, dehyrox-
ymethylation, demethoxylation. and chain scission that occurred in cellulose, hemicellu-
lose, and lignin distributed at the wood surface [Eq. ( l ) ] [52].The presence of oxygen in
the system provides the opportunity for oxygen molecules to react with free radicals in
wood to generate hydroperoxy radicals [Eq. ( 2 ) ] ,which in turn abstract protons to produce
hydroperoxides [Eq. (3)]. This can be seen from the transformation of a multiplet signal
of ESR, due to the various carbon radicals, to an asymmetric singlet signal, due to the
hydroperoxy radicals. Singlet oxygen, resulting from the interaction of ultraviolet light
and molecular oxygen, and its subsequent attack on wood surfaces, has been proved as
another possible initiation route for hydroperoxide formation [571.
L
c
Earlywood
.,
10 20 30 40 50
Moisture Content (“h)
FIGURE 12 Relationship between ESR relative signal intensity and moisture content of cellulose
irradiated with a high-pressure mercury lamp at 77 K for 60 min.
molecules to create singlet oxygen. The participation of singlet oxygen in the photooxi-
dation ofwood wasevidenced by usingsinglet-oxygengeneratorsandsinglet-oxygen
quenchersduring irradiation [57].Iodometrystudiesrevealed that hydroperoxidewas
formed in wood photo-irradiated in the presence of oxygen. The formation rate of hydro-
peroxide at the wood surface increased when singlet-oxygen generators suchas rose bengal
solutions were added to the wood prior to irradiation (Fig. 13). Peroxide radicals involved
in the interimweredetected by an ESR spectrophotometer, i.e., anasymmetric single
signal of peroxy radicals with an average g value of 2.021 (g,,= 2.034; g, = 2.007) was
Wood (control)
Wood lrradlated In N,
wood lrrodlated In air
1
Wood lrradlated In oxygen
. . . . . .
Wood nl th Rose Bengal
Wood (control1
IWoodwlthRoseBengalirradiated in air 1
Wood wlth Rose Bengal andTEM Irradlated I n air
detected. On the other hand, when singlet quenchers, such as triethylamine or 1,4-diazo-
bicyclo[2,2,2]-octane (DABCO), wereusedunder identical experimentconditions, the
hydroperoxide content was reduced in some cases, even in the presence of rose bengal
(Fig. 14). This evidence supports the theory that singlet oxygen is formed during photoir-
radiation and that it interacts rapidly with free radicals of wood to produce hydroperoxides.
Due to its instability against heat and light, the hydroperoxide decomposes rapidly under
ambient conditions to create chromophoric groups, such as carbonyl and carboxyl groups.
These groups contribute to the discoloration of wood surfaces.
The deleterious effects of acidrain on lakes, aquatic ecosystems, vegetation, forests, build-
ings, and artifacts have come to the fore in the past three decades and are the subjects of
continuing investigation [62]. Wood materials offer an impressive range of attractive prop-
erties and in many of their applications they are exposed to the outdoorenvironment.
Hence, they are subject to sunlight, weathering, and acid precipitation. Their increase use
in outdoor applications has resulted in the needtounderstandweathering reactions in-
volving acid rain. Although ultraviolet light is a major element in degrading wood poly-
mers, it is to be expected that for lignocellulosic biopolymers, for which hydrolytic break-
down can be an important modeof deterioration 1631, acid rain will have a catalytic effect.
In Section 1II.E we discussed that wood interacts readily with sulfur dioxide, one of
the principal elements in acid rain, to trigger a wholeseries of free-radical reactions,
especially in the presence of ultraviolet light. These reactions lead ultimately to discol-
oration and loss of surface integrity.
Many factors are involved in wood deterioration in the polluted environment. The
absorption of UV light and acid rain can lead to chemical changes on wood surfaces, and
deterioration of tensile strength was proved experimentally. When wood surfaces are ex-
posed to ultraviolet light, carbonyl-group content increased and lignin content decreased
simultaneously. These changes are accelerated whenthey are also exposed to acid rain,
Weathering and Photochemistryof Wood 529
3.7
0 3.7.l
6 1
12 5.6I 4.2
24
36 - - - -
i.e., a dilute sulfuric acid solution, especially at 65°C and 65% relative humidity. FTlR
studies demonstrated that carbonyl groups are generated on photo-irradiated wood surfaces
at ambient temperature, and the rate of increase in carbonyl groups is enhanced i n the
presence of acid. When wood was exposed to an elevated temperature, i.e., 65°C and 65%
relative humidity, its chemicalcomponents did not showa noticeable degradation. The
carbonylcontent of untreated wood remained almostconstantas the exposuretime in-
creased. The same is true for lignin content. In the presence of acid under such conditions,
the carbonylgroupincreases slightly as a function of acid concentration (Table 1). No
significant change in lignin is observed (Table 2). The effect of acid and U V light at 65°C
and at 65% relative humidity on carbonyl-group and lignin content is significant (Table
3). When wood was exposed to UV light in the absence of acid at such conditions, it was
noticed that carbonyl-groupcontent increased almost sixfold after 48 h of irradiation.
About 80% loss of lignin content was observed (Table 4). Therate of degradation is further
enhanced in the presence of acid. The higher the acid concentration, the more carbonyl
group is found. Although lignin is resistant to acid attack, significant reduction of lignin
content is observed under such conditions.
The increase i n carbonylgroup and reduction of lignin content at wood surfaces
being exposed to acid and U V light at ambicnt and at high temperature and high humidity
4.4
0 4.4"
3 6 3.8 4.3 .c> 5.2
5.5 12 3.8 4.2 3.8
24 5.8 4. I
3.9 3.8
36 - - - -
48 5.0
530 Hon
signaled that chemical reactions take place in the polymeric components of wood, i.e.,
cellulose, hemicelluloses, and lignin. Since holocellulose is sensitive to acid hydrolysis,
there is little doubt that the strength of wood will be reduced due to the chain scission
reaction. The change in color of wood from pale yellow to dark brown and the embrittle-
ment of wood specimens clearly indicate severe degradation of wood quality. Experimental
results showed that oxidative degradation of wood surface is initiated by UV irradiation.
However, photooxidation by itself appeared not to be a contributor to the loss of tensile
strength. Acid rain appeared to be the major culprit contributing to the loss of strength
[80]. Effects of acid rain on tensile strength at ambient temperature and at 65°C are shown
in Figs. 15 and 16.
As discussed earlier, the photo-induced free radicals in wood are capable of reacting
with sulfur dioxide to produce various oxidized products. They, in turn. further influence
the color and surface properties. The presence of UV light may also accelerate oxidation
of sulfur dioxide to sulfur trioxide, which will subsequently react with water to produce
sulfuricacid.Sulfuricacid may also form by othercatalyticprocessesinvolvingsulfur
dioxide, water, and a catalyst present in the atmosphere. Hence, three plausible mecha-
nisms that can lead to surface degradation of wood are worth considering. They are sum-
marized below.
"
t d
IO 20 30 40 50
Exposure Time (H)
FIGURE 15 Effect of acid rain on tensile strength of wood at ambient temperature in the presence
of UV light: (a) UV-irradiated without acid; (b) UV-irradiated with 0.002% acid; (c) UV-irradiated
with 0.009% acid; (d) UV-irradiated with 0.074% acid.
100
-
y 80 0
v
.-0
60
W
t
Q
a
40
c
c
0
C
al
2 20
m
0
10 20 30 40 50
Exposure Time (H)
FIGURE 16 Eftect o f acid rain 011 tensile strength of wood at 65°C: ( ; I ) control at 65°C: ( b ) UV-
irradintctl withoutacid: ( c ) UV-irradiated with 0.002% acid: (cl) UV-irradiatedwith 0.009%' acid:
( c ) UV-irrxliatcd with 0.074% acid.
532 Hon
1. Wood +h u + woodfreeradicals
+
Wood free radicals SO, + oxidized products
11. hu + so, + O? -3 so,
SO,+ H,O H,SO,
H,SO, + wood + degradation products
111. SO, + O2 + catalyst -+ [SO,. H 2 0 ]
[S02-HrO]+ H2S0,
H,SO, + wood -+ degradation products
B. MicroscopicChanges
Microscopicchangesaccompanythegross physical changes in wood during ultraviolet
irradiatioll 13.65-671. The tirst sign of deterioration i n softwood surfaces is enlargemcnt
Weathering and Photochemistryof Wood 533
100
K
.-0
U
K
Q)
+ Southern Yellow Pine
50
80
loo i
20
0
60 120 240 180 360 300 420 480
Weathering Time (Days)
wood; ., .,
FIGURE 18 Change i n color of outdoor weathered wood. Key:
Douglas fir; western red cedar.
0, southern yellow pine; D, red-
534 Hon
A" Bb
of apertures of bordered pits in radial walls of earlywood tracheids. Futo [68] observed
that the degradation begins at a relatively low irradiation intensity with an attack on the
compound middle lamellae. Higher intensities and longer exposure also degrade the sec-
ondary walls, as is made visible by the formation of cavernes. Elevated temperature in-
tensifies the photolytic degradation process. The degradation of the wall substance during
UV irradiation effects a contraction of the cell walls, resulting in microchecks along the
compound middle lamellae and, particularly in latewood, along the border between S , and
S , [69].Diagonal fissures that follow the fibril orientation of the Sz layer also have been
observed. The apertures of the bordered pits in softwood were enlarged or ruptured by
microchecks.
The scanning electron microscope (SEM) is frequently used to study the breakdown
of the structureof wood due toweathering. The surface texturesof woods from Norwegian
stave churches and other wooden construction several hundred years old were studied by
Borgin using an electron microscope [70,71].
Deterioration of wood surfaces after exposure to artificial W light was observed
after wood was exposed for only 500 h [65]. Photodegradative effects on transverse, radial,
and tangential surfaces of a typical southern yellow pine are described in the following
sections.
1. Transverse Section
The transverse sectionof southern yellow pine is normally quitesimple and homogeneous.
Its axial system is essentially composed of wood tracheids, with only a relatively small
number of parenchyma cells. An SEM micrograph of a transverse southern pine surface
before exposure is shown in Fig. 19.
A microtomed transverse wood surface was exposed to W light for 500 h. Surface
deterioration of the exposed wood surface was observed readily from the SEM micrograph
FIGURE 20 Cross section of southern yellow pine exposed to UV light for 500 h (700X).
536 Hon
FIGURE 21 Cross section of southern yellow pine exposed to UV light for loo0 h (700X).
(Fig. 20). The cell walls were separated at the middle lamella zone. In the extreme case,
the secondary wall almost collapsed. Roughening of the surfaces could be observed vi-
sually. Surface deterioration further developed when specimens were exposed for a total
of lo00 h (Fig. 21). Bordered pits located at the tracheid walls were totally destroyed.
The color of the exposed wood changed from pale yellow to light brown and then dark
brown after 500 and 1000 h of W light exposure, respectively.
2. Radial Section
Bordered pits in southern yellowpine could be observed at radial walls in both earlywood
and latewood. Generally, bordered pits located in the earlywood were larger and more
numerous than those in the latewood. Dpical SEM micrographs for half-bordered pitsand
bordered pits at radial walls before W exposure are shown in Figs. 22 and 23.
The first perceptible change in the anatomical structure of the radial sectionof south-
em yellow pine upon exposure appears to take place at the pits. After 500 h of UV
exposure, half-bordered pits were damaged. Bordered pits also interacted with light, but
to a lesser extent (Fig. 24). The bordered pits could still be recognized.In addition, check-
ing and void formation in radial walls occasionally could be seen from the exposed spec-
imen. After 1000 h of exposure, however, severe deterioration of the bordered pits was
chemistry
andWeathering of Wood 537
FIGURE 22 Half-bordered pit structures of southern yellow pine on radial section (700X).
observed. The SEM micrograph (Fig. 25) shows that the apertures of bordered pits were
enlarged to the limit of the pit chambers. The pit domes were destroyed completely. At
the extreme, the deterioration also spread over the radial surface of the tracheid wall.
Complete degradation of these cell walls would probably take place at a longer exposure
time. Disappearance of bordered pits also has been observed in redwood exposed to UV
light.
3. Tangential Section
Bordered pits are rarely found in the tangential surfaces observed. SEM studies revealed
that diagonal microchecks passing through bordered pits in tracheid cell walls were the
most conspicuous anatomical change at the tangential section upon UV exposure. The
narrow microchecks were oriented diagonally to the axis of the cell wall, thus indicating
that microchecks occur at the fibril angles of the S2 cell wall (Figs. 26 and 27). Similar
observations have been reported. The common appearance of the diagonal microchecks
during UV exposure was suggested to be the resultof local concentrations of tensile stress
at right angles to the fibril direction of the S , layer. Relatively wide diagonal checks were
observed in the tangential section of tracheid walls of latewood.
538 Hon
FIGURE 23 Bordered pit structures of southern yellow pine on radial section (700X).
C. ChemicalChanges
The consequences of photodegradation and photooxidation of wood are changes in chem-
ical and physical properties. As discussed earlier, irradiated wood may exhibit a form of
discoloration, loss of lightness, checking, cracking and rougheningof surfaces, damage of
microstructure, and loss of weight. It is believed that these changes are caused by severe
chemical modification of the structures of cellulose, hemicelluloses, and lignin.
Over a century ago, Wiesner [72] reported that the intercellular substance of wood
had been lostandtheremainingmembranes,consisting of chemicallypure or nearly
chemically pure cellulose, were observed. As discussed in the earlier section on micro-
scopic changes, it is clear that absorption of UV light by lignin in the middle lamella as
well as in the secondary cell walls results in preferential lignin degradation. Most of the
solubilized lignin degradation products are washed out by rain [54].Careful analyses of
surface layers of a southern pine that had been weathered for about 10 years revealed that
the top gray layer consistently exhibited a very low lignin content. The grayhrown layer
had a higher lignin content than the outer gray layer but less than the brown underlayer
and in the centered portion (see Table 5).
Accordingly, it is obvious that ultraviolet light initiates significant modification to
the wood polymeric system. From the ESCA study [73], the increase in signal intensities
of carbon-oxygen bonds and oxygen-carbon-oxygen bonds (or unsaturated carbon-ox-
ochemistry
andWeathering of Wood 539
FIGURE 24 Deterioration of half-borderedpits and cell wall of southern yellow pine atradial
section after exposure to UV light for 500 h (700X).
ygen bonds) and oxygen-to-carbon ratio, and the decrease in carbon-carbon and carbon-
hydrogen bonds of weathered and UV-irradiated wood surfaces, suggested that the wood
surface was oxidized. The oxygen-to-carbon ratio data also revealed that weathered wood
surface was rich in cellulose and poor in lignin. FTIR [55] studies showed the increase in
carbonyl and hydroperoxide groups and the decrease in lignin content of UV-irradiated
wood surfaces. On the whole, cellulose, hemicelluloses, and lignin are degraded as illus-
trated by increase in solubility and reducing power of cellulose and formation of carbonyl,
carboxylic, hydroperoxide groups, quinone, and conjugated double bonds in lignin. The
changes in carbonyl, carboxyl, and hydroperoxide groups in cellulose are shown in Table
6 [36]; and the changes in carbonyl and hydroperoxide groups in wood are shown in Ta-
ble 7.
During photoirradiation, in the initial stages up to 1 h, only CO, COz, Hz,and HzO
were detected as gaseous products. At longer exposure times, however, methane, ethane
and ethylene hydrocarbon gases were found. Figure28 shows the rateof formation of CO,
COz, and H2 in a southern pine irradiated with ultraviolet light. Volatile products of for-
maldehyde, methanol,acetone,methylformate,acetaldehyde,propionaldehyde,vanillic
acid, vanilin, and syringyl aldehyde also were found after longer exposures [57].
The photo-irradiated wood increased its solubility in water, benzene, alcohol, and
alkaline aqueous solutions. Carbohydrates and phenolic compounds are detectable from
540 Hon
FIGURE 25 Deterioration of bordered pits and cell wall of southern yellow pine atradial section
after exposure to UV light for lo00 h (700X).
the solutions. The contents of holocellulose, cellulose, and lignin in wood were decreased
as a function of irradiation time (100 days) with different wavelengths. Table8 shows the
results of such degradation.
A study of the loss of weight by UV irradiation (A > 340-320 nm) was carried out
by Futo [68,74]. He found that the loss of weight is highly influenced by the temperature
and the irradiation energy. The loss of weight is much higher when wood is irradiated in
the presence of water, which indicates that water-soluble products are formed in addition
to gaseous and volatile products. The collection of water-soluble fragments was charac-
terized using UVhisible spectroscopy by Hon and Chang [54]. They found that the low-
molecular-weight, water-soluble products are derived mostly from lignin. The degradation
products contained carbonyl-conjugated phenolic hydroxyl groups and had a weight-av-
erage molecular weight of about 900 as determined by gel permeation chromatography.
The degradation of cellulose under the influence of ultraviolet light is indicated by
a decrease in strength and degree of polymerization, and an increase in alkali solubility
and copper number. When a bleached softwood pulp was irradiated in vacuum and in
oxygen for only 10 h, the DP of a-cellulose was reduced from 850 to about 380 and 260,
respectively. The content of a-cellulosedecreasedfrom 88% toabout 50% and 40%,
respectively [32]. Furthermore, ultraviolet light causes yellowing and browning and for-
Weathering and Photochemistryof Wood 541
l "
FIGURE 26 Microchecks of cell wall of southern yellow pine at tangential section (earlywood)
after exposure to UV light for 500 h (700X).
mation of carbonyl, carboxyl, and hydroperoxide groups along the cellulose chain, and a
fragmentation of molecules to diversities of neutral and acidic nonvolatile, volatile, and
gaseous products. Among the volatile degradation products of photo-irradiated cellulose
are acetaldehyde, propionaldehyde, methyl formiate, acetone, methanol, ethanol, methane,
and ethane [76]. Glucose, cellobiose, cellotriose, xylose, xylo-oligomers, arabinose, and
3-P-D-glucosido-D-arabinose wereidentified fromthesoluble degradation products
[10,11,77].
The photodegradation of lignin also has been observed. The reduction of the meth-
oxy1 content and the splitting of monomeric units were reported [15,75,78,79].
VII. CONCLUSIONS
The deterioration of wood materials upon weathering involves a very complex reaction
sequence. Penetration of W light into wood does not traverse deeper than 75pm. None-
theless, wood surface reactions initiated or accelerated by light can be observed by dis-
coloration, loss of brightness, and change in surface texture after artificial W light irra-
diation or long-term solar irradiation.
542 Hon
FIGURE 27 Microchecks of cell wall of southern yellow pine at tangential section (latewood)
after exposure to UV light for 500 h (SOX).
Hydroperoxide
Carboxyl
Carbonyl
group
Irradiation
group
time
group
(h) (mmoV100 g) (mmoV100 g) (mmOV100g)
0 0.2 0.0 0.0
3 5.6 1.1 0.0
5 0.1 10.1 1.9
10 15.9 4.2 0.3
15 18.5 6.5 0.5
20 9.7 0.6
'Irradiated with a hlgh-pressure quartz mercury lamp (h > 253.7 nm) at amblent temperature.
Weathering and Photochemistryof Wood 543
IrradiationCarbonyl
time
Hydroperoxide
group group
(mmoll100(h) 100 g)
~~
74 0 1.20
30 I .71
60 3.10 4.25
85 90 4.32
120
24 150 8.26
180
210
"Irradiated with a high-pressure quartz mercury-xenon compact arc lamp (A >
223 nm) at ambient temperature.
m
0)
H
0
E
I
0 50 100
150
200
250 300
Irradiation Time (H)
FIGURE 28 Rate of formation of (a) carbon dioxide, (b) carbon monoxide, and (c) hydrogen in
a southern yellow pine irradiated with ultraviolet light of A > 254 nm.
Content (%)
Chemical Unirradiated
constituent
wood A > 340 nm A > 280 nm A > 253 nm
Various types of free-radical species, such as phenoxy, alkoxy, and carbon radicals,
are readily generated in wood by light. Phenoxy radicals are quite stable at ambient tern-
perature, whereas alkoxy and carbon radicals decay rapidly at that temperature. Carbon
radicals rapidly interact with oxygen to produce hydroperoxide impurities that are decom-
posed easily to produce chromophoric groups such as carbonyl and carboxyl groups. A
competitive reaction between formation and destruction of hydroperoxide appears to be
occurring during the photoirradiation, and the hydroperoxide is unstable above 65°C. The
hydroperoxidation of wood surfaces can be readily detected by using the DRIFT technique
without any sample preparation and destroying the surface. The ESR technique also pro-
videdvaluableinformation on the free-radical formationmechanism that explains the
formation of hydroperoxide. The use of singlet-oxygen generators, such as rose bengal
andmethylene blue, aswell as singlet-oxygenquenchers,such as 1,4-diazo-bicyclo-
[2.2.2]octane and triethylamine, suggests the participation of singlet oxygen as an effective
intermediate in photooxidation reactions at the wood surface. Phenoxy radicals are inert
toward sulfur dioxide and nitric oxide. All carbon radicals or alkoxy radicals are capable
of reacting with nitric oxide to form nonradical products such as nitroso and nitrite groups.
Somecarbon radicals are sensitivetoward sulfur dioxide to form sulfonyl and sulfite
radicals andconvertedinto sulfinic acidandsulfonate ester. The presence of water in
wood also influences the rate of free-radical formation. When moisture content in wood
is increased from 0 to 6.3%. more free radicals are formed. Beyond this stage, the rate of
radical decay increases. Infrared studies reveal that carbonyl groups are generated in cel-
lulose and lignin. Water-soluble fractions of degraded wood exhibit characteristics of phe-
nolic absorptions due to the loss of lignin. ESCA studies show that oxidized wood surfaces
contain higher oxygen contents than unexposed wood surfaces.
In addition to UV light, acid rain seems to be an important element contributing to
deterioration of wood surface quality and tensile property of wood. Experimental results
showed that while carbonyl group was generated on wood, lignin content simultaneously
decreased when it was exposed to light; and this process was further enhanced when it
was also exposed to acid rain. Either with or withoutultraviolet, at 65"C, wood deteriorated
slightly faster than that at ambient temperature. Acid further accelerated the degradation
process. Wood lost its tensile strength in the presence of acid both at ambient temperature
and at 65°C.
REFERENCES
1. D. N.-S. Hon and W. C. Feist, Wood t r n d Fibel; 12: 121 (1980).
2. D. N.-S. Hon, J. Appl. Polymer Sci., Appl. Polymer Sywp., 3 7 845 (1983).
3. H. Turkulin and J. Sell, Bericht 115/36 (1997).
4. D. N.-S. Hon, Durability and Serviceability of Wood and Wood-based Products, presented at
The New Tropical Timber Crops: Challenges in Processing and Utilization, International Trop-
ical Wood Conference, June 17-20, 1997, Kuala Lumpur. Malaysia.
5. D. Fengel andG. Wegener, Wood: Cl~en~istty, lJltrtrstrut~ture9
Recrctiorl, Walter de Gruyter,
Berlin-New York ( 1984).
6. T. Yoshinloto, Mokuztri Gtrkkniski, IS: 49 ( 1 972).
7. W. C . Feist and D. N.-S. Hon, Ad\!. Cl~enz.SL'K,207: 401 (1984).
8. H. Norrstrom, Sver~.Ptrpperstidn., 72: 25 (1969).
9. N.-S. Hon, J . Polyrnt'r Sci., Polwwr Cllent. Ed., I S : 1347 (1975).
10. A. Beelik and J. K. Hamilton, Pupel; 13: 77 (1959).
11. A. Beelik and J. K. Hamilton, J . Org. Cllern., 26: 5074 ( 1961).
Weathering and Photochemistryof Wood 545
57. D. N.-S. Hon. S.-T. Chang, and W. C. Feist, Wood Sri. Trchnol., 16: 193 ( 1982).
58. D. N.-S. Hon and W. Y. Chao. in Cellulose and Wood: Chemistnland Technology (C. Schuerch,
ed.), Wiley, New York, pp. 1037-1057 (1989).
59. D. N.-S. Hon and W. C. Feist, Wood Fiber Sci., 25: 136 (1993).
60. D. N.-S. Hon and W. C. Feist, Wood Sci., 14: 41 (1981).
61. M. L. Kapalan and A. M. Trozzolo, in Singlet Oxygen (H. H. Wasserman and R. W. Murray,
eds.). Academic Press, New York, chap. 11 ( 1 979).
62. E. T. Paparozzi, in Materials Degradation Caused by Acid Rain (R. Baboian, ed.), American
Chemical Society, Washington, DC, pp. 332-342 (1986).
63. I. S. Goldstein, in Organic Chemicals from Bion~uss(I. S. Goldstein, ed.), CRC Press. Boca
Raton, FL, pp. 101-124 (1981).
64. T. N. Kleinert. Hol&r.sch Holzver-wert.. 22: 21 (1970).
65. S.-T. Chang. D. N.-S. Hon, and W. C. Feist, Wood and Fiber, 14: 104 (1982).
66. M.-I. Kuo and N. Hu, Hol@orsch., 45:347 (1991).
67. L. Paajanen, Structural changes in primed Scots pine and Norway spruce during weathering,
Maret: Srruct., 2 7 237 (1994).
68. L. Futo, Holz Roh-Werksr., 32: 303 (1974).
69. V. P. Miniutti. Paint Technol., 45: 27 (1973).
70. K. Borgin, J. Microsc.. 92: 47 (1970).
71. K. Borgin, J. Inst. Wood Sci., 5: 26 (1971).
72. J. Wiesner, Akad. Wiss. Wien, 49: 61 (1864).
73. D. N.-S. Hon, J. Appl. Polyrner Sci., 29: 2777 (1984).
74. L. Futo. Holz Roh-Werkst, 34: 31 (1976).
75. G . J. Leary, Zrppi, 50: 17 (1967).
76. R. L. Desai and J. A. Shields, Mcrkrotnol. Chem., 122: 134 (1969).
77. B. A. Gringras, D. Cooney, K. A. Jackson, and C. H. Bayley, Textile Res. J., 33: 1000 (1963).
78. W. Sandermann and F. Schlumbom, Holz Roh-Werksr., 20: 245, 285 (1962).
79. G. J. Leary, Tcrppi, 51: 257 (1968).
80. D. N.-S. Hon, Woocl Fiber Sei., 26: 185 (1994).
81. D. N.-S. Hon. J. WoodChem.Technol.,12: 179 (1992).
12
Microbial, Enzymatic, and Biomimetic
Degradation of Lignin in Relation
to Bioremediation
1. INTRODUCTION
Next to cellulose, lignin is the most abundant renewable resource on Earth. Although it is
an important structural component in wood, together with cellulose and hemicellulose, its
utilization in isolated form as a chemical material has not yet been successfully achieved
[ l ] . The main reason for this is the fact that lignin molecules lack stereoregularity and
repeating units in the molecule, andthey aretooheterogeneousandcomplex.Current
research interests focus instead on environmentally benign delignitication processes with
biocatalytic systems,includingwhite-rotfungi, their enzymesystems, and biomimetic
catalysts.
Since the cellulase-less mutant of Phanerochoete chr~y~osporiumwassuccessfully
isolated by Ander and Eriksson [ 1 a], more than 20 years have passed, and it is more than
10 years since discovery of lignin peroxidase (Lip) [2,3] and manganese peroxidase (MnP)
[4] and proposal of the one-electron oxidation mechanism for the enzymatic reactions [5-
71. However, it is deceptively simple to apply these bioligninolytic systems to industrial
delignitication processes despite a great deal of recent work on microbial, enzymatic, and
biomimetic degradation of lignin in relation to the roles of lignin-degrading enzymes of
Lip, MnP, and laccase (Lac).
This chapter gives a brief overview of current lignin biodegradation research, fo-
cusing on microbial, enzymatic, and biomimetic approaches to lignin biodegradation chem-
istry and related fields. Shimada and Higuchi [8] reviewed the biochemical approach to
lignin biodegradation research in the first edition of this book, and the present authors
have attempted to include as many recent findings as possible in this chapter. However,
since it is beyondour ability to cover all current publications, readersshould refer to
recent international proceedings [9- 121 and reviews [ 13- 191 to follow the trend of related
research areas.
II. MICROBIALAPPROACH
A microbial approach is basically important not only for application of fungal ligninolytic
systems to biopulping, biobleaching of unbleached pulps [lo], bioremediation of waste
547
Shimada
548 and Hattori
effluents, and bioconversion of agricultural lignocellulosic residues, but also for the basic
study of physiological features of lignin decomposers in natural environments. Since the
discovery of the cellulase-lessmutant, Phanemchaetechrysosporium hasacquired the
most prominent status in lignin biodegradation research. However, it is still important to
screen other lignin-degrading microbes with high ligninolytic activity [20,2 l ] .
As to biomechanical pulping, Eriksson and Vallander reportedthat treatment of wood
chips with cellulase-less mutant for 10 days could save electrical energy by about 30%
[22], and several research teams have also reported that pretreatment of wood chips with
wild-type white-rot fungi could succeed in reducing the energy requirement and also in-
crease paper strength [23]. Similar results were also obtained by Akamatsu et al. [24].
To find a more appropriate fungus for this purpose, Ceriporiopsis subvertnisporu was
selected out of 400 species of the ligninolytic fungi [20]. The biomechanical biopulping
method has been refined [25-281. Nishida and his co-workers have succeeded in isolating
a powerful ligninolytic fungus named IZU-154 [29]. It is noteworthy that they used only
wood meal substrate for screening this fungus, which has, i n fact, a high selectivity for
lignin decomposition during fungal treatment of the wood meals. Coarse beech mechanical
pulp and wood meal were incubated with only the mycelia of IZU-154, Coriolus versi-
color, and Phuneroc~huete chr~ao.sporiunz. The ligninolytic activity of fungus IZU-154 was
muchgreater than that of the other fungi. Theyhave also reported that 7-dayfungal
treatment of coarse hardwood mechanical pulp reduced the refining-energy requirement
by aboutone-half to two-thirdsandimprovedpulp strength properties abouttwofold.
Similar energy savings and improvement of pulp strength properties were achieved for
softwood samples after a little longer period of treatment.
As to biobleachingwork, this funguswasfound to bleachkraftpulps[30]. The
bleaching activity may be due to the action of the MnP system [311, although this fungus
remains to be identified. Paice et al. demonstrated direct biological bleaching of hardwood
kraft pulp with the fungus Coriolus (Trametes) versicolor [32]. Kondo and Sakai and co-
workers have isolated Phatlerochaete sordidu (YK-624 strain), which has strongkraft pulp
bleaching activity [33]. They also demonstrated that the MnP system catalyzes pulp bleach-
ingand that there is agood correlation between the level of MnP activity andfungal
bleaching ability [34].
Decolorization of pulping waste liquors was first investigated by Fukuzumi et al. in
Japan [35]. Quite recently, Pallerla and Chambers [36] reported the continuous decolori-
zation and AOX reduction of bleach-plant effluents by freeandimmobilized Trumetes
versicolor. They achieved successful results of about 89% continuous decolorization and
about 70% AOX reduction. A number of papers have reported for fungal treatments of
effluents from pulp mills 137-441.
Bioremediation of environmentally hazardous compounds by use of wood-degrading
fungi and bioreactor systems will also be important for removing carcinogenic pollutants
from agricultural and industrial wastewatersbefore letting them flow into rivers and
oceans. For detailed information readers should refer to the recent publications cited [43].
OMe L
CHOH 0 CH
”+”,
Me0
OMe
OMe
OH
guaiacylglycol
I
OMe Me0
OMe OMe
veratraldehyde
benzaldehyde
A. Lignin Peroxidase
After purification andcharacterization, lignin peroxidase (Lip) isolated from Phanero-
chaete chrysosporiunz was found to have a molecular weight of 41,000-43,000 daltons
[48,6OJ, with an iron porphyrin IX as the prosthetic group [4,48,51,52,60]. The enzyme
catalyzes oxidation of veratryl alcohol (VA) or 3,4-dimethoxybenzyl alcohol substrate to
form a veratraldehyde product in the presence of hydrogen peroxide as shown in Fig. 2.
The enzyme was found to have a broad substrate specificity to catalyze oxidations of a
wide variety of substrates [48,51].
At present, the most convenient assay method for Lip is probably to use VA as a
substrate instead of lignin to measure the increase in absorbance at 310 nm of veratral-
dehyde formed [48].
It is interesting to note that tobacco plant peroxidase can catalyze the oxidation of
VA in the presence of calcium and magnesium ions and hydrogen peroxide with a pH
optimum of 1.8 [61]. These dicationic metal ions plays an important role in stabilization
and expression of activity under these extremely low-pH conditions. Similar stabilization
of Lip by calcium ions was also observed [62]. Distribution of Lip among white-rot fungi
seems to be restricted to certain consortia of white-rot fungi and it may serve to detect
the Lip activity among others with low pH value around 2.0 for further survey of Lip
activities [21,45,46].
1. The ReactionMechanism
n. C a - C P B o n d C l e a v a g e qf'P-Z Model Substrate. A one-electron oxidation mech-
anism has been proposed by several different laboratories [6,7,63,64] to explain the C a -
C p bond cleavage reactions of the side chain of lignin.
Before ligninase had been identified as a hemoprotein, Shimada et al. [53] demon-
strated the Ca-c@ bond cleavage of the p-l lignin model substrate under both aerobic
and anaerobic conditions with biomimetic model catalysts of cytochrome P-450 and per-
oxidase, which were established to form compound I in the presence of iodosylbenzene
or tert-butylhydroperoxide. They also established that the porphyrin catalyst used virtually
mimicked lignin peroxidase i n may respects as described below.
Alternatively, Kersten et al. [ 5 ] proved for the first time the ligninase-catalyzed for-
mation of aryl cation radical of methoxybenzenesusing electron spin resonance (ESR)
techniques, indicating that compound I in two-electron-deficient oxo-iron porphyrin com-
plex stepwise abstracts two electrons from the aromatic moieties. Further, as shown in Fig.
3, Hammel et al. [63J reported that the aryl cation radical formed from diarylethane- I ,2-
CH20H
H0
OCH,
OCH3
+i
OCH,
Aerobic
02<
7
L i p , ,, ?
Anaerobic
H+,e-
CH0 OH 4 ?l-
Shimada
552 and Hattori
guaicoxyethanol, and benzyl alcohol derivatives have long been problematic in the early
work [70-731, because of a paucity of firm evidence to conclude the mechanism until the
above one-electron oxidation mechanism had been established.
Alternatively, Umezawa and Higuchi clearly indicated that guaiacol and arylglycerol
were produced via different pathways and these two products were not counterparts to
each other in the degradation of the p - 0 - 4 model substrates in the culture of P. chryso-
sporium [74].
According to the one-electronoxidationmechanism,the initial step of the bond
cleavages also seems to be the aryl cation radical formation [75] in ring A or B of the
substrate by compound I of Lip. Thus, C a - C p bond cleavage of p - 0 - 4 model substrate
caused @-ether bond cleavage to give rise to veratraldehyde, guaiacol, and glycoaldehyde
[74-781 as one of the main degradations of a p - 0 - 4 model compound catalyzed by Lip
(Fig. 4). Furthermore, a wide variety of cleavage reactions of p - 0 - 4 compound clarified
by Umezawa et al. with "C,"O-labeled p - 0 - 4 substrates [78,79] were summarized pre-
viously [8].
Although the mechanism of enzymatic degradation of p - 0 - 4 model compounds is
more complex than that of the p-l model, the results obtained indicate that the initial step
of these diverging fragmentations is the one-electron transfer to form the aryl cation radical
in ring A or B of the substrate, yielding both side-chain and ring-cleavage products. Lignin
contains other dimeric substructures such as pinoresinol @-p),phenylcoumaran ( p - 3 ,and
biphenyl (5-5) structures. The enzymatic degradation of these model compounds has not
yet been investigated in detail, except for the oxidative breakdown of the biphenyl model
compound [80,81].
c. Aromatic Ring Cleavage qf Lignin Model Substrates. Aromatic ring cleavage of
lignin prior to its depolymerizationwas postulated by Chen et al. [82] from chemical
analyses of decayed wood lignin, and ring-cleavageproductsformed from the p - 0 - 4
substrate were first reported by Umezawa et al. [83,84]. In this case also, a one-electron
oxidationmechanismwasemployedtoexplaintheringopening reaction as shown in
Fig. 5.
Thus, the "direct" aromatic ring cleavage of nonphenolic p - 0 - 4 model compounds
was demonstrated in an in-vitro system with Lip in the presence of hydrogen peroxide,
OMe OMe
Veratraldehyde Guaiacol
and the p,y-cyclic carbonate, y-formate, and diesterof oxalic acid (a-arylglycerol-p-meth-
yloxalate) were identified by Umezawa and Higuchi, who rationally explained the reaction
mechanism for the ring cleavage reactions [84].
Alternatively, Hattori et al. [85,85a] successfully elucidated a new type of oxygen-
ation mechanism for the ring cleavage of veratryl alcohol(VA); the unprecedented re-
giospecific oxygenation with either water or dioxygen was demonstrated to occur during
the enzymatic ring rupture (Fig. 6).
d. C a - C p Bond Cleavage of Lignin Carlwhydrate Complex Model Substrcrte. In
spite of the importance of the enzymaticbreakdown of lignin carbohydratecomplex
(LCC), it hasbeen little investigated. However,Tokimatsu et al. havesynthesizedfour
stereochemically pure isomers of p - 0 - 4 LCC model compound [86] and confirmed that
the LCC model substrates also undergo the Ca-Cp bond cleavage by the Lip system as
a result of breakdown of the benzyl ether bondings between glucose and the p - 0 - 4 lignin
model compound moiety [87] (Fig. 7).
CH0
OCH3
OEt
-
- OEt
at 22Oc
Erythro-f o m
or
Threo-f o m
FIGURE 8 Proposed mechanisms for the degradation of lignin by veratryl alcohol cation radical
as a redox mediator [92].
CH0
OCH3
veratraldehyde
@ OC
OCH3
YOOH
VA’
’COOH
FIGURE 9 Proposed mechanisms for the degradation of oxalate by veratryl alcohol cation radical
[96]. The lower the level of oxalate in this reaction system, the more efficiently VA is oxidized to
form cation radical, which is converted to veratraldehyde(route a) or attacks lignin and other organic
compounds. The higher the level of oxalate, themoreefficiently scavenged is the cation radical
intermediate back to VA.
Degradation
Bioremediation
Relation
to ofinLignin 557
B. Manganese Peroxidase
Manganese peroxidase (MnP) was first discovered by Kuwahara et al. [4]. Purification and
characterization of MnP were described by Glen and Gold [ 1011 and Paszcsynski et al.
[102]. Like Lip, MnP is a hemoprotein of about 46,000 daltons with protoporphyrin IX
as a prosthetic group. This enzyme catalyzes oxidation of Mn'+ in the presence of hydro-
gen peroxide to form Mn7+ as a direct oxidant, which is reduced back to Mn" substrate
levels after oxidation of lignin and other oxidizable organic substrates present (Fig. 10).
Thus, manganese acts as a redox mediator like VA. However, Mn7'. is not strong enough
to oxidize the nonphenolic lignin moiety. Nevertheless, in the presence of Tween 80 (as
a source of unsaturated fatty acid) in the reaction mixture, nonphenolic lignin substructure
was cleaved by MnP 11031. Similarly, polycyclic aromatic hydrocarbons such as phenan-
threne was oxidized to yield diphenolic acid by MnP catalysis [104].
It appears that MnP is more widespread than Lip among white-rot fungi and plays
a key role in the initial stages of lignin catabolism.
At present, MnP has been considered enzyme with the most potential in the bio-
bleaching of kraft pulps, because this enzyme system from Phanerochaefe chtysosporium
has been shown to break down high-molecular-weight chlorolignin [ 1051 and because a
good correlation between fungal biobleaching activity and MnP activity was also observed
132,1061. Thus, the MnPsystemsappear to bemoreimportantthan Lip for practical
application [ 1071. Actually, Kondo et al. [l081 succeeded in brightening hardwood kraft
pulp by 43 points by six treatments with the MnP system and alkaline extractions. How-
ever, the mechanism for the enzymatic bleaching reactions catalyzed by the MnP system
is not clearly understood.
C . Laccase
Laccase (EC 1.10.3.2), secreted as an extracellular enzyme from white-rot fungi[ 109,1101,
is one of the copper-containing enzymes, catalyzing oxidation of p-diphenol to p-benzo-
Oxidized
Lignin
FIGURE 10 Proposed mechanisms for the oxidation of lignin catalyzed by MnP in the presence
of HzOz and Mn" 141.
558 Shimada Hattori and
quinone in the presence of dioxygen. Laccase (Lac) has been called polyphenol oxidase,
or phenolase, similar to urushiol oxidase isolated from the Japanese lacquer tree. Like the
lignin peroxidase, Lac is not specific to organic substrates and oxidizes not only phenolic
compounds such as lignin-related compounds, coniferyl alcohol, vanillic acid, and p-cresol,
but also ascorbic acid, p-phenylenediamine [ 109,tl l], and 2,2'-azinobis(3-ethylbenzthia-
zoline-6-sulfonate) (ABTS) [ 11 21. However, it differs from tyrosinase (EC 1.14.8. l), which
is a copper-containing o-diphenol oxidase with monooxygenase activity acting on tyrosine.
Although the correlation between occurrence of Lac and white-rot fungi has been
well known as the Bavendamm reaction since 1928 [ 1131, the direct participation of Lac
in lignin biodegradation has been questioned, because the Lac system preferentially re-
polymerizes lignin rather than depolymerizes it. Furthermore, the enzyme is incapable of
oxidizing nonphenolic moieties oflignin because of the low redox potential of the oxidized
enzyme. However, Bourbonnais et al. first reported that Lac catalyzes both the preferential
depolymerization of lignin [ 1141 and oxidation of the nonphenolic lignin model substrate,
[115], i.e., VA, in the presence of ABTS(Fig. l l ) , whichmightserve as the electron
carrier like VA or Mn(I1) involved in the Lip or MnP system, respectively.
They found that the two Lac isozymes gave similar qualitative effects on kraft lignin
and residual lignin in kraft pulp, with no evidence for a marked preference to depoly-
merization by either isozyme alone. However, the presence of the mediator ABTS pre-
vented and reversed the polymerization of kraft lignin by either of the Lac isoforms. When
the delignification of hardwood and softwood kraft pulps with the two isozymes and the
mediator was compared, either Lac was found to be able to reduce the kappa number of
pulp, but only in the presence of ABTS.
Interestingly, Eggert et al. have recently found that a secondary metabolite, 3-hy-
droxyanthranilate, which was produced by the white-rot fungus Pycnoporus cinnabarinus,
served as a mediator in Lac-catalyzeddegradation of a nonphenolic P-0-4-type lignin
model compound and synthetic lignin [ 1161. They have further reported that laccase-less
mutants of Pycnoporus cinnabarinus were greatly reduced in their ability to metabolize
I4
C-ring-labeled DHP [ 1171. In fact, this white-rot fungus does not contain Lip nor MnP.
Thus, it is clearly demonstrated that Lac is essential for lignin degradation by white-rot
fungi which lack Lip and MnP.
Alternatively, Marzullo et al. [ 1 1 81 reported that the cooperative actions of Lac and
VANA oxidase systems prevented the repolymerization of phenolic compounds and re-
duced the molecular weight of lignosulfonates to a significant extent. In this system, the
FAD-containing VA oxidase can prevent the recondensation of the phenolic reaction prod-
ucts, since the flavoprotein oxidase, like cellobiose-quinone oxidoreductase in Phanero-
chaete chrysosporiunz, is able to reduce laccase-generated quinoids and phenoxy radicals
[ 1191.
0 2
H20
x
At present, no unifyingconceptwhich
accaseox
~ B ~ s
mediation
o x
may explain a wide variety ofoxidative
reactions catalyzed by Lip, MnP, andLachas yet been established, but it is becoming
x ~ L i g
Oxidized
Lignin
n i n
common that each of these three lignin-degrading enzymes seems to require its own redox
mediator. And it would be interesting to find a possible occurrence of a common mediator
for Lac systems among white-rot fungi, if any.
As to industrial use of an enzyme system for pulping and kraft pulp bleaching, it is
still challenging how to solve serious cost problems in practice. However, there are only
one or two successful examples of introducing a soft enzyme catalyst into a hard pulping
industry. The most successful one is the utilization of commercial lipase [ 1201 to eliminate
the pitch problem in papermaking and xylanases [ 121,1221 for bleaching of unbleached
pulps at the industrial and commercial level.
HO'
1
0 2
FIGURE 12 Various forms of active oxygen species produced during reductive oxygen metabo-
lism or light (hv) irradiation. The dashed arrow indicates a conceptual path for formation of the
two-electron-deficient oxygen atom of ''oxene,"which is almost equivalent to the active oxygen
species in high-valent oxo-ironporphyrincomplex(compound I of peroxidaseandcytochrome
P-450).
FIGURE 13 Proposed mechanisms for the production of active oxygen species mediated by ver-
atryl alcohol cation radical and oxalate in the LIP system 1451.
Degradationof Lignin in Relation to Bioremediation 561
V 1/ LFW)
Lactone ZFe(I1)-Oxalate
Cellobiose Autoxidation
dehydrogenase
W. BlOMlMETlCAPPROACH
We reported that synthctic and natural iron-porphyrins mimicked the Lip enzyme catalyz-
ing Ca-CP bond cleavage and ring-opening reactions of the nonphenolic lignin model
substrates [S.?,1381, proposing the one-clectron oxidation mechanism for the cleavage rc-
action (61. Since then, ligninase-mimetic systems have been receiving widespread interest
from basic and practical viewpoints 1 1 39- 14 I 1.
Recently, a wide variety of model systems with water-soluble synthetic metallopor-
phyrins, phthalocyanines, and other metallocomplexes have been reported for lignin deg-
Goo-
y coo-
c0,-+c0,"~ Mn (11)
Reactive
Species 1 .
FIGURE 15 A proposcd role o f oxalate i n brcnking down ol'thc recalcitrant lignm with the MnO,/
oxalntc system undcr acrobic conditions.
562 Hattori and Shimada
radation. Among them, polyoxometalate (POM) complex has been focused on as quite a
promising system in many respects for bleaching unbleached kraft pulps [ 1421.
electron-withdrawing group which are not oxidized by Lip. The results indicated that the
muconic acid esters a s the ring-opening products are even selectively produced in the 3S%
yield based on the substrates, which might be promising for further application to the ring
openings of modified kraft lignin.
They also succeeded in synthesizing binuclear [Mn (11) (bipyridyl) Fe(II1) (porphy-
rin)] complex as a MnP model which showed the first evidence for reversible formation
of Mn (111).
Alternatively, metallophthalocyanines (azaporphyrins) have also rcceived much at-
tention as catalysts mimicking lignin-degrading enzymes. Zhu and Ford have reported that
VA was oxidized in the presence of dioxygen or hydrogen peroxide in water [ 1 SS, 1561.
Tanihara et al. [ 1571 reported oxidative cleavage of p-0-4 lignin nlodel compounds
with water-soluble iron-octacarboxyphthalocyaninei n the presence of terr-butylhydroper-
oxide in the alkaline medium. Their phthalocyanine system simulated Lip activity in C a -
C p bond cleavagc,p-ethercleavage.ring-openingcleavagereactions. However, these
water-soluble phthalocyanines were unstable and easily bleached and arc not useful for
application of delignitication.
Degradationof Lignin in Relation to Bioremediation 565
Quite recently, Weinstock et al. [l421 have developed a promising pulp bleaching
technology, based on the use of a polyoxometalate (POM) such as SiVW,,O,Oand oxygen.
As the delignification reaction cycle shown in Fig. 18, the reaction system is reminiscent
of the biomimetic system of Lac, which requires oxygen as oxidant.
They claimed that the approach is unique by intrinsic design and implementation in
that the versatile capabilities of a single POM species are utilized to facilitate multiple
operations that sum to the selective conversion of wood pulp to paper and nontoxic prod-
ucts ( H 2 0 and CO,), using air as the oxidant and water as the only solvent.
V. CONCLUDING REMARK
Since the first edition of this book was published in 1991, tremendous numbers of reports
onmicrobial,enzymatic and biomimeticdegradation of lignin and seemingly unrelated
compounds, including pollutants and xenotiotics, have been reported. Although the one-
electron oxidation mechanism for lignin degradation now allows us to understand more
comprehensively oxidative degradation processes of lignin, it is still deceptively simple to
apply, in practice, bioligninolytic systems to the pulp and paper industry. Nevertheless,
new vital ligninolytic fungi were screened, investigated, and characterized. New aspects
of radical reactions of lignin and other compounds catalyzed by manganese peroxidase
and laccase in the presence of Tween 80 and ABTS, respectively, have been receiving
much attention. Another intcresting aspect of free-radical reactions caused by ligninolytic
systems has focused on the generation of superoxide anion radicals and hydroxyl radicals
as the result of oxidative degradation of oxalate in the presence of dioxygen and ferric
ions. These nonspecific radical species may play a key role i n degradation of lignocellu-
Degradation
Bioremediation
Relation
to ofinLignin 567
REFERENCES
I. K. V. Sarkanen and C. H. Ludwig, Lignins: Occurrorlce. Fonncrtion, Structure t l l ~ dR ~ ( K -
tiom. Wiley-Interscience, New York, pp. 1-9 1 1 16 ( I97 1 ).
I a. P. Ander and K. E. Eriksson. Svensk., 78: 643 (1975).
2. M. Tien and T. K. Kirk, Scirr~ce.221: 661 (1983).
3. J. K. Glenn, M. Morgan, M. B. Mayfield, M. Kuwahara. and M. H. Gold, Biochenr. Biophys.
Res. Comnm., 114: 1077 (19x3).
4. M. Kuwahnra. J. K. Glenn, M. A. Morgan. and M. H. Gold. FEBS LcW., 160: 247 (1984).
S . P. J. Kersten. M. Tien. B. Kalyanaraman. and T. K. Kirk. J. Bioi. C / I ~ I260: I I . 2609 (19x5).
6. T. Habe, M.Shimada. T. Okamoto, B. Panijpan,and T. Higuchi, J . C1zer11.Soc. Chern.
c o l ~ l l l l L l l l . , 108.7: I323 (1985).
7. H. E. Schoemaker, P. J . Harvey, R. M. Bowen. and J. M. Palmer. FEBS Lett.. 183: 7 (1985).
X. M.Shimadaand T. Higuchi, in Wood andCellulosicChemistry(D. N. S. Hon and N.
Shimishi, eds.), Marcel Dekker, New York. p. SS7 (1991 ).
9. T. K. Kirk and H . M. Chang. in Riotcxhrlology i n Pulp tuld Pnpc’r Mmzufc~ctwc(T. K. Kirk
and H.-M. Chang. eds.). Butterworth-HeinetIlann, Boston. p. 666 (1990).
IO. M. Kuwahnra and M. Shirnada. in B i o t e c h o l o g y i r l Pulp m e / Ptrprr lnt/u.stry (M. Kuwahara
and M. Shimada. eds.), Uni Publishers, Tokyo. p. 544 (1992).
1 I. P. Ander. M. C. Ferreira, and J. C. Duarte, FEMS Microbiol. Rev. 13: 121 (1994).
12. E. Srebotnik and K. Messner, B i o t e c ~ h ~ z o l oirlg ~Pulp m t l P q w r Imht.sr,y. Facultas Univ-
ersitatverlag. Vienna, pp. 1-66 1 ( 1997).
13. J. A. Buswell and E. Odier. CRC Ctit. Re\: Biorechrzol., 6 : 1 (1987).
14. T. K. Kirk and R. L. Fnrrell, Annu. Ret: Microbiol., 41: 465 (1987).
IS. R. A. Blanchette. A m u . REI: Phytopcrthol.. 29: 38 I (199 l ) .
I 6. B. Kurek, i n Pltrrlt Prrosidtrtr.sc~.s 1080-1900. lbpics t r n d Dc.tcrile.t/ Litc.rtrttrw 0 1 1 Mo/eclf/(lr;
Aspec~t.s(C. Panel, Th. Gaspar, and H. Greppin, e&.), Uni-
Hiochc~rrlic.crl,t r r l d Ph~.siologicc~/
versity of Geneva. p. 139 ( 1992).
17. M. H. Gold and M. Alic, Microhiol. Kc,\:, 57: 605 (199.3).
1x. M. Shimada. D. B. Ma, Y. Akamatsu. and T. Hattori, F E M S Microhiol. Rc.1:. 13: 285 (1994).
I 9. D. P. Barr and S. D. Aust. Emiron. S c i . T~c~hrrol., 28: 78A (1994).
20. L. Otjen. R. A. Blanchette. M. Eftland, and G. Leatham. Hol;fifor.sc.h.. 41: 343 ( 1987).
21. A. Hatakka. FEMS Microbial. Ret:. I S : 25 ( 1994).
22. K.-E. Eriksson and L.Vallander. i n Lignirr Biodc.Rrtrt/trtiorl: Mic~robio/ogy,C/lc,/ll;,ytl:\’ cl,l(/
Potcrltictl Appliccttiorl. Vol. 11 (T. K. Kirk, T. Higuchi, and H. M. Chang. e&.). CRC Press,
Boca Raton, FL, p. 2 I3 ( 1980).
13. S. S. Bar-Lev. T. K. Kirk, and H. M. Chang. nrpp;,65: 1 I (1982).
24. I. Akamatsu. K. Yoshiharn, H. Kamishima. and T. Fujii, Moku:tri GtrMtr;,s/f;,-30: 697 ( 1984).
25. G. F. Leatham. G. C. Myers, T. H. Wegner. and R. A. Blanchette. in Biotechtzo/o,yT ;!I P L ~ / / )
t r t d Pnpc.r Mtrr?ufirc.tlrrc (T. K. Kirk and H. M. Chang. eds.), Butterworth-Heinelnann, Bo+
ton. p. 193 ( 1990).
26. M. Akhtar, M. C. Attridge.R. Blanchette, G. C. Nyers. M. B. Wall, M. S. Sykes, J. W,
Koning, Jr.. R. R. Burgess. T. H. Wagner. and T. K.Kirk. in Bjotty./frlo/og? pit//) c l r l c /
Shimada
568 and Hattori
59. H. Wariishi,K. Valli, and M. H. Gold, Biochenz. Biophys. Res. Cornrnur7., 176: 269 (1991).
60. V. Renganathan, K. Miki. and M. H. Gold, Arch. Biochern. Biophys., 241: 304 (1985).
61. I. G. Gazarian, L. M.Lagrimini, S. J. George,andR. N. F. Thorneley, Biockenz. J., 320:
369 ( 1996).
62. T. L. Poulos, S. L. Edwards, H. Wariishi, and M. H. Gold, J. Biol. Chem., 268: 4429 (1993).
63. K. E. Hammel, M. Tien, B. Kalyanaraman, and T. K. Kirk, J. Biol. Chern., 260: 8348 ( 1985).
64. V. Renganathan, K. Miki,and M. H. Gold, Arch. Biochern. Biophys., 246: 155 (1986).
6.5. F. Nakatsubo, I. Reid, and T. K. Kirk, Biochern. Biophys. Res. Comrnurz., 102: 484 ( 198 1 ).
66. T. K. Kirk, F. Nakatsubo. and I. D. Reid, Biochen7. Biophys. Res. Cormzurz., 111: 200 (1983).
67. P. Hall, Erz,-?,rne Microhinl. Techno/., 2: 179 ( 1980).
68. L. J. Forney, C. A. Reddy,M. Tien, and S. D. Aust, J. B i d . Chenz.. 257 1 1455 (1982).
69. H. Kutsuki andM. H. Gold, Biochenl. Biophys. Res. Cornrnun.. 109: 320 (1982).
70. H. lshikawaand T. Oki, Mokrczni Gnkkaishi, 12: 101 (1966).
71. T. Fukuzumiand T. Shibamoto, Mokuzcri Gnkknishi, 11: 248 (1965).
72. T. Fukuzumi, H. Takatsuka, and K. Minami. Arck. Biochern. Biophys., 129: 396(1969).
73. A. Enoki, G . P. Goldsby, and M. H. Gold, Arch. Microbiol.. 129: 141 (1981).
74. T. Umezawa and T. Higuchi, FEMS Microbiol. Lett., 26: 123 ( 1 985).
75. T. K. Kirk, M. Tien, P. J. Kersten, M. D. Mozuch, and B. Kalyanaraman, Biochern. J.. 236:
279 (1 986).
76. K. Miki, V. Renganathan,M.H.Gold, Biochemistry, 25: 4790(1986).
77. T. K. Kirk and M. Tien, in Recent Advnrzces i n Ligrzirz Biodegrudotior7 Resecrrch (T. Higuchi,
H.-M. Chang, and T. K. Kirk, eds.), Uni Publishers, Tokyo, p. 233 (1983).
78. T. Umezawa, Wood Res., No. 75: 21 (1985).
79. T. Umezawa and T. Higuchi, in Erzzyrnes irz Bion~cr.s.sCm~versiot~ (G. F. Leatham and M. E.
Himmel, eds.), American Chemical Society, Washington, DC, p. 236 (1991).
80. T. Hattori and T. Higuchi, Mokuzcri Gnkknishi, 37: 542 (1991).
81. S. Yokota, T. Umezawa, T. Hattori, and T. Higuchi, Mokuzcri Gctkke~ishi,37: 644 (1991).
82. C. L. Chen, H.-M. Chang, and T. K. Kirk, J. Wooe/ Chem. Techr7ol., 3: 35 (1983).
83. T. Umezawaand T. Higuchi. FEBS Lett., 182: 257 (1985).
84. T. Umezawaand T. Higuchi, FEBS Lett.. 218: 255 (1987).
85. M. Shimada, T. Hattori, T. Umezawa, T. Higuchi,and K. Uzura. FEBS Lett., 221: 327
(1987).
85a. T. Hattori, Wood Res., No. 78: 15 (1992).
86. T. Tokimatsu. T. Umezawa. andM. Shimada, Hol;for.sch., 50: 156 (1996).
87. T. Tokimatsu. S. Miyata,S.-H. Ahn. T. Umezawa, T. Hattori,andM.Shimada, MokLczcri
Gukkcrishi, 42: 173 ( 1996).
88. I. Chet, J. Trojanowski, and A. Hutterman, Microbiol. Lett., 29: 37 (1985).
89. J. K. GlennandM. H. Gold, Appl. Environ. Microbial.. 45: 1741 (1983).
90. D. K. Joshi and M. H. Gold, Biochernistn), 33: 10969 ( 1994).
91. J. A.Bumpusand S. D.Aust, Scierzc~,228: 14 (1985).
92. P. J. Harvey. H. E. Schoemaker, and J. M.Palmer: FEBS Lett., 195: 242 (1986).
93. K. Valli, H. Wariishi. and M. H. Gold, Biochemistry, 29: 8535 (1990).
94. A. Khindaria, T. Grover. and S. D. Aust, Bioc.h~rrzi.str~\;34: 6020 (1995).
95. A. Khindaria, 1. Yalnazaki, and S. D. Aust, Biocl~~wzi~try. 35: 6418 (1996).
96. Y. Akamatsu, D. B.Ma. T. Higuchi, and M. Shimada. FEBS Lett., 269: 261 (l9c)O).
97. J. L. Popp. B. Kalynaraman,and T.K. Kirk. Bioclwrnistn, 29: 10475 ( 1990).
98. M. Samejima and K. E. Eriksson, FEBS Let/.. 2Y2: 15 1 ( 1991).
99. D. B. Ma. T. Hattori, Y. Akomatsu. M. Adachi. and M. Shimada, B i o s c i . H i o c ~ h c ~ r r Biorec4-
r.
no/., 56: I378 ( 1992).
100. A. Khindaria. D. P. Barr,and S. D. Aust. Uioc~lrc~rr7i.str~~. 34: 7773 ( 1995).
1 0 1 . M. H. Gold and J. K. Glenn, Metl7otl.s Erzzyno/., 161: 258 (1988).
102. R. L. Paszcsynski, R. L. Crawford.and V. B. Huynh. Mc,thoc/.s En:yrzol. 161: 264 (1088).
103. K. A.Jcnsen, Jr.. W. Bao, S. Kawai. E. Srebotnik,and K. E. Hummel. App/. Erl17irorz.
Mic,robio/.. 62: 3679 ( 1996).
570 Shimada Hattori and
Misato Norimoto
Kyoto University, Kyoto, lapan
1. INTRODUCTION
Wood is strong along the grain for its weight, moderately hard, and easy to work. Its grain
pattern, texture, and color are beautiful and give a natural feeling. Light reflecting from
the wood surface does not glare and provides a mild stimulus to our eyes. As wood is
highlyhygroscopicandhas large specific heat, the air in a roomsurrounded by it is
moderatelyconditioned.These excellent properties of woodsas structural and interior
finishing materials for residential construction result from wood structures ranging from
macroscopic to microscopic and the molecular level.
In many situations, however, the difficulties encountered in wood utilization are
related to its dimensional instability to moisture, biodegradability, and flammability. In the
case of softwoods, especially when used for flooring boards and furniture, hardness and
abrasion resistance may also become a problem. Among these difficulties, the changes of
wood properties caused by ambient humidity variations are of great importance.
Dimensional stability has been evaluated by measuring dimensional changes in load-
free wood specimens. In actual practice, however, wood components are often subjected
to humidity variations under mechanical force. This is the main factor causing an excep-
tionally high creep strain of structural members and tonal instability in wooden musical
instruments. Generally speaking, any type of instability of wood caused by humidity var-
iations originates in its high hygroscopicity. One of the main reasons for chemical modi-
fication of wood is to reduce its dimensional instability.
Chemical modification of wood is defined as any chemical reaction between some
reactive part of a wood cell wall component and a simple single chemical reagent that
forms a covalent bond between the two components [ l ] . However, in this chapter, it is
understood in a wide definition and includes the use of heat, steam, resins, or other chem-
icals. In most cases, theyaffect the amorphous components of the cell wall, the cell lumens
sometimes being filled by resins or chemicals.
Some modifications of wood reach the cores of cellulose microfibrils, destroying the
crystalline structure and eliminating most of the composite structures of the wood. Con-
sequently, the resulting material does not have any of the characteristic properties of wood.
In compensation, it may be provided with other properties such as thermoplasticity 121.
Chemical modification of wood as discussed in this chapter excludes such radical modi-
fications. It may reduce some defects relative to wood utilization, enhance its properties.
573
574 Norimoto
and create new performance or functions, while keeping the bulk of the superior mechan-
ical properties of wood.
In this chapter, a classification of chemically modified woods will be proposed and
the properties of typical modified woods will then be analyzed in relation to their structure.
B4 55
represented as two independent members for the sake of clarity, but in reality, they belong
to a fully interconnected framework including not only the crystalline core of the micro-
fibrils, but also some of the surrounding molecules that do not react to water. The chain
has a water-reactive zone at its boundary, illustrated by an OH group ( h ) represented by
a small open circle, which is available for hydrogen bonding with an adjacent OH group
or with water. The ring ( c ) refers to the outer layer of the secondary wall (S, layer), which
preventsmoisture-inducedexpansion(swelling) of the S2 layer fromexceedinga limit
(hoop effect). Neighboring chains are linked to each other by the hydrogen bond. Water
sorption on the OH group has a double effect. First, the change of matrix volume forces
lateral displacement of the twochains ( a ) in the direction indicated by d. Second, it
weakens the connection between the chains and facilitates slippage in the direction indi-
cated by e, resulting from local shear stresses. At the macroscopic level, the first effect
results in swelling or shrinkage and the second in creep or stress relaxation.
The small filled circles in B-2 and B-3 indicate that the OH group has been substi-
tuted withchemicalbondingand isnot available anymore.The large filled circles in
B-3 andB-4 indicate the bulking effect caused by theintroduction of large molecules
between the constituents. A line shows a covalent bond. The cross-linking effect is thus
pictured by an unbroken sequence of lines and small circles linking the two chains shown
in B-2, preventing both lateral expansion ( 6 ) and molecular movement ( e ) . In B-3, the
reactant establishes a stable bond on one side and a bulking effect occurs. If the reactant
is hydrophilic, it suppresses hydrogen bonding on one side but simultaneously creates a
new sorption site at the other. In B-4, no stable bond is established between the reactant
and the constituents. If the bulking agent is extremely hydrophilic, it establishes extensive
hydrogen bonding with the constituents. However, if the bulking agent is hydrophobic, it
does not interact with the constituents or with water. The pyrolysis and oxidation of the
wood constituents as well as the crystallization of cellulose reduce the water-reactive OH
groups. These structural changes may be represented by B-5. Any type of chemical mod-
ification can be characterized by a combination of the two criteria (A) and (B).
B. ChemicalModificationTreatments
This chapter deals with eight kinds of chemical treatments: vaporous formalization (F),
acetylation (A), propylene oxide (PO) treatment, phenol-formaldehyde (PF) resin impreg-
nation, polyethyleneglycol(PEG)impregnation,wood-methylmethacrylatecomposite
(W), heating (H), and steaming (S).
Formalization was done in a closed vessel with tetraoxane as well as paraformal-
dehyde and SO, (catalyst) at 120°C for various lengths of time. It is a reaction involving
the formation of OCHz bridges between the OH groups of wood constituents by HCHO.
A small molecular bridge is made while the chains are close to each other, so the structure
of the resulting product is (A-2) +(B-2).
Acetylation was performed in neat acetic anhydride at 120°C for 10 h, followed by
leaching for 10 h in boiling water and oven drying overnight. As the OH group is substi-
tuted with the hydrophobic and bulky OCOCH, group, the resulting structure belongs to
(A-2) + (B-3).
Propylene oxide treatment was done using PO and TEA (catalyst) inan autoclave
at 120°C for 2 h. The addition of PO provides wood with the structure of (A-2) + (B-3).
The OH group is substituted with the hydrophilic and bulky OCH2CH(OH)C2H, group.
I
576 Norimoto
ASE = ___
- x 100 (%)
S,,
where S is the volumetric swelling of treated specimen between a dry state and a wet
state, and S,, is the value measured for the untreated wood under the same experimental
conditions. An ASE of 100% means that the specimen is perfectly stabilized, whereas an
ASE of 0% means that the treatmenthas no effect at allon dimensional stability. A
negative ASE has the opposite effect on stabilization. The experiment to evaluate dimen-
sional stability by ASE involves measuring dimensional changes of load-free specimens.
Therefore,ASEalonedoesnotgiveany indication of the modification of mechanical
properties that involve the action of external force on the material.
The formation of a bridge between adjacent OH groups represented by B-2 in Fig.
1 may reduce the swelling of the cell wall. In vaporous formalization, short inflexible
cation Chemical 577
578 Norimoto
U
FF
0 1 2 3 4 O 0 1 2 3 4
. 2 3 4
l
.'"
I I I
PG
U
G
0
0 1 2 3 4
W
U
0 1 2 3 4 O 0 1 2 3 4
t (day)
PO
e PG
e
efficient treatment regarding ACE [3,19,20]. The same results were obtained for the spec-
imens treated with dialdehydessuchas glyoxal and glutaraldehyde, which form cross-
links between the OH groups [20]. In the case of treatments that induce cross-linking,
good correlation could be expected between ASE and ACE [20].
Acetylation yielded high values of both ACE and ASE. Long-term creep tests con-
ducted in an uncontrolled room showed a remarkable reduction in creep deformation after
acetylation [ 191. The creep deformation under humidity changes for particleboards with
the acetylated particles was also suppressed to a large extent. The retention of the modulus
of elasticity (MOE), modulus of rupture (MOR), and internal-bond strength after the creep
test increased with increases in the WPG of particles 1211. To stabilize piano tones under
humidity changes, a pin block must have dimensional stability. Since the string is under
great tension, any movement of the tuning pin causesa reduction in the pitch of the
vibrating string. Measurement on a full-scale model of the string-sustaining part using
acetylated and control wood pin blocks showed that changes in the resonant frequency of
strings with humidity change for the acetylated blocks were much less than that for the
control wood blocks 122,231.
The comparison between acetylation and propylene oxide treatment is of special
interest. These treatments have high ASEs. However, the most striking difference between
these treatments is the opposite effect on mechano-sorptive creep. Acetylation yielded a
positive ACE, but epoxide treatment yielded an extremely negative value. This meant that
epoxide treatment stabilized stress-free wood but destabilized loaded wood. Although the
reactant saturates the OH groups, it produces new ones during the reaction. Furthermore,
as a bulking agent, the reactant increases the accessibility of water molecules to the water-
reactive region. This results in more water sorption and more creep.
Although most PF resin is not supposed to react with the constituents of the cell
wall and thus should act as a pure bulking agent, a small proportion of linkages may be
expected. The hydrophobic nature of the bulking agent resulted in a good ACE for a quite
modest ASE.
PEG impregnation yielded results similar to those of epoxide treatment, resulting in
an excellent ASE but a negative ACE. The creep deformation of the PEG-treated wood
was extremely large even at a constant humidity condition, because PEG acts as a plas-
ticizer 1231.
A low ASE of wood-MMA composite showed that only B small proportion of A "
penetrated into the cell wall. Therefore, neither ASE nor ACE can be affected unless the
cell wall is modified at the molecular level.
given to vibrate the beam and is proportional to the square root of (E'y-l)l(tan 6 - y)
[24].
A large E ' l y and small tan 6 give a high converting efficiency which characterizes the
acoustical quality of wood used for soundboards 125,261.
The dynamic mechanical properties of Glehn's spruce (Picea glehnii) were measured
at constant humidities of 0%, 35%, 60%, and 85% RH at 20°C by means of the free-free
vibration method [27,28].
Figure 4 shows the effect of chemical treatment on the relationship between loga-
rithms of tan 6 and E'ly [29]. A linear regression line indicated by a dotted line on each
graph was obtained for the untreated wood using all of the values measured before the
treatment. This linear relationship was found regardless of species [30-331, measuring
direction [34], or moisture content [35]. Some comment should be made concerning this
regression line. The E ' l y along the grain characterizes the average rigidity of the cell wall
[33,36], whereas the tan 6 represents the relative amount of viscous strain to elastic strain
and the participation of the matrix in the deformation process. The existence of a corre-
lation between E ' l y and tan 6 for the untreated woodisby no means accidental. Both
quantities depend mainly on the mean microfibrillar angle of the S, layer and very little
on density [33]. As variations in both E ' l y and tan 6 originate from the same ultrastructural
factors, they should not be independent of each other.
When different viscous behavior is apparent between a treated specimen and the
untreated reference. we do not know a priori whether the difference is due to the treatment
or originates from the bad matching of the two specimens. However, we did confirm that
it is not the latter case by observing that the point representing the modified wood data
not only differs from that of the untreated wood but also lies far from the dotted line.
According to this interpretation, a perfect correlation between E'ly and tan 6 at a given
level of humidity corresponds to the ideal case where all structural parameters would be
-2.4
1.00.8
' I
1.2 1.6 1.4
I
A.'\\',
-2.21 -A, $,'
-2.4
0.8 1.01.4 1.2
1.6
FIGURE 4 Relationships between logarithms of dynamic loss tangent (tan S) and specific dynamic
modulus ( E ' l y ) for chemically modified woods at various relative humidities. Dotted lines represent
experimental correlation for untreated wood. See legend to Fig. 2 for abbreviations.
582 Norimoto
either constant or exactly correlated with the microfibrillar angle. The deviance of a given
plot relative to the reference line is thus explained by the disturbing action of these other
structural parameters.
The tan 6 decreased as the humidity changed from 0% to 35% RH and then rose
with increasing humidity. The E ' l y remained about the same from 0% to 35% RH and
then decreased with increasing humidity. In an oven-dried state, molecular chains in the
amorphous regions of the cell wall are unnaturally distorted. This causes a smaller E ' l y
and a greater tan 6. With increasing moisture content from oven drying, the distorted
molecular chains are rearranged. As a result, at 35% RH, a more stable state for the cell
wall structure is obtained. Above 35% RH, water acts as a plasticizer that allows more
molecular movement and the cohesive forces between molecules are decreased, resulting
in decreasing E'ly and increasing tan 6.
In vaporous formalization, cross-linking occurs in dry state of the cell wall and
prevents swelling. This decreases the mobility of molecular chains, thus reducing tan 6 at
all humidity levels. This treatment decreased tan S as much as 40-45% in the frequency
range from 150 to 500 Hz [37]. The reduction was smaller at high frequencies and larger
at low frequencies. The E ' l y at 65% RH remained unchanged in the grain direction and
increased 10% in the direction perpendicular to the grain [37]. By a sensory evaluation
test, a violin having a treated bridge was judged to be comprehensively better than the
same violin having an untreated bridge [37]. Vaporous formalization is effective not only
for improving the acoustic properties of wooden instruments but also for their dimensional
changes resulting from humidity variations.
The E ' l y and tan 6 of the acetylated wood decreased as the WPG increased [38].
The E ' l y decreased slightly, while the tan 6 increased greatly as the frequency increased,
but no differences in frequency dependence were observed between the acetylated and
untreated woods [38]. The reduction in E ' l y and tan 6 at 20% WPG was 10-15%. The
hydrophobic OCOCH,groups were subjected to hydrogen atoms in the cell wall. This
reduced the E'ly and tan S owing to the steric hindrance of chains imposed by the bulky
groups. This effect was seen at all humidity levels. However, the humidity-induced changes
of E'ly and tan 6 were greater than those of the formalized wood. Because all available
OH groups were not acetylated, water still was able to bond with the cell wall polymers
and act as a plasticizer. The acoustic properties of wood changes more in a nonequilibrium
moisture condition than in an equilibrium moisture condition. However, acetylation greatly
suppressed these changes, especially in the process of adsorption [23,35,38]. Acetylation
was shown to stabilize tone quality under conditions of changing humidity [23,39].
Propylene oxide treatment (WPG; 22%) results in bonded cell wall bulking as in the
case of acetylation,except that the introduced group is hydrophilic. This made a big
difference as humidity increased. At 0% RH, the treated wood had a slightly lower E ' l y
because of the increase in specific gravity and the tan 6 was lower thanthat for the
untreated wood. At 35% RH, the tan 6 of the treated wood was almost the same as that
for the untreated wood, but at 60% and 85% RH, it was much larger. This was due to the
flexibility introduced into the cell wall polymers because hydrophilic ether allowed water
to act as a plasticizer.
The changes in both E'ly and tan 6 varied with average molecular weights of PF
resin [40]. As the molecular weight of resin decreased, the E'ly increased and tan 6 de-
creased. This result was ascribed to the resin content maintained in the cell walls: the
higher the molecular weight, the less the resin was maintained. Similar results were also
obtained in treatment with MF resin with a low molecular weight. The treatment with
low-molecular-weight PF resin (WPG; 45%) introduced a bulky and hydrophobic group
Chemical
Wood Modification of 583
into the cell wall. Because of the presence of a rigid benzene ring in the resin backbone
as well as reduced moisture sorption, molecular mobility was reduced, which lowered tan
6 for all humidity levels tested.
PEG with a low molecular weight is a hydrophilic polymer that can enter the cell
wall. This very flexible hydrophilic polymer swells and plasticizes the cell wall even at
0% RH, resulting in very large tan S. The hydrophilic nature of the cell wall does not
change as the humidity increases because it does not matter whether PEG or water is
acting as a plasticizer.
When only the lumen was filled or coated with chemical, there was no deviation
from the untreated reference regardless of humidity conditions. The treatment with MMA
(WPG; 138%) indicates that a small amount of this polymer may enter the cell wall, but
in general this treatment only fills the lumen space. The increase in specific gravity de-
creased E’ly. Water can still enter the cell wall and act as a plasticizer, so that the tan 6
increased with increasing humidity.
From a practical point of view, it is of importance to predict the stability to changes
in humidity for chemically modified wood with respect to E’ly and tan 6. To characterize
how E ’l y and tan 6 change with increasing humidity conditions, and E‘ly stability (S,)
and tan 6 stability (S,.) are defined by
A, - B, AT - B,.
S,: = ~ x loo(%), S,.= ___ x loo(%),
A, A,.
where A and B are the relative changes of specific modulus ( E ) or loss tangent ( T ) for
the untreated and treated woods, respectively, A is the change of E ‘ l y or tan 6 between
35% and 85% RH, and E ’ ly and tan 6 are the values at 35% RH.
Figure 5 shows the relationship between SE and ST for the chemically modified
woods. The formalized wood, acetylated wood, and PF resin-impregnated wood had large
positive values of S, and S,., while the epoxide- and PEG-treated woods had large negative
FIGURE5 Relationship between tan S stability (S.,) and E’ly stability (S,) for chemically modified
woods. See legend to Fig. 2 for abbreviations.
584 Norimoto
values of S, and &.These results shows that the greatest stability to changes in acoustic
properties with increasing RH is achieved with vaporous formalization, acetylation, and
PF resin treatment.
IV. FIXATIONOFCOMPRESSIVEDEFORMATION
A. Mechanism of Drying Set
In coniferous woods and fast-growing woods with low density, surface hardness and abra-
sion resistance may become a problem, especially when these woods are used for flooring
boards, furniture, and interior finishing materials. Softwood with useful surface properties
can be obtained by compressing the wood in the direction perpendicular to the grain.
Figure 6 shows the result of radial compression with loading-unloading cycles at
100°C for wet sugi (Cryptomeria japonica) [41-451. Almost complete recovery was ob-
served provided the strain level of the first loading remained small within the linear range.
The curve of the second loading was nearly superimposed on that of the first loading. As
soon as the linear limit (yield point) was exceeded, the strain increased remarkably with
increasing stress. After unloading, there remained a slight residual deformation, so that
most of the strain was not plastic (apparent plasticity). Thesubsequent loading never
followed the previous path. Instead, the curve had a lower initial slope, followed by a new
plateau for a much lower yield stress, indicating a reduction in stiffness. As soon as the
point before unloading was reached, the apparent plasticity started again, as if no unloading
had occurred in between. As an interpretation for this plateau, it is suggested that the
portions of wood with lowest density (earlywood) are crushed first, followed by the next
in terms of density, and so on, until the portions of wood with highest density (latewood)
are crushed as well [44]. When the cells collapsed so much that opposing cell walls
touched at large strains, the curve showed a steep rise (densification) by the deformation
of the cell wall itself. Eventually, the highest previous stress and strain were reached again
simultaneously, so that the original curve corresponding to monotonous loading was
reached exactly where it had been left previously. When a portion has been crushed to
some level, it will be damaged and behave differently during subsequent loading. As far
as the observations by scanning electron microscope (SEM) were concerned, no indications
2
-52
b
0
25 50 1 3
E (%)
FIGURE 6 Nominal stress (+strain ( E ) curve with increasing loading cycles in radial compres-
sion for wet sugi wood at 100°C. Solid and dotted lines represent loading and unloading, respec-
tively. Numbers indicate unloading points.
Chemical Modification of Wood 585
of damages were observed in the cell wall [46]. However, the irreversibility of the defor-
mation process suggests the occurrence of some structural changes inside the cell wall.
Much the same results were obtained for wet specimens at 20°C. On the other hand, a
large residual strain occurred after a loading beyond the yield point in an air-dried con-
dition [45]. However, the strain was largely recovered through hygrothermal treatment.
When a compressed wet specimen is dried under restraint, the stress gradually de-
creases until it has disappeared. The deformation is fixed in the deformed state (drying
set). However, this drying set is only apparent, because it is almost recovered by boiling.
Figure 7 shows the cross sections of the noncompressed specimen (A), the compressed
specimen (B), and the specimen recovered by boiling [47].
The drying set results from the rehardening of the matrix upon cooling and drying
[48]. The elevation of temperature in wet condition softens the matrix, and its two main
constituents, hemicelluloses and lignin, shift from the glassy state to something near the
rubbery state. On the other hand, the microfibrils remain inthe glassy state and are almost
unaffected by moisture or heat. When a load is applied to the material, most of it is
supported by the microfibrils at the local level. Softening of the matrix enables the relative
displacement of the microfibrils in order that the whole framework deforms elastically to
adjust to the local loading. As lignin is a slightly cross-linked high polymer,
its deformation
should be viewed as increased viscoelastic strain rather than plastic flow. The departure
of water molecules due to drying induces the re-formation of hydrogen bonds between
the molecules of the matrix constituents. Together with the temperature decrease, this
process leads to return to the glassy state, where the elastic deformation of the microfibrils
and the matrix are frozen. Accordingly, the set will not be recovered provided no resoft-
ening of the matrix occurs. However, as soon as the matrix is softened again through
remoistening and heating, most of the set is recovered due to the liberation of the energy-
elastic strain stored in the microfibrils and the entropy-elastic molecular movements in the
matrix.
The percentage of size reduction from the original state to the compressed state
(compression set: CS) is calculated by
" -*--- -7
1
A B C
FIGURE 7 Cross sections of noncompressed specimen (A), compressed specimen (B), and spec-
imen recovered by boiling (C) [47].
586 Norimoto
where T,, is the oven-dried thickness before compression and Tc is the oven-dried thickness
after compression. The recovery of set (RS) is calculated by
B. Fixation by Cross-Linking
In order to utilize the compressed wood, permanent fixation of the compressed deformation
is essential. One of the most effective ways of fixing deformation is the formation of cross-
linkings between the wood components. Vaporous formalization is applicable to this pur-
pose. Compressed sugi specimens (CS: 50%) were treated with formaldehyde vapor gen-
erated from paraformaldehyde and SO, in a dry condition at 135°C for 20 min. Figure 8
shows the relationship between RS and treating time 1491. The RS rapidly decreased with
increasing treating time. It should be noted that the deformation was completely fixed by
the treatment in about 5 min. The deformation was also completely fixed by the reaction
with tetraoxane and SO, at 120°C for 2 h [SI]. If the mechanism of fixation by formali-
zation arises from the formation of interchain covalent bonds, the deformation should be
recovered by the scission of the cross-linkings. To confirm this, the formalized compressed
specimen was soaked in a 10 N H,SO, solution. The result showed a remarkable recovery
of set deformation [49]. In liquid-phase formalization. the RS decreased with increasing
treating time, although complete fixation was not realized within the treating time of 250
min [49].
Formaldehyde treatment is also effective in improving the dimensional instability of
medium-density fiberboard (MDF). The untreated MDF swelled infinitely in its thickness
direction when boiled, whereas the treated MDF swelled only about 20% and recovered
its original thickness when dried again [SO]. These results suggested that the dimensional
stabilization of MDF is attributable to the formation of chemical bonds between fibers.
'0 20 40 60 80 100
t (min)
1
li
FIGURE 8 Relationship between recovery of set (RS) and treating time (1) for formalized com-
pressed sugi wood (compression set, CS: 50%).
Chemical Modification of Wood 587
C. Fixation by ResinTreatment
Almost complete fixation of deformation can be achieved by using PF resin that is poly-
merized during the deformation stage.
Oven-dried sugi specimens were soaked in resin solutions of increasing concentra-
tions. After air drying for 24 h, the impregnated specimens were slowly warmed over a
12-h period to 105°C and then compressed in the radial direction using a hot press. The
treated specimens were soaked in water until saturated and their thickness measured. After
drying at 40°C for 20 h and then at 105°C for 4 h, their thickness was again measured.
This procedure was repeated. After the final soaking cycle, the specimens were placed in
a boiling water bath for 2 h, measured, and then oven-dried.
Figure 9 shows the results of soaking-drying cycle test for the compressed speci-
mens (CS: 55%) treated with PF resin (molecular weight: ca. 200). The untreated speci-
mens showed a large increase of RS during the cycle test, while the specimens soaked in
15% and 20% PF solutions effectively resisted swelling. The SEM micrographs showed
no remaining PF resin in the cell lumen. The resin in the cell wall resulted in ASEs of
60-70% at about 20% WPG. Hardness of the products increased with increasing CS and
resin concentrations. A resin concentration of 15% and a CS of 60% increased hardness
by a factor of 3, compared to that of the noncompressed wood; and a 40% resin concen-
tration, by a factor of 4. Therefore, larger values of hardness than those of high-density
hardwoods can be obtained. Abrasion resistances of the specimens compressed to 60%
were about 60% regardless of resin concentrations. Abrasion resistances of the products
increased with increasing CS. The MOE increased with increasing resin concentrations.
The MOR decreased in low concentrations, but increased above 25% concentration.
The compressed specimens (CS: 54%) soaked in 2% and 4% MF (molecular weight:
380) solutions showed a large increase of RS during eight soaking-drying cycles. How-
ever, the specimens soaked in 8%, ls%, and 25% MF solutions effectively resisted swell-
ing. After four boiling cycles, the specimens treated with 2% and 4% solutions recovered
in almost the same way as the untreated specimens. Thespecimens treated with 25%
solutions retained high dimensional stability. This method achieved a maximum BE of 5%
and an ASE of 45%, showing that the chemical had not completely penetrated the cell
wall. Hardness increased in proportion to concentration. Starting with a hardness of 0.24
MPa at 0% concentration, it increased to 0.48 MPa and 0.72 MPa at 15% and 25%
concentrations, respectively, at a compression of 54%. A remarkable decrease in abrasion
resistance with increasing concentrations was observed. Increases in MOE and MOR at a
25% concentration were about 10% and 18%, respectively [ 13,141.
100
80
- 60
2
2 40
20
n
" ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Cycles
FIGURE 9 Soaking/drying-cycle test for compressed sugi wood (CS: 55%) treated with PF resin.
D, oven-drying; W, soaking; B, boiling; RC, resin concentration.
588 Norimoto
D.Fixation by Heating
Heating while under deformation is another effective way of fixing. After wet sugi, radiata
pine (Pinus rdutl1), and albizia (Pamsrrienthes,fnlcatrr) specimens were irradiated with
microwaves at 2.4 kW, they were compressed in the radial direction to about SO% of their
original thickness (CS: 50%), followed by oven drying at 150°C. The compressed speci-
mens were then heated at 160, 180, or 200°C for various lengths of time by three methods:
under molten metal (MH), in the presence of air (CH), and under evacuation of air (VH).
The treated specimens were soaked in water until saturated, placed in boiling water for
30 min, then oven dried.
The RS increased in the order of CH, MH, and VH when compared at the same
heating temperature and time. However, it was independent of heating method and wood
species when compared at the same weight loss (WL)as shown in Fig. 10 [Sl]. The
relationship of RS to WL was expressed by a following hyperbolic equation:
53.02
RS = - 11.6 0 5 WL 5 4.0
WL + 0.579
Almost complete fixation was achieved at about 4.0% WL. Conventional heating for 20
h at 180°C or 5 h at 200°C achieved complete fixation. The retentions of MOE and MOR
at perfect fixation were about 88% and 78%, respectively. As for color changes in the
L*-a*-b* color system, the decrease in shade was 28% and the color difference was 29%
1141.
E. Fixation by Steaming
The fixation of compression set in wood can be also achieved by steaming. Wet sugi
specimens irradiated with microwaves were compressed to about SO% CS in the radial
direction and then oven dried under restraint at 105°C. The compressed specimens were
fitted between stainless steel plates that were placed inside stainless side restraints. They
were steamed in an autoclave at 140, 160, 180, or 200°C for various lengths of time.
Figure 1 1 shows the effect of steam treatment on RS [52].As the temperature increased,
the RS decreased. Almost no recovery of set was observed after only 1 min of steaming
at 200°C or 8 min at 180°C. Hardness resulting from compression increased from 0.07
1
Chemical Modificationof Wood 589
140
t (mid
FIGURE 11 Relationship between RS and treating time ( t ) for compressed sugi wood (CS: 50%)
steamed at 140, 160, 180, and 200°C.
MPa to 0.25 MPa, an increase of about three times. A small reduction in hardness was
observed when steaming the compressed specimens. Figure 12 shows changes in MOR of
the steamed noncompressed specimens [52]. Only a small change in MOR resulted from
steaming specimens at either 180°C for 8 min (3.3%) or 200°C for 1 min (8.6%). Only a
slightyellowing and darkeningoccurred in the specimenssteamed at 180°C for 8 min
-60 1
590 Norimoto
FIGURE 13 Relationship between degree of crystallinity (DC) and treating temperature for heat-
(*) and steam- ( 0 ) treated sugi wood powder.
powder increased with increasing temperature while that of heated powder decreased. The
HW of heated powder increased with increasing temperature while that of steamed powder
decreased. Significant changes in the bands assigned to the CO and COOH groups at 1736,
1719, and 1698 cm” in IR spectra were detected, which were more apparent in heat
treatment than in steam treatment. The stresses stored in the matrix and microfibrils may
be released, because hemicelluloses degrade by both treatments at 180°C [ 16,541. At the
same time, the cross-linking reactions in the matrix and the crystallization of the micro-
fibrils are likely possible. The results of X-ray diffraction and IR absorption measurements
showed that the former mechanism was dominant in heat treatment and the latter in steam
treatment. However, compressive deformation was also fixed to some extent by preheating
or steaming. Therefore, it is reasonably supposed that the fixation of compressive defor-
mation by heating resulted mostly from the release of the stresses stored in both micro-
fibrils and matrix by the degradation of the cell wall polymers. On the other hand, the
fixation by steaming is considered to occur probably because of the formation of cohesive
structure in the cell wall as well as the relaxation of the stresses stored in both microfibrils
and matrix.
V. DYNAMICVISCOELASTICPROPERTIES
A. ViscoelasticModeling
The dynamic mechanical properties of the cell wall of chemically modified woods were
analyzed by using a model shown in Fig. 14 [ S ] , in which an amorphous isotropic matrix
is disposed in parallel along the axis of cellulosic fibrils ( 1 direction) inclining at 8 to the
grain direction of wood. The complex dynamic modulus of the model along the grain
(E?) is expressed by
where ET and ET are the complex dynamic moduli in the 1 and 2 directions, G” is the
complex shear modulus in the 1-2 plane, and p12 is Poisson’s ratio. If 8 is small enough
to ensure sinJ@= 0, cos‘e = 1, sin’O.cos’8 = e’, and p12 is much smaller than the real
part of E ? , the E’ and tan 6 of wood along the grain, as the first approximation, can be
expressed by
Chemical
Wood Modificationof 591
FIGURE 14 A model of the S, layer in the cell wall.A, side view; B, cross section;f, microfibril;
m, amorphous isotropic matrix; 8, microfibrillar angle; 1, longitudinal direction of microfibril; 2,
direction perpendicular to 1 direction.
where A is the volume fraction of the S, layer in the cell wall, y and yw are the specific
gravities of wood and the cell wall, are the dynamic and loss moduli in the 1 direction,
and G’ and G” are the dynamic shear and loss moduli in the 1-2 plane, respectively. E;
and E:’ are expressed by
E: = WE, + (1 - W)E,,, - WE, E’,’ = (1 - *)E,,, tan S,,,
where W is the volume fraction of the fibrils in the cell wall, E, is the dynamic modulus
of the fibrils, and E,,, and tan S,,, are the dynamic modulus and loss tangent of the matrix,
respectively. In the model, the fibrils with square cross section are embedded in the matrix,
so that the fibrils and matrix are aligned partly inseries and partly in parallel tothe
direction of shear force. According to the law of mixtures [56], G‘ and G’’can be expressed
by
tan S,,,
The experimental values of E’, tan S, and y at 20°C and 60% RH for the untreated
and chemically modified woods were fitted to the equations [27,28]. For the untreated
wood, W = 0.5, yw = 1.45, A = 0.84, 8 = 0.09 rad, E, = 134 GPa, and E,,, = 2 GPa were
used [33,36]. For the modified wood, W = 0.33-0.45 and yw = 0.92-1.29 were estimated
from both weight gains and volume swellings.
The values of E,,,and tan S,,, were quantified by eliminating the swelling contribution
to the E‘ and tan S of five kinds of chemically modified woods, including untreated wood.
The relationship between logarithms of tan S,,, and E,,,are shown in Fig. 15 [ S ] . Propylene
oxide treatment is similar to acetylation in terms of conventional stabilization, but tends
to increase the molecular mobility instead of reducing it because of the hydrophilic nature
of the bulking agent. Although the two treatments induced the same amount of volume
and weight increases, pronounced drops of tan S,,, implies a strong reduction of matrix
I
592 Norimoto
-1.3
-1.4
-1.5
1 0
p~
-1.9
-l8 L"!
0 0.1 0.2 0.3 04
log Ern
FIGURE 15 Relationship between logarithms of loss tangent (tan S,,,) and dynamic modulus (E,,,)
for matrix substances of chemically modified woods. See legend to Fig. 2 for abbreviations.
mobility because of the hydrophobic nature of the bulking agent. PEG treatment is rather
similar to propylene oxide treatment in terms of chemical modification of the cell wall,
except that the hydrophilic nature of the molecule is even more pronounced. A slight
decrease oftan S,,, in formalization impliessmaller mobility, which is what the cross-
linking would be expected to produce, but a slight decrease in E,,, suggests some degra-
dation of the matrix components.
B. Relaxation Process
One of the useful ways of understanding the mechanical relaxation processes of chemically
modified woods is through the temperature dependence of E' and tan S at fixed frequencies.
The temperature variations of E' and tan 6 along the grain for five kinds of chemically
modified woods as well as untreated wood were measured over a temperature range from
- 150°C to 200°C with an automatic dynamic viscoelastometer (ORIENTEC Co. Ltd.,
Rheovibron: DDV-2SFP).
The variations of E' and tan 6 with temperature at 11 Hz for the untreated (U) and
formalized woods (F) in a dried condition are compared in Fig. 16 [57,58]. With respect
to tan 6 for the untreated wood, four relaxation processes were detected within the tem-
perature range examined. The tan 6 increased with increasing temperature above 100°C.
This relaxation process. labeled a(,,shifted to lower temperatures with increasing moisture
content and exhibited a clear peak at high moisture contents below 100°C [59,60]. This
process was attributed to the micro-Brownian motions of the cell wall polymers in the
noncrystalline region [S9]. A small tan 6 peak at about 70°C has been observed by many
investigators, who assigned it to the local motions (torsional oscillations) of the cell wall
polymers. However, its location remained unchanged in measurements atdifferentfre-
quencies. Besides, the peak was not observed in the absolutely dried specimens [ S i ' ] . It
may be ascribed to the segmental motions of the cell wall polymers activated during water
desorption. The loss peak at around -40"C, labeled P(,,was not detected i n the absolutely
dried specimens. The same process was also observed in the dielectric measurements for
specimens containing a small amount of water [61]. The results showed that this process
was due to the rotational motion of the adsorbed water itself [61-64]. The tan 6 peak at
around - I lO"C, labeled y(,, was also detected at the same temperature-frequency location
in the dielectric measurements. The apparent activation energies obtained by both mea-
surements showed the same value of 9.8 k c a l h o l [61]. This process has been reported by
Chemical Modification of Wood 593
0.001
-150 -100 -50 ' l EO xx)
0 50 E
T ("C)
FIGURE 16 Temperaturevariations of dynamicmodulus ( E ' ) andlosstangent (tan S) for un-
treated (dotted line) and formalized (solid line) woods.
many investigators, who assigned it to the motion of the CH,OH group of the cell wall
polymers in the noncrystalline region [61,65-671.
In formalization, three relaxation processes were observed in tan 6, labeled aI.,PI.,
and yI; in order of decreasing temperature at which they detected [57,58]. This treatment
involves the cross-linking of chains by OCH, bridges between the OH groups of the cell
wall polymers. The bridges may restrict a part of the micro-Brownian motions of the cell
wall polymers to some extent, resulting in a decrease in tan 6 above 100°C. The PI. process
is comparable to the Dl, process. The loss peak due to the motion of the OCH, group was
observed at almost the same temperature location as that of the CH,OH group [68]. There-
fore, the y,: process may involve both motions of the remaining CH20H groups and in-
troduced OCHz groups.
The variations of E' and tan 6 with temperature at 1 1 Hz for the acetylated wood
are compared to that of the untreated wood in Fig. 17 159,601. The E' was lower than that
of the untreated wood over the temperature range tested. Two relaxation processes, labeled
a , and P;l, were recognized in tan 6. The OH groupsaresubstituted with the bulky
OCOCH, groups, which reduces the cohesive forces between the main chains of the cell
wall polymers and probably facilitates their backbone motions. This effect is similar to
that produced by the addition of plasticizer and is usually known as internal plasticization.
The marked decrease in E' and increase in tan 6 over 100°C are ascribed to the internal
plasticization.Thus, the process is assigned to themicro-Brownian motions of the
acetylated cell wall polymers. The loss peak comparable to the p(, process related to water
desorption was not recognized in acetylated wood because of a reduction in hygroscopicity.
As most of the OH groups i n the noncrystalline region are replaced by the OCOCH,
groups at about 20% WPG, the motion of the OCOCH, group instead of the CH,OH group
ought to occur. The peak location agreed well with that due to the motion of the
OCOCH, group i n acetylcellulose 1691.
594 Norimoto
0.1
Y
5 0.01
0.001
-150-100 -50 0 50 m 1 5 0 200
T ?C>
The variations of E' and tan 6 with temperature at I 1 Hz lor the propylene oxide-
treated wood areshown in Fig. 18 [57,58]. Thistreatment results in bonded cell-wall
bulking as in the case of acetylation, except that the introduced group is hydrophilic. The
E' was lower than those of the untreated samplesover the temperature range tested,
especially above 100°C. Three relaxation processes, labeled a,,(,to 'y,.(,,were detected in
""-
kll
"""""
Chemical
Wood Modification of 595
tan 6. As the introduced OCH,CH(OH)C,H5 group is bulkier and more flexible than the
OCOCH, group. the cell wall polymers are much plasticized compared to those of acet-
ylation and the a/.,,peak was observed within the temperature range tested. The treatment
reduceshygroscopicity at low relative humidity levels 1291, so the &, processdue to
adsorbed water became less distinct. The introduced side chain contains the OCH, group
whose motion may be responsible for the y,.(, process as in the y/..process.
The temperature dependence of E' and tan 6 at I 1 Hz for the wood-MMA composite
is shown in Fig. 19. The E' was larger than that of the untreated wood within the tem-
perature range tested. With respect to tan 6, five relaxation processes were observed,
labeled aIV,PI,,, ylv, 6,,.,
and E,,. in order of decreasing temperature. A very low ASE shows
that the cell wall remains untreated, while PMMA fills the lumen. Therefore, both relax-
ation processes observed in the untreated wood and PMMA are expected. Two relaxation
processes assigned, respectively, to the micro-Brownian motion of the main chain and the
motion of the OCOCH, side group were detected in PMMA 1701. These locations corre-
sponded, respectively, exactly to the p,,. and yl,. processes. Thus. the remaining al,,, 6,,,,
and cl,, processes can be ascribed to the same motions as those of the a(,to yI, process.
Thetemperaturedependence of E' and tan 6 at 1 1 Hz for the PEG-impregnated
wood is shown i n Fig. 20 [57,58]. The E' was larger than that of the untreated wood
below O"C, but became lower above that temperature. Three peaks were observed in tan
6, labeled p,.(;,and y/.(;i n order of decreasingtemperature. I t is well known that PEG
molecules with molecular weights of 1000 can penetrate into the cell wall [71,72] and act
as a plasticizer, which leads to large reductions in the cohesive forces between the cell
wall polymer molecules. The a/.(;was attributed to the micro-Brownian motions of the
cell wall polymers plasticized with PEG molecules [60,72]. The tan 6 of the PEG-im-
pregnated glass fibers had a peak between -50°C and 0°C. The apparent activation energy
for this process estimated from the measurements at five different frequencies was about
40 kcal/mol, which suggested a mechanism involving chain backbonemotions of PEG
. l
-
9
0.01
596 Norimoto
2 -51 """-
(3
v
lo
W
5
FIGURE 20 Temperature variations of E' and tan 6 for untreated (dotted line) and polyethylene
glycol-impregnated (solid line) woods.
molecules. On the other hand, the tan 6 increased remarkably above 20"C, which corre-
sponds almost to the melting point of PEG. Therefore, this sharp increase in tan 6 may
be related to the flow of PEG molecules. The p,.(; process was not recognized in PEG.
PEG molecules in the cell wall plasticize the cell wall polymers, but at the same time
their micro-Brownian motions must be restricted to some extent by the cell wall polymers.
This may lead to a shift of the peak due to the micro-Brownian motions of PEG molecules
toa higher temperature.The peak due to CHzOHgroup occurred about 20°C below
the y,, peak temperature, and its value was rather large.
REFERENCES
1. R. M. Rowell, in Wood r r n d Ce//u/o.srChemistry (D. N.-S. Hon and N. Shiraishi, eds.), Marcel
Dekker. New York and Basel, p. 703 (1991).
2. N. Shiraishi, T. Aoki, M. Norimoto, and M. Okumura, C h e n ~ t c ~ h13: . 366 (1983).
3. M. Norimoto, J . Gril, and R. M. Rowell, Wood Fiber Sci., 24: 25 (1992).
4. M. Norimoto and J . Gril, in R w d Resenrrh U I I Wood m d W O O ~ - B ~ ~Mnrerids
. S C > L ~ (N. Shiraish.
H. Kajita, and M. Norimoto, eds.). Elsevier Applied Science, London and New York, p. 13.5
(1993).
S M.Norimoto, in C / W I ? I ~ C Modificrcriorl
U/ o f Ligrzoce//[r/o.sicMrrfericcls (D. N.-S. Hon,ed.).
Marcel Dekker. New York, Basel. and Hong Kong, p. 3 11 (1996).
6. M. Inoue, N. Kadokawa, J. Nishio, and M. Norimoto. Wood Res. Techno/. Notes, No. 29: S4
( 1993).
7. R. M.Rowell and P. Konkol. FPL Gen. f i 4 . Rep. (USIIA). No. 55: 1 (1987).
8. A. J. Stanm and H. Tarkow, J. Phys. Co//oir/C / ~ e m .31: , 493 (1947).
9. M. Norimoto. Wood Res. 7 k ' / ? I l . Nures, No. 24: 13 ( 1988).
IO. R. M. Rowell. A.-M. Tillman. and R. Simonson. J . Wood C/7rw7. f i d t r r o l . , 15: 293 (1986).
II. R. M. Rowell. Y. Imamura, S. Kawai. and M. Norimoto, Wood Fiber Sci.. 21: h7 (1989).
12. R. M. Rowell and W. D. Ellis, FPL Res. Pup. (USDA), No. 451: 12 (1984).
Chemical
Wood Modificationof 597
Akira lsogai
The University of Tokyo, Tokyo, Japan
1. INTRODUCTION
Looking back on the history of the science and technologies of cellulosic materials, we
realize that numerous technologies for modifying cellulose by chemical and mechanical
methods have been developed. Some of these technologies have been practically applied
to produce modified cellulosic materials for our daily and industrial necessities. Cellulosic
materials are generally strong, hydrophilic, insoluble in water, stable to chemicals, safe to
living bodies, reproducible, recyclable, and biodegradable. With these specific and advan-
tageous characteristics of cellulose, “modification” techniques to reinforce these original
properties or to add new functionalities to cellulose have been investigated, and have
contributed to the development of cellulose science and technologies. Chemical treatments
have been positioned at the center of the field of cellulose modification.
The purposes of chemical modifications or derivatizations of cellulose are divided
into three major groups: ( l ) to produce modified cellulosic materials having specific prop-
erties at commercial levels, ( 2 ) to characterize cellulosic materials at laboratory levels, and
( 3 ) to approach the nature of cellulose for scientific interests. Commercial cellulose deriv-
atives having water solubility, organic-solvent solubility, ion-exchanging groups, or hy-
drophobic groups are used as aqueous thickeners, plastics, column-supporting materials
for chromatography, and others. Cellulose carbamates and nitrates are used for determining
molecular mass and molecular mass distribution of cellulose to evaluate cellulosic mate-
rials by size-exclusion chromatography. Furthermore, since there are still many unsolved
and mysterious subjects in cellulose science, chemical modifications are sometimes used
to study fundamental research subjects of cellulose, such as solid-state structures of cel-
lulose (hydrogen bonding patterns, chain conformations, crystal and amorphous structures,
etc.), interactions with other substances at molecular levels, and molecular dynamics in
solution states.
The typical modifications of cellulose are esterifications and etherifications at hy-
droxylgroups of cellulose. Most water-soluble and organicsolvent-solublecellulose
derivatives are prepared by these substitution reactions, and drastic changes in the orig-
inal properties of cellulose can usually be achieved by thesechemical modifications.
Othersare ionic and radical grafting,acetalation,deoxyhalogenation,andoxidation.
Since the usual cellulosic materials originating from wood and cotton pulps have alde-
hyde and carboxyl groups in quite small quantities, depending on the purity of the pulps,
599
600 lsogai
these minor groups are also target positions for chemical modifications. Figure l shows
schematic representation of positions in the cellulose structure for chemical modifica-
tions.
The methods for chemical modifications of cellulose are various, depending on the
reactions. Cellulose esters such as cellulose acetate and nitrate are generally prepared from
cellulose pulps with reagents in the presence of concentrated sulfuric acid as an acid
catalyst. In the case of cellulose nitrate preparation, degree of substitution (DS) is con-
trolled by water content in the reaction medium, the H,SO,-HNO, system, and the shape
of the solid cellulose pulp is maintained during the nitration. Cellulose triacetate is pre-
pared from solid cellulose pulp in a mixture of acetic anhydride, acetic acid, and sulfuric
acid.The reaction mixture becomes transparent by dissolving celluloseacetate in the
medium, as the DS values are increased by acetylation. Water is then added to the mixture
to hydrolyze acetyl ester groups, in part for preparing cellulose diacetate, and this partial
1
Substitution reaction Esterification
Etherification
Deoxyhalogenation Acid hydrolysis
Oxidative cleavage
Oxidation To carboxylic acid
To aldehyde
Acetalation
H0
OH OH
Etherification
Deoxyhalogenation
Acetalation
Reactions
at minor groups Carboxyl group Esterification
Amidation
Reduction to alcohol
Aldehyde group Oxidation to carboxylic acid
Reduction to alcohol
and other organic solvents, these are also examined as the starting materials for derivati-
zations. From industrial and environmental points of view, the possibility of using lower-
a-cellulose-content pulps, especially for producing cellulose acetate, is a quite significant
subject.
Since detailed and systematic information about chemical modifications of cellulose
have been already reviewed by Professor Ishizu in the first edition of this book [3], sci-
entific and technical reports in this research field published after 1990 are supplemented
in the following sections.
Chemical Modificationof Cellulose 603
EDA,ethylenediamine;MC,methylcellulose;HPC,hydroxypropylcellulose;HEC,hydroxyethylcellulose;
CMC,carboxylmethylcelluloseNasalt; PF, paraformaldehyde; DMSO,dimethylsulfoxide;DMAc,N,N-di-
methylacetamide; DMF, N,N-dimethylformamide.
II. ESTERIFICATION
Conventional process
...............
~. , ..................
~
Dissolvlng pulp
Lmter
or pulp
:H
................
/ t
H,O i
:ii ..............
I
1
cellulose
De-acetylation
Crude
diacetate
T
I
I oniatA;; process
I
HP1 trlacetate
I ,
purification
Separation process
and 1 4 I
.-
~.................... . ........................................
. . ~ 4
.............. ..................
~
i DilutedAcOH j
..................... I .......... ........., ..................... I
Ac,O production
orocess
;
L.....
AcOH
............... I { ACOHprocess
New process
....................... :
IChernc
la1
l1 I ....................
L
Steam i >
v
Acetylation
at
4
~ ..........................................
. . . ...................
L ~
...................
, ..................
:......H,SO, t i
.................c
Ac,O
Ac,O production
AcOH
....., .....................
l
I
-.......... ........
iL.. ..................
AcOH I process
with sodium acetate, pyridine, and TsCl in DMF, cellulose acetates with DS 1.4-2.8 were
obtained, depending on the reaction conditions. The effect of molar ratios of TsCl/AcONa
on the DS values was investigated in detail. Tri-0-substituted cellulose benzoates having
various substituents at the benzene ring were also prepared by a similar method without
using DMF. The substituents at the benzene ring had little influence on the DS values
obtained. In these systems, the organic acid anhydrides or organic acid chlorides seem to
be formed in situ in the organic acid-TsCI-pyridine mixture, and thus these esterifying
reagents do react with cellulose hydroxyl groups in the presence of pyridine [ 121.
The LiCl-DMAc cellulose solvent system was often used for homogeneous esteri-
fications of cellulose in solution states. When N,N-dicyclohexylcarbodiimide,organic acids
(or organic acid anhydrides) and 4-pyrolidinopyridine were added to acellulose/LiCl-
DMAc solution, cellulose esters having DS values lower than 2.5 were prepared [ 13,141.
In the case of cinnamoylation of cellulose in the LiCl-DMAc system, cinnamoyl groups
were introduced almost selectively into C6 hydroxyl groups at DS values lower than about
1 [ 151. Cellulose propionates with DS values lower than 1.76 were prepared using the
LiCI-DMAc system and propionyl chlorideasacellulose solvent and an etherifying
reagent. Solution-state properties including liquid crystalline behavior of the products were
studied, and rigidity of cellulose propionate molecules in solution states decreased with
an increase in the DS values [16]. When cellulose is dissolved in N,O,-DMF or N,O,-
DMSOsystems, all hydroxyl groups finally form unstable nitrite esters in the solution
[ 171, where the ease of nitrosation among the hydroxyl groups is in the order of C6 > C2
> C3. On the other hand, when the cellulose solutions are poured into water, all nitrite
ester groups are finally removed from cellulose by hydrolysis to produce regenerated
cellulose, where the ease of denitrosation is of the order of C6 >> C3 > C2 [ 181.
Some water-soluble polysaccharides having sulfate esters have blood anticoagulant
activity and also sometimes an anti-HIV one in the medical field. Preparation of water-
soluble cellulose sulfate esters has been tried from these points of view. as well as ion
exchanging and other specific properties suchasCMC in aqueoussolutions.Since the
conventional sulfonation of cellulose with S03-DMF complex leads to severedepoly-
merization of cellulose, even though the DS values reach up to about 2, a homogeneous
solution-state sulfonation with various reagents was examined [191. The results obtained
showed that SOzClz gavethe highest efficiency of sulfonation when N,O,-DMF was used
as the cellulosesolvent.Thesulfate groups were distributed asC6-OH > C2-OH on
sulfonation at 20°C and C2-OH = C6-OH at -20°C. Permethylation analysis of cellulose
sulfate with DS 0.52- 1.37 was carried out for elucidating distribution of sulfate groups
among the three hydroxyl groups of one anhydroglucose residue by gas chromatography
and I3C-NMR 1201.
Since p-toluenesulfonate ester groups (tosyl esters) can be introduced selectively into
primary hydroxyl groups of carbohydrates under certain conditions and also can be sub-
jected to the following deoxy-substitution reactions, preparations of cellulose tosyl esters
have been examined for a long time. When microcrystalline cellulose was reacted with
TsCl and pyridine under heterogeneous solid-liquid phase conditions, the corresponding
esters with DS 2.3 were obtained [21]. Arai and Aoki reported preparation of cellulose
tosylates with DS 0.79-0.92 under heterogeneous (i.e.,solid-liquidphase)conditions,
and the cellulose tosylates obtained were further reacted with sodium bisulfite to prepare
cellulose deoxysulfonic acid with DS 0.065-0.1 [22]. However, the products obtained still
contained tosyl groups tosome extent. The LiCI-DMAc cellulose solvent system was
introduced to the solution-state homogeneous tosylation of cellulose by Heinze et al.
[23.24]. Their DS values were in the range 0.4-2.3, and were controllable by the reaction
dificationChemical of Cellulose 607
conditions. Thermal properties of these cellulose tosylates and further esterifications were
examined using the cellulose tosylates with pyridine, sodium acetate, and esterifying agents
(organic acid anhydrides). Regioselective tosylation at C6 hydroxyl groups of cellulose
does not seem to be achieved even by the homogeneous tosylation, and severe depoly-
merization may occur on cellulose chains during the reactions.
CF3COOH
Cellulose
Dissolution
U
(CF3CO)zO t CF3COOH
Cellulose W II
Cellulose-0-CCF,
Dissolution
DS = 1.5-2.1
-q
perfluorononenyloxyphtharic anhydride F& cF3
+
pyridine or triethylarnme
Cellulose * Cellulose-o
LiCI-DMAc
DS < 2.1 COOH
1 .l ,2,2,3-pentafluoropropyloxy-
22-difluoroporpionly fluorlde
t
pyridine or triethylamine 0
I1 Fz Hz
Cellulose * Cellulose-o-c, /c, /c, ,CH~F
LiCI-DMAc C O C
DS <F2
2.1 F2
4-perfluoropropyloxy-
2.2-difluoroporp~onyl
fluorlde
Cellulose
t
pyridine or trlethylarnine
LiCI-DMAc
- Cellulose-0-c,
DS < 2.1
0
II Fz Hz
C
/c, /c,
F,
O C
/CF,
F2
m o c , 3
Cellulose-0 Cellulose-0
y e c l
Cellulose-0O m Cellulose-0
Cellulose-0
>CH3
Cellulose-0
>Q
dification Chemical
CellUlOSe of 609
R-CHXH-CH-R*
I
*Cellulose-OH
o=y
l l
F=O
R-CH=CH-CH-R*
OH 0-Cellulose
ASA-cellulose half ester
CH, -CH
I I
o=c, R-CH=CH-CH-R*
0
HO
, (OH- )
ASA
y=o
o=y
OH OH
ASAcid
R-CH,-C=O
Cellulose-OH
I
i
R- CH
I
F=O
R-CH=F: -P 0- Cellulose
AKD-cellulose P-keto ester
-c=o
R*
HO
, (OH-) R-CH,-fi-CH,-R*
AKD 0
CO,' CO, Ketone
mechanisms of paper by AKD and ASA were carried out using some analytical techniques,
and the following conclusions were obtained 155-611.
1. Synthesis of "C-labeled AKD and the following solid-state "C-NMR analysis
of the handsheets prepared thereby revealed that nearly no covalent bonds are
present in the AKD-sized handsheets, when resonance peaks were assigned cor-
rectly. Thus, the hydrolyzed products of the reactive sizes, i.e., ketones for AKD
and alkenylsuccinic acid for ASA, must consequently contribute to sizing per-
formance of paper.
2. Most size components, i.e., hydrolyzed products, are physically anchored to pulp
fiber surfaces, and these components are extractable from fibers when the paper
sheetsare defibrated at 60°Cin1% Tween 80, which is one of the nonionic
surfactants.
3. The structures of AKD and ASA that are reactive with water are necessary for
achieving homogeneously distributed hydrophobic sizingcomponents with
smaller coagulants as possible on hydrophilic cellulosic pulp fiber surfaces by
in-situ hydrolysis of the sizing components in paper.
4. In the pulpsuspensions,cationicsize emulsion particles are adsorbed on pulp
fiber surfaces by forming ionic bonds between the cationic groups on the sizing
emulsion surfaces and carboxyl groups on pulp fiber surfaces, which are present
as minor groups of 0.02-0.08 mmol/g for bleached kraft pulps.
612 lsogai
From the aspect of esterifications of cellulose, therefore, the sizing of paper with
AKD or ASA is not included in the story of this chapter, “chemical modification of
cellulose.” However, such efficient surface modification techniques of cellulose as paper
sizing with reactive sizes may give some clues to developingnew techniques for modifying
cellulosic materials efficiently with reagents that are reactive with “water” under usual
conditions, where water is always present and has strong interactions with cellulose.
111. ETHERIFICATION
Since ether linkages are more stable than ester linkages to acid and alkaline conditions,
various cellulose ethers have been prepared for both practical and fundamental purposes.
Especially in the case of commercial cellulose etherifications, control of DS values and
distribution of substituents at various levels (Fig. 2) are quite important subjects for man-
ufacturing, because generally etherifications proceed heterogeneously to solid cellulose
pulps through swollen alkalicellulose.Thus,some modified processes to prepare com-
mercial cellulose ethers having more homogeneous distribution of substituents have been
developed at industrial levels. Also, regioselectively substituted cellulose ethers were pre-
pared, sometimes using cellulose solvent systems or multistage reactions, to obtain fun-
damental information about cellulose ethers and cellulose itself. As to new cellulose ethers,
hydrophobically modified water-soluble cellulose ethers have been found to have unique
viscosity behavior in aqueous solutions.
A.New CarboxymethylationProcess
Carboxymethylcellulose sodium salt(CMC) is produced in the largest quantity among
water-soluble cellulose derivatives, and is utilized as a thickener or dispersant in various
fields. Taguchi et al. studied the effect of distribution of substituents of CMC samples on
stability of their solution properties, and a new carboxymethylation process of cellulose
was developed on the basis of the results obtained thereby [62]. Distribution of substituents
of CMC samples along one cellulosechain and that among cellulose chains were evaluated
by cellulase degradation and electrophoresis, respectively. The results showed that CMC
samples having more homogeneous distribution of substituents had higher stability of their
solution properties to enzymes and salt concentrations, even at the same DS values and
the same distribution of substituents among the three hydroxyl groups of one anhydroglu-
cose residue. The new process had two key points for producing CMC having more
homogeneous distribution of substituents: ( I ) the use of i-propyl monochloroacetate ester
as the reagent instead of monochloroacetic acid or (2) separate additions of NaOH solu-
tions to the reaction mixture. This process has been already utilized at practical levels.
Other new carboxymethylation methods were reported for fundamental aspects. 6-
0-triphenylmethylcellulose (tritylcellulose) was first prepared under homogeneous condi-
tions using the LiCI-DMAc cellulose solvent system, and residual hydroxyl groups at C2
and C3 of tritylcellulose were partially or completely carboxymethylated.Carboxy-
methylcellulose partially substituted at C2 and C3 hydroxyl groups were obtained by
detritylation of the carboxymethylated tritylcellulose with HCI.Liu et al. reported that
these CMC samples became water-soluble at DS values more than 0.3. thus suggesting
that distribution of carboxymethyl groups strongly influences the solubility behavior in
water [63]. 2,3-Di-O-carboxymethylcellulosewas prepared from tritylcellulose by ( l ) re-
peated carboxymethylation with monochloroacetic acid sodium salt and powdered NaOH
Modification
Chemical of Cellulose 613
using a tritylcellulose/DMSO solution and (2) the following detritylation [64]. 6-0-Car-
boxymethylcellulose has not been prepared yet.
C.New CelluloseEthers
Preparation of hydrophobically modified water-soluble cellulose ethers and characteriza-
tion of their solution properties are significant current topics for cellulose ethers [74]. For
example, commercially available hydrophobically modified hydroxyethylcellulose (HM-
HEC) is prepared from HEC by reactions with epoxides having long alkyl chains of C12-
C24. This HM-HEC contains l-2% hydrophobic groups, i.e., DS 0.01-0.03, and thus is
614 lsogai
Hydrophobic
15 I I
l
/
/
Hydrophoblcally modified 1
/
hydroxyethylcellulose 1
/
-
U! 1 0
Kl /
l
/
a
v /
/
.a
m
/
0 /
0
v)
/
/
5 5 /
/
/
Normal
/
"00
"
"
0 1 2 3
Concentration of polymer in water ("h)
factants in water and have some interactions with other compounds in water. Therefore,
the solution properties of these HM-cellulose ethers have been studied in terms of tem-
peratures, phase-separation behavior, and interactions with water, surfactants, or latex par-
ticles, which have hydrophobic sites on their surfaces [76-791. The unique solution prop-
erties due to HM-cellulose ethers may be further utilized in the future in various fields.
Pentyl ether of hydroxypropylcellulose and its cross-linked gels were prepared in
nonaqueous solutions, and their thermotropic liquid-crystalline behavior was studied on
the basis of changes in the optical pitch by heating [80]. Also, liquid-crystalline properties
of cellulose ethers having long alkyl chains (cellulose-0-C,,,H,,,,-OH:m = 12-24) were
studied by Liu et al. [81]. Comb-shaped amphiphilic cellulose derivatives, 0-(2-hydroxy-
3-butoxypropyl)cellulose, with MS 0.4- 1.4 were prepared with butylglycidilether in cel-
luloseLiC1-DMAc solutions. The substitution occurred at the three hydroxyl groups in
the order C6 > C2 >> C3, and the products were soluble in water or DMSO, depending
on their DS values [82]. When cellulose was reacted with p-methoxytriphenylmethyl chlo-
ride in a cellulose/LiCl-DMAc solution at 25-70°C for 4-96 h, the corresponding cel-
lulose ether-like tritylcellulose with DS of about l was obtained. This cellulose ether is
soluble in DMF, DMAc, and DMSO [83]. Amine group-containing cellulose ethers were
prepared fromepoxide group-containing cellulosederivatives by reaction with hexa-
methylene diamine or polyethylene imine, and the products obtained had some antimicro-
bial activity [84]. Cellulose ethers containing Michler's ketone groups as substituents of
DS 0.56 were prepared, and their photoregulable properties were investigated [ U ] . SUI-
foethyl groups were introduced into cellulose by reacting cellulose with BrCH,CH,SO,Na
and powdered NaOH in the cellulose/SO,-diethylamine-DMSO solution or by reacting
celluloseacetate with the same reagents in the celluloseacetate/DMSO solution. The
water-soluble products prepared by the latter method had some anti-HIV activity [86].
Further derivatizations of CMC have been studied in some reports. Etherifications
of CMC with diethylaminoethyl chloride HCl salt or 3-chloro-2-hydroxypropyltrimethyl-
ammonium chloride gaveamphoteric cellulose derivativescontaining both carboxyl groups
and either tertiary amine or quaternary amine groups. Size-exclusion chromatography of
these amphoteric cellulose derivatives indicated that some hydrophobic interactions were
present among the molecules in the diluted aqueous solutions [87,88]. Carboxyl groups of
CMC were partly amidated by hexadecylamine in DMSO containing dicyclohexyl-
carbodiimide, and the products obtained were analyzed by I3C-NMR after acid hydrolysis
[89]. CMC xanthate was prepared from CMC by the addition of NaOH and CS, in the
aqueous CMC solutions, and it had some metal-adsorptive properties [90]. CMC is soluble
in formic acid, forming CMC formyl ester with DS0.4-2.0[91].Some new cellulose
ethers reported recently are illustrated in Fig. 9.
NMe2
l
Cellulose - 0 -
OH
Cellulose- 0 - CH
, 2, -OH (m = 12-24) NMe,
-(
Cellulose- 0 CH2CH2 f OCH,
Cellulose-O-CH2CHCH2-N+-CH3
I
OCH2COONa
OH
I I
Hydroxyethylcellulose - 0 - CH2CH
I
OH
Hydroxyethylcellulose- 0 -CH2CH2CH2CH2CH3
IV. GRAFTING
Grafting is also one of the chemical modification methods of cellulose. In order to intro-
duce fluorine-containing groups into cellulose by grafting, cellulose was reacted with per-
fluorooctylethylacrylate as a reagent and either ammonium persulfate or AIBN as a catalyst
using acellulose/PF-DMSOorcellulose/LiCl-DMAc solution [97]. The grafting effi-
ciency was evaluated for the products, and the PF-DMSO system with ammonium per-
sulfate gave good results. On the other hand, aqueous and heterogeneous solid-liquid
phase reactions gave poor efficiency in cellulose grafting. Thermal and chemical properties
of the grafted products were studied from various aspects [98].
Grafting of cellulose acetate with N-vinylcarbazol or epoxide-containing cellulose
derivatives with various vinyl monomers was also investigated [99,100]. When cellulose
was reacted with diethylaminosulfur trifluoride (DAST) in a cellulose/LiCl-DMAc solu-
tion, a polysaccharide having C6-0-C1 branch structures of the cellulose backbone was
obtained [ I O I 1.
V. DEOXYHALOGENATION
Preparation of cellulose derivatives having C-X (X: halogen) groups have been studied
often using nonaqueous cellulose solvent systems(Fig. lo), and in some reports these
deoxyhalogenated cellulose derivatives were further subjected to substitution reactions to
add functionalities to cellulose.
Deoxychlorination of cellulose was carried out in a cellulose/LiCl-DMAc solution
with N-chlorosuccinimide-triphenylphosphine (TPP). At the early stage the reaction oc-
curred only at C6-OH, and then it also occurred at C3-OH with Walden inversion as the
deoxychlorination proceeded. The maximum DS value was 1.86 by this chlorination [ 1021.
When cellulose was reacted with sulfuryl chloride in a cellulose/LiCl-DMAc solution,
deoxychlorination occurred at C6-OH and C3-OH, with Walden inversion, up to DS 1.8.
In this case, however, sulfur-containing groups were also introduced into cellulose [103].
Since cellulose is soluble in the LiBr-DMAc system, deoxybromination was performed
with N-bromosuccinimide-TPP in acellulose/LiBr-DMAc solution. The products ob-
tained had deoxybromide groups only at C6 with DS 0.9 [104]. Although cellulose was
reacted with tribomoimidazole-TPP under heterogeneous solid-liquid phase conditions,
the products had deoxybromide groups with DS less than 0.6 [ 1051. On the other hand,
when the homogeneous deoxybromination was adopted with the same reagents using a
cellulose/LiBr-DMAc solution, DS values of deoxybromide groups reached up to 1.6 by
reacting at C6-OH and C3-OH with Walden inversion [106]. Halodeoxycelluloses thus
prepared were further subjected to nucleophilic substitution reactions with tiols. inorganic
compounds, or long aliphatic amines [ 107- 1091.
6-Deoxyfluorocellulose acetate with DS less than 0.6 was prepared without depo-
lymerization of cellulose by reacting cellulose acetate with diethylaminosulfur trifluoride
(DAST) in dioxane [ 1 IO]. Regioselectively substituted 6-deoxyfluorocellulose with DS 0.9
was prepared through the following steps: ( I ) preparation of 6-0-tritylcellulose using the
LiCI-DMAc cellulose solvent system, ( 2 ) acylation (benzoylation or acetylation) of the
tritylcellulose to prepare 2,3-di-0-acyl-6-0-tritylcellulose,( 3 ) HBr treatment for detrity-
lation, (4) reaction with DAST to prepare 6-deoxyfluoro-2,3-di-O-acylcellulose, and (5)
saponification of the acyl groups with methanolic sodium methoxide [ 11 11.
618 lsogai
N-Cl +O
QP
CH2CI
OH
LiCI-DMAc
%? Cl OH
DS < 1.86
di-2
CH20H CH2CI
-Q 1 1 1 1 1 l l l l l
SOpCI,
LiCI-DMAc
c
OH Cl OH
DS < 1.8
6*o
+ e
&-p
N-Br p
CH20H CH@
O
LiBr-DMAc
OH OH
DS = 0.9
Br
c2
CH20H
OH
TH20H
Br
1) Tritylation
LiCI-DMAc
in
H
LiBr-DMAc
5) De-acylation in
MeONdMeOH
*
CH2Br
Br
DS < 1.6
CH2F
OH
2) Acylation with
AcPOor BzpO
3) De-tritylation
HBr with F l
OH 4) Fluorinationwith
F-S-NEt, OH
I
DAST F DS= 1
VI. OXIDATION
Many studies about cellulose oxidation have been done, concerning both chemical mod-
ifications of cellulose and oxidative bleaching of residual lignin in chemical pulps. Oxi-
dation reactions applied to cellulose for chemical modifications in this decade are sum-
marized in Fig. 11. Some oxidation reactions occur on cellulose selectively at particular
Chemical Modificationof Cellulose 619
CH20H
C CH O YHC
OH
I I
I I
NaBH, NaCIO,
fQ
CH20H
N204in CHCI,
c2 + side reactions
*a
OH OH
CH20H
pH 10-11
TEMPO-NaBr-NaCIO
OH OH
Pulp-CH0
HCIO,, pH 4-5
* Pulp-COOH
positions. The periodate oxidation is typically involved in this category, and has been used
to prepare dialdehyde cellulose at laboratory levels. Generally, the oxidation requires sev-
eraldays at room temperature in the darktopreparedialdehydecellulose from solid
cellulose samples, whose C2-C3 bonds are mostly cleaved, and thus depolymerization is
inevitable during the periodate oxidation. When acelluloseRF-DMSO solution was
poured into methanol, partly methyloylated cellulose was obtained as a precipitate. Since
I
620 lsogai
this methyloylcellulose was soluble in water, the periodate oxidation proceeded homoge-
neously in the aqueous solution, and almost completely oxidized dialdehyde cellulose was
obtained within 20 h [ 1121. Since the aldehyde groups of dialdehyde cellulose form intra-
and intermolecular hemiacetal linkages, the isolated and dried dialdehyde cellulose, even
having 100% oxidized structures at the C2-C3 bond, is insoluble in water. The dialdehyde
cellulose was oxidized to the corresponding dicarboxyl cellulose with sodium chlorite, or
reduced to the corresponding dialcohol cellulose with sodium borohydride. These oxidized
and reduced products were soluble in water, and characterized from various aspects [ 1 12-
1 1 81. Dialdehyde cellulose was further modified to the corresponding hydroxamic acid
derivatives by three-step reactions, and their behavior to form metal complexes in water
was investigated [ l 19,120). Periodate oxidation was also applied to cellulose beads for
chromatography [ 12 1 1.
It is well known that primary alcohol groups of cellulose are partly converted to
carboxyl ones by oxidation with N20, in chloroform. However, side reactions are inevi-
table during this N20, oxidation. On the other hand, recently a new water-soluble reagent,
2,2,6,6-tetramethylpipelidine- I -oxy1 radical (TEMPO), has become commercially avail-
able, and TEMPO can oxidize primary alcohol groups of water-soluble polysaccharides
such as starch to carboxyl ones in good yields and selectivity in the presence of a co-
oxidizing agent at pH 9- l l [ 1221. The TEMPO-NaBr-NaC10 system was first applied
to native cellulose by Chang and Robyt [ 1231, although their method could not give water-
soluble cello-uronic acid [ 1241. Then the TEMPO-mediated oxidation of cellulose samples
under various conditions was examined to prepare water-soluble cello-uronic acid having
high DP. Although only small amounts of carboxyl groups were introduced into native
cellulose samples by this oxidation, water-soluble cello-urionic acid sodium salts were
obtained quantitatively by using regenerated and mercerized celluloses as starting materials
(Fig. 12) [ 1241. DP values of the cello-uronic acids thus prepared were greatly influenced
by the oxidation conditions, and partial depolymerization occurred on cellulose chains by
p-elimination under alkaline conditions. Since cello-uronic acids prepared by the TEMPO
system regularly consist of the glucuronic acid repeating unit, differing from the conven-
-C,OONa
l I I I I I I I I
220
200 180 160 140
120 loo 80 60 ppm
tional water-soluble cellulose derivatives, they may have some unique solution properties
or bioactivity.
Cellulosic fibers usually contain carboxyl and aldehyde groups in quite small quantities,
depending on the purity of the fibers. Usual bleached kraft pulps have carboxyl contents
of 0.02-0.08 mmol/g and aldehyde contents of 0.01-0.03 mmol/g. Cotton linter pulps
have much lower values. The carboxyl groups originate from those present in the original
hemicellulose and/or those formed during pulping (especially anthraquinone-added kraft
pulping) and bleaching. The C 1 groups at the reducing ends of cellulose and hemicellulose
chains are accounted forasaldehydegroups.Carboxylgroups in cellulosic fibers play
significant roles in fiber processing treatments: dyeing and surface treatments in fiber
processing, and efficient adsorption of wet-end additives on pulp fibers at the wet end in
papermaking.
Aldehyde groups in cellulosic materials can be oxidized with ClO; to carboxylic
acids, and thus their carboxyl contents are increased by this oxidation [60]. Since water-
soluble carbodiimide, WSC: 1-ethyl-3-(3-dimethylaminopropyl)carbodiimideHCI salt, has
become commercially available, carboxyl groups can be selectively converted to amides
by WSC and compounds having primary amine groups under aqueous conditions at pH
4.75. More than 95% of the carboxyl groups in a bleached kraft pulp were converted into
nonionic or cationicamides with WSC and the corresponding amine in aqueous pulp
suspensions (Fig. 13) [57-60,1251. Here, the cationic pulps prepared by this method can
be regarded as true cationic pulps, because the cationic groups are introduced into pulps
by blocking the anionic carboxylgroups. On the other hand, the usual cationic pulps
prepared from normal pulps by etherification with, for example, diethylaminoethyl chloride
HCI salt at hydroxyl groups, should be called “amphoteric” pulps. Fundamental properties
such as pulp freeness were unchanged by the amidation, and thus only slight chemical
modifications of pulp fibers can be achieved by this amidation. However, when handsheets
were prepared from the nonionic pulp, which was prepared from normal bleached kraft
pulp by amidation with methylamine and WSC, effects of wet-end additives, such as sizes,
wet-strength resins, and retention aids,decreased drastically. This is because retention
values of these additives clearly drop by blocking the anionic carboxyl groups with non-
ionic methylamide groups in the pulp fibers. Thus, carboxyl groups in pulp fibers, even
though their quantity is quite low, behave as essential retention sites for wet-end additives
in paper stock by ionic interactions [57-60,1251.
VIII. CONCLUSION
t
PuI~-COOH" -----
Solvent exchange
from
water to ether
through methanol and n-hexane
CH2N2
in
ether
PuIP-COOCH~
&N+H-Et
NH(CH,),NMe,
o=c
RNHz, pH 4.75
Pulp-CONH-R
R : -CH3 (nonionic)
R : -CH2(CH2)&H3 (hydrophobic)
(hydrophobic)
FIGURE 13 Methylation and amidations of carboxyl groups in bleached kraft pulp [125].
REFERENCES
1. A. Isogai and M. Usuda, Mokuzai Gakkuishi, 37: 339 ( 1 99 1 ).
2. A. Isogai and R. H. Atalla, J . Polymer Sci., Polymer Chem. Ed., 29: 113 (1991).
ificationChemical of Cellulose 623
3. A. Ishizu, in Wood and Cellulosic Chenlistly (D. N.-S. Hon and N.Shiraishi, eds.). Marcel
Dekker, New York, p. 757 (1991).
4. H. Yabune, Cellulose Commun., 4: 114 (1997).
5. S. Z. Rogovina, V. A. Zhorin, and N. S. Enikolpiau, J . Appl. Polymer Sci., 57: 439 (1994).
6. K. Ueda, S. Saka, Y. Funaki, and S. Soejima, Mokuzai Gakkaishi. 34: 356 (1988).
7. H. Matsumura and S. Saka, Mokuzai Gakkaishi, 38: 270 (1992).
8. H. Matsumura and S. Saka, Mokuzui Gakkaishi, 38: 862 (1992).
9. S. Saka and K. Takahashi, Preprints '97 Cellulose R&D, Nagoya, p. 44 (1997).
IO. Y. Shimizu, A. Nakayama, and J. Hayashi, Cellulose Chem. Technol., 25: 275 (1991).
11. Y. Shimizu, J. Nakayama, and J. Hayashi, Sen 'i Gakkuishi, 49: 352 (1993).
12. Y. Shimizu, J. Nakayama,and J. Hayashi. Preprints '94 Cellulo.se R&D. Tokyo, p. 165
( I 994).
13. G. Samaranayake and W. G. Glasser, Carbohydr. Polymers, 22: 1 (1993).
14. G. Sarnaranayake and W. G. Glasser, Carbdzydr. Polymers, 22: 79 ( I 993).
15. A. Ishizu, A. Isogai, M. Tomikawa, and J. Nakano, Mokuzai Gakkaishi, 3 7 829 ( I 99 I).
16. E. Bianchi, E. Marsano, L. Ricco, and G. Conio, Carbohydr. Polymers, 34: 91 (1997).
17. A. Isogai, A. Ishizu, and J. Nakano, Sen'i Gakkaishi, 42: T654 (1986).
18. W. Wagenknecht, I. Nehls, and B. Philipp, Carbohydr. Res., 237: 21 1 (1992).
19. W. Wagenknecht, I. Nehls, and B. Philipp, Carbohyd,: Res., 240: 245 (1993).
20. M. Gohdes, P. Mischnick, and W. Wagenknecht, Carbollydr. Polymers, 33: 163 (1997).
21. C. Roussel, C. Popescu, and L. Fabre, Carbohydr. Res., 282: 307 (1996).
22. K. Arai and F. Aoki, Sen 'i Gakkaishi, 50: 5 I O ( l 994).
23. T. Heinze, K. Rahn, M. Jaspers, and H. Berghmans, J . Appl. Polymer Sci., 60: 1891 (1996).
24. T. Heinze, K. Rahn, M. Jaspers, and H. Berghmans, Macromol. Chem., 197 4207 (1996).
25. T. Itoh, Y. Tsuji, H. Suzuki, T. Fukuda, and T. Miyamoto, Polymer J., 24: 641 (1992).
26. J. E.Sealey, G.Samaranayake, J. G.Todd,and W. G. Glasser, J . Polyner Sci., Polymer
Chem. Ed., 34: 1613 (1996).
27. T. Iwata, J. Azuma, K. Okamura. M. Muramoto,and B. Chun, Carbohvdr. Res., 224: 277
( 1992).
28. T. Iwata, K. Okamura, J. Azuma, H. Chanzy, and F. Tanaka, Cellulose, l : 67 (1994).
29. T. Iwata, K. Okarnura, J. Azuma, and F. Tanaka, Cellulose, 3: 91 ( 1996).
30. T. Iwata, K. Okamura, J. Azuma, and F. Tanaka, Cellu/o.se. 3 : 107 (1996).
31. M. Hasegawa, A. Isogai, F. Onabe, and M. Usada, J. Appl. Polymer Sci.. 45: 1857 (1992).
32. T. Liebert, M . Schnabelrauch, D. Klemm, and U. Erler, Cellulose, l : 249 (1994).
33. M. Muramoto, M. Yoshioka, and N. Shiraishi, Sen'i Gukkaishi. 46: 496 (1990).
34. K. Goto, M. Yoshioka, and N. Shiraishi, Mokrczai Gakkaishi, 37: 57 (1991).
35. F. Guittard, T. Yamagishi, A. Cambon, and P. Sixou, Macromolecules, 27: 6988 (1994).
36. A. Ohnishi and T. Shibata, Cell~clow Conmun.,4 : 2 (1997).
37. S. Ciovica, V. Sunel, and N. Asandei, Cellulose Clwm. Technol., 28: 58 1 (1994).
38. S. Ciovica, V. Sunel, and N. Asandei, Cellulose Chetn. Technol., 28: 493 (1994).
39. R. Pernikis and B. Lazdina, Cellulose Chern. Techno/., 30: 187 (1996).
40. M. Karnath, J. Kincaid, and B. Mandal. J. Appl. Polytner Sci., 59: 45 (1996).
41. K. Arai and H. Coda. Sen 'i Gakkuishi. 49: 482 (1993).
42. S. Vogt, T. Heinze, K. Rottig, and D. Klemm, Carbohydr. Res., 266: 315 (1995).
43. K. Kowsaka. K. Okajirna, and K. Kamide, Polymer J., 23: 823 (1991).
44. Y . Tezuka and Y. Tsuchiya. C d m h y d r : Res., 273: 83 (1995).
45. Y. Tezuka, C~wlx~hydr. Res., 241: 285 (1993).
46. Y . Tezuka, K. Imai. M. Oshima, and K. Ito, Carbohydr. Res., 222: 255 (1992).
47. L. R. Schroeder and F. C. Haigh. Ttrppi J., 62: 103 (1979).
48. R. Evans, R. H. Wearne and A. F. A. Wallis, J . Appl. Polyrner Sci., 42: 8 13 ( 1991 ).
49. R. Evans. R. H. Wearne and A. F. A. Wallis, J. Appl. Polymer S r i , 42: 821 ( l 991 ).
50. H. Shibazaki, S. Kuga and T. Okano, Cellulose, 4: 75 (1997).
51. T. Lindstrom and G. Soderberg, Nordic P~rlp Paper Res. J.. I : 26 (1986).
624 lsogai
Fumiaki Nakatsubo
Kyoto University, Kyoto, japan
I. INTRODUCTION
Cellulose is the most abundant natural organic polymer existing as a main plant cell wall
component and is important as a biodegradable and renewable organic resource [l]. The
study of cellulose, therefore, has continued for more than 150 years. However, there are
still several problems which should be solved: biosynthesis, crystal structure, chemical
synthesis, regiospecific substitution reactions, structure-function relationships of their de-
rivatives, and so on [ 1,2].
The synthesis of cellulose has been a very important but extremely difficult problem
to solve, sinceSchlubach first tried the synthesis in 1941 [3]. Recently, Kobayashi and
co-workers reported enzymic synthesis of cellulose [4]. Their synthetic method with cel-
lulase is important and interesting as the first in-vitro synthesis using enzyme. The method,
however, does not satisfy the recent molecular design of cellulose derivatives having spe-
cial functions,because it may not enable us to regiospecifically introduce the special
functional group into only the desired hydroxyl groups in the repeating pyranose units of
cellulose.
There are many functional cellulose derivatives: cellulose esters and ethers having
liquid crystalline properties [5] and chiral recognition ability [6], sulfated cellulose with
anticoagulant activity like a heparin [7], branched cellulosederivatives with antitumor
activity [8], and so on. Much is still unknown about the relationship between the structure
and properties, which derivatives are more functional or active among those substituted
at 2,3- or 6-positions. For these studies and,furthermore,for molecular design of the
advanced materials from cellulose, methods which make it possible to prepare cellulose
derivatives with functional groups at special positions among 2,3,6-hydroxyl groups in the
repeating pyranose unit of cellulose are indispensable.
Polycondensation and ring-opening polymerization methods using glucose derivative
as a starting monomer satisfy the above requirements [9], but all trials synthesizing cel-
lulose by these methods have been unsuccessful.
Husemann and Muller [lo] and Hirano [ I l l reported the condensation of 2,3,6-
glucose tricarbanilate with phosphorus pentoxide in a mixture of chlorofoddimethyl-
sulfoxideto give a cellulose-like polymer containing branched polymer and about 1%
phosphorus. Micheel et al. [ 121 and Uryu et al. [ 131 tried cationic polymerization of 1,4-
anhydro-2,3,6-tri-O-benzyl-a-~-glucose initiated with various Lewis acids, but stereo-
regular (1 +4)-P-D-glUCOpyranan was not obtained. Furthermore, Uryu et al. [ 141 reported
627
628 Nakatsubo
A. BasicSyntheticStrategy
In principle, the synthesis of cellulose seems tobe simple because only three synthetic
problems need to be solved: regiospecific control and stereospecific control, and increasing
the molecular weight (Fig. l ) .
Chemical Synthesis of Cellulose 629
U
Regiospecific control (1,4-bond)
OR 1 -CHpCH=CHp (All)
CICHpCO- J -COCH,( Ac)
[
R = -CO-C(CH& (Pi")
-CH2C& (Bn)
which is further converted to the activated compound, a-imidate (3). The glycosylation of
the acceptor (2) with the donor (3) catalyzed by BF, etherate in CH,CIZ at -70°C gave
the expectedcellobiosederivative (4) in 85% yield without any production of the a -
anomer, via a completely S,2 reaction mechanism as proposed by Schmidt et al. [33].
Repeating the set of two reactions, deprotection of the PMB group and subsequent
P-glycosylation with donor (3), six times, finally affords a cello-octaose derivative (10)
which is converted to an acetyl derivative after deprotection and subsequent acetylation.
The cello-octaose acetate obtained was interestingly compared with the acetate from hy-
drolysis product of cellulose reported by Buchanan et al. 1341.
Thus, the selection of both combinations of the protective groups, allyl, PMB, and
Bn groups, and the P-glycosylation method (imidatemethod)werefound to be quite
suitable for the linear synthesis of cello-oligosaccharides. However, it was revealed that
selection of the protective groups is not suitable for the convergent synthesis, because of
instability of the dimeric and tetrameric imidates: imidoylation of the donor derived from
cellobiose (4) and cellotetraose (6) derivatives always gave a mixture consisting of a - and
P-anomers which cannot be obtained as a pure compound, i.e., these imidates are very
unstable, resulting in decomposition upon separation by silica gel TLC.
Generally, an electron-donatingprotectivegroup, such as a Bn group,accelerates
reactivity of the glycosyl donor, but an electron-withdrawing group such as the acyl group
reduces it [ 3 5 ] .Consequently, it is expected that we must design suitable starting materials
with both optimum reactivity and stability by a partial exchange of several Bn groups with
acyl groups in compound (1).
”. . O
, Bn OBn
s
m 0
0
-4
OBn OBn OBn
no a-Glucoside
-0 HO, --
-/ -U
- 12 [a-Glucoside (29%)
<OAc
that the 3-0-Bn group is indispensable for obtaining highly stereoselective P-glycosylation
in high yield. The effect of the benzyl group introduced into the 6 - 0 position on the alp
ratio seems to be small. These results agree well with those 011 the glycosylation ofN-
acetyl glucosamine derivatives with acetyl galactosyl bromide reported by Sinay 1371.
Consequently, the positions in compound (1) i n which the protective groups can be
replaced by acyl groups for the stabilization of a-imidate are limited only to the 2 - 0 and
6 - 0 positions for molecular design of the starting material for the convergent synthesis.
Then, Nishimura et al. [ 38) considered the most suitable combination of the 2 - 0 and
6 - 0 protective groups of glycosyl acceptor and donor, as summarized in Table 2.
Reactions 1-4 were conducted by the general glycosylation method. and reactions
S-7 used a simple vacuum system (Fig. 6). A high-vacuum system is conveniently used
Yield (%)
Reaction Reaction
- Ac - Ac 2 16.4 29.0
-Ac -Bn 2 13.0 33.0
-Bn -Ac I 94.2
-B11 -Bn I 96.0
-0Bn
CCI,
634 Nakatsubo
Donor Acceptor
Reaction 0-glycoside Stability
R2 no. R, R2 yield (96) of imidate
1 -Ac - AC - AC -Ac 15 Q
2 -Piv -Ac -AC -AC 17 8
3 -Piv -Piv -Ac -AC 32 0
4 -Piv -Piv -Piv -Piv S1 Q
S -Piv
-Piv -Piv -Piv 85 Q
6 -Bn
-Bn -Piv -Piv 60 ( ( ~ 2 0 ) 0
7:' -Piv -Bn -Piv -Bn 86 0
8 -Bn -Bn -Bn -Bn 84 X
for the synthesis of higher cello-oligosaccharides to escape the hydrolysis of the moisture-
sensitive imidate. A sing-opening polymerization, which must be carried out under com-
pletely anhydrous reaction conditions, is also conducted in the high-vacuum system as
described in a later section.
All of these reactions gave only P-glycosides, except for reaction 6. Thesedata
indicate that the introduction of acryl (acetyl or pivaloyl) groups fairly stabilizes the un-
stable imidate to be purified by silica gel chromatography as expected, and that pivaloyl
groups are more useful than acetyl groups for the P-glycosylation. Pivaloyl groups have
a larger electron-donating effect than do acetyl groups [39], resulting in higher reactivity
of the glycosyl acceptors and imidates. In addition, pivaloyl groups at the 0 - 2 position
prevent the formation of by-products such as an orthoester because of the larger steric
hindrance.
OPiv
PMB' B 8 F o d
OPiv OBn
13 14
removal of protective groups and subsequent acetylation. The ‘H-NMR spectrum of the
synthesized acetyl cello-eicosaose derivative was found to be almost identical to that of
authentic cellulose triacetate (CTA).
Generally, in carbohydrate chemistry the term oligosaccharide refers to a substance
having two to ten monosaccharide units 1171. According to this terminology, derivatives
with degree of polymerization (DP) above 12 are not oligosaccharides, although they may
not be macromolecules. Concerning this point, Kobayashi et al. regarded their synthesized
product having a DP value of 22 as a polysaccharide, that is, “cellulose” [4]. In fact, at
around this DP, several characteristics of cellulose such as crystal structure seem to appear
[4]. Isogai et al. also regarded samples having DP values of 15 as low-molecular-weight
cellulose [42]. Thus, the synthesized compounds (29)-(33) may be regarded as “cellu-
lose,” more strictly as low-molecular-weight or medium-size cellulose. Here, the stepwise
synthesis of cellulose was achieved for the first time.
Interesting changes in several of the properties accompanying an increase in DP of
a series of the synthesized medium-size molecular weight compounds were found. Peak
patterns of both ‘H- and “C-NMR spectra were gradually simplified with an increase of
DP, leading to peaks attributable only to the internal repeating units. Thespectrum of
acetyl cello-eicosaose derivative was almost identical to that of CTA.
Upon gel permeation chromatography (GPC) analysis, the plots of logarithm of mo-
lecular weight (log M> versus elution volume are clearly linear, as shown in Fig. 9. Segal
reported measurement of the molecular weight of cellulose trinitrates by GPC, but the
values obtained were fairly large in comparison with those obtained viscometrically [43].
In this investigation it was found that, at least with high-DPcello-oligosaccharidesor
medium-size cellulose up to eicosaose, the molecular weight can be determined almost
exactly using this calibration curve.
According to Freudenberg [44], if the linkages in a polymer-homologous series are
uniform, the plot of[M] n/n against ( n - I)/n yields a straight line. Here, [m n =
molecular rotation and 11 = DP. Figure 10 shows a plot of molecular rotations versus DP.
The Freudenberg theory was found to be applicable to a series of a-D-acetates of
cello-oligosaccharides up to DP 7 [45]. For the synthesized cellulose derivatives,the theory
appears to fit up to DP 16. The point for DP 20 deviates from this straight line in spite
of uniformity of the glycosidic linkages. It may be assumed that a conformational change
due to a polymer effect starts to appear at around cello-eicosaose. Kajiwara et al. proposed
an extended zigzag structure with an inflection point at DP 20, by the Monte Carlo sim-
ulation method based on the molecular mechanics [46]. For the proof of this possibility,
synthesis of cellulose derivatives above eicosaose and their properties have to be studied.
1 ooooc
0: ~o~ys~yrene
10000
2
0,
-
0
1000
c
\
c
-5000 -1 /4
-10000 I I I I I
0.5 0.6 0.7 0.8 0.9 1
(n-1 )/n
FIGURE 10 Relation between degree of polymerization and molecular rotation: n, degree of po-
*,
lymerization; IM] ! l . molecular rotation; 0.2,h-di-O-pivaloyl-3-O-benzyl cellulose series; (Y-D-
acctatc of cello-oligosaccharides series rcported by Wolforrn and Dacons 1451.
A*,,
36: (1-4)-a-P 37: (1+4)-p-P
On the other hand, a ring-opening polymerization requires only one reaction for the
syntheses of cello-eicosaose and also polysaccharides from the starting monomer, although
usually it gives a polydispersed polymer. Thus, ring-opening polymerization is very useful
for the preparation of high-molecular-weight polysaccharides. This method, however, has
to overcome two extremely difficult problems, regiospecificity and stereospecificity, at the
same time during the polymcrization. That is, ring-opening polymerizations of 1,4-anhydro
glucose (34) and glucose 1,2,4-orthoester (35) monomers usually yield four possible struc-
tural units, that is, 1,4-anhydro glucose monomer (34) usually gives the (1+4)-a (36) and
(1+4)-p (37) -D-glucopyranosidic units and the ( I + S ) - a (38) and ( I "+S)-p (39) -D-
glucofuranosidic units. And glucopyranose l ,2,4-orthoestermonomer (35) gives the
642 Nakatsubo
(1-+4)-a (36) and (1+4)-P (37)-~-glucopyranosidicunits, and (1 -+2)-a (40) and (1+2)-
P (41)-~-glucopyranosidicunits. That is, for synthesis of cellulose, glucan consisting of
only (1-4)-P-~-glucopyranosidic unit in four possible units has to be made.
Furthermore, the stereo- and regioselectivity must be extremely high during the po-
lymerization. For example, 99% stereoselectivity yielding 99% P- and l% a-glucosides
upon glycosylation is a satisfactory result in the stepwise synthesis, because the undesired
a-anomer can be separated by chromatography after glycosylation. However, 99% stereo-
selectivity is not acceptable for the ring-opening polymerization. This is because the ei-
cosaose derivative obtained by polymerization with 99% selectivity is theoretically a mix-
ture consisting of 524,288 (= 2'2')"') molecular species containing 82.6% (= 0.99(20"))of a
stereoregular cello-eicosaose derivative. This is why no one so far has succeeded in the
synthesis of cellulose by ring-opening polymerization.
Yield
Time
Temp [a],
Exp.
Monomer
Initiator
no. (“C) (h) (%) (deg) 10-’MGKe DP,,
Generally, the electron-donating benzyl group accelerates reactivity. but the electron-
withdrawing acetyl group retards both the glycosylations and the polymerizations of an-
hydro-sugars. For example, Zachoval et al. [51 ] reported that I ,6-anhydro-P-D-glucopy-
ranose triacetate was less reactive than 1,6-anhydro-P-1~-glucopyranose tribenzylether.
Monomers (42) and (43), having two benzyl groups. expectedly afforded polymer
with high molecular weight. Monomer(45), having one benzyl group, did not afford
polymer under various reaction conditions (Table 3, experiments 20-25), but unexpectedly,
monomer (44), having only one benzyl group, afforded polymer with high molecular
weight. Consequently, the present results indicate that the important factors for accelerating
polymerizability are not only the number of the substituent benzyl groups but also the
positions where benzyl groups are attached. The comnlon substituent among monomers
(42), (43), and (44) is a benzyl group at the 3 - 0 position: monomer (45) does not have a
benzyl group at the 3 - 0 position. Thus, the benzyl group at the 3 - 0 position is indispens-
able for yielding glucan with high molecular weight.
Comparing monomers (43) having a benzyl group at the 6-0 position and (44) having
a pivaloyl group at the 6-0 position, there is little difference in molecular weight under
optimum reaction conditions (Table 3, experiments 8 and 15). Consequently, the benzyl
group at the 6 - 0 position did not have much effect on polymerizability.
benzyl group at the 3 - 0 position has a special function for yielding ;I stereoregular poly-
saccharide with high molecular weight in a ring-opening polymerization [results from
monomers (43) and (45)). It was reconfirmed that the presence of the pivaloyl group at
the 2 - 0 position makes the polysaccharide take the P-configuration [results from mono-
mers (42) and (44)]. Polysaccharides with high molecular weight tend to have high stereo-
regularity as shown in Table 3 (experiments 8, 9, 15, and 16). Consequently. both the
pivaloyl group at the 2 - 0 position and the benzyl group at the 3-0 position are indispens-
able for yielding stereoregular ( 1 +5)-/3-~-glucofuranan derivatives with high molecular
weight.
Furthermore, polymerization of 1,4-anhydro-a-~-glucopyranose was found to always
preferentially afford ( 1 +5)-~-glucofuranose units, not ( 1 ~4)-P-D-giucopyranoseunits.
These results agreed with the cases of the ring-opening polymerization of 2,7-dioxabicy-
clo-[2.2.1]heptane [52] and that of 1,4-anhydro-2,3,6-tri-O-benzyl-a-~-glucopyranose [ 131.
Thus, it is also concluded that the 1,4-anhydro-a-~-glucopyranoseskeleton is not suitable
for yielding a ( 1 +4)-P-~-gIucopyranan, that is, a cellulose molecule. It was revealed from
the results of the polyrnerizations of 1,4-anhydroglucose derivatives that at least one of
the most important problems i n the synthesis of cellulose, i.e., stereospecificity giving a
stereoregular P-glucosidic polymer, could be successfully solved by applying substituent
effects of 2-O-pivaloyl and 3-O-benzyl groups obtained from synthetic studies of cello-
oligosaccharides.
OR1
43 t
\ poly(43): R1 = Piv, R2 = R3 = Bn
46
OPiv OBn
BnO /
\ poly(43)': R1 = R2 = R3 = H
X"",,
47
poly(47): R, = Piv, R2 = R3 = Bn
48:R1 = Ac, R2 = R3 = Bn
49: R1 = R2 = R3 =Ac
50: R1 = R2 = R3 H
FIGURE 11 Synthetic route for cellulose by a cationic ring-opening polymerization. "p-TsOH/
benzene/reflux/55%,"PF,/toluene/-3O0C, 'N,N'-carbonyldiimidazole/benzene/reflux,
62.8%. "Ph,CBFJ
CH2Cl,/r.t.
100 95 90 85 80 75 70
6 (PPW
FIGURE 12 2D-NMR spectra of poIy(47) from C-H COSY experiment (CDCI? as solvent).
Chemical Synthesisof Cellulose 647
C2-OAc
t
.-,OAc I
. .oO
*- AcO
OAC
Authentic CTA
SyntheticCTA t 1 t ~ - t
120
140
l60
180 100 80 60 2040
6 (PPW
FIGURE 13 'H- and I3C-NMR spectra of authentic and synthetic cellulose triacetate (CDCI, as
solvent).
13. Finally, the cellulose triacetate (49) thus obtained was converted into cellulose (50)
by deacetylation with NaOCH, in THF/methanol. The IR spectrum and X-ray diagram
(Fig. 14) of cellulose prepared in this way were completely identical with those of regen-
erated cellulose with the cellulose-I1 crystal structure. Thus, the first chemical synthesis
of cellulose by a cationic ring-opening polymerization had been achieved [ 151.
Nakatsubo
10 20 30 40
28 (deg.)
FIGURE 14 X-ray diffractograms of (A) Whatman cellulose CFI 1. (B) regenerated celluhe. and
(C) synthetic cellulose.
The molecular weights of poly(47). poly(51), and poly(52) were almost equal, but
those of poly(53) were low under all reaction conditions tried. The polymerizability of
four Inonomers is in the order (47) (51) >> (53), as judged from the highest molecular
weight and yield obtained from these monomers (Table 4, experiments 3, S , 8, and 12).
It can be said that polymerizability of (53) could notbe compared exactly with that of
(47), (Sl), and (52) because polymerization of (53) was conducted at lower monomer
concentration than that of (47), (Sl), and (52), but the above-mentioned trend must be
true and substantial.
In cases of ring-opening polymerizations of I ,4-anhydro-c~-~-glucopyranose deriva-
tives, the benzyl group at the 3-0 position is indispensable for yielding glucan with high
molecular weight. Similarly, in the present cases of cY-D-glUcOpyranOSe 1,2,4-orthopivalate
derivatives, the benzyl group at the 3-0 position is indispensable for realizing stereo-
regularity, butis dispensable for yielding glucan with high molecular weight: that is,
monomer (52), having a pivaloyl group at the 3-0 position, polymerized well but without
yielding stereoregularity.
Monomers having one benzyl group and two pivaloyl groups (including an ortho-
pivaloyl group), that is, monomers (51) and (52), had the same polymerizability as mono-
mer (47), having two benzyl groups and one pivaloyl group. Monomer (53), however, had
Chemical Synthesis of Cellulose 649
remarkably low polymerizability among the four monomers. Consequently, at least one
benzyl group is indispensable for yielding a polymer with high molecular weight.
Polymerizations of (47) and (51), having 3-0-benzylgroups, gave stereoregular
( I +4)-p-D-glUCOpyranan derivatives (Table 4, experiments 1-6). However. polymeriza-
tions of (52) did afford a stereo-irregular polysaccharide consisting of ( 1 +2)-a-P, ( 1 +2)-
p-P, and ( 1 -94)-p-P units. The (l+2)-bond formation increased with an decrease in tem-
perature. That is, while the probability of coordination of Ph,CBF., with a C., oxygen was
equal to that with a C., oxygen at 0°C and at 20°C, the coordination with the C? oxygen
took place in preference to that with the C, oxygen at -30°C: the product ratios were
( I +2)-a-P/( 1 +2)-p-P/( 1+4)-p-P = 1 : I :2 [Table 4, experiment 8 (0°C) and 9 (20"C)I and
(l+2)-a-P/( 1+2)-p-P/( 1 +4)-p-P = 3:3:4 [Table 4, experiment 7 (-30"C)l. The products
ratio was calculated from the peak areas of the anomeric carbons of the corresponding
constitutional units. The results indicate that in the case of coordination of a catalyst with
the C, oxygen, the probability of an a-side attack resulting in the formation of (l+2)-a-
P was equal to that of a @-side attack resulting in formation of (1+2)-p-P.
After all, the benzyl group at the 3 - 0 position is indispensable for yielding a stereo-
regular cellulose derivative, i.e., a ( 1 -4)-P-~-glucopyranan derivative. This fact agreed
with the case for yielding stereoregular ( 1 +5)-/3-~-glucofuranan derivatives from 1,4-
anhydro-a-glucopyranose derivatives.
droxyl groups on the glycosylation. The electron-withdrawing abilities of acyl groups are
associated with the pK,, values of the corresponding carboxylic acids.
OBn
47: R=-C(CH&
54: R=-CHPCH~
55: R=-CH3
56: R=-CBHS
R
Hori et al. [57] selected three additional orthoesters, 1,2,4-orthopropionate (54), or-
thoacetate ( S ) , and orthobenzoate (56), as starting monomers for cationic ring-opening
polymerization and prepared from propionic (pK, 4.88), acetic (pK, 4.76), and benzoic
(pK,, 4.20) acids to investigate the electronic effects of the orthoester groups on the cationic
ring-opening polymerization. The results were compared with that of orthopivalate (47)
(pK,, of pivalic acid, 5.05), whose polymerization gave a completely stereoregular cellulose
derivative as described previously.
Polymerizations of 3,6-di-O-benzyl-a-~-glucopyranose 1,2,4-orthoester derivatives
(54)-(56), were camed out under the same reaction conditions as those giving cellulose
derivative with BE 19.3 in the polymerization of (47). That is, all polymerizations were
carried out at 20°C in the presence of triphenylcarbenium tetrafluoroborate as an initiator,
in 5 mol% of initiator concentration and in 100 g/100 mL of monomer concentration.
The results are summarized in Table 5. Specific rotations of polys(54)-(56) were
+ + +
positive values, 1.56", 12.9", and 11.8", respectively, as compared with poly(47), a
cellulose derivative which has a large negative specific rotation, -37.2". These specific
rotation values suggest that not all polymers newly obtained from monomers (54)-(56)
are stereoregular. In fact, all these polymers were found to be nonregioregular, consisting
of ( 1 -2)-and ( 1 +4)-P units, although with P-glycosidic linkage, that is, with stereo-
specificity but without regiospecificity.
These data indicate that there is no distinct relationship between the production of
(1-+4)-p-Punit and the pK,, values of the carboxylic acid groups introduced at the 2-0-
positions of the monomers. Only the 2-0-pivaloyl group has a characteristic effect on the
fate of ring-opening polymerization, resulting in the formation of stereoregular (1+4)-p-
~~
"Initiator, Ph,CBF,; initiator concentration, S mol%; solvent, CH,CI,; monomer/solvent, 100 g / l O mL; reaction
temperature, 20°C.
hpK,,values are those of the carboxylic acid groups introduced into 2-0-pos~tionsof monomers.
'Polymer is insoluble fraction in chloroformh-hexane (ca. 1/S. v/v).
'Calculated from polystyrene standard.
"Calculated from the anomeric peak ratios in "C-NMR spectra of polymers.
thesis Chemical of Cellulose 651
IV. FUTUREPROSPECTS
In spite of all efforts at chemical synthesis of cellulose for about55 years since Schulubach
first started synthetic study, no one reached the difficult goal. Recently, the author and his
co-workers succeeded in the first chemical synthesis of cellulose. In this chapter, their
story of search for synthesis for a period of 15 years was described briefly. In this story,
extremely important findings are on the substituent effects on P-glycosylation. That is,
both 2-0-pivaloyl and 3-0-benzyl groups were found to be indispensable for obtaining P-
glycoside with high stereoselectivity, from the results obtained on both synthetic methods,
i.e., stepwise synthesisgiving cello-eicosaose derivative (DP20) and cationic ring-opening
polymerization giving cellulose derivative with DP about 20 (DP 45 in the recent experi-
mental results). Furthermore, interesting results were obtained that the 2-0-pivaloyl group
also participates in the highly regioselective 1,4-bond formation in ring-opening poly-
merization.
However, the way that both stereo- and regioselectivity can be overcome by the
effects of both 2-0-pivaloyl and 3-0-benzyl groups are still unknown, especially highly
regioselective control by the 2-0-pivaloyl group on ring-opening polymerization of the
orthoester derivatives, although the neighboring participation of the 2-0-pivaloyl group
giving P-glycoside is well known. Furthermore, additional problems which cannot be
understood are that the present ring-opening polymerization method developed for the
synthesis of cellulose cannot be applied for the synthesis of (1 +4)-P-xylopyranan. The
ring-opening polymerization of 3-0-benzyl-xylopyranose 1,2,4-orthoester always gives a
polymer consisting of both (1+2) and (1+4)-P-xylopyranose units [58]. We expect that
research and the solution of the aboveproblems in synthesis may lead to the logical
molecular design for the synthesis of other natural polysaccharides. We have succeeded
in the syntheses of stereoregular galactofuranan and arabinofuranan [59].
On the other hand, one of the future aims in cellulose application is development of
functional derivatives with special functions. For this, preparation of highly regioselec-
tively substituted cellulose derivatives is a key. At present, 6-0-substituted cellulose de-
rivatives can be synthesized, but the cellulose derivatives substituted at only the 2-0- and
3-0-positions cannot be obtained. Cellulose derivatives (44) in Fig. 8 and poly(47) and
poly(51) may be theoretically used as starting materials for the preparation of regioselec-
652 Nakatsubo
tively 2-0- and 3-0-substituted cellulose, because the 2-0- and 3-0-positions of these
derivatives are protected with ester and ether groups, respectively, which can be depro-
tected independently. At present, we have succeeded in the synthesis of all kinds of pos-
sibleregioselectivelymethylatedcellulosederivativesfrompoly(47)andpoly(51) 1591.
The studies on the relationship between structure and function of these regioselectively
substituted cellulose derivatives may enable us to design new functional cellulose deriv-
atives in the future.
Thus, the present chemical synthetic methods for cellulose are expected to apply to
the solution of basic problems of cellulose chemistry and technology.
REFERENCES
I. See, for example, T. P. Nevell and S. H. Zeronian, in Crllulosc~Chenris/r;yemel I t s Applicertion.s
(T. P. Nevell and S. H. Zeronian, eds.), Ellis Horwood-Wiley, NcwYork, p. IS (1985).
2. See books on cellulose: (a) R. H. Atalla, in ACS S y n p Sex 340 (R. H. Atalla, ed.), American
Chemical Society, Washington, DC, p. I (1987); (b) R. P. Millane et al.. In Cel/ulose ( r n d
Woot/"Chet?listry c n ? d Tec.hr~ology,Proc. Tenth Cellulo.sc~ Cor$ (C. Schuerch. ed.). Wiley. New
York, p. 39 ( 1989).
3. H.M. Schlubach and L. Luhrs, Ann, 547: 73 (1941).
4. S. Kobayashi. K. Kashiwa, T. Kawasaki, and S. Shoda. J. An?. Chetn. Soc., 113: 3079 (199 l ) .
S. (a) P. Sixou ct al., in Cellulose-Str~~ctl~re, Modification crnd Hydro1y.si.s (R. A. Young and R.
M. Rowell, eds.), Wiley, NewYork,p.203 (1986); (b) M. Siekmeyer et al.. in Cellulose-
Structure c u d Functioned Aspecrs (J. F. Kennedy, G. 0. Phillips, and P.A. Williams, cds.).
Ellis Horwood, NewYork,p.34.5 (1989).
6. (a) G. Hesse and R. Hagel, Chrotncr/ogrrrphia, 6: 277 (1973): (b) T. Shibata, I. Okamito, and
K. Ishii, J. Lie/. Chromatog., Y: 313 (1986); (c) Y. Okamoto, M. Kawashima, and K. Hatada,
J. A n . Chetn. Sac.. 106: S357 (1984); (d) T. Shibata, T. Sei, H. Nishimura, and K. Deguchi.
Chror~~c~togrc~phia~ 24: SS2 ( I 987).
7. (a) K. Kamide, K. Okajima, T. Matsui, and H. Kobayashi, Po/ytnc,r J.. 15: 309 (1983); (b) K.
Okajima, in Cel/L~lose-Stnlcrllr-r),se-Sfr~~ct~~re
c ~ n c lFunctional Aspects (J. F. Kennedy. G. 0. Phillips, and
P.A. Williams, cds.). Ellis Horwood. New York, p. 439 ( 1989).
8. (a) K. Matsuzaki. I. Yamamoto,and T. Sato. Makrornol. Chern.. 186: 449 (1985); (b) 1. Ya-
mamoto, 1. Takayama. K. Hornma, T. Gonda, K. Matsuzaki, K. Hatanaka, T. Uryu, 0. Yoshida.
H. Nakashirna, N. Yamamoto, Y. Kaneko; and T. Mimura, Carhohydr: Polymers, 14: S3 ( 1991).
9. General reviews on polysaccharide synthesis: (a) A. F. Bochkov and G. E. Zaikov, Chemistt?
of the 0-Glycosiclic Bond-Fornution m e 1 CIlwwge (C. Schuerch. trans., cd.). Pergamon
Press. Oxford, U.K., p. 130 (1979); (b) N. K. Kochetkov, 7ktrc~hc~clron. 43: 2389 (1987); (c)
N.K. Kochetkov. in Studies in Nutuml Products Chernistty, Vol. 14. Elsevier. Amsterdam. p.
20 I ( 1994).
10. E. Husemann and G. J. M. Muller, Makrornol. Ch~tn..01: 212 ( 1966).
11. S. Hirano, Agric. Biol. Chetn.. 37: 187 (1973).
12. (a) F. Micheel, 0.-E. Brodde, and K. Reinkmg. LiehigsAnnu. Chcw~,124 (1974); (b) F.
Micheel and 0.-E. Brodde, Liebigs Annu. Chcwr., 702 (1974); ( c ) F. Micheel and 0.-E. Bordde,
Liehigs Annu. Chern., I 107 ( 1975).
13. T. Uryu. C. Yarnaguchi. K. Morikawa, K. Terui, T. Kanai. and K. Matsuzaki. Mercror~?oleclrl~~s.
/H: 599 (1985).
14. (a) T. Uryu, K. Kitano, K. Ito, J . Yamanouchi,and K. Matsuzaki, Mclc,nrtlo/c~crrlt.s, 14: I
(1981); (b) T. Uryu, J. ydmanouchi, T. Kato. S. Higuchi. and K. Matsuzaki, J . An?. Chenr.
Soc.. 105: 6865 (1983).
IS. F. Nakatsubo, H. Knmitakahra. and M. Hori, J. Am.Chrrn. Soc.. 1fH: 1677 (1996).
Chemical Synthesis of Cellulose 653
49. K. Kobayashi, T. Ishii, M. Okada, and C. Schuerch, Polymer J., 25: 49 (1993).
50. (a) H. Kamitakahara, F. Nakatsubo, and K. Murakami, Mokuzui Gnkkaishi, 40: 302 (1994);
(b) H. Kamitakahara, F. Nakatsubo, and K. Murakami, Macromolecules. 27: 5937 (1994); (c)
H. Kamitakahara and F. Nakatsubo, Macromolecules, 29: 1 1 19 (1996).
51. J. Zachoval and C. Schuerch, J. Am. Chm. Soc., 91: 1165 (1969).
52. H. K. Hall,Jr., F. DeBlauwe, L. J. Carr, V. S. Rao, and G. S. J. Reddy, J. PolymerSci.,
Polymer Symp.. 56: 101 (1976).
53. (a) N. K. Kochetkov, A. F. Bochkov, and I. G. Yazlovertsky, k v . Akad. Nauk SSSR, Se,: Khinz.,
1972 (1966); (b) N . K.Kochetkov, A. Ya.Khorlin, A. F. Bochkov, and I. G. Yazlovertsky,
Curbohydr. Res., 2: 84 (1966); (c) A. F. Bochkov, I. G. Yazlovertsky, and N. K. Kochetkov,
17s. Akad. Nuuk SSSR, Ser. Khim., 1812 (1968); (d) N. K. Kochetkov, A. F. Bochkov, and 1.
G. Yazlovertsky, IZIJ.Akad. Nauk SSSR, Se,: Khim., 1818 (1968); (e) N. K. Kochetkov, A. F.
Bochkov,and I. G . Ydzlovertsky, Curbohydr. Res., 9: 49 (1969); (f) A. F. Bochkov, V. N.
Chernetsky,and N. K. Kochetkov, lzv. Akad. Nuuk SSSR, Se,: Khinz., 465 (1975); (g) A. F.
Bochkov, I. V. Obruchnikov, and N. K. Kochetkov, Zh. Obshch. Khim., 42: 2766 (1972); (h)
A. F. Bochkov, I. V. Obruchnikov, and N. K. Kochetkov, Zh. Obshch. Khinz., 44: 1197 (1974);
(i) A. F. Bochkov, V. N. Chernetsky, and N. K. Kochetkov, Curhohydr. Res., 43: 35 (1975);
ti) A. F. Bochkov and A. V. Rodionov, Izv. Akacl. Nauk SSSR, Sex Khim., 2789 ( 1 976).
54. (a) J. P. Kutney and A. H. Ratcliffe, Synthefic Commun., 5: 47 (1975); (b) S.-K. Kang, J.-H.
Jeon, K.-S. Nam, C.-H. Park, and H.-W. Lee, Synthetic Commun., 24: 305 (1994).
55. E. Katchalski and M. Sela, in Advances in Prorein Chernistq (C. B. Antinsen, Jr., M. L.
Anson, K. Bailey, and J. T. Edsall, eds.), Academic Press, New York, p. 243 (1985).
56. H. Kamitakahara, M. Hori, and F. Nakatsubo, Muc,rornolecules, 29: 6126 (1996).
57. M. Hori, H. Kamitakahara, and F. Nakatsubo, Macromolecules, 30: 2891 (1997).
58. M. Hori and F. Nakatsubo, unpublished data.
59. F. Nakatsubo, Proc. 10th ISWPC, Yokohama, Japan, 1999, Volume l , p. 20.
16
wood Plasticization
Nobuo Shiraishi
Kyoto University, Kyoto, lapan
1. INTRODUCTION
Methods for processing wood are very limited. While metals, plastics, and glass can be
processed in the liquid phase at high temperatures, natural wood cannot. This difference
is due to the lack of plasticity of wood, so that it cannot be melted, dissolved, or softened
sufficiently for molding. Consequently, the scope for utilizing wood is restricted, which
sometimes makes wood less valuable as a material. If plastic properties could be imparted
to wood, it would become a more useful material.
This need, and the desirability of using wood waste and renewable forest product
resources better, have resulted in extensive new studies on wood chemicals, modified
natural polymers, new pulping methods, and reconstituted wood products. Inherent in this
work is the plasticization of wood by simple chemical processing [ 1-71.
The name “plastics” is given to the numerous macromolecular organic materials that can
be softened, melted, and molded by heat and pressure or by mechanical means, such as
mixing and calendering.
Wood, which is also composed of macromolecular organic materials, differs from
the plastics in such basic properties as plasticity, thermoplasticity, and dissolubility in
organic or aqueous solvents. In other words, wood lacks plasticity, which means that it
cannot be softened sufficiently for molding, melted, or dissolved. A summary of the pre-
vious literature on thermoplasticities of wood reveals that only the phenomena up to the
second-order transition and/or thermal softening have been investigated, but not the phe-
nomena of thermal flow [&lo].
Wood is composed of 50-55% cellulose, 15-25% hemicellulose, and 20-30% lig-
nin, with small quantities of ash and extractives. The main components make up an in-
terwoven network in the cell walls and middle lamella. The minor components are mostly
in cell lumens or special tissues such as resin canals, and are directly or indirectly related
to the physiology of trees.
Thus, the components that are directly related to the fundamental properties of wood,
such as thermoplasticity, are considered to be the main ones, and such properties are
655
656 Shiraishi
formed not as the simple summation of properties of individual components but as the
integration of these properties as a result of mutual interactions by these components.
This has been observed in previous studies on thermoplasticity of wood. Lignin and
hemicellulose are amorphous polymers and are much more thermoplastic than cellulose,
which is a highly crystalline polymer. Goring's data [S1 concerning the thermoplasticity
of wood components are typical examples of measurements of this kind. They show that,
in a dry state, while lignin and hemicellulose have thermal softening temperatures around
127-235°C and 167-2 17"C, respectively, for cellulose the thermal softening temperature
is around 231 -253°C. The variation in the value of the softening temperature found in
each wood component is caused chiefly by the difference in the method of isolating the
components from wood. On the other hand, the thermoplasticity of lignin, hen~icellulose,
and cellulose can be increased by wetting the samples. When the water content is increased
to 20%, lignin and hemicellulose show softening points around 72- 128°C and 54- 142"C,
respectively, much lower than for the dry state. By contrast, cellulose shows a decrease
of only about 6-9°C in a wet state. These differences indicate how water can act a s a
plasticizer and how the softening temperature depends on the two distinct states, crystalline
and amorphous, of the components.
On the other hand, an examination of thermal softening of wood as a whole reveals
that no thermal softening can be identified similar to that of individual components. Wood
does not show thermal softening until it is heated to a much higher temperature.This
suggests that interaction among the main wood components plays an important role. Gor-
ing [8] found an insignificant softening at temperatures over 200"C, and Chow [9] reported
that thermal softening of wood begins at 180°C and reaches the maximum rate of softening
at 380°C. Back et al. proved that the second-order transition point of hardboard made
from wood fiber is around 230-350°C. Baldwin and Goring [ 101 found that steamed wood
has a lower softening point, around 200°C. This may be because steaming reduces the
interactions among the main components of wood, causing amorphous portions of the
main components to soften independently of the cellulose. This implies that the thermo-
plasticity of wood is governed by cellulose. The thermal behavior of lignin and hemicel-
lulose is restricted by interactions due to secondary intermolecular bonding with cellulose.
The difference in the effect of sorbed water on thermal softening of each of the main
components implies that the crystallinity of cellulose contributes greatly to the degree of
thermoplasticity of wood.
As described above, although wood shows thermal softening, it occurs only at temperature
over 200°C. Thermal fluidity is not observed for wood. Thus, wood is known as a non-
plastic material. The fact that wood is insensitive to heat and shows no thermal fluidity
and plasticity can be attributed to the following reasons:
1. Cellulose is a crystalline polymer with about 50-70% crystallinity.
2. Lignin has a three-dimensional molecular structure with very high molecular
weight.
3. Chemical bonds are formed even between main components of wood, such as
in lignin-carbohydrate complexes (LCCs).
Cellulose is a linear high polymer having glucose residue as a repeating unit. Glucose
has three hydroxyl groups, which means that cellulose has the ability to form significant
Wood Plasticization 657
The preceding section suggests that it maybe difficult to change wood intoplastics;
however, the author has recently found that wood can be converted into a thermally flow-
able material by chemical modifications such asesterification,etherification, and some
other dcrivatizations 11 -7,l I , 121. Chemical modification does not necessarily require spe-
cial techniques, and therefore conventional and simplemethods can be used for this
purpose.
This phenomenon can be explained most simply in terms of the internal plasticization
of wood. That is, the internal plasticization of wood through chemical modification can
producea meaningful change in fundamentalpropertiesincluding thermoplasticity. The
degree of change in the plastic properties of wood is dependent on the molecular size of
the substituentgroupsintroduced, the degree of substitution, as well as the reaction
method.
A. Large-Substituent Modification
Accordingly, the introduction of large substituent groups into wood can result in a chem-
ically modified wood with high thermoplasticity. Since the change in thermoplasticity is
significant in this case, early studies dealt with thermoplasticization of wood reacted with
large groups. Plasticization of wood was first observed after esterification with a series of
higher fatty acids in a nonaqueous cellulose solvent medium (N,O,-dimethylformamidc)
[ I , 121. In this case, acylation of more than one-third of the available hydroxyl groups in
the wood is sufficient for the products to show thcrmofluidity. The thermofluidity was
proved with a thermomechanical analyzer, a flow tester, and scanning electron microscope
(SEM) observations.
Figure 1 shows the SEM of heat-treated birchwood meal and that of lauroylatcd
wood meal [ 121. While the untreated wood meal which was heat-treated at 270°C (Fig.
658 Shiraishi
FIGURE 1 (a) Untreated birch sawdust and (b) lauroylated sawdust with 93% ester content, both
heat-treated at 270°C. Observed under scanning electron microscope (SEM).
la) shows normal wood tissues such as parenchyma and wood fibers, the corresponding
lauroylated wood meal (Fig. lb) reveals clear melting with the disappearance ofwood
tissue. In the latter case, about one-third of the hydroxyl groups in the wood have been
lauroylated. It has also been proved that this thermal flow of the modified wood is not
caused by chemical degradation during chemical modification or thermal deterioration of
the sample during molding under heat and pressure [ 1l]. Films molded by heat and pres-
sure were crushed and reexamined by a thermomechanical analyzer to show that their
thermodiagrams were not different from those of the samples prior to hot-press molding.
The thermal softening behavior of samples first acylated and then completely saponified
were also found to be essentially the same as that of untreated wood meal.
It was thus shown that the acylated wood, obtained underspecial reaction conditions
by using a nonaqueous cellulose solvent as a reaction medium, showed thermofluidity.
Subsequently, studies were continued to determine whether wood can be plasticized
in the same manner by more general methods of higher-aliphatic-acid acylation.
Preparations of cellulose esters of a series of higher aliphatic acids are generally
carried out by reaction with a trifluoroacetic anhydride-fatty acid mixture (TFAAmethod)
or a fatty acid chloride-pyridine system (acid chloride method). We performed the former
reaction under conditions of temperature 30-50°C and reaction time 0.5-24 h; and the
latter under conditions of temperature lOO"C, reaction time 2-8 h with DMF as a solvent.
Both methods yield acylated woods with thermofluidity.
Typicalthermograms of the acylated woodmeasuredwith the thermomechanical
analyzer are shown in Fig. 2 [13,14]. The figure compares the thermoplasticity of the
Wood Plasticization 659
I I I 1
0
0
0
lauroylaled wood 0
I I I
0 100 200 300 400
T ('C)
FIGURE 2 Thermomechanicalcurvesfor (a) untreated, (b) acetylated. (c) propionylated, and (d)
lauroylated sawdust.
lauroylated and, as examples of lower fatty acid esters, the acetylated and propionylated
wood meal, with that of untreated wood meal. Acylation in all the above cases was carried
out by the TFAA method. Measurements were carried out by following the collapse of a
column of powder sample under a constant load of 3 kgfkm' in a heated glass capillary
tube in the temperature range 20-350°C at a programmed heating rate of l"C/min. The
untreated wood meal shows no thermal softening until it is heated to a relatively high
temperature (around 200°C) and seems to have a softening point around 270°C. Measure-
ments covering carbonization temperature ranges prove that the untreated wood sample as
a whole does not show any thermal flow.
On the other hand, the lauroylated wood meal shows a sharp drop dueto the complete
flowof the sample at 195°C. Itis alsorecognizable in the figure that acetylated and
propionylated woods, both prepared by the TFAA method, can have complete flow state.
At this stage, it should be clarified whether the thermal flow is due to the complete
melting of all the components of the acylated wood or of only some of them, with others
remaining in a solid phase and flowing together with the former. To examine this point as
well as to make other observations, films were molded from the acylated wood meals
under heat and pressure. An example of the results is shown in Fig. 3. Conditions for the
hot-pressing in this case were 140"C, 150 kgfkm', and 2 min. Transparent films, as shown
in the figure, were obtained, which did not show any solid residue when observed under
an optical microscope. This suggests that the flow behavior observed in the thermome-
chanical analyzer is caused by the thermal flow of all the main components of the chem-
ically modified wood. The complete melting of all the components of lauroylated wood
was also confirmed by measurement with a flow tester. It was found that lauroylated wood
could be easily extruded from the flow tester under a rather low pressure (10 kgfkm')
660 Shiraishi
from a nozzle with a diameter of 0.5 mm, when heated to about 150-200°C. If any one
of the major acylated woodcomponents could not reacha flow state, then it was impossible
to extrude the whole material through the nozzle. The conclusion that the thermal flow of
the chemically modified wood accompanies the flow of all of its main components should
be reasonable when the IPN structure of wood components described above is considered.
In addition, the temperature actually appliedin the molding mentionedabove is much
lower than the flow temperature obtained with a thermomechanical analyzer for the same
acylated wood shown in Fig. 2. In connection with this, it was proved experimentally that
this flow temperature decreased with the increasing load applied. In general, amorphous
polymers, such as higher-fatty-acid esterified wood, incontrast to the crystalline polymers,
do not have a definable melting point but do have an apparent melting point. This point
varies depending on the conditions of measurement, such as the load pressure or the rate
of heating.
If different substituent groups are introduced into wood, the apparent melting point
(flow temperature) will be different. Even if the same groups are substituted, the apparent
melting points of the products may be different, depending on the method and conditions
of the reactions.
In this regard, and also as another example of plasticization of wood by introducing
large substituent groups into wood, results of benzylation are shown in Fig. 4 W]. This
figure shows how the extent of benzylation affects the thermoplasticity of the etherified
wood. Only sample 4, which has an ether content of 43%, gives a transparent film. Other
filmsfromtheetherifiedwoodof lower substitution comparedwiththat of sample 4
contain solid granules. These results agree wellwiththe other experimental data [l61
obtained with the thermomechanical analyzer. Benzylated wood with ether content less
than 40% does not show complete flow, but that with ether content within the 40-50%
rangerevealsanapparentmeltingpointofabout 300°C. Further increasein the ether
content results in a steady decrease in the melting temperature, which finally reaches about
200°C with an ether content of more than 60%.
FIGURE 4 Benzylated wood films. Conditions for benzylation were as follows:
40%Sodium
hydroxide
aq. Benzylchloride Reaction
time"
Sample (ml/g wood) (ml/g wood) (h)
1 3.5 3.6
2 3.5 3.6
3 3.5 3.6
4 7.0 7.2
aeactlon temperature 100°C.
B. Small-SubstituentModification
Larger substituents are not easily introduced using practical chemical modification meth-
ods. Therefore, as the next stage of this study, the author has been examining the plasti-
cization of wood by introducing smaller groups within wood [ 13,14,17].
One of the most practical chemical modifications of wood is acetylation. The ther-
moplasticity of acetylated wood was found to vary with the acetylation method adopted.
The plasticity of acetylated wood preparedby the TFAA method is high and shows a clear
melting phenomenon at about 300-320°C even under a low pressure of 3 kgf/cm2, as
shown in Fig. 2. However, thisis an exceptional case, as acetylated wood samples prepared
662 Shiraishi
by other conventional procedures do not show thermal flow. For example, acetylated wood
samples prepared by an acetic anhydride-acetic acid in the presence of catalyst (sulfuric
acid or perchloric acid) did not show thermal flow under the above pressure.
Techniques for converting thermally nonmeltable acetylated wood prepared by the
conventional method into plastic material have been investigated.The first method in-
volves partial saponification subsequent to acetylation, which enhances the plasticity of
cellulose acetate within wood. Figure 5 shows the thermodiagram of peracetylated wood
and that of partially saponified wood. Although the peracetylated wood does not show
thermofluidity, complete thermal flow can be seen for the sample after partial saponifica-
tion. Considering that cellulose triacetate is a polymer that can be partially crystallized
and has an apparent melting temperature as high as 300°C, and that cellulose acetate with
high thermoplasticity is a diacetate with a degree of substitution of around 2.4, we can
understand the effect of saponification as found above.
The second method is mixed esterification with other acyl groups such as butyryl or
propionyl groups. As shown in Fig. 6 , thermofluidity is conferred by this kind of mixed
acetyl-propionylation. Mixed esterification is known to reduce the crystallinity and en-
hance the plasticity of cellulose as compared to cellulose triacetate.
The third method for obtaining thermofluidity in acetylated wood isby replacing
acetic acid by trifluoroacetic acid (TFA) in the pretreatment step of acetylation [ 1S]. This
1.0
0 100 200 300
Temperature ( O C )
Untreated
0
BBaDpo
A.A. :
0 :
1:
2:
3:
L
0 100 200 300
method is considered as a simulation of the TFAA method. which can be used to prepare
thermaily flowable acetylated wood asdescribed previously. Figure 7 shows that this
method of acetylation can also give thermofluidity to acetylatedwood.Whatis more
interesting is that the acetylated wood obtained shows an apparent melting temperature of
about 210°C. This temperature is almost 90°C lower than the apparent melting temperature
of ordinal acctylated wood or cellulose triacetate.
Actually, the apparent melting temperature of cellulose triacetate is about 300°C (Fig.
8). The figure shows the differential thermonlechanical diagrams of a series of fractionated
664 Shiraishi
(3-
h Ace.arhyd(O.26).TFA(L.6). 50°C
0.. 2 -
0..4 -
.6-
0.
0.8-
1.10-
0 200 300
FIGURE7 Effect of pretreatnlent time on thermal softening and flow behavior of acetylated wood.
Pretreatment: TFA (4.6 lnllg wood)-acetic anhydride (0.28 mllg wood). S O T , 2-6 h; acetylation:
50°C. 3 h.
cellulose acetates. The apparent melting temperature is shown as the peak at the higher
temperature. Even the cellulose acetate with very low molecular weight (0.61 X 10') flows
around 300°C. I n this connection, the findings of the previous figure that the acetylated
wood showed apparent melting temperature of about 210°C are quite characteristic. The
compatibility and/or the mutual plasticization among the acetylated wood components can
be considered to account for the shift of the flow temperature.
Then, plasticization of celluloseacetate with an interaction with acetylated lignin
could bc postulated. In order to confirm this. variations i n the thermofluidity of acetylated
wood were investigated in relation to a stepwise extraction of acetylated lignin from the
sample. The results are shown in Fig. 9. The curve for the acetylated wood before methanol
extractionshowsa flow at almost 200°C. Thecurvefor the sampledelignitied by the
peracetic acid method behaves like cellulose triacetate. The thern~omechanical diagrams
for the acetylated wood partially delignitied by methanol extraction appear between the
above-mentioned two curves. The methanol extracts were characterized by IR and NMR,
and were found to be acetylated lignin contaminated with a small amount of acetylated
hemicellulose.
Wood Plasticization 665
Cellulose triacetate
k
666 Shiraishi
0.4 k
0.6
tI o : Not ext. withMeOH
0 : 12 h MeOH ext
IA : 24 h MeOH ext.
I A : 48 h MeOH ext.
0 . 8 t o : 7 2 h MeOHext.
x : Delignification
From these results, it can be concluded that acetylated wood that was prepared aftcr
TFA pretreatment flows at a rather low temperature. and this is caused by the plasticization
of cellulose acetate by acetylated lignin within the acetylated wood cell wall.
The splitting of the benzylaryl ether and, less frequently, the existing ester bond in
the lignin macromolecule by the action of trifluoroacetic acid could play a large role in
making the acetylated wood prepared after TFA pretreatment flowable at a n unexpectedly
low temperaturc. I t can also be said that the same effect makes acetylated wood prepared
by the TFAA method thermally flowable. I n the latter case, however, acetylated and par-
tially cleaved lignin was removed during the purification stage, i.e., during washing with
methanol. Thus, acetylated wood prepared by the TFAA method cannot flowat around
200°C but reveals an apparent melting temperature of about 300°C as mentioned above.
These findings, in addition to the above-mentioned finding that acetylated wood prepared
by the traditional method does not show thertnofluidity, suggest that the lignin inherently
suppresses thermal f o w i n acetylated wood if i t is not partly split.
Wood Plasticization 667
0.2
0.4
0.6
0.8
1.0
T( O c 1
FIGURE 10 Thermofluidity (thermoplasticity) of the acetylated woodpreparedafter explosion
pretreatment at 179°C for I O min. Acetylation was conducted at 50°C for 3 or 6 h in the presence
of ;I small amount of P-naphthyl ethyl ether (0.1 m o l per mole of phenyl propane unit estimated to
exist in exploded wood n x A ) .
668 Shiraishi
i
l
FIGURE 11 Appearance of moldings from acetylated wood (wood pulp), acetylated wood-poly-
methyl methacrylate (PMMA) [ 1:1 (wt)], and acetylated wood-PMMA-dimethyl phthalate (DMP)
[5:3:2 (wt)] blends.
In the upper part of Fig. 12, untreated wood meals are shown on the left, allylated
wood meals in the middle, and carboxymethylated wood meals on the right. In the lower
part thecorresponding molded materials made by hot-pressing are shown in the same
order. It can be seen that none of these could be molded into a transparent film.
Examinations were carried out to see if these allylated or carboxymethylated wood
meals could be converted into thermally flowable material by mixing with plasticizers,
synthetic polymers, or both.
Figure 13 shows the results for allylated wood. The molded allylated wood (left), a
film molded after mixing 25% with dimethylphthalate (DMP) (middle), and a film molded
after mixing 100% with polymethylmethacrylate (PMMA) and 25% with DMP (right) are
shown. The spots found in the film mixed only with DMP are bubbles formed during hot-
pressing. The film on the right is transparent, showing the complete thermofluidity of the
sample. These results also indicate that conventional allylation does not cause sufficient
internal plasticization and hence cannot confer thermofluidity on wood. The addition of
adequate external plasticizer can confer thermofluidity on the allylated wood.
Similar results were obtained for carboxymethylated wood (CM wood) (Fig. 14).
CM wood powder is shown on the left, and after molding in the middle. A transparent
film could not be obtained in this case. CM wood was converted to a thermally flowable
material, however, after mixing with resorcinol (right).
In other experiments it was found that hydroxyethylated wood meal can be molded
into a transparent film after mixing and kneading with a small quantity of water.
In all of these examples, it should be pointed out that the appearance of chemically
modified wood meal is not very altered from that of untreated wood meal, as can be seen
in Figs. 12-14.
Wood Plasticization 669
FIGURE 12 Appearance of etherified wood meal and moldings. Upper: untreated wood (left),
allylated wood (center), and carboxymethylated wood (right). Lower: moldings from untreated wood
(left), allylated wood (center), and carboxymethylated wood (right).
I
i
FIGURE 13 Plasticization of allylated wood by blending with DMP and PMMA-DMP. Appear-
ance of moldings of allylated wood (left), allylated wood-DMP blend (center), and allylated wood-
PMMA-DMP blend.
670 Shlraishi
From the above observations andfindings, it can be said that, in principle, regardless
of the molecular sizeof the substituent group introduced, wood can be given thermoplastic
properties by internal plasticization, that is, chemical modification, supplemented, if nec-
essary, by external plasticization [17].
The sixth method for conferring thermofluidity on acetylated wood is thatof grafting
vinyl monomerssubsequenttothe acetylation [13,14]. Inthiscasetheresultant vinyl
polymer acts as a plasticizer within wood. Figure 15 shows that acetylated wood prepared
by the perchloric acid catalyst method does not reveal thermal flow, whereas subsequent
grafting with styrene changes the acetylated wood to a thermally flowable material.
This grafting method is similar in principal to the method supplemented by external
plasticization with mixing with synthetic polymer. In that case it was not easy to find
appropriatesyntheticpolymers as suitableplasticizersforacetylated wood. A suitable
polymer-plasticizer should be one that is compatible with the polymer component of the
acetylated wood at the molecular level. Compatibility between polymers, however, rarely
occurs. To overcomethis difficulty, graftingtechniquescanbe used. Compatibility of
acetylated wood components and synthetic polymers is attainable by the grafting tech-
nique. Even with a combination of acetylated wood and synthetic polymers, with which
a homogeneous transparent film cannot be effected merely by blending, the use of grafting
techniques can result in a transparent molded film. The method of grafting is also impor-
tant. It is desirable to select a method by which grafting or polymerization of monomer
compounds occurs homogeneously within acetylated wood to produce products with high
compatibility. A representative experimental result (Fig. 16) shows how the grafting tech-
nique is superior to blending in yielding high compatibility among the components of
Wood Plasticization 671
O~
a 0.5
1 .o
acetylated wood and synthetic polymers. The figure illustrates the temperature dependence
of loss tangent obtained by a viscoelastic measurement for composites prepared by blend-
ing acetylated wood [acetylated radiata pine refiner ground pulp (RGP)] with PMMA with
a kneading technique or by grafting MMA onto the acetylated wood. This figure shows
that grafting results in forced compatibility of acetylated wood with PMMA. Although the
blended material of acetylated wood and PMMA shows two T, peaks ascribable to the
two constituting components, the corresponding grafted material reveals only one inter-
mediate T, peak. This shows that graft products of PMMA onto acetylated wood com-
672 Shiraishi
ponents can be partially produced which can act as colnpatibilizers and which enable the
total mixing and interaction of the acetylated wood and PMMA to be enhanced.
On the other hand, some investigations have been advancing from the point of view
of not only conferring but also enhancing the thermofluidity of chemically modified wood.
Among them, the method of using TFA as a reagent for the pretreatment, to split selectively
the parts of bonds combining lignin units [IS], a s well as methods chlorinating lignin's
aromatic ring parts of chemically modified wood [ 19,201, are very effective and specific.
The former method has already been explained i n detail as the third method of offering
thermoplasticity to acetylated wood. It should be noted here that since a significant shift
of the apparent melting temperature toward the lower temperature can be effected as
mentioned previously, the former method is also useful as a method for enhancing the
thermofluidity of modified wood.
The latter method is concerned with a postchlorination of the chemically modified
wood. Morita, Sakata, and co-workers have been investigating the effect of postchlorina-
tion of cyanoethylated wood on the thermofluidity of modified wood [ 19,201. The chlo-
rination was conducted by soaking the modified wood in a delute aqueous chlorinesolution
(0.1-0.2%), at a low temperature, for a short time (O'C, 5 min) under stirring, followed
by washing with water and drying. Under these mild conditions of chlorination, chlorine
is known to react selectively with the aromatic ring of lignin. The chlorination of lignin
can be considered to cause ( I ) chlorine substitution reactions at the S - and 6-positions of
lignin aromatic rings; (2) chlorination substitutions at the I-position of the aromatic ring,
Wood Plasticization 673
eliminating the propane side chain; (3) hydrolysis of ether bonds (demethylation of me-
thoxy groups and hydrolysis of phenyl ether bonds); and (4) oxidation of the aromatic
rings. Among these, the former two reactions are primary reactions, while the latter two
are side reactions.
As representative results of these studies, it has been shown that the flow temperature
ofthe cyanothylated wood can be shifted downward by 100- 120°C (Fig. 17). At this
point, decrystallization of the samplcs is known to proceed without degradation of the
cyanoethyl cellulose fraction [19,20]. The lowering oftheflow temperature of cyano-
ethylated woodis thought to be attributable to a loosening of the lignin structure by
chlorine substitution, action ofthe chlorinated lignin as an external plasticizer for cy-
anoethyl cellulosc within the modified wood (mutual interaction among modifiedwood
components), a s well as decrystallization produced by the chlorination [X)].Lowering the
flow temperature is considered not to be caused by the degradation or depolymerization
of the components of the chemically modified wood. This is concluded from the obser-
vations that moderate chlorination, as described above, results in a tremendous drop in
the flow temperature of the chemically modified wood, as well as the fact that the viscosity
change of the aqueous solution of cyanoethyl cellulose before and after the chlorine treat-
ment is very low (7%) 1201.
Similar phenomena have been found with other chemically modified woods. For
example, chlorination causes the lowering of the flow temperature of benzylated wood,
and results in thermally flowable hydroxyethylated wood that does not show thermofluidity
before the chlorination.
I I l
.-
E
d
l00 200
Temperature ( "C l
FIGURE 1 7 Thermal softening and flow curves for CE-wood treated with chlorine.
The temperature dependences of dynamic modulus E' and loss tangent tan S for the
caprylated wood film are shown in Fig. 19 [22]. For comparison, the same relation was
obtained for an untreated Beihi cypress specimen (fiber direction) and is shown as a dashed
line in the figure. The dynamic modulus E' for the untreated wood shows a small decrease
within the sameorder of magnitude when the temperature increases from - 180°C to
(U
E
V
\
m 400
-Y
" I 200
C
I I I I
20 40 60 80
E: (%)
FIGURE 1 8 Stress (n)-strain ( E ) curve for caprylated wood film (c) and wood specimens (Hinoki)
with two different edge cuttings (a and b).
Wood Plasticization 675
E'
-----------------------------""~~""""""""""""~"""
"""
1010- -1
rt
2
S 109- -0.1
C
z
E
iJl
-
"-" "_" - 0.01
108
107,
-200 -100 0 100 200
T (C)
FIGURE 19 Temperaturedependenceof E' and tan 6 forcaprylated wood film and wood
specimen.
250°C. Although the dynamic modulus E' for caprylated wood film at - 170°C is very
close to that for the untreated wood, it decreases with temperature and, especially, is
dramatically lowered between -90 and -40°C and between 30 to 100°C. As a result, the
E' value becomes three orders of magnitude lower than that of the untreated wood at
140°C. With respect to tan S , three relaxation processes are detectable within the experi-
mental conditions. They are labeled as a, p, and y process in order of decreasing tem-
perature. These processes were identified on the basis of the information obtained from a
series of cellulose acetates. The a process is assigned to be due to a micro-Brownian
motion of the main chain of the caprylated wood (the process corresponding to the glass-
rubber transition); the p process to the motion of the capryl side chain introduced, and
the y process to the motion of a part of the side chain (the motion initiated by a minimum
of three methylene groups in addition to the oxycarbonyl group of the side chain). Figure
19 shows that a large decrease in E' is caused by the motion initiated by the whole acyl
side chain ( p process) and that the glass-rubber transition temperature of the caprylated
wood is higher than room temperature. These results are interesting in relation to the results
shown in Fig. 18. That is, the observation in Fig. 18 that the caprylated wood film behaves
as though it is an elastomer at the measuring temperature (room temperature) can be
directly explained by the effect of the micro-Brownian motion of the acyl side chain, and
not by micro-Brownian motion of the main chain.
As another example of the mechanical properties of a modified wood with large
substituent groups, stress-strain curves for a benzylated wood film measured at various
temperatures are shown in Fig. 20 [15]. The stress-strain curve varies widely with the
temperature. It clearly shows a yield point when measured at 55 and 81°C while the film
reveals brittleness at room temperature. The breaking strength of the film decreases with
increasing temperature, and the breaking elongation increases rapidly with temperature up
to around 1 17°C and then decreases beyond it.
676 Shiraishi
E ob)
I n Fig. 2 I , the temperature dependence of dynamic modulus E' and loss tangent tan
6 for the benzylated wood film is shown [ 151. Compared with the corresponding results
with a Beihi cypress specimen (fiber direction), results with the benzylated wood specimen
are similar to those obtained with general amorphous synthetic polymers. Within its glassy-
state region, where the E' value is more than 10"' dynes/cm2, a secondary dispersion ( p
process), accompanied by a very slight decrease in E', is recognized around -9O"C, and
the tan 6 curve reveals a corresponding peak. This process is assigned to be due to the
motion of the benzyl side chain introduced. The effect of the motion of the benzyl side
chain on the decrease i n the E' value is considerably smaller compared to that caused by
the motion of the capryl side chain mentioncd above. On the other hand, a primary dis-
persion ( a process) appears around 118°C in this benzylated wood. The findings in Fig.
21 explain the observations shown in Fig. 20;i.e., benzylated woodfilm behaves asa
brittle and glassy polymer film at room temperature, and the breaking elongation increases
with the measuring temperature up to around 118°C (the glass transition temperature) and
then decreases beyond it.
In Table I , mechanical properties of the films from certain benzylated woods are
compared with those of several common synthetic polymers [ 161. It can be seen that there
are no fundamental differences between the two.
The physical properties of benzylated wood can be altered or, more desirably, en-
hanced by blending with synthetic polymers and/or low-molecular-weight plasticizers.
Based on similarity in chemical structure between the benzylated wood and polysty-
rene, good compatibility can be expected between them. Thus, the related blending has
been studied [23]. In Fig. 22, apparent melting points (T,) for the benzylated wood, the
polystyrene, their blends in several ratios. as well as a typical chlorinated benzylated wood
(Cl-BzW) are shown. From this figure, it is apparent that the thermoplasticity of benzylated
Wood Plasticization 677
llT 1
lO$O
71-3 l I l
- 100 0 100 200
FIGURE 21 Temperaturedependence of E' and tan S for benzylatcd wood film and wood
specimen.
TABLE 1 Comparison of Mechanical Properties of Benzylated Wood Films (C-l to C-7) with
Those of Common Synthetic Polymers
, , I . , . , .
FIGURE 22 Apparent melting temperature (TJ for benzylated wood (BzW), polystyrene (PS), and
their blends with various ratios. Cl-BzW should be referred to the text.
wood can be markedly enhanced. Especially, the samples obtained by blending more than
50% of polystyrene are known to have T , values almost equal to that of the polystyrene.
In this connection, Morita and Sakata [l91 reported that the apparent melting temperature
of the chemically modified wood can be considerably lowered by a weak chlorination
aftertreatment, which gives the Cl-BzW mentioned above.
Dynamic viscoelastic properties of the benzylated wood, the polystyrene, and a blend
of the two in equal weight ratio were measured, and the obtained temperature dependencies
of logarithmic decrement (aT)are shown in Fig. 23. In general, the comparison of this
kind of viscoelastic relaxation curves is suitable for judging qualitatively the compatibility
of a polymer blend. That is, in the case of the compatible polymer blend, the glass tran-
sition temperature (T,) appears as a single peak in a position proportional to the existing
amounts of component polymers, whereas in the case of an incompatible polymer blend,
the individual phase domains retain the glass transitions of their respective parent homo-
polymers. These viscoelastic relaxation data were correlated with electron microscopic
observation results and the following have been pointed out. In the case of the compatible
blend system, the component polymers were found to exist as domain mixtures, whose
grain diameter was less than 15 nm (150 W),
whereas in the case of the incompatible
blends, they were often existing as individual phase domains whose diameters were larger
than 100 nm. I n the case of the semicompatible blend, in which even though a number
of T, peaks can be observed, the transitions are broadened and their temperatures become
closer together, and the individual phase domains have grain diameters in the range 20-
100 nm.
Consequently, the experimental results shown in Fig. 23 indicate that relatively sig-
nificant molecular mixing takes place in the benzylated woodpolystyrene blend. When
the transparency of the molded sheets from this benzylated woodpolystyrene blend is also
taken into consideration, the two- or multicomponent polymer system is concluded to be
a compatible polymer blend, and the individual phase domains have grain diameters of
less than15 nm.
Figure 24 shows the results of tensile tests obtained for films prepared from ben-
zylated wood, polystyrene, and their blended mixtures. From this figure, it can be seen
that the tensile properties obtained for these blending composites are not very good. That
is, even though the benzylated woodpolystyrene blends are miscible, the interfacial bond-
ing of the related individual phase domains is not enough for the corresponding molded
sheets to have desirable mechanical properties.
This result is not necessarily unusual, and in order to overcome this situation, com-
patibilizers can be used in various ways. In this case, a styrene-maleic anhydride copoly-
mer (SMA: Arc0 Chemical, Inc., Dylark 232; maleic anhydride content 8 wt%) was tested
in order to make clear whether it can work as a compatibilizer ornot when certain amounts
of it are kneaded with the benzylated wood and polystyrene at a temperature as high as
160°C.
The result is shown in Figs. 25 and 26, in which the amount of benzylated wood
was fixed to 50 wt% and the mixing ratio of polystyrene to SMA, the total amount of
which was 50 wt%, was changed. From the figures, it is seen that both the tensile strength
and the breaking elongation take the maximum values when SMA is added at 5 wt% to
the total weight of the blended composites. This result shows clearly that benzylated w o o d
SMA graft copolymers, which have been produced by kneading [blending at high tem-
perature (16O"C)], can act as a compatibilizer. In other words, due to the localized existence
of thus-produced benzylated wood components/SMA graft copolymers at the interfaces
between the benzylated wood and the polystyrene phase domains, as shown in Fig. 27,
their interfacial bonding is considered to be improved, giving materials with excellent
physical or mechanical properties.
In Fig. 25, when the amounts of SMA added to polystyrene exceed 1096, the tensile
strength of the composite films obtained shows a monotonic decrease. This is considered
15
14
13
12
11
515 1010
BzWPS (wtlwt)
FIGURE 24 Tensile properties for benzylated wood, polystyrene, andtheir blending mixtures. 0 ,
tensile strength (um1.,J;
A, breaking elongation m, Young's modulus ( E ) .
680 Shiraishi
L
, . , . , . , .
1010 812 614 416 218 0110
PSISMA (wtlwt)
FIGURE 26 Polystyrene-to-SMA ratio dependency of the breaking elongation of the molded films
prepared from their blends with benzylated wood (the benzylated wood content is fixed to 50 wt%).
SMA Chain P
FIGURE 27 Schematic diagram showing the localized existence of benzylated wood components/
SMA graft copolymer at the interface between the benzylated wood and the polystyrene phase.
Wood Plasticization 681
to be caused not by the decline of the mechanical properties of products with increasing
SMA content, but from the formation of crosslinking within the benzylated wood com-
ponents phase, the amounts of which increase with the amount of SMA added. The for-
mation of crosslinking during kneading or blending results in products with less moldable
thermomechanical properties.
Films can sometimes be molded directly even from modified wood with small-mo-
lecular-sized substituents. For example, acetylated wood prepared by the TFAA method
and the TFA pretreatment method, as well as acetyl-butyrylated wood, can be molded into
films.
Table 2 is a n example of the experimental data. Even though the films can be molded
from wood substituted by small groups, the films show brittleness. In order to improve
the physical properties of these moldings and to enhance the thermofluidity and moldability
of modified woods having small-molecular-sized substituent groups, these modified woods
have also been blended and plasticized with plasticizers, synthetic polymers, or both. Graft-
copolymerization with synthetic polymers has also been tried, and the resulting physical
properties have been measured.
Mechanical properties of films prepared after the external plasticization of allylated
as well as carboxymethylated wood meals with synthetic polymers andor low-molecular-
weight plasticizers [ 171 are shown in Table 3. Corresponding data for each of the synthetic
polymers used as the external plasticizer and for a benzylated wood film are also included
for comparison. It is known that the films from the externally plasticized allylated as well
as carboxymethylated wood reveal mechanical properties comparable to those of the cor-
responding synthetic polymers. Reinforcement because of the presence of modified wood
can be recognized.
It can be concluded that the physical properties of chemically modified wood alloyed
with synthetic polymers with or without other plasticizers are dependent on the species of
synthetic polymers and the composition of the alloys. Generally, external plasticization by
graft-copolymerization results in greater enhancement of thermofluidity (moldability) as
well as better physical properties of the moldings than those provided by mere mixing
(blending). In the case of graft-copolymerization, the method of grafting also plays an
important role. Wood-synthetic polymer composites with higher thermofluidity and better
physical properties can be obtained when grafting techniques resulting in a uniform dis-
tribution of polymers within the wood cell wall are adopted (seeFigs. 16 and 28). As
shown in Fig. 28, however, even though the graft-polymerization technique is adopted,
two peaks assignable to the respective primary dispersion of acetylated wood and that of
polymethyl methacrylate (PMMA) can be found in the tan S-versus-temperature curve of
a graft product prepared by use ofan azobisisobutyronitrile initiator (AIBN catalyst
method). This result is quite different from that of Fig. 16, where a curve for a grafted
product prepared by a vinyl polymerization without initiator in the presence of water is
included. In this case, only one primary dispersion peak appears, suggesting a high com-
patibility at the molecular level among the components of the acetylated wood and PMMA.
As already mentioned, when individual phase domains composing the blended composite
have grain diameters less than 15 nm, the composite can exhibit one principal glass tran-
sition. Itis also known that the tensile strength and the elongation at the break of the
molded films become larger when higher interfacial bondings among the individual phase
domains are developed. Further addition of adequate plasticizers, such as dimethyl phthal-
ate (DMP), di-2-ethylhexyl phthalate (DOP), tricresyl phosphate (TCP), etc., in suitable
quantities, to chemically modified wood-synthetic polymer composites can substantially
enhance the thermofluidity of the polymer alloys. More homogeneous films are thus ob-
tainable by hot-pressing.
From the standpoint of preparing molding materials with excellent properties from
chemically modified wood, it is possible to make use of chemicals that can act as plasti-
cizers for the modified wood as well as that can react with the components of the modified
wood during kneading and hot-pressing. One example of this is the preparation of films
from blends of acetylated wood-resorcinol paste with various synthetic polymers (poly-
vinyl acetate, ethyl acrylate, acrylonitril-butadiene rubber, etc.), to which a definite amount
of formalin was added.
The mixture was reacted under heat and kneading, and then molded into films by
hot-pressing. Three-dimensionalcuringoccursduring molding. Physical properties are
shown in Fig. 29. In this case, formalin was added in a quantity of 0.5 mol of formaldehyde
for each mole of resorcinol. Although the mechanical properties of the films are largely
dependent on the kind of synthetic polymers added, tensile strengths up to 770 kgf/cm2
could be obtained even within the small number of experimental trials.
Another example of the use of chemicals as plasticizers as well as reacting agents
for chemically modified wood is that of preparing three-dimensionally cured plasticlike
Wood Plasticization 683
PMMA+AcP
X.method
N.method
c F.method
t!
A.method
Blend ( DMP)
Blend
FIGURE 28 Temperature dependence of ay(= tan 6) for various PMMA-acetylated wood graft
products prepared by different procedures and that for PMMA-acetylated wood blends. X.method,
xanthate method; N.method, noncatalyzed grafting method; Emethod, redox method with Fenton's
reagent; A.method, catalyzed method with AIBN; Blend (DMP), blending method using kneader in
the presence of DMP; Blend, blending method using kneader.
wooden board, developed by Matsuda and Ueda [24]. The special feature of this method
is the combined use of carboxyl group-bearing esterified wood, such as maleoylate wood
(maleic acid half-esterified wood), phthaloylated wood (half-ester), and bisphenol A di-
glycidyl ether. In this case, a crosslinking reaction accompanied by plasticization occurs
during the hot-press molding stage, resulting in new types of cured wood. Various molded
boards having plasticlike appearance as shown in Fig. 30 were obtained. The color of the
board depends on the species of the esterified wood. Most common are red-brown, yellow-
brown, or black-brown colored boards. When the wood content of the boards falls to 60-
70%, materials with high water resistance can be obtained. The physical properties, such
as strength, elongation, etc., of the boards are superior to those of conventional boards
(fiberboard, particle board, etc.). Cured board from meleoylated wood is superior in blend-
ing strength, while boards from phthaloylated wood have excellent compressive strength
684 Shiraishi
l
aoo
Ac -Res - PVAc
7: 3 1 3
(0.5rnol HCHO)
56oo
-
\
U
0
x
m
m
E
I
v1
k-Res-NBRI
7:3:3
(0.5mol HCHO 1
L
0 20 40 60
Strain(%1
FIGURE 29 Stress-strain curves for the acetylated wood-resorcinol-formaldehyde-synthetic
polymer blended and cured films.
e
e e
e
m
e m e
150 m
e e
~~
hesive strength found in Fig. 32 is more than 150 kgf/cm2, which is 1.5 times higher than
the JIS specification demands.
Chemically modified wood has been found to liquefy and/or dissolve in various neutral
aqueous solvents, organic solvents, or organic solutions, depending on the characteristics
of the modified wood [2-7,28-341. So far, three methods have been found for this
purpose.
The first trial of the liquefaction and/or dissolution of chemically modified wood
was accomplished by treating chemically modified wood in the presence of organic or
aqueous solvents, or solutions at various temperatures in the range 80-290°C [34]. One
example used wood samples esterified with aseries of aliphatic acids which could be
liquefied in benzyl ether, styrene oxide, phenol, resorcinol, benzaldehyde, aqueous phenol,
Wood Plasticization 687
FIGURE 33 Dissolution of fatty acid-acylated wood prepared by TFFA process in benzyl ether
(C = number of carbon atoms in the acyl group). Dissolution conditions: 250°C and 20-40 min.
688 Shiraishi
by only 9.25%. However, once chlorinated, it can dissolve almost completely in the same
solvent at room temperature.
There are many potential applications for the liquefaction and/or dissolution of chemically
modified wood.Examples include the fractionation of modified wood components
[34,36,37], the preparation of solvent-sensitive and/or reaction-sensitive wood-based ad-
hesives 123 1,35331, the preparation of resinified wood-based moldings such as the foam
type 131 1, and the preparation of wood-based fibers and their conversion to carbon fibers
(391.
To fractionate modified wood components, the dissolution-precipitation technique
has been successfully used [34,36.37].
Attempts to prepare wood-based adhesive as well a s their curing based on the re-
action of modified wood and the reactive solvents are reported in the literature. In these
cases, phenols, bisphenols, and polyhydric alcohols have been used as reactive solvents
[31,35,38,40,41 I. Combined use of these reactive solvents with reactive agents, crosslink-
ing agents, and/or hardeners has given rise to phenol-formaldehyde resins (such as resol
resin), polyurethane resins, epoxy resins, etc., all of which contain meaningful amounts of
chemically modified wood. The chemically modified woods are not designed merely to
dissolve and disperse in the final resins, but to react chemically and bond to thc resins.
This can be achieved by liquefaction of the chemically modified wood into the reactive
solvent using solvolysis techniques. In the case of epoxy resins. it can also be achieved
by reacting various alcoholic hydroxyl groups remaining in the modified woods with
epichlorohydrin, resulting i n introduction of glycidyl groups. Crosslinking within and be-
tween degraded wood components, especially between degraded polysaccharide compo-
nents during the last stage of resinification, by reaction with crosslinking agcnts, can also
be used. In order to prepare wood-based resins with meaningful amounts of the wood
component, it is very important to liquefy or dissolve the chemically modified wood into
reactive solvents in high concentrations (more than 50% is preferable).
When hydrophilic chemically modified woods, such as carboxymethylated woods,
hydroxyethylated woods, or ethylated woods are used in wood-based adhesives, aqueous
resol resin adhesives that maintain their solution state during the preparation-that is,
from the period of the completion of the liquefaction or dissolution in phenol to the final
stage of prepolymer formation-are obtainable[35,38,40,41 I, Whcn a phenol solution
with a concentration more than 50% is obtained, the chemically modified wood powder
cannot bc completely immersed in phenol but can be only partly penetrated by the phenol
during the first stage of dissolution (see Fig. 34). However, when the heterogeneous mix-
ture is allowed to stand for about 30 min at 80°C (without stirring in the presence of
appropriate amounts of hydrochloric acid as catalyst), a homogeneous paste can be ob-
tained. Subsequent stirring of the paste for about I - 1 .S h enhances the liquefaction.
In this liquefaction process, a certain degree of phenolysis ofwood components.
especially that of lignin, takes place, which makes it easy to liquefy and dissolve them in
phenol. After neutralizing the paste with aqueous sodium hydroxide, il definite amount of
formalin and sodium hydroxide are added and resinified in accordance with the conven-
tional procedure for preparing the resol resin adhesives. The appearance of thc resin ob-
tained (Fig. 3 5 ) is similar to that of the corresponding commercial phenol resin adhesive.
Wood Plasticization 689
FIGURE 34 Early stage of liquefaction of hydroxyethylated wood meal into phenol. Photograph
was taken after 10 min of liquefaction at 80°C.
The wood-based resol resin adhesives have superior gluability and workability. The ad-
hesives can be used with fillers, thickeners, and fortifiers such as wheat flour, coconut
shell, walnut flour, and polymeric MD1 (4,4’-diphenyl methane diisocyanate). It has been
found that by using mildconditions for the first liquefaction and dissolution process, which
does not completely dissolve all of the wood meal, the addition of fillers and thickeners
into the adhesives become unnecessary. On the other hand, the addition of appropriate
fortifiers, especially crosslinking agents such as polymeric MDI, into the wood-based ad-
hesives enhances their dry-bond and waterproof gluabilities remarkably. In the cases of
utilizing hydrophobic chemically modifiedwoodssuch as acetylated wood, butyrylated
wood, etc., the liquefaction and dissolutionof these chemically modified woods into phenol
is possible, but the products tend to become solids when they are subsequently resinified.
In these cases, resinification should bedone within the kneader so that the products become
solid powders (Fig. 36, center). The products can be dissolved in suitable solvents such
as ethyl acetate, resulting in solutions as shown on the right side of Fig. 36, and can be
used as reaction-sensitive liquid adhesives.
For the preparation of wood-based polyurethane as well as epoxy resin adhesives,
the above-mentioned hydrophilic chemically modified woods, prepared by conventional
methods, are liquefied and dissolved in polyhydric alcohols or bisphenol A in a manner
similar to the dissolution in phenol [31]. Concentration of the modified wood is usually
more than 50%. Diluents such as ethanol or methanol are also often added to the disso-
lution system according to the requirements. After the liquefaction, the pastes are neu-
tralized and the diluents are distilled off. When the pastes are used in combination with
suitable polyisocyanatecompounds,theybecomewood-based polyurethane adhesives.
When the pastes are further reacted with epichlorohydrin, glycidyl etherified resins are
formed. These can be used with hardeners such as amines and acid anhydrides, and they
690 Shiraishi
become wood-based epoxy resin adhesives. Figure 37 shows epoxy resin adhesives pre-
pared under various conditions after liquefying allylated wood into the same weight of
bisphenolA. These are comparedwith a commercial epoxyresinadhesive. Figure 38
shows the fluidity of one of the epoxy resin adhesives prepared from allylated wood.
Generally, the wood-based epoxy resins tend to become very viscose or solid, depending
on the conditions of preparation, and require dilution or dissolution with solvents such as
ethyl acetate, acetone, etc. These resins make satisfactory adhesives, which can be used
in waterproof glues.
Molding materials such as foams or shaped moldings can be obtained from chemi-
cally modified wood solutions of polyhydricalcohols, phenols, and bisphenolsas described
above [31]. An example is shown in Fig.39. These can be prepared by adding anadequate
amount of water as a foaming agent and a polyisocyanate compound (polymeric MDI) as
a hardener to the 1,6-hexanediol solution of allylated wood, mixing well, and heating.
When heated at 100°C, foaming and resinification of the resins are initiated within 2 min
and are completed within several minutes. If promotors such as triethylamine are added,
rapid reactions occur even at room temperature and foams can be obtained within several
minutes. The foam shown in Fig. 39 has a very low apparent density (0.04 g/cm’). It also
Wood Plasticlzatlon 691
FIGURE 39 Appearance of polyurethane foams from allylated wood, prepared by liquefying and
dissolving in 1,6-hexanediol, adding a foaming agent (water) and reacting agent (polymeric MDI),
and resinifying.
Wood Plasticization 693
has substantial strength and a restoring force for compression deformation, as can be
presumed from Fig. 40. The apparent density of the foams is dependent on the amount of
foaming agent as well as the kind of reactive solvent (such as species of polyhydric
alcohols, phenols, bisphenols, anddiluents) used. Foams preparedfrom the modified wood
solution of bisphenol A using a similar procedure tend to have apparent densities around
0.1 g/cm3 and considerable strength.
In order to elucidate the role of the liquefied chemically modified wood within the
foams, comparative experiments of preparing the foams without the chemically modified
wood have also been conducted. It was found that foaming actually occurs during the
resinifying processbut, immediately after that, a contraction in volume of the foam occurs,
resulting in resin moldings with apparent densities around 0.2 g/cm3 with little foaming-
cell structure remaining. This result is understandable because the open-mold, one-shot
process used here is said to require rather high-molecular-weight polyhydric alcohols as
one raw material of the foam. This result also reveals that the liquefied chemically modified
wood plays a positive role in maintaining the shape of the foam during formation.
One other application of the phenomenon of liquefying modified wood is in the
formation of filaments or fibers produced by spinning methods. Tsujimoto [39] prepared
wood-based fibers from acetylated wood. Acetylated wood is first liquefied and dissolved
in phenol, and hexamethylenetetramine is added to the solution in various proportions.
This is followed by heating to 150°C to promote addition-condensation, resulting in a
resinified solution with high spinnability. From the spinning solution, filaments are spun
and hardened in a heating oven at a definite heating rate (maximum temperature 250°C).
An example of the filament is shown in Fig. 41. These filaments can be carbonized to
give carbon filaments. Carbonization is carried out in an electrically heated furnace at a
temperature of 900°C with a heating rate of 5.5"Umin. The strength of carbon filaments
is measured according to JISR7601, and tensile strengths up to 120 kgf/mm' has been
FIGURE 40 Deformation and restoring property of foam prepared from allylated wood.
694 Shiraishi
The liquefaction and dissolution of chemically modified woods have been reviewed so
far.Morerecently,untreatedwood has also beenfound to liquefyin several organic
solvents [3,5-7,33,42-551. For example, after treating at around 250°C for 15-180 min,
wood chips and wood meals were liquefied in phenols, bisphenols, alcohols (benzyl al-
cohol), polyhydric alcohols (1,6-hexanediol, l ,4-butanediol, and glycerin), and hydroxy-
ethers (methyl cellosolve, ethyl cellosolve, diethylene glycol, triethylene glycol, and poly-
ethylene glycol).
The liquefaction of untreated wood can also be achieved at a lower temperature of
150°C and at atmospheric pressure inthe presence of an acid catalyst, phenolsulfonic acid,
sulfuric acid, phosphoric acid, oxalic acid, and hydrochloric acid having been used [3,5-
7,33,42-551.
It is possible to obtain pastelike solutions with a high concentration of wood solute
of up to 70%. After liquefaction, the wood components were found to have degraded and
became reactive, which will be shown more detail in the section on the liquefaction mech-
anism for wood. The obtained wooden solute can be used to prepare adhesives and other
moldings, opening a practical new field for utilization of wood materials, the details of
which will also be explained later.
During the liquefaction of wood, especially in the presence of acid catalyst, recon-
densation of degraded wood components occurs, as has been also observed inthe explosion
and autohydrolysis process for wood. Because of the recondensation, it becomes very
Wood Plasticization 695
difficultto obtain a liquid with a large wood concentration, which is often considered
undesirable from the viewpoint of biomass utilization. On the other hand, starch is very
easy to liquefy even at a very small liquid ratio and catalyst concentration, when a catalyst
is even necessary. Based on this background, a combined liquefaction process for wood
and starch was proposed as a practical method for preparing large-biomass-content liquids
[45]. That is, a stepwise liquefaction procedure in which the wood could be preliquefied
alone at a relatively large liquid ratio, followed by the addition and liquefaction of the
starch, was proposed. By this procedure, a large-biomass-content liquid was prepared with
relatively small unliquefied residue [45].
In order to find an appropriate method for accurately determining the amounts of
unliquefied residues, the soluble properties of liquefied wood and starch were investigated
using a series of diluent solvents [47]. It was found that the solubility behavior of a
liquefied biomass in a certain solvent was a kind of fractionation of the degraded and
liquefied biomass components. In most cases, any single solvent could not dissolve all of
the liquefied components completely. Several binary solvent mixtures composed of sol-
vents considerably different in polarity were found to be good diluent solvents for liquefied
biomasses. These phenomena can be illustrated by consulting previous works on physi-
cochemical properties of binary solvent mixtures. Among several satisfactory binaries, the
binary of dioxane and water has been studied in detail and found to be widely suitable
for liquefied biomasses prepared in different liquefaction solvents. The range of dioxane/
water mixing ratio usable for the complete dilution of liquefied biomasses was wide
enough for practical usage. Especially, a binary with a dioxanelwater composition of 812
was recommended as a universal diluent for liquefied biomass [47].
Phosphoric acid and even oxalic acid were found to be usable as catalysts for the
liquefaction of wood [46,48-501. In the latter case, a small amount of hydrochloric acid
was used simultaneously. These uses of catalyst were evaluated in connection with the
flow properties and reactivities or curing properties of the liquefied wood. In this extension,
phenolated woodlphenollformaldehyde co-condensed resins were proposed [53]. Wood
was first liquefied in the presence of phenol by using an acid catalyst to produce a phen-
olated wood, and after the liquefaction, formalin was added to conduct a condensation
reaction forconverting the remaining nonreacted phenol into resin components. It was
found that this procedure can convert almost all the phenol remaining after liquefaction
into resins, and therefore significantly upgrades the practical value of the liquefaction
technique. Another advantage of this co-condensation isthat it can greatly improve the
thermofluidity of the phenolated wood resins and the mechanical properties of their molded
products. The flow temperatures and melt viscosities of the co-condensed resins were much
lower than those of the phenolated wood resins. That is, these two properties were more
or less similar to those of the conventional Novolac resin, resulting in excellent process-
ability. The flexural properties of the molded products made from the co-condensed resins,
although this point should be discussed in the next section, were much higher than those
of the phenolated wood and also somewhat superior to those of the conventional Novolac
resin [53].
of 140-240 atm, with synthesis gas CO/H2 1581. Treatment for 1 h at 300-350°C resulted
in a 40-60% yield of benzene solubles(oil) and 95-99% conversion of the starting
materials. This type of liquefaction can be more precisely called the oilification of ligno-
cellulosics.
The review presents recent progress on lignocellulosic liquefaction under milder
treatment conditions, that is, at temperatures of 80-150°C with an acidic catalyst. One
special group of chemically modified woods can be dissolved in cresols even at room
temperature, as shown previously.
and phenol is more complicated compared with that between polysaccharide and alcohols,
because of the multifunctionality of phenol. As a result, liquefaction products prepared in
the presence of phenol tends to convert to higher-molecular-weight substances with in-
crease in the reaction time.
On the other hand, the liquefaction mechanism for lignin in the presence of phenol
was studied in relatively wider ranges, that is, without and with acidic catalysts [61-631.
As the model compound for lignin, guaiacylglycerol-P-guaiacylether (GC) was used and
the range of the liquefaction studied was as follows: ( I ) under elevated temperature (200-
250°C) without catalyst; (2) under an elevated temperature of 200-250°C in the presence
of acetic acid (catalyst); (3) under a moderate temperature of 150°C in the presence of
acetic acid (catalyst); (4) under a moderate temperature of 150°C in the presence of sulfuric
acid (catalyst), which corresponds to the study on the liquefaction of cellobiose described
above.
Conclusions obtained from this study are as follows: ( 1 ) The liquefaction of GC in
the presence of phenol under elevated temperature without catalyst proceeds very rapidly
through homolysis, producing coniferyl alcohol radical and guaiacol radical through qui-
none methide as the initial main intermediate. However, various homolytic cleavages oc-
cur, which give various kinds of radical compounds. As the result, considerable compounds
are produced through reaction among these radical species, with reactions among coniferyl
alcohol radical, guaiacol radical, and phenoxy or phenyl radicals resulting in dominating
reaction pathways. (2) Acetic acid can greatly promote the homolysis reaction of GG, but
does not alter the reaction mechanism; that is, in the presence of acetic acid, homolytic
cleavage and coupling can occur even at a mild temperature of 150"C, and the resulting
reaction products resemble those obtained under elevated temperature without catalysts.
(3) Under catalysis with sulfuric acid, GG is first transferred into mainly benzyl cation.
Benzyl cation rapidly condenses with phenol to give four condensed products as the initial
reaction intermediates, which are no sooner produced thanthey are further subjected to
extensivecleavage in their P-0-4 linkages and C&, bondings. The resulting cleaved
fragments further react with phenol to form various phenolated products. The characteristic
of this liquefaction reaction is heterolytic, that is, ionic, giving relatively small numbers
of products when compared with the homolytic reactions [61-63].
IX. CONCLUSION
The present state of studies on wood plasticization has been briefly reviewed. The scope
of the description has been somewhat limited, that is, somewhat focused on works of the
author and his colleagues. This is because the study of wood plasticization isnewand
immature and sometimes tends to confusion. However, it can be said that this is a new
field for chemical processing of wood with high future potential, although more studies
need to be undertaken.
REFERENCES
Wood Plasticization 699
41. M. Morita, Y. Yatnawaki, M. Shigematsu. and I . Sakata, Mokrrxri Gokkrrishi, 36: 659 (1990).
42. S. Pu. and N. Shiraishi, M o k u x i Gcrkktrishi, 39: 446, 453 (1993).
43. S. Pu. M. Yoshioka, Y. Tanihara, and N. Shiraishi. in C.-Y. Hse and B. Tomita (eds.),Adho.si~~e.s
m d Ror~dedWood Proc/wts, Forest Products Society (USA). Madison, WI, p. 344 (1994).
44. S. Pu, and N . Shiraishi, Mokuzcri G~rkktrishi.40: 824 (1994).
45. Y. Yao. M. Yoshioka. and N. Shiraish, Mokrrz(ri Gtlkkniski. 39: 930 ( 1 993).
46. L. Lin, M. Yoshioka, Y. Yao, and N. Shiraish, J. A/?/)/.P d y r ~ r t ~Sci.,
r 52: 1629 (1094).
47. Y. Yao. M. Yoshioka. and N. Shiraish, Mokrc=.ni Gtrkkaiski, 40: 176 (1994).
48. M. H. Alma. M. Yoshioka, Y. Yao, and N. Shiraishi, M o k ~ t z Gtrkktrishi,
~i 4 / : 1122 (1995).
49. M. H. Alma. M. Yoshioka. Y. Yao, and N. Shiraishi. J. A p p / . Po/yrn~r:Sci.. 61: 675 (1996).
50. M. H. Alma, M. Yoshioka, Y. Yao, and N. Shiraishi. Ho/$~r.sc~I~.. 50: 85 (1996).
51. Y. Yao, M. Yoshioka, and N. Shiraish. Mokrzrri Gnkktrishi, 41: h59 (1995).
52. L. Lin. M. Yoshioka, Y. Yao. and N. Shiraish. J . A/)/?/. Po/yr~rer:Sci.. 55: 1563 ( 1995).
53. L. Lin, M. Yoshioka, Y. Yao, and N. Shiraish, J . A/>/>/.Po/yrrrtv: Sci.. 58: 1297 ( I 995).
54. Y. Yao, M. Yoshioka. and N. Shiraish. J . AI?/>/.Po/yrrwc Sci., 60: 1939 (1996).
55. M. H. Alma, M. Yoshioka, Y. Yao. and N. Shiraishi, Mokrrztri Gttkkoishi, 4 / : 741 (1995).
56. C. Vanasse. E. Chornet, and R. P. Overend, C m . J . Cl~cwt.Ens., 66: I12 (1988).
57. H. R. Appel. I. Wender. and R. D. Miller, U.S. Bur: Mirles. Ech. Prog. Kcp., 25: S (1969).
58. H . R . Appel. Y. C. Fu, E. G. Illig, F. W. Stetfgen, and R. D. Miller, U.S. H r o : M i r ~ c xHI 801.3,
p. 27 ( I 975).
59. N. Shiraishi. in H. Inagaki and G. 0. Phillips (eds.). Ct4lctln.sic.s Ufiliztrfiorl; Htwrrrch t r r r d
Rcwnrcls in Ce//rr/o.sic.s.Elsevier Applied Science, London, p. 97 ( 1989).
60. N.Shtraishi, Japan Patent Publication Unexamined. 1989-179483.
61. L. Lin. Y. Yao. M. Yoshioka. and N. Shiraishi, Ho/$)r.sch.. 51: 316 (1997).
62. L. Lin, M. Yoshioka. Y. Yao. and N. Shiraishi. Ho/$orsch.. 51: 325 (1997).
63. L. Lin. M. Yoshioka, Y. Yao, and N.Shimishi, Ho/$orsc./~..51: 333 (1997).
17
Wood-Polymer Composites
Hiroshi Mizumachi
The University of Tokyo, Tokyo, lapan
1. INTRODUCTION
In composite systems consisting of polymers and other solid materials, polymer molecules
are in close contact with surfaces of solids, and it is quite reasonable to think that there
are some interactions between the two materials near the surface. Polymer segments in
the vicinity of the surface will be less mobile than those of the bulk polymers beyond the
range of influence of the surface. Physical properties of polymers filled with various solid
particles have been extensively studied in the past, and aresummarized by Kraus [ l ] ,
Flory [ 2 ] , and Nielsen [3].
There is much literature showing that carbon black particles in vulcanized natural
rubber modify the mobility of the polymer chains as a result of chemical linkages formed
between the two materials. This modified polymer layer is estimated to be about SO W i n
thickness. Therefore, the modulus and the ultimate strength of the filled rubber are greater
than those of the unfilled one. Kraus and Gugone [4] studied the adsorption of elastomers
onreinforcing fillers from solutions,analyzedtheadsorptionisothermsaccording to a
statistical mechanical theory of polymer adsorption developed by Simha, Frisch, and Eirich
[S-71, and obtained valuable information on the polymer-filler interaction responsible for
reinforcement. Kwei [S] studied the sorption of water vapor by Ti02-filled and unfilled
epoxy polymer, and performed a thermodynamic analysis that showed that polymer seg-
ments at a distance less than IS00 A from the surface of the filler are under the influence
of the filler. Kambe and Kat0 191 studieddynamicmechanicalproperties of a series of
amorphousepoxy resin films filled with homogeneous polyethyl methacrylate (PEMA)
beads with different diameters, and showed that larger increases of T,of PEMA are found
for smaller particles, and that the loss peak per unit volume of PEMA is lowered with
increase in the particle size, thus indicating some interactions between beads and the matrix
phase. I t is evident from these several examples that we need detailed knowledge of the
interfacial interactions between the two phases if we want to study the physical properties
of composite materials. We need to measure the physical or physicochemical properties
of the composite matcrials a s a whole, as well as those of the individual constituents, in
order to know the properties of the bound polymer near the surface, which reflect the
interactions between the component phases.
Physical properties such as dimensional stability, bending strength, abrasion, etc., of
wood-plastic composites have been widely studied from a practical point of view, as
701
702 Mizumachi
reviewed by Kent et al. [IO], Siau et al. [ l l ] , and Burmester [12]. However. the inter-
molecular interactions between the two materials through the boundary surface has not
been studied in detail from the physicochemical standpoint. In this chapter, some typical
studies of the interactions between polymers and woods or wood components are reviewed.
E = ell exp(iwr)
where is the amplitude of strain, W is the angular frequency of vibration, and r is time,
then the stress U also will vary sinusoidally with the same angular frequency W, with the
amplitude of v,,, and with a certain phase angle 6, namely,
where E' and E" are called storage modulus and loss modulus, respectively.
Mechanical energy applied to a material is split into two parts; one is stored in the
material as the mechanical energy, which is responsible for the elastic response of the
material, and the other is converted to heat as a result of internal friction. The former is
proportional to E' and the latter to E". Both E' and E" are functions of frequency and
temperature. Typical curves of E' and E" at a constant frequency for an amorphous un-
crosslinked polymer are shown schematically in Fig. 1 as a function of temperature. Gen-
erally, we find a few dispersions (stepwise drops of E ' ) accompanied by the same number
of mechanical absorptions (peaks in E"). Each dispersion and absorption, which must occur
simultaneously, correspond to a mode of molecular motion. For example, the primary
absorption of a-absorption (or a-dispersion) that appears at the highest temperature is
assigned to the initiation of micro-Brownian motion of polymer backbone segments.
Around the temperature of this primary absorption, E' of the material changes drastically
from IO"' to lo7 dyneskm'. Below the temperature of this absorption, backbone chains of
the polymer are frozen. and the material is glassy. However, some kinds of local motions,
such as rotation of side groups or local twisting of the chains, can happen at low temper-
ature, and this iswhy some secondary absorptions-P-absorption, y-absorption, and so
on. appear.
We can get information on molecular motions of polymers from their dynamic me-
chanical properties measured over a wide temperature range. Therefore, if we measure the
dynamic mechanical properties of polymers in the composite systems and compare them
with those of the bulk polymers, useful information on the interactions between the com-
ponent phases can be obtained.
Wood-Polymer Composites 703
log E '
l o g E"
TEMPERATURE
FIGURE 1 Schematic representation of E' and E" at constant frequency as a function of temper-
ature for a linear amorphous polymer.
10' I
3.5
-100
.
-50 0 50 100
TEMPERATURE "C
interested in the interactions between woods and polymers-in other words, in the extent
to which molecular motions of polymers are influenced by solid materials.
When polymers are impregnated into woods or filter papers by immersing the spec-
imens in polymer solutions, a very thin polymer layer is formed on the very complicated
surfaces of the wood or filter paper. These surfaces are porous and have such heteroge-
neous structures that the effective surface areas for polymer adsorption differ from spec-
imen to specimen. Therefore, it is difficult to control, or even to measure, the thickness
of the polymer layer or the surface of the material. Because an interaction between the
solid surface and polymer molecules, if any, must be the consequence of a very short-
ranged molecular force (van der Waals force), it may be predicted that the thinner the
polymer layer, the greater the influence of the solid surface. On the contrary, as the polymer
layerbecomes thicker, the properties of the layer approach those of the bulk polymer
which lies beyond the region of influence of the surface. Figure 3b shows variation of
temperature of a-absorption, 7&!&x), with the amount of polymer. Obviously, T(E,',:c,x) of
a-absorption of polymer in the composite systems is higher than that of the polymer itself,
suggesting that polymer chains on the surfaces of cellulosic materials are somewhat im-
l
-40 0 40 80
TEMPERATURE "C
0 5 10 100
w t .'lo of polyrncr
(b)
FIGURE 3 Variation of dynamic loss moduli at 110 Hz with the amount of polymer impregnated:
(a) log E" as a function of temperature; (h) T(E:i,,,x)as a function of the amount of polymer. Here
a commercial NBR adhesive polymer is used.
Wood-Polymer Composites 705
mobilized, and are falling toward that of the bulk polymer as the amount of polymer
increases. But when the polymer content is very small, i.e., below 10 wt% in the case of
filter paper, T(E::,:,Jvaries very slightly. In other words, it is substantially constant within
the experilnental error a s shown by the upper dashed line in Fig. 3b. When wood speci-
mens are immersed into 5 % polymer solutions and evacuated to remove solvent, the
amount of polymers impregnated in the wood generally falls within the region where T
(E:;,~!,)is substantially constant. AT(E::,.d,or the difference between of polymer
under the influence of the solid surface and that of the bulk polymer, which corresponds
to the difference between the two dashed lines in Fig. 3b, will be one of the appropriate
parameters to express the degree of interfacial interactions between the two components.
AT(E::,,,)is approximately equal to ATv or the difference between T V of polymers in the
colnposite and in the bulk phase. Typical data on the dynamic mechanical properties of
wood-polymer composites are shown in Figs. 4-7. Data for the respective bulk polymer
and for wood are shown for comparison. Generally, in composite systems the drop of E'
is smaller and the E" peak shifts to higher temperature than in the respective bulk polymer.
Dynamic mechanical properties of the composite materials can be calculated by
means of the theory of polymer blends. Takayanagi et al. [ 141 proposed mechanical models
like the ones shown in Fig. 8 to describe the viscoelastic behavior of polymer blends in
terms of the known properties of the two component polymers.
If phase cl is dispersed in phase M', there are two possible equivalent models for the
system, model I and model 11. Complex moduli of these models are expressed as follows:
E2' (model I) = (+/[AE$ + (1 - A)E:l + ( 1 - AYE:) '
E* (model 11) = A[@E$ + (1 - A)/E,T]" + ( 1 - A)E:
where E;: and E: are complex moduli of phase N and phase W , respectively; the values
of A and 4 correspond to the thickness fraction and the length fraction, respectively, of
phase N as shown in Fig. 8, and the product A4 is equal to the volume fraction ( U , , ) of
phase N. Practically, A and 4 are called the parameters of the mixing state, because the
values of these parameters naturally vary with the change of mixing state for a system
with the same value of U<,. For an actual composite system having separate phases of W
and (I, the behavior can be described by either of the models with appropriate values of
A and 4. It is experimentally verified that these equations can be applied to many polymer
blends when the component polymers from two separate phases in the texture of the system
without any interfacial interaction. Examples of numerical calculations for some wood-
polymer composite systems according to the above equations are shown in Figs. 9 and
10, where W and N refer to wood and adhesive, respectively. As A approaches 1.0 (both
of the models are reduced to the series connection of w and a ) , T(E:;,.,Jincreases somewhat,
but at the same time the corresponding E' falls drastically. Because the experimentally
observed E' curve shows very little dispersion in every case, a mechanical model with (b
approaching 1.0 (the parallel connection of LV and a ) must be employed for these wood-
polymer composite systems. Because the elevation of T(E::,L,x) cannot be accounted for, we
have to take some interfacial interactions between the two phases into consideration in
order to interpret the fact that the experimentally observed T(E::,J shifts to higher tem-
perature than that of the bulk polymer without any sizable dispersion of E'. It is evident
that polymer molecules adsorbed on the surface of cellulosic materials are immobilized
by the intermolecular forces at the boundary surface and that the composite material con-
sists of wood texture and a thin layer of the immobilized polymer on the surface, which
is constructed almost in parallel combination.The E" peak of the composite materials
I
-.
r - -
E'
c
E'
-
-0
E'
z
! s'
.
i <:
E" # E" n?
=* l E" .
b .
job
10'
t -50 0
TEMPERATURE
50
"C
-100 -50
TEMPERATURE
0
"C
-100 -50
TEMPERATURE
0
t
"C
50 I00
TEMPERATURE
150
"C
FIGURE 4 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and NBR co-polymer
(acrylonitrile 50%). Data on the constituent bulk materials are also shown for comparison. 0,composite system; 0 , polymer; @,
wood.
FIGURE 5 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and SBR co-polymer
(styrene 23.5%). Data on the constituent bulk materials are also shown for comparison. 0,composite system; 0 , polymer; @, wood.
FIGURE 6 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and SBR co-polymer
(styrene 46.0%). Data on the constituent bulk materials are also shown for comparison. 0,composite system; 0 , polymer; @, wood.
FIGURE 7 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and polystyrene. Data
on the constituent bulk materials are also shown for comparison. 0.composite system; 0 , polymer; @, wood.
Wood-Polymer Composites 707
/
t
Model I Model Il
FIGURE 8 Mechanical equivalent models for the composite system where phase a is dispersed
in phase M' without interfacial interaction between the two phases. A and 4 are the parameters of
mixing states, and A4 is equal to the volume fraction of phase a.
E' E'
.
J 0900 0151
IOW C138
'O'
-100 -50 0 50 -100 -50 0 50
TEMPERATURE "C TEMPERATURE "C
FIGURE 9 E' and E" versus temperature at 110 Hz for the composite system composed of wood
(Sugi) and SBR (styrene 23.5%). The experimentally obtained curves are compared with the cal-
culated ones. Numerical calculations are carried out according to the equations based on (a) model
I and (b) model 11, using the properties of the constituent bulk materials, which are also shown in
the figure. Solid lines are the calculated curves with different parameters of mixing state.
708 Mizumachi
FIGURE 10 E’ and E” versus temperature at II O Hz for the composite system composed of wood
(Sugi) and polystyrene. The experimentally obtained curves are compared with the calculated ones.
Numerical calculationsare carried out accordingto the equations based on (a) model 1 and (b)
model 11. using the properties of the constituent bulk materials which arc also shown in the figure.
Solid lines are the calculated curves with different parameters of mixing state.
becomes broader than that of the bulk polymer, which suggests that there are various kinds
of segmental environments in the adsorbed polymer layer, and this broadening cannot be
interpreted by any of the above equations.
Dependence of AT(E;,;,,)or ATg on the polymer composition is shown in Fig. 11 for
co-polymers of styrenehutadiene (SBR) and acrylonitrilehutadiene (NBR). AT(Et,t,x)ob-
viously varies systematically with polymer composition, but unfortunately there is no
molecular theory to interpret these characteristics. It is well known that two materials with
similar values of solubility parameter (S), which is a square root of the cohesive energy
density (C.E.D.), are compatible with each other. The values of 8 for the materials are 8.4
20
15
d i
‘0.0 0.5 1.0 4 0 65 0
PBD PSt PBD N
(a>
FIGURE 11 Dependence of ATx on polymer composition in wt%: (a) SBR co-polymers; (b) NBR
co-polymers. 0, Sugi; 0 , filter paper.
Wood-Polymer Composites 709
(callmol)”’for polybutadiene, 9.1 for polystyrene, 12.5 for polyacrylonitrile, and 15.65
for cellulose; there is a possibility that as the 6 of polymer approaches that of wood (or
filter paper), the two materials become more compatible and polymer molecules will be
adsorbed more tightly on the solid surfaces, and consequently AT(E::,.,) will become larger.
Iyengar et al. 1151 studied the role of adhesive-substrate compatibility in adhesion and
measured peel strength of Mylar/adhesive/Mylar systems using various kinds of adhesives.
They found that maximum peel strength is obtained when the solubility parameter of the
adhesive is close to that of the substrate; in other words, peel strength becomes greater as
the two materials become more and more compatible.
AT(E::,,,) of a Sugi-polymer system is slightly larger than that of the corresponding
filter paper-polymer system, but the general tendency of AT(E::,ctx) variation with polymer
composition is almost the samefor the two materials. It is possible to think that Sugi
immobilizes the polymer slightly more than filter paper does, but this small difference will
notbe significant because the amount of polymer impregnated is different for the two
solids and there is a large scatter of experimental points in Fig. 1 1 . Studies of this kind
have been advanced also by Tadokoro et al. [ 161, Okumura et al. [ 171, Handa et al. [ 181,
and Motohashi and Tomita 1191, who made some interesting observations concerning the
dependence of the rheological properties of various wood-polymer composite systems on
such factors as polymer loading (the amount of polymer impregnated), the kind of solvent
used for impregnation, the molecular weight of polymers, the method of preparation of
the composite system (blend or graft), and so on. It is interesting to note that they have
double peaks in E” curves of wood-polymercompositesystems when the amount of
polymer impregnated is extremely large. Figure 12 illustrates some typical data of the
dynamic mechanical properties of wood-PVAc emulsion composites plotted as a function
of temperature. One peak which appears on the higher-temperature side is assigned to the
initiation of micro-Brownian motion of the immobilized polymer segments in the vicinity
of the solid surface. The other peak corresponds to that of the ordinary segments which
lie in the bulk phase.
It has been repeatedly ascertained in these reports that polymer molecules are im-
mobilized near the surface of cellulosic materials as a result of molecular interactions.
t
20 oi 60 80 100
Temperature (‘C)
FIGURE 12 E” versus temperature at 110 Hz for the composite system composed of wood (Sugi)
and PVAc emulsion with various polymer contents (A, 7.0; B. 15.1; C, 22.1; D, 37.0%).
calcium carbonate varies with the filler content in accordance with the theory of composite
systems, where the interfacial interaction between the dispersed phase and the matrix phase
is not taken into consideration [2].
Shimbo and Ochi [20]also showed that the modulus of polyethylene filled with
calcium carbonate falls drastically above the melting temperature of polyethylene crystals
without revealing any rubbery plateau region.
We obtain information on the interaction between polymers and fillers by comparing
the dynamic mechanical properties of filled polymers with those of the respective unfilled
ones. In cases where polymer and fillers are mixed in solvent at room temperature, no
chemical reaction is expected to occur at the boundary surface. Therefore, the rubbery
plateau region is not observed in the modulus-versus-temperature curves, but the temper-
ature of the absorption peak T(E:;,:,.Jis generally altered if the two components interact
with each other. Figure 13 shows the loss modulus (E”) at 100 Hz versus temperature of
filled (50 wt%) and unfilled polystyrene [21]. T(E::,;,,)of polystyrene filled with calcium
carbonate is almost the same as that of the unfilled polymer, reflecting the inertness of the
filler. Strictly speaking, T(E::,itx)
of the filled polymer is slightly lower than that of the
unfilled polymer in this case. The exact reason is not known, but it may be due to the
incorporation of free volume during the process offilm forming. On the other hand,
T(E::,rtx)of polystyrene filled with carbon black is the highest of all the composite systems.
Wood-Polymer Composites 711
CaCO,
(unfilled)
lignin
cellulose
xylan
I~~
carbon black
This means that carbon black alters the physical properties of polystyrene molecules in
the matrix phase to the greatest extent.
Similar data for tilled and untilled SBR and polybutadiene are shown in Figs. 14
and 15. Variations of T(E::,:,x)with tiller content for the three polymers are summarized in
Figures 16- 18. In every case, calcium carbonate and carbon black play similar roles in
the rheological behavior of the tilled polymers. Therefore it would be reasonable to think
that the elevation of T(E::,ltJin the compositesystems, as compared with that of the
respective bulk polymer, corresponds well to the extent of interfacial interaction between
the component phases. Yim et al. [22], who referred to T(tan S,,,,) as “Tx” in their report,
showed that the shift in “TK”is directly proportional to the extent of polymer-filler in-
teraction energy estimated from the heats of adsorption of the model compounds of the
polymers on the tiller surfaces, which is in agreement with the previously mentioned
speculation.
The extent of interfacial interaction in polymer-wood component composite systems
can now be compared. As the content of wood component tilled in the polymer increases,
7‘(E::rdx)
of the composite system increases in every case. This means that the three com-
ponents are more active than calciumcarbonate, so it is clear thatwood components
modify the physical properties of nearby polymer molecules. Fillers of wood components
elevate T(.E::,z,x)
of polystyrene to a greater extent than that of polybutadiene. This is con-
sistent with the informations obtained from Fig. 11, where AT(,?::,,,) of a series of SBR-
impregnated woods are plotted as a function of copolymer composition. Among the three
wood components, xylan is slightly more active than the other two in every case.
712 Mizumachi
carbon black
carbon black
\
I
-150 -130 -110 -90 -70 -50 -30
temperature 'C
FIGURE 15 Comparison of the temperature dependence of the loss moduli of polybutadiene tilled
with calciumcarbonate,carbonblack,cellulose. lignin, and xylan to SO76 by weight. Data for
unfilled polybutadiene are also shown.
Wood-Polymer Composites 713
105
0 10 20 30 40 50
fillcr content W 1. "1.
FIGURE 16 Dependence of T(.Ei:,ax) on filler content for polystyrene. Fillers: @. calcium carbonate:
U, carbon black; 0 , cellulose; +, lignin: 0 , xylan.
"
0
-15
0 10 20 30 40 50
fillcr contont wt.%
FIGURE 17 Dependence of T(E::,,,,)on filler content for SBR (styrene 46.0%). Fillers: 0 , calcium
carbonate: U, carbon black; 0 , cellulose; +, lignin; 0 , xylan.
714 Mizumachi
V
- 85
e
-90
:z
W
Y
0
t" -95
-100
10 20 30 40 50
filler contont W 1. @l.
FIGURE 18 Dependence of T(E::,J on filler content for polybutadiene. Fillers: Q, calcium car-
*,
bonate; a, carbon black; cellulose; +, lignin; 0, xylan.
where C is the capacitance of a pair of capacitors when the space between them is filled
with the material, and C,, is the capacitance of the same arrangement of conductors in
vacuo.
When a material is placed in an electric field, the centers of gravity of the negative
charges are displaced from those of the positive charges-in other words, dipolesare
induced instantaneously, and the charge on the conductors becomes greater than that in
the case of conductors in vacuo. In addition, if the material contains permanent dipoles,
they tend to orient in the direction of the electric field, and the charge will appear on the
conductors. However, the orientation of permanent dipoles is associated in general with
some elementary processes of the passage over potential energy barriers as a consequence
of Brownian motion, and therefore it isnot only time-dependent, but also temperature-
dependent. Therefore, E of a material must be expressed as follows:
E = E(t) at T
The variation of E ( t ) with application of voltage is analogous to that of compliance J ( t )
with application of stress.
If the voltage varies simultaneously with time, then the dielectric constant is ex-
pressed as a complex number:
&* = &l - is"
and the ratio E"/E' is usually called the dielectric loss tangent tan 6,. Electrical energy
applied to a material is resolved into two components: one is stored in the material, and
the other is dissipated as heat. The former is proportional to E' and the latter to E". Typical
curves of E' and E" at a constant temperature are shown schematically as a function of
angular frequency in Fig. 19. At infinite frequency, permanent dipoles cannot orient in
accordance with the fast sinusoidal variation, and E' is equal to a certain constant value
E,, which is related only to the electronic depolarization of the material. As frequency
Wood-Polymer Composites 715
I c'I
approaches zero, on the other hand, all the permanent dipoles can orient in the direction
of the electric field at any moment without delay, and E' approaches a larger limiting value
E(,. A stepwise change or dispersion of E' occurs at the frequency L,,,,
which is comparable
to the time scale within which permanent dipoles move, and simultaneously a peak of
E", or absorption, is observed at the same frequency.
These characteristics are described mathematically as follows:
If E" is plotted as ordinate against E' as abscissa, which is called a Cole-Cole plot,
a circular arc is obtained as shown in Fig. 20. The length of the cord of the circular arc
is equal to ( E ( ) - E,) = A s , which is proportional to the concentration of the dipoles
contributing to the orientation polarization; PT is equal to the angle between the radii of
the arc drawn to the points E( ) and E, from the center of the circle, and the value of P (0
< P < 1 ) represents the degree of broadness of a dielectric absorption curve. The smaller
P corresponds to the broader absorption peak.
The relaxation time
E'
CO P,
- €CO cosec -
E'
varies with temperature, according to either the Arrhenius equation or the WLF equation.
When E' and E" of polymeric materials are measured over a wide range of temperature
and frequency, there appear multiple absorption peaks (a, P, 7, . . .) in many cases, such
as are shown in Fig. 2 I .
Each absorption peak corresponds well to the dynamic mechanical absorption phe-
nomena, as shown in the same tigure. Dielectric properties and dynamic mechanical prop-
erties give us substantially the same information on the molecular motions in the materials,
the former being amplified by polar or polarizable groups.
TEMPERATURE
TEMPERATURE
absorptions of PMMA. Typical data by Handa et al. on the dielectric properties of a wood
(Buna)-PMMA composite system are shown in Figs. 22 and 23. It is interesting to note
that woods impregnated with polymers and those impregnated with monomers and then
irradiated have different properties. Figure 24 shows the Cole-Cole plots for control wood
and wood-PMMA composite systems at -60°C. The parameters A s and 0 are plotted
against polymer content in Fig. 25. In composite systems where polymers are impregnated
into woods, A s and 0 are constant and almost the same as the respective parameters of
woods. However, in cases where monomers are impregnated into woods and then irradi-
2.4 -
2.3 -
.c,
-
-
2.2
2.1 -
2.0
-60 40 -20 0 20
(a) Temperature ( C )
O-O~I
0.04 -
c,
-
-
0.03
0.02 t
-60 40 -20 0 20
(b) Temperature ( C 1
1.7
60 -40 -20 0 20
Temperature (C)
003 - l
c,
9
002 - c
f
bl B
a d
OOlL
60 -40 -20 0 20
Temperature ( C 1
ated, A& decreases as polymer content increases, which indicate that graft co-polymeri-
zation onto the methylol group occurs in wood, and as a consequence the concentration
of dipoles contributing to the orientation polarization within the temperature and frequency
range of this experiment decreases during irradiation. At the same time, the p-value also
decreases, indicating the increase of nonuniformity in the environment around the side
groups of the polymer.
Wood-Polymer Composites 719
0
0.031 L , ,
0.03 1 A
0.031
0
7%
. ,
0.03
0 I
0.03 0 04-
I
O A
2.0 1.9 1.8 1.7
€*
FIGURE 24 Cole-Cole plot forcontrolwoodandPMMA-WPCsystems at -60°C. X , control
wood; 0, PMMA-WPC (injection) system (polymer content 21.2%); A , PMMA-WPC (injection)
system (13.3%);0 , PMMA-WPC (irradiation) system (19.2%); A, PMMA-WPC (irradiation) sys-
tem (9.2%).
0.4 I
0.2 -
l 1 l
20 15 10
0 5 25
Polymer Content ( % )
FIGURE 25 (a) A E versus polymer content. (b) p versus polymer content. x, control wood; 0.
PMMA-WPC (injection) system; 0 . PMMA-WPC (irradiation) system.
720 Mizumachi
or
where AH* is the apparent activation energy of the relaxation process, R is the gas con-
stant, and T is the absolute temperature. Figure 29 shows the plots of logarithm of ,A,,.L,
versus I/T for unfilled and filled PVAc. A straight line is obtained in every case, and AH*
can be calculated from the slopes of the straight lines. The values of AH* can be regarded
as the energy required for a mole of polymer segment to initiate micro-Brownian motion.
A H N thus evaluated for untilled PVAc is 53 kcal/mol, which is in good agreement with
the literature values 1301. It is evidence that AH* for filled PVAc is larger than that of the
- 3 - 2 - 1 0 1 2 3 4
log( f / f r n a x )
5.0
** .
..
4.0
......
*\-.
\
log f max
3.0
2.0
73 K c a l h o l e
1.o
k
2.80
3.00 2.90 3.10
lo3 I T
FIGURE 29 Plots of logA,,asx versus 1/T for unfilled andfilledPVAc: 0, unfilledPVAc; 0, PVAc
filled with cellulose (40%); X, PVAc tilled with lignin (40%).
unfilled one, and that the average relaxation time for the former system is longer than that
for the latter. This means that the bound polymer that lies within the region of influence
of the solid surfaces is less mobile than the polymers in the bulk phase.
which the free energy AF,,, of the transformation of filled polymer to the unfilled polymer
can be evaluated. Kwei described his thermodynamic cycle and the corresponding equa-
tions on a mole basis, but here they are represented on a gram basis. The free energy
changes AF and AF' in the thermodynamic cycle can be evaluated by analysis of the
vapor sorptions of filled and unfilled polymer, respectively, using the Gibbs-Duhem
relation:
AF = w , A F , + wzAt',
w,dAF, + w,dAF2 = 0
where
and R is the gas constant, T is the absolute temperature, and M ,is the molecular weight
of the solvent. From the values of AF and AF' thus obtained, the difference in the free
energy of the filled and unfilled polymers can be evaluated as
AF,,, = AF' - AF
It is obvious from the thermodynamic cycle that the AF,,, value must be determined
independently of the kind of solvent used in the experiment, and therefore we come to
know the difference in the free energy of the filled and unfilled polymers through the
value of AF,,,.Figure 31 shows the sorption of benzene at 50°C by PVAcIcellulose ( 5 % ) ,
and unfilled PVAc as a function of p,lpy. At very low p,lp';,there is a slight difference
between the sorption of unfilled PVAc and that of the filled one, but the two isotherms
coincide after a critical vapor pressure (p,Ip~= 0.1) is reached. Data in Fig. 31 were
applied to the Kwei's thermodynamic cycle and AF,,, is calculated to be 0.018 callg poly-
mer at 50°C. This means that cellulose lowers the free energy ofPVAc molecules near
20
U
S 10
U
U
cn
E
0
0.00 Q05 0.10 0.15 P,/P,O
the surface or, in other words, the latter is slightly stabilized as compared with the same
polymer molecules in the bulk.
The amount of benzene sorbed at 50°C and 60°C by NBRkellulose (50%) is com-
pared with that of unfilled NBR in Figs. 32 and 33, respectively. In this case, there is no
difference between the sorption isotherms of filled and unfilled polymer at both temper-
atures in the whole range of relative vapor pressure, in spite of the fact that the filler
content is much increased as compared with the case of PVAc. This means that cellulose
surfaces do not alter the thermodynamic properties of the surrounding NBR molecules in
the filled system.
Results of the thermodynamic analysis are summarized in Table 1. If the value of
AF,,, is evaluated at several temperatures, the enthalpy (AH,,,)and the entropy (AS,,,)com-
ponents in AF,,, can be separated by the following equation:
AF,,, = AH,,, - TA&,
where
Kwei made this analysis on epoxy polymer tilled with TiO?, which has higher energy
surfaces than cellulose has, and obtained AH,,, = 1.81 (at 80°C) to 2.47 (at 70°C) cal/g
polymer, and AS,,, = 4.9 X 10.' (at 80°C) to 6.7 X (at 70°C) cal/deg/g polymer. He
pointed out that these values are about 30 times smaller than the corresponding values for
the melting of polyethylene crystal, but it is clear that the positive values of AH,,, and AS,,,
indicate that there exist some structures with local ordering of polymer segments in the
filled state. From the data in Figs. 32 and 33, AF,,, is zero at SO-6O0C, and therefore AH,,,
= 0.0 cal/g polymer and AS,,, = 0.0 cal/deg/g polymer in this temperature range, meaning
that there is no difference in enthalpy and entropy between filled and unfilled NBR within
the experimental error. As there are no data of AF,,, at different temperatures in the case
of the PVAckellulose system, AH,,, and AS,,, cannot be calculated, but it can safely be said
that the thermodynamic interaction between cellulose and polymer is greater in the PVAd
cellulose system than i n the NBWcellulose system, because AF,,, has a positive value in
V.-
the former case at 50°C, while it is zero in the latter case. This trend is consistent with
the results obtained in rheological studies on the interaction between wood and polymers:
the elevation of T(E;:,;,Jin the PVAc/filter paper system is 10°C, which means that the
interfacial interaction between the component phases is greater in the former system.
When solid powders are mixed with polymer solutions, polymer molecules will be ad-
sorbed on the solid surfaces until adsorption equilibrium is attained. Adsorption behaviors
are governed by balance of affinities between polymers and solid surfaces, between sol-
vents and solid surfaces, and between polymers and solvents (i.e., solubility). So, if we
can compare the adsorption isotherms with the other conditions being equal, it is possible
to obtain information on the interactions between polymers and solid surfaces. It is some-
times pointed out that the adsorption isotherm for polymers in solutions is similar to the
Langmuir isotherm or to the Freundlich isotherm within experimental error. For example,
Luce et al. [3 1 1 studied the adsorption ofPVAc from benzene solution on filter pulp
swollen i n various solvents, and obtained isotherms as shown in Fig. 34.
TABLE 1 ThermodynamicQuantities
"Kwci (1964).
Mizumachi
80 -
FORMAMIDE
METHANOL
L
BENZENE
0
1 2 3 4 5 6
EQUILIBRIUM
CONCENTRATION (dl)
FIGURE 34 Typical adsorption isotherms for PVAc from benzene solution onto filter pulps swol-
len in dimethylsulfoxide (DMSO), formamide, water, methanol, and benzene.
Surface area available for adsorption of polymers can be altered by swelling and
solvent exchange prior to adsorption. It is found that the greater the degree of swelling in
the pretreatment, the larger is the apparent saturation value of the adsorption. Although
adsorption is initially rapid, the rate decreases gradually and it takes a long time to attain
equilibrium adsorption, particularly when porous material is used as adsorbent. It has also
been shown that the solvent plays an important role in the adsorption process. Poor solvent
generally favors adsorption, as shown in Fig. 35, due to the facts that polymer molecules
are tightly coiling in the poor solvent and therefore the more efficient packing of polymer
molecules on the solid surfaces is realized, and also to the fact that polymer molecules
are thermodynamically more stable in the adsorbed state than in the solution state when
poor solvents are used. On the contrary, polymer molecules have extended conformations
in good solvent, and they are more stable in the solution state than i n the adsorbed state,
which iswhy the relative adsorption becomes smalleras the solubility parameter of a
solvent comes closer to that of the solid as shown in Fig. 35. Adsorption isotherms in Fig.
34 are apparently conformable to both Langmuir and Freundlich isotherms, as illustrated
in Figs. 36 and 37, respectively. Adsorption of polymers onto wood surfaces is usually
analyzed according to the Langmuir equations 132-341.
However, it is quite reasonable to imagine that polymer molecules are adsorbed on
the solid surfaces at ;I number of segments along the molecular chains, with the remainder
of the molecules extending into the solution phase. Luce and Robertson have pointed out
that although there is some uncertainty about the surface area available for adsorption, the
extent of adsorption is such as to suggest that the amount of polymer adsorbed is at least
Wood-Polymer Composites 727
2.0L
0.5 9.0 9.5 10.0
SOLUBILITY
PARAMETER
0.4 /
I/ METHANOL f-
0 2 4 6
EQUILIBRIUM
CONCENTRATION (g/l)
FIGURE 36 Langmuir isotherms for adsorption of PVAc from benzene solution onto filter pulp
swollen in various solvents. The ordinate is the equilibrium polymer concentration divided by the
corresponding equilibrium adsorption.
728 Mizumachi
WATER
cl
0.1 1 10
one order greater than could be accommodated if all segments occupied adsorption areas.
There must be a close relation between conformation of adsorbed polymer molecules and
adsorption isotherm.
Simha et al. [6] have developed theories of the adsorption of linear flexible polymers,
and derived the following equation:
where 8 is the fraction of the surface covered by polymer segments; (v) is the average
number of segments adsorbed per molecules as shown schematically in Fig. 38; K is a
constant dependent on (v), the heat of adsorption, solvent interaction, molecular weight,
and temperature; k , is a measure of the free energy of interaction between adsorbed seg-
ments in excess over the interaction in the solution phase; and C is the concentration of
polymer in the solution phase that is in adsorption equilibrium.
Kraus and Gugone 141 studied the adsorption of elastomers on various carbon blacks
in detail, and tried to analyze the adsorption isotherms according to the aforementioned
theory. They neglected the excess lateral interactions ( k , = O), and transformed the equation
into a form suitable for plotting experimental data:
where LZ isthe adsorption at concentration C, and a, isits saturation value. To fit the
experimental data. the left-hand side of the equation was plotted against log C for several
assigned values of a,, and the best straight line through the experimental points chosen
for the calculation of K and (U). They could fit all the isotherms to the equation, among
which some typical examples are shown in Fig. 39, and the values of the parameters are
summarized in Table 2. If (v) is equal to unity, the isotherm of Simha et al. is identical
with that of Langmuir, but when (v) is larger than unity, amount of polymer adsorbed ( a )
appears to level outa very low concentration and then keeps on increasing gradually
toward a, as the concentration increases. This means that the isotherm deviates from the
Langmuir equation. Although the isotherms are not very sensitive to the exact choice of
parameters, it is interesting to be able to get information on the state of adsorption of
polymer molecules through this kind of analysis.
REFERENCES
1. G.Kraus, RubberChem. Techn., 38: 1070(1965).
2. P. J. Flory, Principles of Polymer Chemistry, Cornell University Press, Ithaca, NY (1953).
3. L. E. Nielsen, Mechanical Properties of Polymers and Composites, Marcel Dekker, New York,
1975.
4.G.Krausand J. Gugone, Ind. Eng. Chern., 47: 1809 (1955).
5. H. L. Frisch, R. Simha, and F. R.Eirich, J. Chem.Phys.. 21: 365 (1953).
6. R. Simha, H. L. Frisch, and F. R. Eirich, J. Phys. Chem., 5 7 584 (1954).
7. H. L. Frisch and F. R. Eirich, J. Phys. Chem., 58: 507 (1954).
8. T. K. Kwei, J. Polymer Sci. A, 3: 3229 (1965).
9. H. Kambeand T. Kato, Rep.Prog.PolymerPhys. Japan, 12: 317 (1969).
IO. J. A. Kentand N. Nash, Woodworking Digest, 67: 31 (1965).
11. J. F. Siau, J. A. Meyer, and C. Skaar, Forest Prod. J., 15: 162 (1965).
12. A. Burmester, Holz-Zenrralbl., 93: 1191 (1967).
13. H. MizumachiandM.Fujino, Holiforsch., 26:164 (1972).
14. M. Takayanagi, H. Harima,and Y. Iwata, Mem. Fac. Eng. Kyushu Univ., 23: I (1963).
15. Y. lyengar and D. E. Erickson, J. Appl. Polymer Sci., 11: 231 1 (1967).
16. K. Tadokoro, T. Sadoh, and K. Nakato, Mokuzai Gakkaishi, 22: 309 (1976).
17. M. Okumura, N. Shiraishi, T. Sadoh,and T. Yokota, Proc.26th Ann. Meeting Japan Wood
Research Society, p. 215 (1976).
18. T. Handa, S. Yoshizawa, M. Fukuoka, M. Suzuki, Y. Ikeda, and M. Saito, Proc. Rheol. Symp.
Japan Wood Research Society, p. 3 (1976).
19. K. Motohashiand B. Tomita, Mokuzai Gakkaishi, 26: 87 (1980).
20. M. Shimbo and K. Ochi, Proc. 24th Ann. Meeting Chemical Society of Japan, 4 : 2044 (197 I).
21. H. MizumachiandM. Tsukiji, Holzjlorsch., 34: 122 (1980).
22. A. Yim, R. S. Chahal, and L. E. St. Pierre, J. Colloid Interjke Sci., 43: 583 (1973).
23. Y. Ishidaand K. Yamafuji, Kolloid-Z., 177 97 (1961).
24. T. Tanakaand Y. Ishida, J. Phys.Soc. Japrm, 15: 261 (1960).
25. K. Kitsuta, Mokuzai Kogyo (WoodIndustry), 25: 14 (1970).
26. T. Handa, S. Yoshizawa, andM.Fukuoka, KobunshiRonbunshu, 34: 617 (1977).
27. H.Mizumachiand M. Kamidohzono, Holdorsch., 29: 229 (1975).
28. Y. Ishida, M. Matsuo,and K. Yamafuji, Kolloid-Z., 180: 108 (1962).
29. S. Saito, Kolloid-Z., 189: I16 (1963).
30. H. A. Stuart, Die Physik derHochpolyn~eren, Springer-Verlag, Berlin/Gottingen/Heiderberg
( 1956)
31. J. E. Luce and A. A. Robertson. J. Polymer Sci., SI: 317 (1961).
32. K. Nakato, N. Shiraishi,and K. Yokoo, Zairyo, 16: 839 (1967).
33. N. Shiraishi, Y. Ishimaru, K.Nakato, and T. Yokota. Moku;ui Gakktrishi, 15: 20 (1969).
34. Y. Ishimaru. Mokuzcti Gnkknishi, 22: 22 ( I 976).
This Page Intentionally Left Blank
18
Adhesion and Adhesives
Hiroshi Mizumachi
The University of Tokyo, Tokyo, lapan
1. INTRODUCTION
Most of the materials currently used are composites, and they are manufactured by means
of adhesives and adhesion techniques. Recently, increasingly diversified and advanced
technical requirements have been imposed on the functions of those composite materials,
and it is well known that there are many cases where adhesion between the constituents
within the composite materials plays an important role.
There are various types of performance evaluation tests forcomposite materials,
among which tests for evaluating the bond performance may be regarded as the most
interesting ones for those engaged in work related to adhesion.
We can find numerous material characteristics concerning both adhesives and ad-
herends, and also many factors concerning adhesion processing techniques that are ex-
pected to have some effects on adhesion. These factors must interact with each other in a
complicated manner, resulting in development of various material functions, including a
necessary level of bond performance. Engineering difficulties would be reduced if all the
relations among these factors were clarified theoretically. However, as a matter of a fact,
the greater part of them are hidden in a large black box. and a vast number of trials will
have to be carried out to develop a new composite material. The present situation is far
from one where we are ready to perform successful molecular design for optimum adhe-
sives to produce a composite material with required physical product characteristics, and
at the same time we can easily establish the optimum processing procedures. However, it
is an important responsibility of adhesion scientists and engineers to continue efforts to
accomplish basic or scientific approaches one by one until the final goal is achieved, even
if it is a long way off. In this chapter, attention is concentrated on the relation between
structure/properties of adhesives and adhesive strength, and adhesion mechanism is dis-
cussed rheologically.
II. ADHESIVESTRENGTH
We adhere a pair of adherends by applying an adhesive between them, and after adhesion
is completed, we have to ascertain to what extent the bonding is satisfactory. In what way
do we perform the test? Usually, we measure the breaking strength of the composite system
733
734 Mizumachi
-
thermal expansion coefficient, etc.). And it is true that the apparent adhesive strength is
different if a different method of testing is employed, even for the same combination of
m L I
H-
m
t 4
-F
Pee 1
00 4
Shear
t
E
t
Tens i le
adherend and adhesive. The typical types of adhesion tests most frequently employed are
tensile test, shear test, and peel test, which are shown schematically in Fig. 2. Also, some
types of tack tests are employed for pressure-sensitive adhesives.
It is interesting to note that adhesive shear strength of a steel/epoxy residsteel system
is usually 200-300 kg/cm’, while peel strength of the same combination of the materials
is only 0-2 kg/25 mm. Although the two quantities have different dimensions and they
cannot be compared in a simple manner, it is a kind of common sense that when adhesive
shear strength is high for an adhesive, peel strength is low for the same adhesive, and
vice versa. Adhesive tensile strength has the same dimension as shear strength, namely
kg/cm’, but the two quantities are not necessarily the same. If the testing method is dif-
ferent, the situation is quite different, and therefore, it is absolutely imperative for devel-
opment of the science and technology of adhesion to clarify what kinds of factors are
dominant for each type of adhesive strength. So far, numerous data have already been
accumulatedconcerning the relations between adhesive strength and various factors of
woods, i.e., specific gravity, moisture content, morphological factors, physical properties,
etc., and we can easily find them in handbooks [ 1,2], but few researchers have concentrated
their attention on the effects of the structuresand physical properties of adhesives on
adhesive strength of wood. This is also a very important problem, and in this chapter the
dependence of adhesive tensile strength, shear strength, and peel strength, especially on
rheological properties of adhesives, is reviewed. In addition, tack of pressure-sensitive
adhesives is also discussed.
The molecular weight of PVAc and the grafting efficiency or the amount of grafting
of VAC monomer on PVA vary largely according to the differences of reaction conditions,
some of which are listed in Table 1 .
Figure 4 shows that there is no significant relation between adhesive tensile strength
of cross-lap wood joints at room temperature and the molecular weight of PVAc. However,
the adhesive tensile strength increases significantly asgrafting efficiency increases, as
shown in Fig. 5.
Adhesive films were prepared by casting the emulsions on Teflon sheets and drying
at room temperature, and the ultimate strengths were measured. Obviously, the ultimate
strength ofan emulsion film increases linearly with grafting efficiency, as illustrated in
Fig. 6, so it is concluded that the adhesive tensile strength of a PVAc emulsion synthesized
using PVA as emulsifier has a significant relation to grafting efficiency. This means that
adhesive strength is dependent to a great extent on the mechanical properties of the
adhesive.
Polymeric materials are generally viscoelastic in nature, and physical or mechanical
properties are time-dependent as well as temperature-dependent, especially around the
glass transition region. Motohashi et al. [4] chose two typical PVAc emulsions, differing
in grafting efficiency (see Table 2), measured adhesive tensile strength at a constant speed
of crosshead separation as a function of temperature in the range from - 130 to + 140°C,
and obtained the results shown in Fig. 7. At extremely low temperature, micro-Brownian
motion of polymer segments is frozen and the adhesive is glassy. The tensile strength in
this region is almost constant and the value is approximately 10 times as smallas the
ultimate strength of the adhesive film.
The adhesive tensile strength increases gradually as the temperature is raised, and
reaches a maximum near the glass transition temperature of the polymer, where micro-
Brownian motion of backbone chain segments is initiated. At higher temperatures, molec-
ular cohesion of the polymer decreases greatly and accordingly adhesive tensile strength
also becomes very low.
TABLE 1 Kinds of PVAc Emulsion and All Experimental Results
0
40, 0
0
A o
n
0
d 15 e
A
I
Y
& A
A
L,
m 0
C 0
2 3.0,
L
cn 00
U V V V
C V V
0 V
m Gralling eflicirncy ( )
2.5.
0 -
48.4 57.6
A 30.6 - 40.8
0 20.3 - 27.0
2 .o,*
V 10.3 - 17.1
1 J
0 1 2 3 4 S
FIGURE 4 Relation between A?,, and adhesive tensile strength of cross-lap joint at each level of
grafting efficiency.
Hatano et al. [5] have investigated the relationship between the adhesive strength
and the dynamic mechanical properties as well asthefailure characteristics of epoxy
polymers at room temperature. Epikote 828 (epoxy equivalent 189 2 5) and Epikote 871
(epoxy equivalent 41 3) were blended at various ratios and cured with a stoichiometric
amount of diethylenetriamine. Asthe weight fraction of Epikote 871 in the Epikote 828/
Epikote 871 blends increased, the temperature of viscoelastic absorption peak T(E:ax)at
l10 Hz, which is about 15-20°C higher than the glass transition temperature of the cured
polymers, was varied from 135 to 20°C. At the same time the state of the adhesive phase
at room temperature changed from glassy (storage modulus E' = 1.5 X 10" dyneslcm2)
to rubbery (E' = 10' dyneslcm'), and the ultimate tensile strength of the film decreased
monotonically. Especially when Epikote 828 content was rich, the curing reaction did not
proceed far enough, which was evident from the fact that there appeared somewhat anom-
alous behavior in the dynamic mechanical properties of the film at 55°C as shown in Fig.
8. The films were hard but brittle. However, the anomaly disappeared after the same sample
was kept for a few hours at elevated temperature, and the film became tougher.
In the practical case of wood adhesion, where epoxy resins are used as adhesives,
they must be cured at room temperature, although they are usually cured at extremely
high temperature in the case of adhesion of metals, etc. Figure 9 shows the datafor
adhesive tensile strength of epoxy resins cured at room temperatureas a function of
composition. Tensile strength was low either when the adhesive was very hard (Epikote
Adhesion and Adhesives 739
4.O
2
I 3.5
Y
f
0
C
c
2
3.0
0
C
0
m
0 io 20 30 40 50 60
Grafting efficiency (%l
FIGURE 5 Relation between grafting efficiency and adhesive tensile strength of cross-lap joint at
each level ofA?,$,.
87 1 -rich region) or when the adhesive was very soft (Epikote 87 1-rich region), and reached
a maximum when the adhesive was moderately hard (theintermediate composition of
Epikote 828Epikote 87 1 blends).
Similar characteristics have also been clarified by Ishii and Yamaguchi [6], who
adhered such adherends as steel,aluminum alloys, copper, polycarbonate, etc., with an
epoxy resin, and measured adhesive tensile strength as a function of both temperature and
deformation velocity. The master curve that they obtained shows a peak around U = lo4-
10' mm/min, and tensile strength became lower either at higher or lower velocity. This is
in agreement with the above-mentioned features of adhesive tensile strength.
B. AdhesiveShear Strength
In adhesive shear tests, the external force is applied to the adhesion system in the direction
parallel to the adhesive layer.
Motohashi et al. [4] measured adhesive shear strength of woods joined with the same
PVAc emulsions as they had used in the adhesive tensile tests, and obtained the results
shown in Fig. 10. Here again, maximum adhesive strength is seen around the glass tran-
sition temperature of the adhesive, and it naturally decreases as temperature is raised
further, which is a reflection of the decrease of molecular cohesion in the adhesive layer.
However, adhesive strength in shear tests at lower temperatures is two or three times larger
than in tensile tests in the same temperature range. Figure 11 shows similar data for epoxy
740 Mizumachi
50
LI
40.
5
m
C
L
L
30.
m
-.-E
e
101
v 0.81 - 1.54
-
0 10 20 30 40 50 60
Grafting efficiency (%)
FIGURE 6 Relation between graftingefficiency and ultimate tensile strength of PVAc emulsion
films.
resin adhesives, which were obtained by Hatano et al. [ S ] , and the same trends are ob-
served. These characteristics are commonly seen in many types of adhesives. It isvery
interesting to note that even in cases where such polymers as polystyrene and polyvinyl
chloride, which are not expected to have affinity to solid surfaces, are artificially used as
adhesives, adhesive shear strength for wood adherends is about 20-50 kgkm' at low
temperature, although that for metal adherends is almost zero. An example is shown in
Fig. 12. Of course, there are some cases where specific adhesion at the boundary surface
is so strong as to retain a state of intimate contact even when the internal stress due to
Monomer
Grafting
Molecular weight of benzene-soluble parth
conversion efficiency"
Emulsion ("/.) M,, x 10" M,,,x l 0 M%?
lM,8
A 54.4 4.06
5.12 20.8
B 96.2 12.2 2.18 6.52 2.98
"Percentage of unextracted PVAc amount with benzene to total PVAc in drled film.
hCalculatedfrom gel permeationchromatogram. h?,,, number-averagemolecularweight; h?,$, weight-average
molecular weight.
741
Adhesion and Adhesives
7.
Ernulslon A
6
5. 8
:i,:
I
a 7
FIGURE 7 Temperature dependence of adhesive tensile strength of cross-lap wood joints bonded
with two emulsion adhesives.
i
4
d
1. ~
Epikole828 Epikote87l
Weightfraction
FIGURE 9 Plot of adhesivetensile strength of cross-lap wood joints bonded with epoxy resin
adhesives against weight fraction of Epikote 871 in Epikote 828/Epikote 871 blends. Wood failure
is also shown.
the difference of thermal expansion coefficients of materials becomes great. For example,
Koizumi and Matsunaga [7] measured adhesive shearstrength of steels bonded with a
certain epoxy resin adhesive in the temperature range 20-80°C and in the velocity range
10-3-10" c d s . The master curve that they obtained by reducing the datato 20°C is
sigmoid in shape and there is no region where adhesive shear strength decreases as velocity
increases within their experimental conditions. Shimbo et al. 181 have studied a variety of
epoxy resins and obtained similar results. They found that as the adhesive becomes harder,
adhesive shear strength becomes greater, in spite of the fact that internal stress becomes
larger.
Mizumachi [9,10] has summarized the general features of wood adhesion in relation
to the dynamic mechanical properties of adhesives as shown in Fig. 13.
When the temperature of the adhesion test is lowered, when the rate of strain is
elevated, or when polymers of high glass transition temperature are used as adhesives
without plasticizers, adhesive strength for wood adherends has a certain positive value,
even when specific affinity at the boundary surface is not expected. This seems to be due
to the fact that adhesive anchors formed near the surfaces of woods can resist the external
force. This anchor effect or mechanical interlocking effect plays an important role in wood
adhesion when the adhesive isin glassy state. In adhesion shear tests, all the anchors
formed are effective for adhesive strength; in tensile tests, however, some fraction of the
anchors is withdrawn without any resistance, and that is the cause of the difference be-
tween adhesive shear strength and tensile strength. This region (E' > 10"' dynes/cm') is
Adhesion and Adhesives 743
2
ul
121 Emulsion B
0 . 0
.
D
' 0 0,
0 . - . .1 . . - . 1 . . . . ~ . " 1 . ~ - - - ~ l - @ - ~ l
-150 -100 -50 0 50 l00 IS0
Temperature ('C)
FIGURE 10 Temperature dependence of adhesive shear strength of woods bonded with two emul-
sion adhesives.
called A-region (A refers to -Anchors), and failure modes in this region are illustrated
schematically in Fig. 14.
When the test temperature is raised a little more or the rate of strain is lowered, or
when glassy polymers are plasticized so as to make E' of the adhesive between 10" and
10"' (mostly when 5 X 10') dyneskm', adhesion through the boundary surface between
adherend and adhesive becomes mostly effective, and in this region both adhesive tensile
strength and shear strength for woods as well as for metal adherends become maximum.
This region is called the B-region (B refers to -Boundary surface), meaning that adhesion
at the boundary surface predominates there.
When the adhesive becomes extremely soft (E' < 10' dyneskm?), failureoccurs
always in the adhesive phase, which means that adhesion at the boundary surface exceeds
the cohesive strength of the adhesives. This region is called the C-region (C refers to
Cohesion), because the experimentally measured adhesive strength is determined solely
-
by the cohesive strength of the adhesives.
C. Peel Strength
In peel tests, an adherend is largely bent and the external force is applied so as to pull it.
Therefore, it is not possible to perform peel tests on wood adherends.
744 Mizurnachi
I
I
10
1
Epikote828 Epikate87I
Weight fraction
FIGURE 11 Plot of adhesive shear strength of woods bonded with epoxy resin adhesives against
weight fraction of Epikote 871 in Epikote 828/Epikote 871 blends. Wood failure is also shown.
FIGURE 12 Adhesive shear strength of polystyrene as a function of temperature, where Kaba (0)
and aluminum ( 0 ) are used as adherends.
Adhesion and Adhesives 745
5x109 dyne/crn2
B region
-# E of adhesives
( Degree of plasticization
etc.
*-
m n
I
S hear Tensile
FIGURE 14 Comparison of failure modes for shear and tensile tests in the case when the adhesive
is in glassy state (A-region).
746 Mizumachi
P = k[-COOH]"
where k and n are constants ( n = 0.5-0.75). They believed that as the concentration of
the "COOH group increases in the adhesive, the probability of interaction between the
functional group and " O H group on the surface of the cellophane increases, and con-
sequently peel strength also increases. However, if cohesive failure must be partly involved
in adhesion tests, their discussion does not seem reasonable.
Mizumachi et al. [ 121 studied the influence of chemical structures and dynamic
mechanical properties of a series of methacrylic co-polymers on peel strength of cello-
phane. Monomers used in the study are listed in Fig. 15. One can easily prepare adhesive
polymers with various viscoelastic properties and with various concentrations of functional
groups such as hydroxyl groups and epoxy groups by combining these monomers. Mon-
omer mixtures are polymerized in contact with adherend in the glass cell shown in Fig.
16, and after polymerization is completed, a peel test is performed. Some of the typical
data are shown. Figure 17 shows the cases of co-polymers of 2-ethylhexyl methacrylate
(EH) and 2-hydroxyethyl methacrylate (HO) or glycidyl methacrylate (G). It is true that
?43
CHt' F cnZ= c
C.0 $90
0
FH2 42M25
Lauryl methacrylate
'4'9'2'5 (L)
2 Ethyl hexyl methacrylate
~
(EH)
P
F"2
P 2
OH
2 - Hydroxy ethyl methacrylate Glycidyl methacrylate
WO) (G)
FIGURE 15 Monomers.
Adhesion and Adhesives 747
UV light
111
monomers
cellophane
1.0
EH H0
G
FIGURE 1 7 Peel strength of cellophane as a function of co-polymer composition for EH/HO (0)
and EH/G ( 0 ) systems.
748 Mizumachi
120 I
0
0 0 .
8 0 0 0
0.5 1.0
E H0
C
FIGURE 19 Peel strength of cellophane as a function of co-polymer composition for E/HO (0)
and E/G ( 0 ) systems.
Figure 19 shows the cases of ethyl methacrylate and the two functional monomers
( H 0 and G ) . These co-polymers are glassy at room temperature over the whole range of
co-polymer composition as shown in Fig. 20. Peel strength is almost zero, in spite of the
fact that the concentration of functional groups increases to a great extent. Figure 2 1 shows
the case of lauryl methacrylate (L) and methyl methacrylate (M), where no functional
group is involved and only viscoelastic properties of the copolymers are varied, as shown
in Fig. 22. Here again, maximum peel strength is obtained when the temperature of the
viscoelastic absorption peak is near room temperature. Master curves of the rheological
functions of some copolymers were obtained and compared with the data on peel strength.
It was concluded that the viscoelastic properties of adhesives are the dominant factor in
adhesion. Peel strength becomes maximum when the viscoelastic absorption peak of the
adhesive at room temperature in the test appears at the frequency that corresponds to peel
velocity.
These characteristics are summarized as follows. The absolute value of the peel
strength can be somewhat different if the chemical structure of the copolymerization sys-
tem is different, but maximum peel strength is always obtained when the storage modulus
E‘ of the adhesive is about 10’ dyneskm’ as shown in Fig. 13. If Hofrichter et al. had
performed their experiments over a much wider range of concentration, they might have
come to somewhat different conclusions.
120
100 - 0.
0
n **.**-
u
0 0.
-80 * .m - 0"
60
be-" \o-o
h
8
::
W
"10
c
20 .
O0oA a A '
0.5 a ' a 1.0
E H0
G
FIGURE 20 Temperature of viscoelastic absorption peak at 100 Hz as a function of co-polymer
composition for E/HO ( 0 ) and E/G ( 0 ) systems.
L M
FIGURE 21 Peel strength of cellophane as a function of co-polymer composition for L h 4 system.
Adhesion and Adhesives 751
120
100 ’
n
u
-
0
-eo
60
n
I
0
: E
W
g40 -
/
0
Oak- ’ ’ ’
0.5 * m 1.o
L M
FIGURE 22 Temperature of viscoelastic absorption peak at 100 Hz as a function of co-polymer
composition for L/M system.
be classified as one of the adhesive tensile tests, only the latter is discussed here. It is
believed that rolling motion of a ball on a pressure-sensitive adhesive reflects tackiness of
the adhesive, and therefore rolling-ball tests have been employed in many countries for a
very long time. In J. Dow’s method of measuring rolling-ball tack, balls are rolled on an
inclined surface. A pressure-sensitive adhesive 10 cm in length is placed on a part of the
surface. Angle of inclination is 30°, and leading distance is 10 cm. If a ball is too large,
it may roll out across the pressure-sensitive adhesive zone and go down farther. Then a
little smaller ball is rolled, and so on. If a ball of a certain diameter, namely, 11/32 in.,
stops within the pressure-sensitive adhesive zone, tack of the pressure-sensitive adhesive
is expressed as “ball number 17.” In the PSTC-6 method, a ball of 14/32 in. diameter rolls
down on an inclined path and onto a horizontal surface of a pressure-sensitive adhesive.
Here, tack of the pressure-sensitive adhesive is expressed in terms of the rollout distance,
because the reciprocal of it is considered to be proportional to the tack of the pressure-
sensitive adhesive.
These ways of expressing tack are useful in some practical cases, but the physical
meaning of the value is not necessarily clear. If the angle of inclination is different, or if
the length of path on a rigid substrate is different, the ball number or the rollout distance
might be different for the same pressure-sensitive adhesive.
It is hoped that a method can be developed by which tack of pressure-sensitive
adhesive can be expressed in terms of significant physical meaning. Mizumachi [ 14,151
pointed out that tack must reasonably be described in terms of the rolling friction coeffi-
752 Mizumachi
cient ( f ) of the pressure-sensitive adhesive, because rolling motion of a ball can be ex-
pressed by a set of equations of motion, where f is involved and ,f does not depend on
such parameters as the angle of inclination of the surface or the leading distance, but
depends only on the physical properties of the materials on which the ball rolls. He solved
the equations of motion of a ball, assuming that
f = 40 + 4 l V
and derived the following equations:
(x - x(,),, = [(7R(R sin a - cos a)/(5g4?cos’a)]
X log[(R sin a - (+(, + ~ I v , ) c o sa)l(R sin a - 4(,cos a ) ]
+ 7Rv,,/(5g4f COS a )
where (x - x(,),, is the rollout distance of a ball on a pressure-sensitive adhesive, R is the
radius of a ball, a is the angle of inclination of the surface, and v,, is the initial velocity
of the ball in the pressure-sensitive adhesive zone, which is given as follows:
v,, = ( X - X(,)”’( 1Og/7R)”’(R sin a - A) cos a)”’
where ( X - X(,) is the leading distance, and.f;, is the rolling friction coefficient of the rigid
substrate, which is nearly equal to zero. Urushizaki et al. [ 161 made extensive experiments
on rolling motion of a ball on pressure-sensitive adhesives, and showed that the data can
be analyzed successfully according to the above equations. A typical example is given in
Fig. 23. Values of the two parameters, 4” and 41,are determined so as to minimize the
sum of the deviations of experimental data from the theoretical curves. Agreement between
the theoretical curves and the experimental data is satisfactory. In this case, +(, = 0.67 cm
and = -0.0043 S , regardless of the leading distance or the angle of inclination of the
surface.
x-x0 (CM 1
FIGURE 23 Plot of rollout distance against leading distancc o f a ball
Adhesion and Adhesives 753
FIGURE 24 Rollingcylindermethod
Mizumachi and Saito [ 171 also clarified that rolling processes of a ball can be ana-
lyzed by the theory over the whole range of velocity, and at the same time rollout distance
can also be analyzed by the same theory. In any case, one must perform a rather compli-
cated analysis in order to determine the values of ,f (or 4,)and because velocity of a
ball changes at every moment and at the same time ,f varies as a function of velocity.
However, Mizumachi [ 14,15,18] point out that it we adopt the pulling cylinder method as
illustrated in Fig. 24, we can easily determine the value offwithout any elaborate analysis.
Suppose that a cylinder of radius R, length 0,and weight M g is pulled by a force P
on a horizontal plane of a pressure-sensitive adhesive at a constant velocity v. Then f of
the material is given as
This type of experiment and also the analysis can be easily made, and we do not
need any postulate concerning the dependence offon U . If we want to know f as a function
of 71, we only have to pull a cylinder at several constant velocities.
Naturally, .f‘depends on the physical properties of a pressure-sensitive adhesive, and
if the value off is plotted against log 7~ for a viscoelastic material over a very wide range
of velocity, it is expected that some curve will be obtained that increases from a relatively
low value to a certain maximum and then decreases as the velocity becomes greater. A
typical example is given in Fig. 25.
If we measure stress and strain at break (c/,, c,,)of a viscoelastic material, and plot c,
against E/,, a curve of common shape such as given in Fig. 26 is obtained. This is the
“failure envelope” of Smith [ 191. Several attempts 120-241 have been made to interpret
the failure envelope theoretically, among which Hata’s theory [23] is the simplest. He
showed that the failure envelope of viscoelastic materials (adhesives) can be reproduced
if we choose a simple mechanical model, a parallel connection of two Maxwell elements,
and assume some failure criteria. Parameters concerning the elements in the model are as
shown in Fig. 27. Failure starts in the weak point (the tirst Maxwell element), and then
the residual part (the second Maxwell element) breaks down, carrying the whole load. The
failure at the first Maxwell element may occur if either of the following two conditions
is fulfilled:
I. Strain of the spring ( E ~ , reaches
) a certain critical value (cIIC).
This corresponds
to the region where strain rate ( r ) is high or temperature ( T ) is low.
754 Mizumachi
lr
FIGURE 25 Plot off measured by pulling-cylinder method against log v for a pressure-sensitive
adhesive.
-
Strain
Results of the numerical calculations, substituting appropriate values for the param-
eters,are shown in Fig. 28, where characteristic features of the failureenvelopeare
reproduced. It is possible to show according to this theory that the failure point (U!,, E!,)
moves counterclockwise along the failure envelope as the temperature is lowered or the
strain rate is increased. In the case of failure of an adhesively bonded joint, the same
treatment must be possible in the region where cohesive failure occurs. However, it is
evident that an additional failure criterion is needed in order to interpret the interfacial
failure. Hata [24] postulated the third criterion: 111. Interfacial failure occurs when energy
stored in the springs of the model (W) reaches a certain critical value (Wc-):
d
“Q
c- -
””“”
&
FIGURE 28 Explanatory diagrams of the model theory for the failure envelope.
756 Mizumachi
Shear
l
Tens i l e
FIGURE 29 Comparison of shear and tensile deformation:
Shear Tcnsile
Strain E =d h E = .r/h
Because we have to select the failure mechanism with the lowest m,, at every velocity,
the solid line in the figure expresses the overall adhesive shear strength as a function of
velocity. According to the rheological principle, larger velocity is equivalent to lower
temperature, or vice versa, and therefore we can imagine how adhesive shear strength
depends on temperature at a constant velocity. Curves I, 11, and 111 in Fig. 30 correspond
to the C-region, B-region, and A-region in Fig. 13, respectively. If we want to take into
758 Mizumachi
10-
il
\‘
0-
\’4
-
g
.
U?
P
-
m
Y
(L
-
6-
4-
h
2-
FIGURE 32 An example of the results ofthe numerical calculations of peel strength, which is
plotted against log 7). Values of the parameters are as follows: b = 1.5 cm, h = 0.025 cm, W,, = 5
X 10’ dyneskm’, E , = 10“’ dyneslcm’, 77, = 10’ poise, E’ = 10“’ dyneslcm’, 77’ = 10’ poise, q I C =
0.02, = 0.005, W , = 10’ ergslcm’. It is assumed that a suddenjumpfromcurve I tocurve 111
occurs at the same strain rate as in the case of shear test.
interesting to notice that the absolute value of P is very low (order of magnitude of 10”
kg/1.5 cm), in spite of the fact that the adhesive shear strength for the same model is
about 10’ kg/cm?. A second interesting point is that the velocity corresponding to maximum
peel strength is much lower than that corresponding to maximum shear strength, which is
in good agreement with the previously described fact that adhesive shear strength becomes
maximum when E‘ is about 5 X 10’ dyneslcm’, while peel strength becomes maximum
when the adhesive is much softer (E’ = lo8 dyneslcm’).
v (cmlsl
FIGURE 34 An example of the results of the numerical calculations of,f. which is plotted agianst
I t has been clarified in the previous section that the values of adhesiveshearstrength,
adhesive tensile strength, and ped strength can be calculated theoretically as a function
of rate of deformation, according to the mechanical model proposed by Hata 123,241. If
we postulate an additional criterion concerning the abrupt transition from cohesive failure
to interfacial failure, we can construct a curve having a peak at some rate for each adhesive
strength. The rate at which peel strength becomes maximum is lower than that at which
both adhesive shear strength and adhesive tensile strength do. The peak in each adhesive
760 Mizumachi
strength shifts toward the higher-rate side as the relaxation time of the viscoelastic material
(adhesive) decreases. or vice versa. In other words, for a particular adhesive, the peak
shifts toward the higher-rate side when the temperature of measurements is raised, and
when we compare the curves for various adhesives at some fixed temperature, an adhesive
of lower T, will have its maximum adhesive strength in a higher-rate region. These features
are in agreement with most of our experimental observations so far. However, there is a
large discrepancy, that is, adhesive tensile strength is calculated to be much greater than
adhesive shear strength, whereas the experimentally measured values are generally in the
reverse order. In spite of this discrepancy, this simplified theory may be useful in describing
qualitatively the complicated phenomena of adhesion, but we have to think seriously of
the case of the discrepancy.
Of course, the biggest problem is stress concentration. It is postulated in the above-
mentioned theory that stress and strain are uniform within the specimen, but this is not
true especially in case where the modulus of the adhesive is high. When an external force
is applied to the adhesive joint, stress is extremely concentrated at the edge or corner of
the adhesive layer, and failure of the joint initiates there and propagates along the bond
line as the force reaches a critical value.
Stress concentration within a material and its fracture are the major objects of the
fracture mechanics which has been developed in the fields of solid materials such as glass,
ceramics, hard plastics, and others, and has also been applied to the fracture of adhesive
joints [27-391.
In fracture mechanics, the critical strain energy release rate G,. is an important pa-
rameter, a measure of the toughness of the material, and is defined as follows:
G,. = (2)(g)
where. P,. and h refer to load at failure and width of the specimen, respectively, and ( K /
ilA) is a partial differentiation of compliance (C) with respect to crack length ( A ) . G , is
the energy required to increase the unit area of the crack surface in the specimen, and is
different if the deformation mode is different. There are three typical deformation modes:
mode I or tensile-opening mode, mode I1 or in-plane shear mode, and mode I11 or antiplane
shear mode, which are illustrated schematically in Fig. 35. Lim et al. [38,393 measured
G, for three deformation modes: G,,. for mode I, GIICfor mode 11, and GI,,,. for mode 111,
using double-cantilever beam specimens of wood joints by the compliance method. Char-
acteristics of both adhesives and adherends are listed i n Tables 3 and 4, respectively.
Adhesion and Adhesives 761
The experimentally obtained GC values for the joints. where the commercially avail-
able adhesives are used, are summarized in Table S. It is shown that G,,. < G,,,,. < G,,,.,
and this tendency is in agreement with that found in the literature 134,361. An interesting
question is what kind of correlation exists between G, value and the corresponding con-
ventionaladhesivestrength.Ancxampleisshown in Fig. 36. whereadhesivctensile
strength of a series of the joints is plotted against the square root of G,,, because they are
obtained by the tests of similar deformation mode, and G,, is proportional to the square
of the load at failure. The correlation coefficientin this case is not necessarily high enough,
but if well-characterized adhesives are employed, a moresignificantcorrelation will be
found. It is quite natural to think that the correlation depends o n the state of molecular
motion of the adhesive (i.e., glassy state, transient state, rubbery state, or fluid state), and
systematic data are being accumulated 1401.
A rheological approach as well as a fracture mechanical approach will be necessary
i n order to clarify the mechanism of adhcsion.
Young's modulus
Specilic gravity Moisture ( 1 Or kgl/cm')
content
Adherends Air Dry (%) E , . E,, Ell,
TABLE 5 Critical Strain Energy Release Rates for Modes I, II, and I11
GllC
Adhesives G,, (kgf cm/cm') G,,,,. GIICGC GlllC/Gl,
Adhesion involves a variety of factors, among which structures and propertiesof adhesives
and/or adherends are most important, because failure of the materials is always involved
in adhesion tests. Takayanagi [41] has pointed out that physical properties of a material
( Z ) are generally expressed in terms of two variables, namely, the state of aggregation ( X )
of molecules and the state of molecular motion (Y). The variable X represents such aspects
of material ascrystalline/amorphous,orientedhonoriented, and homogeneous/heteroge-
neous, as well as superstructures. Information related to X can be obtained through studies
utilizing electron microscopy, light-scattering observation, and X-ray diffraction measure-
ments. The variable X also includes thermodynamic measurements with respect to com-
patibility among the components of the material, and phase separation.
1o+"----
On the other hand, the variable Y represents such aspects as the micro-Brownian
motion of molecular segments and some local modes of molecular motion. Information
concerning these aspectscan be obtained through such observationsas viscoelasticity,
dielectric characterization, and NMR. Let us assume, for example, that we have two ma-
terials, both of which are completely amorphous and located at the same point on the X
axis. If the micro-Brownian motion of the backbone chains is restricted in one material
and nonrestricted in the other-in other words, if they are located at different points on
the Y axis-then the two materials will show completely different physical properties.
The former will be in a glassy state with modulus E' of the order of IO"' dynes/cm', while
the latter will be a rubbery material with E' of about lo7 dynedcm'.
If we take two materials, each having a two-phase structure, the overall material
characteristics of a system consisting of a continuous phase in a glassy state and a dis-
persed phase in a rubbery state would differ extremely from those of the other in which
the state of each phase is reversed. This indicates that a conclusion with generality cannot
always be achieved if relations of the adhesion performance with either of the variables
X and Y are investigated separately. And if some data on the practical performance of
adhesion can be connected to Z as a function of X and Y , it will enhance the understanding
of adhesion phenomena from the viewpoint of polymer science. The so-called adhesion
theories and adhesion principles presented in the past may be said to have clarified some
theoretical or empirical laws to connect practical adhesion performance to physical prop-
erties of materials. Even in the case where a law is only an empirical one, it will not only
be helpful for the efficient development of new materials, it will also serve to stimulate
theoretical advancement in the future if the law is expressed in terms of Z =,fix,Y ) .
In this chapter, adhesive strengths of various kinds are connected to dynamic me-
chanical properties of an adhesive, and it became evident that there exists some common
tendency, which was interpreted by a simplified rheological theory.
We have come to know that although adhesion is a very complicated phenomenon,
lots of common elementary processes are involved in various aspects of adhesion. If we
continue to accumulate data on adhesion systematically, using well-characterized adhesives
(molecular-characterized as well as material-characterized adhesives), the mechanism of
adhesion will be clarified scientifically in more detail.
REFERENCES
1 E. A. Davies, Adhesion t r n d Arlhesives, Furldnrrlentnls rrnd Prcrctice, Society of the Chemical
Industry (1954).
2. Seccyczk~c-Riron to Oqo, Kobunshi Gakkai ( 1 959).
3. K. Motohashi, B. Tomita, and H. Mizumachi, Holdi~rsch..36: 183 (1982).
4. K. Motohashi, B. Tomita, H. Mizumachi, and H. Sakaguchi, Wood Fiber Sci., 16: 72 (1984).
5. Y. Hatano, B. Tomita, and H. Mizumachi, Moklczcri Gakkaishi, 29: 578 (1983).
6. H. Ishii and Y. Yamaguchi, J . Adlwsiorl Soc. Japan, l / : 59 (1975).
7. S. Koizumi and T. Matsunaga, J. Adkesior~Soc. Jtrparl. 6 : 437 (1970).
8. Seccynkrc Hcrrldbook, Nikkan Kogyo ( 1980).
9. H. Mizumachi, Mokuzcri K o g y , 36: 3 (1981).
10. H. Mizumachi, Mokuzai Kogyo, 36: 57 (1981 ).
II. C. F. Hifrichter and A. D. McLaren, / m f . Eng. Chern., 40: 329 (1948).
12. H. Mizurnachi, M. Tsukiji, Y. Konishi, and A. Tsujita, J. Adhesior? Sor. J t p t r r l , 12: 378 (1976).
13. J. Johnston, A d h e s i ~Age,
~ ~ 26: 34 (1983).
14. H . Mizumachi, Zcriry Gijutsu, 2: 72 (1984).
IS. H. Mizumachi, J . Adhesiorl Soc. h p m t , 20: 522 (1984).
764 Mizumachi
Hiroshi Mizumachi
The University of Tokyo, Tokyo, lapan
1. INTRODUCTION
Because pressure-sensitive adhesives (PSA) are used to bond one material to another, they
can be regarded as a kind of adhesive, but in industry they are treated substantially as if
they were materials which are different from the so-called adhesives other than PSA. The
industrial societies are organized separately, and various statistical data are collected sep-
arately. If we look at the situation in detail, we come to know that there surely are some
differences between PSA and other adhesives. First, other adhesives are converted from
liquid state to solid state after they are applied on the surface of the adherends, but PSA
is never hardened. Second, other adhesives are usually supplied to the consumer as liquids
or solids (hot-melt adhesives), but PSA is supplied as a PSA product in which PSA is
coated on some film material. Usually, PSA itself, eitherasa solid or a liquid, is not
available commercially. The most abundant PSA products are PSA tapes, which are used
in the fields of packaging, office uses, electronics, vehicles, buildings, medicine, etc. The
next abundant PSA products are PSA labels or decals, for which no explanation will be
needed about the labels. Decals are sometimes called “sticking paints,” meaning that the
necessary information is printed on a film and PSA is coated on the back side of the print
so that one can stick them anywhere one likes. Decals can be conveniently used instead
of coatings. (For example, in many cases decals are applied on the outsides of cars and
trains or on the walls and windows of buildings.) Many kinds of PSA products for medical
uses have been developed recently. The industrial scale of PSA-related fields in Japan in
1990 is shown in Fig. 1 [ 1,2]. The total scale of the PSA industry was 452,000 million
Y (about $3,480 million U S . ) . On the other hand, the total industrial scale of adhesives
other than PSA in the same year was 246,000 million Y (about $1,900 million U.S.),
according to the statistics of the Industrial Society of Adhesives in Japan. Scale ofthe
former approximately doubles that of the latter. The amount of polymers used as PSA is
about 10%ofthatused as adhesives other than PSA, but nevertheless PSA has such a
large industrial scale not simply because it includes the cost of various substrates such as
papers, fabrics, plastic films, metal foils, etc., but because PSA has a tremendous function
of combining chemical components of PSA and substrates. Values of polymers in PSA are
added greatly.
765
I
766 Mizumachi
60.0%
l4.0%
- 36.0%
42.0%
There are several kinds of PSA, such as rubber-based PSA, hot-melt PSA, acrylic
PSA, silicone-based PSA, and others. Rubber-based PSA includes natural rubber and syn-
thetic rubber, but the former is much more used than the latter. Because rubber alone is
not sticky enough as PSA, we have to blend some tackifier resins with the rubber. It must
be pointed out that solvents must be used in the manufacturing process of the rubber-
based PSA. Polymers mostly used in hot-melt PSA are SIS, SBS (block co-polymers of
styrene-isoprene-styrene or styrene-butadiene-styrene) and SEBS (hydrogenated SBS).
Tackifier resins must be blended with these block co-polymers, too. No solvent is needed
for these co-polymers either in the blending process or in the coating process. This is why
these systems are called hot-melt PSA, and this is a great advantage which this type of
PSA has. There are many kinds of acrylic PSA because they are produced by combining
various acrylic co-monomers (acrylates and/or methacrylates), some of which are made
by solution polymerization and others by emulsion polymerization. Everybody thinks that
any industrial products must be environmentally friendly, and therefore in the field of
PSA, people have tried to reduce solvent-based materials as much as possible, which is
the reason why the production of water-based (emulsion) PSA is gradually increasing.
Tackifier resins used not to be blended with acrylic co-polymers because it is quite easy
to make polymers with any T, and modulus by controlling the kind and composition of
co-monomers. Recently, however,there have been many cases where tackifier resins have
been introduced in this type of PSA for the purpose of modifying the adhesion properties
against polyolefins. Silicone-based PSA is a blend of silicone rubbers and silicone resins.
Because surface tension of this series of polymers is low,they easily stick to Teflon,
polyimides, silicone rubbers, etc., which are known to be inert to most adhesives. At the
same time, this type of PSA is resistant to heat, chemicals, weathering, etc.
All these blends are coated on papers, cellophane, fabrics, plastic film, or metal foil
to produce PSAtapes, labels, or decals, and theyare abundantly used in industry, offices,
in
and in homes.
Pressure-Sensitive Adhesives 767
Figure 2 [1,2] shows the market shares of both PSA products (substrate plus PSA) and
PSA itself in 1990 in Japan. Natural rubber, which is a forest product, has the biggest
share: 43.3% by area of PSA products (tapes and labels) and 53.3% by weight of PSA
itself. SBR (random co-polymers of styrene and butadiene), which is the most popular
synthetic rubber, is used at about 10% of natural rubber as PSA. Block co-polymers such
as SIS, SBS, and SEBS, which are used as hot-melt PSA, are sometimes called thermo-
plastic rubbers, and there are cases where they are classified as one of the rubber-based
PSAs. Anyway, the amount of these block co-polymers used as PSA is much lower than
that of natural rubber. Recently, the legal regulations on environmental problems have
become more andmore severe, and manyresearchers have been seriously trying to develop
some production systems where solvents are not needed. However, according to the an-
ticipation by PSA specialists [3], natural rubber will continue to be used for some period
of time, in spite of the fact that we cannot avoid the solvent problem in this system. This
may be due not only to the low cost, but to the tremendous accumulation oftechnological
data, and also to the fact that this type of PSA has delicate performance which the other
PSA do not have.
It must be pointed out that there is a great difference between the solvent problem
in PSA and that in other adhesives or coatings. In the case of adhesives (other than PSA)
or coatings, solvent will vaporize into the air anywhere they are used. On the other hand,
in the case of PSA, solvent is used within the plant, andwhen the PSA products are
transferred to consumers, the solvent problem does not occur at all. If the solvent is
2500 100
2000 80
(Y
E 1500 60
C c
.-
-
0 0
c,
Y
1000 40
500 20
0 0
PSA PRODUCTS PS ADHESIVES
FIGURE 2 Polymers usedas PSA in Japan in 1990 NR, natural rubber; SBR,random co-polymers
of styrene and butadiene; SIS, styrene-isoprene-styrene block co-polymer; Sol.Acryl., solution of
acrylic copolymers; Em.Acryl., emulsion of acrylic copolymers.
I
768 Mizurnachi
recovered effectively within the plant as is the case in most of the PSA industry, there is
almost no problem. Nevertheless, much effort has been concentrated to developsome
manufacturing processes without any solvents, such as emulsion techniques, hot-melt tech-
niques, radiation-cure techniques, and others.
As mentioned earlier, the modulus and viscosity of rubber itself are usually too high
for PSA, and we have to add some tackifier resins to the system. The most popular tackifier
resins which have been used in PSA are rosins and terpene resins. Rosin is harvested from
pine trees, and its main component is abietic acid. Terpene resins are polymers of CY-
pinene, P-pinene, dipentene, etc. Molecular weight of the resins is around 1000.
There are many kinds of tackifier resins because both rosin and polyterpenes are
chemically modified in various ways. They are hydrogenated, dehydrogenated, dimerized,
or polymerized further, esterified with glycol, glycerol, or pentaerythritol. In some cases,
they are co-polymerized with phenolic compounds. Most of the resins are solid, although
there are some liquid tackifiers. The solids are very brittle, and if mechanical shock is
given to them by a hammer, they are tinely divided into very small fragments or powders,
but if we blend the resins with natural rubber, the viscosity of the system becomes ex-
tremely low, as shown in Fig. 3 [4]. Thisseems to be an anomalousphenomenon. If
molecules of natural rubber and tackifier resins share the free volumes within the material,
then the viscosity of the blend must vary monotonically with the blend ratio [ 5 ] .How to
analyze this phenomenon on a scientific basis is a great problem which has not been
l O’O
I
Q
In
lo9 ‘g v
>.
.-
c,
v)
n
0
io8 g
:
€
Y
v
1
5
c,
C
Jz Q
L
G
c
107 g
2 2
3-;
Q)
Q) 0.5 1 o6
a -Et Polyethylene
0
0 10 20 30 40 50 60 70 80 90 100
Resin Content (%>
FIGURE 3 Viscosity and peel strength of natural rubber-based PSA (a blend of naturalrubber
and polyterpene resin Piccolyte S 125).
Pressure-Sensitive 769
solved yet. Practically,PSA specialists are interested in the blend ratio where the viscosity
and modulus are extremely low. There are tackifier resins -which are synthesized from
petroleum, besides those which come from forest resources, and all of themare competing
in the industrial markets. Table 1 [6] shows the classification of the tackifier resins.
Thus, the fact that not only natural rubber, but also such tackifier resins as rosin
derivatives and polyterpene derivatives, which originate from forest products, are used
abundantly in industry will be attracting the interest of many forest products researchers.
In addition, it must be pointed out that a large amount of kraft paper is used as backing
material for PSA, as shown in Fig. 4. For example, OPP (oriented polypropylene) film is
mostly used inthe United States as the backing material for packaging tapes, but inJapan,
kraft paper is used in the same area as the majority of all the film materials.
,Others ,Others
(4 (W
FIGURE4 Materials used as backings of the PSA products. (a) Fraction by area of PSA products.
(b) Fraction by money of PSA products.
770 Mizumachi
PSA tapes stick easily to any material upon light touch, which is the most important
property of PSA. They are not usually expected to show strong adhesive strength, contrary
to the case of structural adhesives. In some applications low adhesive strength is a great
merit of the products. On the other hand, there are some PSA specialists who are trying
to develop PSA products which stick to various adherends like ordinary PSA does, and at
the same time can be used in some structural applications. So, there are tremendous num-
bers of PSA products which are commercially available now.
The practically important properties of PSA are sometimes called “the three fun-
damental PSA performances.” Testing methods are standardized by the JIS (Japan Indus-
trial Standards), ASTM (American Society for Testing Materials), PSTC (Pressure Sensi-
tive Tape Council), and others. Some of the typical testing methods are shown in Figs.
5-7 [ 6 ] .
A. Adhesion
“Adhesion” expresses a degree of adhesion of PSA under a normal condition, and it is
measured by 180” peel test after PSA tape is thoroughly adhered on an adherend. Theo-
retical approaches have been tried by many PSA researchers [ 7 ] .
B. Tack
“Tack” means the instantaneous adhesion of PSA. There are three types of tack test. The
first one is called the rolling-ball method, where steel balls are rolled on PSA, and a rollout
distance or some related quantities are regarded as a measure of tack of PSA. Mizumachi
et al. [8-191 analyzed the rolling motion of a ball on PSA according to the equation of
motion of a solid ball, and proposed that the rolling friction coefficient f of PSA must be
1 PSA tape
“ 0 0
Steel balls, diameter increasing in 1 /32” steps
taken as a measure of tack. The second method is called the “quick stick” or “loop tack”
method, where a looped tape with the sticky surface facing outside is hung in a tube, and
upon light touch on an adherend it is pulled away. The measured resistance is a measure
of tack. The last test is the “probe tack” test. This is a simulation of the finger test. An
end surface of a cylinder touches the PSA surface and moves away. The resistance is
regarded as the probe tack of the PSA tape.
into account the balance of the three performances which willbe appropriate for the
specific applications.
For example, in the case of memorandum notes which stick on anything, the nec-
essary conditions are that some degree of tack is needed and at the same time the adhesive
strength (peel strength) must not be high so as not to withdraw any fibers from the paper
to which it sticks. Similarly, in the case of bandages where human skins are expected to
be the adherends, the peel strength must be controlled at a relatively low level in order to
avoid damaging the skin. On the other hand, in the case of packaging tape or industrial
tapes for semistructural uses (e.g., VHB), both peel strength and holding power must be
especially great. If the practical requirements for a variety of applications are clarified
quantitatively, we can control the appropriate levels of the three PSA performances by
Pressure-Sensitive Adhesives 773
0 Sheaf
e Peel -
-
E
* Quick lack .-“Ine
-3 m
8
all
U
-2 g
.-
VI
L
L
4
-1 $
I I
-40-30-20-10 0 +l0 +20
FIGURE 8 Relation between thethree fundamental PSA Performances and 7‘, of PSA
choosing chemical structures, composition, molecular weight, and tackifier resins and their
concentrations. There are many empirical guidelines which have been proposed by some
well-known specialists [2 l].
However, if we want to understand the phenomena related to PSA on a scientific
basis, we are usually confronted with some difficulties because the empirical rules of PSA
are expressed in terms of the PSA performances measured within a very narrow range of
experimental conditions (temperature, time scale, stress level, etc.), and we cannot analyze
them according to the principles of physics, physicochemistry, surface chemistry, and/or
rheology. Experimentally evaluated quantities of PSA performances vary systematically in
accordance with the change of the experimental condition, and it is recommended that the
PSA performances be measured over a wide range of experimental conditions, and the
results compared with the physical properties of PSA such as viscoelastic properties, or
other structure/property of the materials. Accumulation of such data will result in clarifi-
cation of the mechanisms of PSA phenomena.
The PSA performances depend greatly on the viscoelastic properties of the PSA, which
may suggest in turn that PSA of the same viscoelastic properties will have the same PSA
performances, but actually the PSA performances are delicately different from one PSA
to another. Here, the rheological aspects of the PSA performances of acrylic PSA are
compared with those of natural rubber-based PSA.
curves are shifted along the temperature axis by the difference of T,, all the curves can
be superimposed. Similar behavior is seen in a series of co-polymers. This means that the
modulus at the rubbery plateau region is higher for a polymer of higher T,q.
On the other hand, Figure I O shows similar plots for a natural rubber-based PSA
which is a blend of natural rubber and glycerol ester of hydrogenated rosin. Crossing over
of the curves is seen in the figure, i.e., the modulus at the plateau region is lower for
higher T,. These trends are sometimes seen in natural rubber-resin systems.
l
2000
E
$1500
L
p 1000
% 500 I
e
0
? * r r r
10 -
2
2
%5-
properties of PSA and substrate materials, but detailed study is needed to clarify some
delicate differences between PSA systems based on natural resources and those based on
synthetic polymers.
V. CONCLUDING REMARKS
It has so far been mentioned that the industrial scale of PSA is approximately double that
of the adhesives other than PSA, that the fractions of forest products such asnatural rubber,
rosins, polyterpene, etc., used as PSA components are very large, and that a tremendous
amount of paper is also used as backing material for tapes and labels. However, very few
researchers in the field of forest products science are interested in PSA. The reason might
be that usually PSA products are not used abundantly in the wood industry. But it must
be pointed out that the study of PSA is indeed an important part of forest products science,
for which the industrial utilization of forest products is the main aim of research.
Especially in the case of natural rubber-based PSA, we have to study the blends of
natural rubber and tackifier resins, and it is quite natural to think that the PSA performances
depend greatly on the physical or physicochemical properties of the blends, such as mis-
cibility between the components, viscoelastic properties of the blends, surface chemical
properties of the materials, and so on. Variation of the chemical structure and composition
of gum and resin may have very complicated influences on various aspects of PSA phe-
nomena, but it is ratherdoubtfulwhether the structure/properties of the blends or the
components, which originallycamefrom woods, have been fully characterized on the
basis of forest products science.
Some PSA specialists used to speculate that the reason natural rubber is compatible
with most of the tackifiers made from forest products is that they initially existed in the
woods. Of course, this explanation is quite emotional rather than scientific, which indicates
that there is a lot of room left in the field of PSA for researchers of forest products science.
They are expected to approachPSA science from the standpoint of forest products science.
REFERENCES
1. K. Fukuzawa, AFERA Conf., Amsterdam (1991).
2. K. Fukuzawa, European Adhesives and Sealants, p. 19 (March 1992).
3. For example, D. Satas, AFERA Conf., Amsterdam (1991).
4. K. Fukuzawa, J . Adhesion Soc. Japan, 6 : 441 (1970).
5. F. Bueche, Physical Properties of Polymers, Wiley, New York (1970).
6. Japan Adhesive Tapes Makers Association, Handbook of PSA (1995).
7. T. Saito, J. Adhesion Soc. Japan, 2 f : 220 (1985).
8. F. Urushizaki, H. Yamaguchi, and H. Mizumachi, J. Adhesion Soc. Jupun, 20: 295 (1984).
9. H. Mizumachi, Matex Technol., 2: 72 (1984).
10. H. Mizumachi, J. Adhesion Soc. Japan, 20: 522 (1984).
11. H. Mizumachi, J. Appl. Polymer Sci., 30: 2675 (1985).
12. F. Urushizaki, H. Yamaguchi, and H. Mizumachi, Yakugaku Zasshi, 106: 491 (1986).
13. H. MizumachiandT. Saito, J . Adhesion, 20: 83 (1986).
14. H. Mizumachi and Y. Hatano, J . Adhesion, 21: 251 (1987).
15. H. Mizumachi and Y. Hatano, J. Appl. Polymer Sci., 3 7 3097 (1989).
16. T. Tsukatani, Y. Hatano, and H. Mizumachi, J. Adhesion. 31: 59 ( I 989).
17. F.Urushizakiand H. Mizumachi, Chem. Pharm. Bull., 3 9 159 (1991).
18. T. Tsukatani, Y. Hatano, and H. Mizumachi, J. Adhesion Soc. Japan, 27: 217 (1991).
19. T. Tsukatani, Y. Hatano, and H. Mizumachi, J. Adhesion Soc. Japan, 2 7 217 (1991).
20. J. W. Hagan, C. B. Mallon, andM. R. Rifi, Adhesives Age, 29: (March 1979).
21. H. Mizumachi, in Adhesion Society of Japan (ed.), Handbook of Adhesion, p. 257 (1996).
22. A. Zosel, in Advances in PSA Technology-2, p. 92 (1992).
23. T. Tsukatani, T. Hata, Y. Hatano, H. Mizumachi, and R. Ramharack, Proc. Eurocoat Congress,
Nice, France, p. 319 (1991).
24. T. Hata, T. Tsukatani, and H. Mizumachi, J. Adhesion Soc. Japan, 30: 307 (1994).
This Page Intentionally Left Blank
20
Wood-Inorganic Composites as Prepared
by the Sol-Gel Process
Shiro Saka
Kyoto University, Kyoto, japan
1. INTRODUCTION
Wood has been used by mankind since ancient times. However, it has some defects as a
natural material. A recent trend in wood research is to remove such defects and add the
new value to woody materials. A study of wood-inorganic composites is also one of its
trials and has been actively carried out to develop new functional woody materials.
In a study of wood-inorganic composites, the double-diffusion process of barium
phosphate must be mentioned. This has been developed at the Wood Research Institute,
Kyoto University, in collaboration with Central Research Laboratory, Matsushita Electric
Works Ltd. From a hint that old electrical wood poles buried in the ground cannot be
burned out, development of fire-resistant wood has been started, trying to improve dimen-
sional stability and biodeterioration together with fire-resistance [ 1,2].
Different from this process, we have developed a method using a metal alkoxide to
prepare wood-inorganic composites by sol-gel process. For this preparation, efforts have
been made to add new value to the wood without losing its characteristic properties such
as porous structure [3- 131.
A basic principle of the sol-gel process involves hydrolysis of the metal alkoxide M(OR),,
(M = Si, Ti, Ba, Zr, etc.; R = alkyl group; n = oxidation number) with water and subsequent
polycondensation by dehydration or dealcoholation reaction to produce metaloxane sols
as described below.
M(OR),, + xH,O + M(OH),(OR),,_,+ xROH
Dehydration reaction:
-M-OH + HO-M- + -M-O-M- + HZ0
Dealcoholation reaction:
-M-OH + R-0-M- + "-0"- + ROH
781
782 Saka
The sols formed are then solidified as wet metaloxane gels in a temperature range between
25 and 80°C and then heated to 120°C and dried [14]. In an ordinary sol-gel process, the
gelled masses are further heated to 800°C to prepare glasses. However, the gels in this
study must be treated under the temperature at which wood is not thermally degraded.
Therefore, in a reaction medium of the metal alkoxide/alcohol/acetic acid (catalyst), the
moisture-conditioned wood or water-saturated wood was soaked at ambient temperature
under reduced pressure or atmospheric pressure. The water present within the wood cells
initiates the reaction of the hydrolysis and polycondensationof metal alkoxide.The soaked
wood was subsequently treated at a temperature between 50 and 60°C for 24 h, and at
105°C for another 24 h to prepare wood-inorganic composites [3,4].
111. MONOCOMPONENTWOOD-INORGANICCOMPOSITES
A. Distribution of Inorganic Substances inWood Cells
Recent studies have indicated that the distribution of inorganic substances in wood cell
walls is dependent on a combination ofthe metal alkoxide species and the water-retaining
conditions of the wood specimens.That is, the distribution of inorganic substances formed
in the moisture-conditioned specimens is completelydifferent from thatin the water-
saturated specimens.
Figure 1 shows the types of distribution of inorganic substances. Table 1 summarizes
the results of its distribution for different metal alkoxide/alcohol reaction systems. The
moisture-conditioned specimens refer to woodwithonlyboundwaterbelow the fiber
saturation point; therefore, water is distributed only within the cell walls. Water-saturated
specimens have free water in the cell cavities in addition to the bound water within the
cell walls. In this way,the use of these specimens with differently distributed water in the
cells can make it possible to prepare wood composites with inorganic substances distrib-
uted differently [3,4,6].
In type I, the metal alkoxides that form metaloxane gels specifically within the cell
walls are silicon alkoxide and boron alkoxide with the moisture-conditioned specimens
(Fig. 2a). The use of water-saturated specimens makes the inorganic substances distributed
0 I
II m
Iv v
FIGURE 1 Distribution of inorganic substances in wood-inorganic composites. (From Ref. 6.)
Wood-Inorganic Composites 783
TABLE 1 Inorganic Gels Formed and Their Distribution in Wood Cells as Prepared in Various
Metal Alkoxide/Alcohol Reaction Media
as type IV, due to the presence of water in the cell cavities in addition to the cell walls
(Fig. 2b).
In these reaction systems,the distribution of the inorganic substances is really parallel
to that of water in the wood specimens. However, metal alkoxides with Ti, Al, andZr are
gelled as type 11 or 111, filled or surrounded with gels only in the cell cavities, by use of
moisture-conditioned specimens (Fig. 2c) but not as in type I. Additionally, the use of
water-saturated wood did not allow inorganic gels in either cell walls or cell cavities, but
only allowed them to cover the outer surface of the specimens with inorganic substances.
FIGURE 2 SEM micrographs (upper) and Si-K, and Ti-& X-ray mapping (lower) over the cor-
responding area of the SEM micrographs: (a) composites prepared from moisture-conditioned spec-
imens (9.5 WPG) with SiO, gels in the cell walls; (b) composites prepared from water-saturated
specimens (122 WPG) with SiO, gels in the cell cavities; (c) composites prepared from moisture-
conditioned specimens (45 WPG) with TiO, gels in the cell cavities. (From Refs. 3-5.)
784 Saka
These distribution results are due to the hydrolysis rate of metal alkoxides and the
subsequent sol rate by the polycondensation reaction, which increases in the order
Si, B < AI, Zr < Ti
A difference in the hydrolysis/polycondensation rate seems to result in the difference in
the distribution of inorganic substances within the wood cells [6]. Evidence for this concept
can be found in a study of TiO, wood-inorganic composites prepared with titanium al-
koxides and titanium chelates [IO]. These agents have different rates of hydrolysis and
polycondensation, and accordingly result in different distribution of TiO, gels in wood
cells.
1. DimensionalStability
Figure 3 shows relationship between antiswelling efficiency (ASE) and weight percent
gain (WPG) in various wood-inorganic composites. Here, the ASE is a measure to eval-
uate the dimensional stability, i.e., an ASE of 0% refers to no control of dimensional
stabilization, whereas 100% ASE refers to its complete control.
In wood-inorganic composites with SiOz or BzO, gels (type I), inorganic substances
are formed selectively within the cell walls and their dimensional stability is effectively
enhanced with a small increase in WPG [4,6]. It is further demonstrated that even higher
50
l
-30
0
‘‘ I
10
l
20
I
30 40
I I
50 60
W P G (%>
FIGURE 3 Relationship between antiswelling efficiency (ASE) and weightpercent gain (WPG)
in various wood-inorganic composites. (From Ref. 6.)
Wood-Inorganic Composites 785
dimensional stabilization can be achieved if the SiO, gels are bound directly with cell
wall components through isocyanate- or epoxy-type silane coupling agents [7]. On the
other hand, for wood-inorganic composites with TiO,, AI2O3,and ZrO, gels (type 11 or
111), in which inorganic substances are present in the cell cavities, dimensional stabilization
cannot be achieved. Interestingly, however, TiO, wood-inorganic composites from tita-
nium chelates (type I) as mentioned above revealed an improvement of dimensional sta-
bilization [lo].
2. Fire-RetardantProperties
Figure 4 shows the combustibility test after 45-S ignition of model houses made from 2-
mm-thick veneers of hinoki (Charnaecyparis obfusa Endl.). It is apparent that the SiO,
wood-inorganic composites (right) show much higher resistance to burning, compared
with the untreated wood (left).
Figures 5a-5c show micrographs obtained by scanning electron microscope (SEM)
of the transverse surfaces after the combustibility test [4].It should be noted that the
untreated wood (a) has thin cell walls, whereas thecomposites prepared fromthe moisture-
conditioned specimens with 9.5 WPG (b) retain rather thick cell walls by the deposition
of SiO, gels and carbonizationof the cell walls. This apparently shows that a small amount
of SiO, gel formed within the cell walls can enhance fire resistance. On the other hand,
the composites with 40 WPG prepared with water-saturated specimens (c) cannot protect
the cell walls and the cell walls are as thin as in the untreated wood (a).
3. TermiteDeterioration
Figure 6 shows the results of termite tests with Reticulitermes sperafus Kolbe. On the
relationship between the number of dead termites and the test periods [4,6]. Compared
with untreated specimens, the composites from water-saturated specimens with 70 WPG
(type IV) show significant resistance to termite attack, due perhaps to the blocking effects
of the gels. However, the composites prepared from moisture-conditioned specimens (type
I) also reveal termite resistance with only 10 WPG. It is therefore noteworthy that a small
amount of SiO, gel formed within the cell walls is as effective against termite attack as
gels formed in the cell cavities.
FIGURE 4 Comparison of combustibility tests of model houses made from untreated (left) and
SiO, composite wood veneer (right) after 45-S of ignition.
Saka
FIGURE 5 SEM micrographs of wood-inorganic composites after combustibility tests: (a) un-
treated wood; (b) composites with Si02 gels in the cell walls (9.5 WPG); (c) composites with Si02
gels in the cell cavities (40WPG); (d) composites with Si0,-P,O, gels (16.9 WC);(e) composites
with Si02-B20, gels (38.7WPG); (f) composites with SiO2-P2O5-B2O3 gels (34.3WPG). (From
Refs. 4.8.)
100
W
a,
.l4
c,
E
8
rd
8 50
a
%l
0
35
= o
0 10 20 30 40
Test Period (days)
FIGURE 6 Relationship between the number of dead termites and the test periods for various
wood-inorganic composites: 0 SiOz composites (type W ) ;0 SiO, composites (type I); 0 TiO,
composites (type 11); ZrO, composites (type 111); A A1,0, composites (type 111); A untreated wood.
(From Refs. 4,6.)
Wood-Inorganic Composites 787
4. WaterRepellency
Figure 7 shows the change of water absorption ratio (WAR) in Si02 wood-inorganic com-
posites prepared with and without a water-repellent agent as a property enhancer. As SiO,
wood-inorganic composites were prepared in the reaction medium of tetraethoxysilane
(TEOS)/ethanol(EtOH)/acetic acid (catalyst) with a molar ratio of 1 :1:0.01, the water-repel-
lent agent silicon alkoxide with a long-chain alkyl residue or perfluoroalkyl residue was
added to the reaction medium in a molar ratio of 0.0 I . These water-repellent agents are, for
example, DTMOS [decyltrimethoxysilane, CH,(CH,),)Si(OCH,),] and HFOETMOS 12-hep-
tadecafluorooctylethyltrimethoxysilane;CF,(CF,),CH2CH2Si(OCH3),],and have a large mo-
lecular weight so that they are mainly distributed over the surfaces of the cell cavities, as
observed by SEM-EDXA study. However, due to the trimethoxysilyl residue present in the
water-repellent agent, it is bound with SiO, gels formed within the cell walls as described
below and thus fixed in the wood cells with low-surface-energy residues (R') exposed over
the surface of the cell cavities.
l
\ SiO, OH + H-0-Si-R'
\
\
Compared with SiO, composites and untreated wood, composites with a water-re-
pellent agent revealed lower WARin Fig. 7, indicating a water-repellent property added
to the composites [ S ] .
5. Antibacterial Properties
Creosote and chromated copper arsenate (CCA) are widely used as preservatives for an-
tibacterial treatment of wood. In spite of their excellence in this property, they have some
Untrealed
TFPTMOS-SiO,
HFOETMOS-SiO,
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0 5 10 15
Test period ( days )
FIGURE 7 Change of waterabsorptionratio in Si02 wood-inorganic composites prepared with
and without a water-rcpellentagent:TFPTMOS, 3,3,3-trifluoropropyItrimethoxysilane;DTMOS,
decyltrimethoxysilane: HFOETMOS. 2-heptadecaHuorooctylcthyltrin~ethoxysilane.
(From Rcf. 9.)
788 Saka
drawbacks in terms of toxicity. Therefore, less toxic and environmentally acceptable chem-
icals are expected to be used. Quarternary alkylammonium salts are among the candidates
for the antimicrobial treatment ofwood.Inthisstudyofwood-inorganic composites,
trimethoxysilylpropyldimethyloctadecylammonium chloride (TMSAC), shown below, was
used as a property enhancer to add an antibacterial property to wood [ll].
L
As in water-repellent agents, TMSAC can be expected to be bound with SiO, gels and
fixed in the wood cells, with quarternary alkylammonium salt residue exposed over the
surface of the cell cavities. Therefore, TMSAC with a molar ratio of 0.005 was added to
the reaction medium of TEOSEtOWacetic acid (molar ratio 1: 1:O.Ol).It is quite apparent
in Fig. 8 that TMSAC-added SiO, composites (TMSAC-Si02) reveal a significant resis-
tance against attack by white-rot fungi [Coriolus versicolor (L.ex Fr.) Quell. It should be
noted further that TMSAC-Si0, composites are more resistant than TMSAC wood. This
result canbeexplained by Fig. 9, inwhichTMSAC-SiO, composites are more water
repellent compared with TMSAC wood, due perhaps to more uniformly distributed
TMSAC with its long alkyl residue, chemically bound with Si02 gels in wood [l l].
In summary, if the composites have inorganic gels distributed selectively within the
cell walls, as in type I, effectivelyenhanced properties can be achieved.However,in
composites of type 11 or 111, improvement of the properties cannot be expected. In types
IV and V, property improvement can be expected to some extent, but the porous structure
characteristic of wood would be diminished. The results obtained therefore indicate that
it is more effective for enhancement of the wood properties to incorporate inorganic sub-
stances into wood in neighbor wood cell wall components rather than to deposite them in
the cell cavities, far from the cell wall components. For the property enhancers of silicon
alkoxide with hydrophobic residue, such as a long-chain alkyl or perfluoroalkyl residue,
a water-repellent property can be added to woodby covering the surfaces of the cell
cavities. Similarly, property enhancers with a quarternary alkylammonium salt residue can
add an antibacterial property to wood. Therefore, it may be concluded that the topochem-
FIGURE8 Comparisons of fungal attack by whlte-rot fungi:SiOz composites (6.7 WPG); TMSAC
wood (0.7 WPG); TMSAC-SiO, composites (4.1 WPG). (From Ref. 11.)
Wood-Inorganic Composites 789
I
~ ~~
l Untreated
1.o *
x v
E ,a
0
.m
Y
e 0.5
c)
I I
0 5 10 15
Test periods ( days )
FIGURE 9 Changes of water absorption ratio for various composites: SiO, composites (6.7 WPG);
TMSAC wood (0.7 WPG); TMSAC-SiO, composites (4.1 WPG). (From Ref. 1 I .)
ical effects of inorganic substances and property-enhancers exist for wood property en-
hancement in wood-inorganic composites.
IV. MULTICOMPONENTWOOD-INORGANICCOMPOSITES
A. Si0,-P,O,-B,O, Wood-InorganicComposites
The preparation of Si02-P,0s, Si02-B203, and SiOZ-Pz0,-B2O3 wood-inorganic com-
posites is based on the reaction medium of TEOS/EtOH/acetic acid (molar ratio 1: 1:0.01)
with trimethylphosphite (TMP) and/or trimethylborate (TMB) with a molar ratio of 0.05
181.
Figure I O shows the results of thermogravirnetric (TG) analyses of the composites
obtained. In the TG curve of the untreated wood (a), an abrupt decrease in its weight can
be observed in a temperature range between 300 and 350°C due to the flaming. However,
in the SiO, composites formed within the cell walls (8.4 WPG) (b),the flaming temperature
was shifted higher with the higher residual weight, compared with untreated wood. The
difference between (a) and (b) in the TG curves would have reflected upon the differences
observed in Figs. 4 and 5. On the other hand, binary and ternary composites (c, d, and e)
790 Sa ka
- Flaming. . Glowing
100-
A
x
v
c,
-c
M
.F-
a
M 50-
2
2
v1
a, -...--- _._-__
e
CT .
"42
""
"."
- d
0-1 I l I I 1
--
_______.__
I
C
-. b
l
""""
I a
0 100 200 300 400 500 600 700 800
Temperature("C)
FIGURE 10 Thermogravimetricanalyses of variouswood-inorganiccomposites: (a) untreated
wood: (b) SiOzcomposites (8.4 WPG); (c) Si02-P20, composites (16.9 WPG); (d) SiO,-BzO2
composites (38.7 WPG); (e) Si02-Pz0,-Bz0, composites (34.3 WPG). (From Ref. 8.)
revealed fairly high residual weight for flaming, showing very high tire resistance. Fur-
thermore, for glowing at a higher temperature over 350°C in (d), or over 300°C in (c) and
(e), both binary and ternary composites show fairly high tire resistance. These composites
have the highest WPG at 38.7, keeping the cell cavities nearly empty.
Figure 11 shows comparisons of the combustibility tests among five plywood veneers
made from untreated wood (left) and inorganic composites (right) [S]. After 30-S of ignition
with a gas burner, the burner was removed and combustibility with the remaining flame
was observed. Compared with SiOLcomposites, binary and ternary composites reveal
stronger fire resistance. Furthermore, as shown in Figs. Sd-Sf, SEM micrographs of these
carbonized composites after this combustibility test apparently show that the thinning of
the cell walls is much less in these binary and ternary composites, compared with untreated
wood (Fig. Sa).
Figure 12 shows the results of differential thermal analyses (DTA) of those binary
and ternary composites [S]. In the untreated wood (a), significant endothermic peaks cor-
responding to flaming and glowing can be observed. However, these peaks weaken i n the
SiO, composites (b). and in binary and ternary composites they disappear and broaden to
the higher temperature in SiO,-B,O, composites (d). In Si0,-P,O, (c) and Si02-P,0,-
B,O, composites (e), the endothermic peaks broaden to both higher and lower tempera-
tures, revealing high resistance to combustion.
For the tire retardance observed above, the mechanism of B,O, gels is different from
that of P,O, gels. In Fig. 10, the flaming temperature is shifted lower for composites with
P,O, gels, whereas for the composites with B,O, gels or SiO? composites, it is shifted
higher. Thegels in the latter composites are believed to be melted during the flaming
process and form a glassy layer to cover the cell wall components. As a result, fire-
retardant properties are added through the physical barrier against heat and oxygen. On
the other hand, for composites with P,O, gels, chemical effect by dehydration with phos-
phorus has promoted the carbonization of the composites which have revealed fire-retar-
dant properties [ IS]. Therefore, as seen in Fig. IO, ternary composites with Si0,-P,O,-
Wood-Inorganic Composites 791
- ~ ~. ~- ~.~
I I I I I I 1 l l
B20, gels would have high fire-retardant properties through effects of both a physical
barrier and chemical reaction as mentioned above.
In a practical sense, TMP needs to be handled cautiously because of its odor and toxicity,
while BzO, gels from TMB are not stable and are readily leachable. Additionally, consid-
Me OMe Me 0
I I I I
MeO-Si-0-Si-0-Si-0-Si-OMe
I l I I
0 Me OMe Me R = CH2CH3 or H
Through their evaluation, the composites had a fire resistance as high as the com-
posites prepared from the reaction system TEOS/EtOH/acetic acid with TMP and TMB.
Furthermore, leaching of the gels was prevented to ;I greater extent, due perhaps to chem-
ical bonding of multicomponents in oligomer levels which could have stabilized inorganic
substances in wood cell walls. Additionally, the oligomers prepared are nontoxic, so en-
vironmental safety in their preparation was achieved. By adding HFOETMOS in a small
quantity as a property enhancer to the oligomer reaction system, the composites obtained
could improve further the antileachability of PzO, and B,O, gels.
To develop wood with high function and remarkable properties, we have tried to prepare
inorganic composites of wood without losing characteristic properties of the wood as seen
in its porous structure. As mentioned already, in spite of the same inorganic substances
used, the observed properties are different if inorganic substances are distributed differently
in the wood cells.This is because topochemical effects exist in wood for property en-
hancement.Our goal fordeveloping ideal “superwood” is to achievewood-inorganic
composites from environmentally friendly materials with their minimal use and maximal
effect on property enhancement. More extensive study of topochemistry in wood property
enhancement will provide a clue to a development of “superwood.”
794 Sa ka
1. INTRODUCTION
A number of books and summaries that deal with wood preservation are currently available
[ 1-51. Consequently, there is little need for another general review of this subject. How-
ever, in the past few years there have been significant changes in the wood preserving
industry, so a review of these trends seems appropriate. Accordingly, in this chapter em-
phasis will be placed on new developments and trends in the wood preserving industry.
New developments in wood preservation have been mainly in the area of new preserva-
tives, so this chapter will focus heavily on these advances.
A. PressureProcesses
The majority of wood products are treated by conventional pressure methods using either
the full-cell or empty-cell process. The full-cell process uses an initial vacuum to evacuate
air from the wood, followed by filling the cylinder with preservative solution under vac-
uum prior to the application of pressure. This process is generally used for water-borne
preservatives, where maximum treating solution retention is desired. In recent years the
lnoditied full-cell process has become increasingly popular. The same basic cycle is used
with the exception that the initial vacuum is reduced by about 40-50% and a final vacuum
is applied after the pressure period. Use of the modified full-cell treatmentprovidesa
795
796 Nicholas
means of reducing the final solution retention. which minimizes the dripping of preser-
vative solution and subsequent holding time after treatment.
Empty-cell processes (Lowry and Rueping) do not employ an initial vacuum and as
a consequence result in much lower net preservative solution retentions. By applying some
degree of initial air pressure and filling the cylinder at this pressure (Rueping process),
the net solution retention can be reduced even further. These treating processes are gen-
erally used with oil-borne preservatives where it is desirable to minimize the amount of
carrier oil used.
B. Vapor-PhaseProcess
A major concern i n the treatment of many wood species is achieving adequate penetration
of the preservative. One possible approach to this problem is to use vapor-phase treatments.
The validity of this concept has been demonstrated in New Zealand, where they success-
fully treated various wood products with borates Barnes and Murphy [6]. In this process,
trimethyl borate is vaporized by heating and then introduced into an evacuated cylinder
containing wood at a low moisture content. The trimethyl borate rapidly diffuses into the
wood and reacts with residual water to form boric acid in situ.
C. Supercritical-Fluid Process
Another approach to the problem of poor preservative penetration is to use supercritical
fluids as carrier solvents in the treating process. Supercritical fluids are capable of pene-
trating the small openings i n wood because they effectively eliminate the interfacial prob-
lems associated with conventional liquids. The feasibility of this process has been dem-
onstrated in laboratory studies 171. However, a substantial amount of research willbe
required before it can be determined whether this process has any commercial applications.
111. PRESERVATIVES
A. Types of Preservatives
Wood preservatives are generally classified into two basic types-oil-borne and water-
borne-which are distinguished by the type of carrier used to solubilize the biocides.
1. Oil-borne Preservatives
Creosote and pentachlorophenol (penta) are the major wood preservatives currently being
used. Another oil-borne biocide that has been used on a limited basis is tributylin oxide
(TBTO). Abrief discussion of the general characteristics of these preservatives is presented
below.
a. Creosote. The use of creosote as a wood preservative was patented in 1838 by
Bethell, and since that time creosote has remained an effective, widely used chemical for
treating wood [S]. Creosote is a complex mixture containing at least 200 identifiable
compounds. However, it is generally agreed that several thousand different compounds
are present in very small amounts [91.
The greater part of the composition of creosote consists of neutral fractions (Table
l ) . Tar acids, such as phenol and the creosols,as well as such tar bases as pyridenes,
quinolines, and acridines, constitute a rather small percentage of the total weight of cre-
osote. Unlike the neutral fractions, the tar acids and bases are usually soluble in water and
hence contribute very little to the efficacy of creosote as a wood preservative. It follows
Naphthalene 3.0
Methyl naphthalene 2.1
Diphenyl dimethylnaphthalene
Biphenyl 0.8
Acenaphthene 9.0
Dimethylnnphthalcne 2.0
Diphcnyloxide -
Dibenzofuran S.o
Fluorene-related compounds 10.0
Methyl Huorenes 3.o
Phenanthrene 21.0
Anthracene 2.0
Carbazole 2.0
Methylphcnanthrene 3.0
Methyl anthracenes 4.0
Fluoranthene 10.0
Pyrene 8.5
Benzofluorene 2.0
Chrysene 3.0
Other components not identified
798 Nicholas
from the foregoing statements that the chemistry of creosote and that of the coal-tar neutral
fractions are quite similar. So, for that matter, is the chemistry of the parent material-
coal tar. Compositional data for coke-oven coal tar produced in the United States is given
in Table 2.
The majority of the compounds in creosote are aromatic hydrocarbons with con-
densed ring systems. In addition, tar acids, which are heterocyclic compounds containing
nitrogen plus some neutral oxygenated compounds, are present [ 101. The tar acids and
bases contain a wide range of fungicidal constituents.
The fungicidal activity of creosote fractions obtained by distillation varies widely.
In this regard, Schulze and Becker [ 1 l ] investigated the fungicidal activity of I5 distillate
fractions boiling between 120 and 360°C against three wood-decay fungi. They found that
the toxicity of the different fractions varied widely, with those boiling between 180 and
Component Wt (95)
240°C being the most active. Within these fractions, thionaphthene, 2-naphtho1, 2-meth-
ylnaphthalene, and isoquinoline were the most active compounds. However, it is likely
that some of the components in creosote exhibit synergism, so the toxicity of individual
components is probably not of major significance.
In recent years attempts have been made to improve the surface characteristics of
creosote-treated wood. Thisconcept was termed “clean”creosote andis achieved by
reducing the xylene insolubles (XI) present in creosote solutions [ 121. By reducing the XI
content to 0.1% or lower, a much cleaner product is produced.
b. Pentachlor-ophenol (Petzta). Penta was first used asa wood preservative in the
1930s and rapidly became established as the most widely used single oil-soluble biocide
in wood preservation. The performance of penta-treated wood for ground-contact appli-
cations is highly dependent on the carrier system used [ 13- 171.
The effect of the carrier oil on performance has been attributed to the following
factors: ( 1 ) its effect on penta depletion; (2) its effect on distribution of penta in the wood
structure; and (3) its intrinsic biological activity [ 171. The fungicidal activity of the carriers
varies considerably among the petroleum oils used for this purpose, with some oils dem-
onstrating reasonably good performance in the soil block test [ 171. With regard to penta
depletion from treated wood, there is considerable variation among the oils. Furthermore,
there appears to be an inverse relationship between the penta depletion rate and perfor-
mance of field stakes [ 171.
Over the years, a number of different carriers other than petroleum oils have been
used for penta treatments. These include transient light solvent systems-liquefied petro-
leum gas(LPG), methylene chloride, and mineral spirits-and water-based emulsions.
The LPG (Cellon) and methylene chloride (Dow process) systems were used extensively
for a number of years to treat utility poles. However, poor performance of the treated
wood due to erratic penta distribution and environmental concerns resulted in abandonment
of these processes.
The water-borne emulsion penta system also failed due to poor performance of the
treated wood products. Inadequate performance was probably due to excessive leaching
of penta from treated products subjected to exterior exposure.Another water-borne system,
the water-soluble sodium salt of penta, was used extensively for dip or spray treatment of
green wood tocontrol stain and mold fungi. Because of environmental concerns this
compound is no longer used for this application.
c. Tributylin Oxide. Organotin compounds were shown to have fungicidal proper-
ties in the 1950s. Subsequently, laboratory and field tests demonstrated that tributylin oxide
(TBTO) was the best wood preservative on the basis of cost, permanence, and mammalian
toxicity [ 181. TBTO has been used extensively in Europe as a wood preservative for above-
ground applications. It has also been used to a limited extent in the United States as a
replacement for penta in millwork applications. However, recent studies have shown that
this compound is not stable in contact with wood and gradually decomposes over a period
of time [19]. As a consequence, TBTO is no longer used in the United States. In Europe,
tributylin naphthenate has replaced TBTO and apparently is equally as effective and more
stable.
2. Water-Borne Preservatives
Chromated copper arsenate (CCA) and ammoniacal copper zinc arsenate (ACZA) are the
major water-borne wood preservatives currently being used commercially in the United
States. A brief description of the general characteristics of these preservatives is presented
below.
800 Nicholas
(1. Chromlted Copper Arsenate. CCA is unquestionably the most important wood
preservative in the United States, representing 78% of all preservatives used in 1993 [6].
Although CCA is water-soluble, it undergoes a series of complex fixation reactions
in wood. These reactions involve both lignin and carbohydratecomplexesas well as
inorganic precipitates (Fig. l ) . As a consequence of these reactions, CCA is highly fixed
in wood and resists leaching even under severe exposure conditions. This interaction with
wood results in a decrease in strength properties, with toughness being particularly sen-
sitive to this treatment. However, these strength losses can be minimized to 10% or less
by drying the treated wood at temperatures of 71°C or less 120-221.
CCA is a very effective wood preservative andis used for numerous applications
suchas lumber fordecks, utility poles, marine piling, etc. When the wood is properly
treated. an extremely long service life can be obtained with these products.
h. Anmoniacal Copper Zinc Arsenrrre. The useof ACZA is limited to the West
Coast area, where it is used to treat Douglas fir and other local wood species. The alkaline
solution provides better penetration of these relatively refractory wood species.
ACZA is not as highly fixed as CCA, and the chemical reactions responsible for this
insolubilization are not clearly understood. The main mechanism of fixation of copper and
zinc is postulated to be the formation of insolublecopper arsenate and zinc arsenate.
However, the overall mechanism is undoubtedly more complex because cuprammonium
ions react by ion exchange with functional groups in wood 1231. In addition, both copper
and zinc complexes can be formed with the wood substrate, but copper and zinc complex
formation do not appear to be related.
Cu2+,Cr3+,As5+
l l
Carbohydrate Inorganic
Precipitates
3-Iodo-2-propynyl butyl H 9
l
carbamate I -c=c-~-o--~-N - C * H ~
A
Ti mbor2'In Disodium octaborate Na2B,0,,.4H20
tetrahydratc
4,S-Dichloro-2-,1-octyl-4-
isothiazolin-3-one
Busan 3 ( P / 2-(Thiocyanomcthylthio)
TCMTB benzothiaxole
Cl
WocosenW (2RS, 4RS)-2-(2,4- I
propiconazole dichlorophenyl)-2-[ 1- 1H -
( I .2,4-triazoIe)methyI1-4-
propyl- 1.3-dioxolanc
Tebuconazole OH
(3RS)-5-(4-~hlorophenyl)-
2. 2-dimethylethyl-3-1H - C I O C H ~ - C H Z - C - C (I C H ~ ) ~
[ I .2.4-triazole)methyl)-3- ?Hz
pentanol
Amical 4 8 O Diiodomethy-/,-tolysulfone
C=N
Chlorothalonill 2,4,5.6-Tetrachloro-
tuffgard@>B isophthalonitrile
C=N
802 Nicholas
naphthenate and Cu-8, have been used to a limited extent as wood preservatives in the
past but are now gaining more popularity because of their relatively low mammalian
toxicity. A brief discussion of each of these biocides is presented below.
N. Polyphusea (ZPBC). IPBC is an organic biocide that exhibits low mammalian
toxicity and has broad-range activity against common wood decay, mold, and stain fungi,
but is not effective against wood-destroying insects. It is currently being used to treat
wood for above-ground applications for millwork and similar products. A combination of
IPBC and DDAC is effective against mold and sapstain fungi and is used extensively for
controlling these microorganisms in freshly sawn lumber.
b . TinzborQ ( B o r d B o r i c Acid). T i m b o e is an inorganic biocide with boron being
the active component. It hasa very low mammalian toxicity and exhibits broad-range
activity against both wood-decay fungi and insects. It is highly soluble in water and readily
diffuses in and out of wet wood. Consequently, its use as a wood preservative is limited
to above-ground applications which are protected from the weather. It is currently being
used to a limited extent in the United States for commercial products.
c. Copper- Nuphthenate. Copper naphthenate is an organometallic compound that
is normally prepared by the direct reaction of copper hydroxide with naphthenic acid at
elevated temperatures in a hydrocarbon solvent. It exhibits low mammalian toxicity and
has broad-range activity against wood-decay fungi and insects.
Copper naphthenate is not a new wood preservative and has been used to a limited
extent as a preservative for a variety of wood and textile products over the years. As a
result of recent environmental concerns with the major wood preservatives, interest in
copper naphthenate has increased. This led to its use by several companies as a preser-
vative for utility poles. However, some utilities have experienced early failure of some of
these poles, and this has resulted in curtailment in the use of copper naphthenate. The
exact cause of these early failures is not known, but it appears that it may be due at least
in part to inactivation of copper naphthenate when it is used to treat green poles that are
steam-conditioned. In any event, this experience will undoubtedly result in limited use of
this preservative in the future.
d. Oxine Copper (Cu-S). Cu-8 is an organometallic compound formed by the re-
action of copper with 8-quinolinol. It exhibits low mammalian toxicity and has broad-
range activity against wood-decay fungi and insects. Cu-8 is very insoluble in water and
most organic solvents, butan oil-soluble form can be made by reaction with 2-ethyl
hexoate. A water-soluble form can be made with dodecylbenzene sulfonic acid, but this
formulation is highly corrosive to metals.
Cu-8 is not a new preservative and has been used to a limited extent as apreservative
for a variety of wood and textile products over the years. It is currently the only preser-
vative that is approved for treating wood that is in contact with foodstuffs. It isused
primarily for treating specialty items such as food pallets and picnic tables.
e. Burduc 2 2 a (DDAC). DDAC is an organic biocide that exhibits low mammalian
toxicity and has broad-range activity against wood-decay fungi and insects. It is a water-
soluble compound, but undergoes an ion-exchange reaction with wood which greatly re-
duces its leachability from wood exposed to water or wet soil.
In order to improve the efficacy of DDAC as a wood preservative, it is generally
combined with other biocides. In this regard, DDAC is currently used in two commercial
wood preservative formulations.As mentioned previously, a formulation composed of
DDAC and IPBC is widely used to control mold/sapstain in green wood. More recently,
a combination of DDAC and ammoniacal copper (ACQ) has emerged as a commercially
viable alternative to CCA for many applications.
Preservation of Wood 803
4. TrendsinWoodPreservativeDevelopment
The current trend in the development of wood preservatives is to use biocide combinations.
These combinations include both inorganic-organic and organic-organic binary mixtures.
Examples of these are ammoniacal copper quat (ACQ), copper dimethyldithiocarbamate
(CDDC), ammoniacal copper citrate (ACC), ammoniacal copper azole, DDAC-IPBC (NP-
l), chlorothalonil-chlorpyrfos,and DDAC-Na omadine.
ACQ has two different formulations-ACQ B and ACQ D. The only difference in
these formulations is the copper-complexing agent. Type B contains ammonia and type D
contains ethanolamine. In both formulations, the active ingredients are copper and DDAC
in a ratio of 2: 1 (Cu0:DDAC). This combination provides broad-range efficacy against
wood-decay fungi and insects in both above-ground and ground-contact applications.
CDDC is formulated with copper ethanolamine and sodium dimethyldithiocarbamate
(SDDC). Since copper reacts rapidly with SDDC to form an insoluble complex, a two-
step treating process is required with this preservative system. Accordingly, the wood is
first treated with a copper ethanolamine solution in one treating cylinder and then moved
804 Nicholas
to a second treating cylinder for treatment with SDDC.The reaction product is a 1.2
c0pper:dimethyldithiocarbamate chelate which is highly insoluble in water. This chelate
appears to have reasonable broad-range efficacy against wood-decay fungi and insects in
both above-ground and ground-contact applications.
ACC is formulated with ammoniacal copper carbonate and citric acid, using a ratio
of 1.6S:l (Cu0:citric acid). Although the AWPA has developed a Standard, the value of
this wood preservative system is questionable because both soil block and field stake test
data indicate that the treated wood is not resistant to copper-tolerant fungi. This weakness
is not surprising, since it does not contain a co-biocide.
NP- I Q is formulated with DDAC and IPBC at an 8.5: 1 ratio of DDAC:IPBC. NP-I
is used mainly for sapstain and mold control in freshly sawn wood but may have potential
for millwork and similar applications.
Copper azole is formulated with a combination of copper (49%), boric acid (49%)
and tebuconazole (2%), using ethanolamine as the complexing agent for copper. This
system is still undergoing evaluation, but shows promise as a viable wood preservative
for some applications.
The trend toward the use of binary and tertiary biocide combinations in wood pre-
servative formulations is expected to continue.This formulation strategy is particularly
significant when the biocides exhibit synergism. Indeed, the development ofwood pre-
servatives on the basis of synergistic mixtures is currently being explored and appears to
have considerable potential [24].
5. Biocontrol
Another approach to wood preservation is to use antagonistic microorganisms rather than
toxic chemicals. The first report concerning the potential to this approach was by Richard
and Bollen [2S], using Scytalidiun~sp. to inhibit Antrodicr cclrhonica in Douglas fir poles.
The results of this study stimulated additional research on the validity of this method for
wood preservation, and this led to the development of Binab AB@. This bioprotectant has
been marketed in Europe a s a preservative for Scots pine [26,27]. Although some trials
were very positive, there is concern about the long-term effectiveness of this formulation
127-291. Binab AB was also evaluated by Morrell and Sexton [30] as a possible preser-
vative for Douglas fir and southern pine. Poor results were obtained in this latter study,
and failure of the system was attributed to the fact that it did not control all the decay
fungi that colonize these wood species. Consequently, it was concluded that biocontrol is
not a currently viable method for preserving wood products against decay fungi.
Although the possibility of using microorganisms for controlling wood-decay fungi
is discouraging, this approach may have potential for controlling sapstain/mold fungi in
freshly sawn wood products [2]. This particular application requires only short-term pro-
tection, so concern about long-term survival of the microorganisms is eliminated. Both
bacteria [31] and fungi 1321 show promise for this application.
IV. FUTUREDEVELOPMENTS
Environmental issues have been the major driving forces behind the development of new
wood preservatives in recent years. This trend will undoubtedly continue into the foresee-
able future because there currently is considerable concern about disposal of treated wood
after it has ended its useful life. This is particularly true for CCA-treated wood, which is
Preservation of Wood 805
not amenable to disposal by incineration. This dilemma will encourage research to develop
practical methods for removing CCA from treated wood so that it can be recycled. At the
same time, efforts will continue toward the development of alternative wood preservatives
that do not pose disposal problems for treated wood. The development of ACQ, which
eliminated chromium and arsenic, was a step in the right direction, but the presence of
copper complicates disposal problems.
The major thrust in wood preservative development in the future will probably be
based on total organic biocide systems. If the mammalian toxicity of these systems is low,
then the environmental concerns will be effectively eliminated. Other, more specific, and
possibly nonbiocidal,approaches to wood preservation may develop as we learn more
about the complex reactions involved in the overall decay mechanisms utilized by fungi.
The problem of poor weathering characteristics of treated wood, especially CCA-
treated wood, will probably alsoreceiveconsiderable attention in the future. If a high
degree of water repellency canbeimpartedtowood, it will minimize the weathering
problem and also minimize the biocide levels required to inhibit biodeterioration. Signif-
icant advancements in this area will help pave the way for development of cost-effective
organic wood preservative systems.
With regard to treating processes, significant improvements have been minimal in
the past, and this trend probably will continue. One of the main problems that needs to
be addressed is providing adequate treatment of refractory heartwood. More research is
needed in this area, and progress in solving this problem could be very rewarding. Con-
tinued progress with the supercritical-fluid treatment process could make a major contri-
bution in this area [ 3 3 ] .Other developments, such as the shock-wave treating process and
a compression/vibration pretreatment of the wood, may also prove to be effective methods
for improving the treatability of refractory woods.
REFERENCES
I. R. A. Eaton and M. D. C. Hale, Wood D e c q Pests trr~dProtectiorl. Chapman & Hall, London
( 199.3).
2. R. A. Zabel and J. J. Morrell, Wood Microbiology: 1)ccay m l d I t s Prclvrltiorl, Academic Press,
New York ( 1992).
3. J. G. Wilkinson, Irldrtstrid Tirnher Preservntiorz, Associated Business Press, London ( 1979).
4. D. D. Nicholas(ed.), Wood Deteriorntion clnd I t s P revention by Presenwtivo Trcyttnlerlts,
Syracuse University Press, Syracuse, NY (1973).
5. G.M. Hunt and G. A. Garrett, Wood Preservntiorz. 3rd ed., McGraw-Hill, New York (1967).
6. H. M. Barnes and R. J. Murphy, Forest Prod. J . , 4.5(9):16 (1995).
7. J. J. Morrell, K. L. Levien, E. S. Demessie, S. Kutnar, S. Smith, and H. M.Barnes, Proc.
Curl. Wood Preservntiorl Assrl., 14:6 ( I 993).
X. M. P. Levi, in Wood Drteriortrtiorl crntl I t s Prevelltion by Presenrrtive Trmtnlents, Vol. l ,
D e g r d r t i o n u ~ l dProtection of Wood (D. D. Nicholas, ed.), Syracuse University Press. Syro-
cuse, NY. p. I83 ( 1973).
9. Anonymous, The Biologic and Economic Assessment of Pcntachlorophenol, Inorganic Arsen-
icals and Creosote, Vol. I. Wood Preservatives. USDA Tech. Bull. 16.58-1:28 (1981 ).
IO. W. H. Hartford, in Wood Deteriorrrtion and I t s Preverltior? by Pre.ser\*nti\v Tremmv1t.s. Vol. 11,
Presenutivcv trrltl Prescwcrtive Systerus, Syracuse University Press,Syracuse, NY, p. I O
( 1973).
11. B. Schulze and G. Becker, Holz$~r.scl~.,2 3 7 (1948).
12. H. M. Barnes and L. L. Ingram, Jr.. Pmc. Am. Wood Pre.srl?.c.r.s’A.s.srl.,Y1:108 (1995).
13. R. H. Baechler and H. G. Roth, Forest Prod. J., 12: 187 (1962).
806 Nicholas
“Ozymandias” ( 1817)
Percy Bysshe Shelley
1. INTRODUCTION
Wood is a versatile material which has been used since the dawn of civilization. At the
basic level wood satisfied human beings’ needs or wants in shelter, defense, transport, and
leisure. Archeologists have discovered, from time to time, weapons, domestic utensils,
tools, building materials, and boats made of wood. It is a durable material under a benign
environment.Since woodis acomposite biological material, it inevitably continues to
deteriorate asa result of physical, chemical, mechanical, and biological processes. To
prolong the service life of wood, many chemical treatments have been developed to protect
it from attack by microorganisms such as bacteria and fungi, and insects such as termites
(see Chapters 12 and 21). Surface treatments are used to protect wood against moisture
and weathering (see Chapters 9 and 11).
Since wood is one of the oldest materials used, many artifacts, from cradlesto
coffins, which were used in the past have been discovered by archeologists. Thus, specific
kinds of treatments have been developed to preserve waterlogged wood which has been
stored under wet conditions such as burial in soil below the permanent water table, at the
bottom of rivers or lakes, or i n a marine environment. More than 40% of all shipwreck
losses in the Western Hemisphere have occurred due to ships wrecking in shallow water.
Wood is a biological material; when it is submerged in the marine environment, it comes
under immediate attack from the teredos, fungi, and different bacteria. The more wood
items are exposed to salt water, the more they suffer and the more quickly some of them
vanish. Oftentimes, wood-based materials, when buried deep under sediment, will suffer
less, or not at all, and on occasion can be discovered in an excellent state of preservation.
Sometimes, wood is completely preserved by being saturated by iron oxide while lying
807
808 Hon
close to iron objects under water, as they tend to become mineralized and hard in texture.
Wood will also be preserved if it is saturated in fresh water.
In 1997, Titanic was a popular movie which broke the box-office sale record for the
century. The story was based on the S.S. Titanic, which went down in the Atlantic Ocean
in April 19 12, eighty-eight years ago. On July 14, 1986, the sunken Tirctrlic was discovered,
and more than 3000 artifacts have been lifted from the debris field. Contrary to many
expectations, the deep ocean did not preserve the Titanic from decay. Organic materials
such as wood, paper, and cloth all perished. The wood decking and furniture were nearly
all gone. Media attention, in print and on television, was usually intense and at times
frenzied, as anyone who remembers the discovery of the wreck of the Titanic in 1985 can
attest. When the Titunic- broke the surface of the water after 87 years of marine burial, we
can only imagine the depth of emotions experienced by those who were there. Hearts must
have been bursting and tears flowing. The sight of the ship must have conjured up many
different thoughts in the people watching: thoughts of its preciousness as an object of
antiquity and the incredible tie it forms between the people of today and those 87 years
ago; thoughts about the ship and its repository of information about past ship-building
technology.
To study shipwrecks is to study human history. To study wood in the past is also to
study human activity. Such studies have brought to light cultures that existed long before
written records, and have transformed featureless prehistory into a fascinating landscape
of evolution, cultural change, and technological advance. They have delineated sequences
of events in the past and have discovered and illustrated the course of human civilization.
Because wood-based artifacts provide a rich and varied record of our early activities and
technology, there is a great need to preserve them in the many forms that remain from
ancient to modern times. It is not only because it will be interesting to future generations,
but to use it to study the cultures, human behavior, and stratagems of intelligent human
minds throughout history and to study the wood aging process itself. Cultural materials
used by society exist in systemic context. Scientists and technologists study waterlogged
woods, including shipwrecks, in archeological context in order to understand the systemic
contexts of past societies. Preservation can be viewed as actions by a contemporary society
to slow the rate of deterioration and destruction of cultural materials. The majority of
waterlogged objects need to be preserved with special chemicals. If this is not done quickly
and properly, the artifacts will rapidly deteriorate. The conservation and preservation of
waterlogged wood tend to be expensive, require specialized knowledge and facilities, and
be complex and time consuming.
Preservation of waterlogged wood depends greatly on the condition of the wood and
the conditions of the environment where the wood will be stored after preservation. Present
techniques range from art to science. The history of preservation or conservation treatments
is a story of success and failures [ l ] . Until recently, many techniques have been based on
empirical approaches rather than hard scientific data. Because of the understanding of
degradation mechanisms of wood and availability of new instrumentations [2], scientists
are able to tackle preservation problems with more scientific approaches and thus acquire
better results. For most waterlogged woods, preservation consists of both careful removal
of water from the wood to minimize shrinkage, and introduction of a substance into the
wood to improve its strength. Generally, the processes developed to preserve waterlogged
woods are classified into two groups. One consists of dehydrating the wood first, before
treating it with a consolidant. The process is commonly employed only on small objects,
because careful removal of water from wood involves solvent exchange, a process which
takes longer as the size of the object increases. The other group, which is employed with
Preservation of Waterlogged Wood 809
larger artifacts such as entire ships, uses injection of a consolidant first and dehydrating
the artifact after treatment. These processes use water-soluble consolidants and introduce
the consolidants into the wood through the time-consuming process of diffusion. In this
chapter. many methods that have been developed to preserve waterlogged wood are re-
viewed. Additional information can be obtained from several excellent monographs [3-
S]. A case study of preservation of the historic gunboat U.S.S. Cairo is included.
As mentioned earlier, wood is a biological material that deteriorates in almost any envi-
ronmental condition. If wood is kept in a very moist or wet environment, it will absorb
water and eventually become waterlogged. Wood normally decays under combined bio-
logical and chemical attack when buried in the ground or submerged in water. Buried
wood will generally not survive unless it becomes waterlogged to create the anaerobic
conditions necessary for protection from decay fungi and inserts. Under this condition,
wood is still vulnerable to chemical degradation and attack by anaerobic bacteria. In time,
changes will occur in the wood structure which will affect its physical and chemical
properties. These changes adversely affect the integrity of the wood.
At the outset, the readily water-soluble extractives and mineral salts in the wood will
diffuse into the surrounding medium, followed by readily hydrolyzed compounds such as
the pectins and pentosans. Then a microbiological degradation of the more stable hemi-
celluloses and cellulose will follow. Finally, mainly lignin remains, which can also be
decomposed slowly by microorganisms in the anaerobicenvironment. Both fungi and
bacteria cause the microbiological breakdown of wood and both use extracellular enzymes
to hydrolyze cellulose, hemicelluloses, and lignin. In the case of afungus breakdown,
degradation spreads throughout the capillary system of the wood, at least as far as the
extracellular enzymes can penetrate into this system. Because of the loss of polymeric
components in the cell structure, voids increase throughout the cell and the wood becomes
more porous and permeable to water. Hence, for highly waterlogged wood, it is not unusual
to have a moisture content of over 800% (based on oven-dried weight). Thehygroscopicity
of waterlogged wood tends to increase, and equilibrium moisture content values as much
as twice those of recent wood have been found [6].
Because of increased hygroscopicity, deteriorated waterlogged wood also exhibits
significant increase in shrinkage, reaching values of the order of 70% for volumetric
shrinkage. The shrinkage in volume is found to be related linearly to maximum moisture
content of wood 171. As long as the waterlogged wood is kept wet, it will retain its shape.
If the wood is exposed to the air, it cannot redistribute internal moisture properly during
drying due to various factors. The most important one is due to capillary tension collapse
above the fiber saturation point in the deteriorated wood. Cell wall shrinkage occurs below
the fiber saturation point because of desorption and resultant dimensional change [ 7 ] .
Typically, moisture attacks the secondary cell wall, reducing the resistant strength and
flexibility of the wood [g]. Hence, when the excess water evaporates, the resulting surface
tension forces of the evaporating water cause the weakened cell walls to collapse. Such
an action subsequently leads to significant warping, shrinkage, and cracking. Drying of
highly deteriorated waterlogged wood therefore results in severe damage and distortion of
artifacts. Shrinkage values of more than 30% in the tangential and more than 10% in the
longitudinal directions have been recorded [6]. Moreover, because of the loss of major
polymeric structural components in the cell, deteriorated waterlogged wood also shows
81 0 Hon
Ill. PRESERVATIONTREATMENTS
Conservation treatments for waterlogged wood have been designed to prevent dramatic
dimensional changes caused by cell collapse and cell wall shrinkage during drying. Wood
artifacts to be conserved are usually physically weak and chemically very complex. The
most active phase of a conservation treatment is in controlling the processes as the wa-
terlogged wood is transferred from its deposition to its new “home.” Nonnally, one of
the first steps that must be taken is to remove contaminants such as sea salt and iron
sulfide which will be unstable in the new environment and damage the wood’s structure.
Hence, the specimens must be first cleaned and kept wet after removal from the discovery
site. If they are allowed to dry, they will shrink and split. Normally it is recommended to
let wood soak for 1-12 months in fresh water, depending on the size of the artifact.
Hardwoods can handle a mild solution ( 5 % ) of muriatic acid. In this case, no steel or iron
should be involved, as muriatic acid will destroy these items. After muriatic acid is used,
the wood must be soaked again in fresh water. This will help remove the smell of acid.
Preservation of waterlogged archeological wood involves stabilization a s tosize,
shape, and durability. The state of the wood surface as seen from an esthetic viewpoint is
also taken into consideration in preservation, whereas questions of restoration, completion,
and related aspects are considered outside of preservation. Most if not all processes de-
veloped to preserve waterlogged wood have as their objectives to achieve dimensional
stabilization and improvement of wood strength. Wood dimensions can be stabilized by
three different means or combination of these means. These are (1) reducing the hygro-
scopicity of the wood so that less water can be taken up, ( 2 ) forming crosslinks between
cellulose chains so as to minimize separation of these units, and (3) depositing a bulking
agent within the swollen wood so as to reduce shrinkage 191.
A. On-Site Preservation
On-site preservation consists of preservation measures that are employed at the excavation
site. Immediate treatments are usually aimed at preventing shrinkage and further deterio-
ration of the wood before it can be treated at a conservation facility. Techniques include
spraying with a 5% solution of sodium borate, boric acid, or any other wood-preservative
Preservation of Waterlogged Wood 811
to arrest and prevent further attack of deteriorating organisms, immersing in fresh water,
or covering with plastic foam sheets saturated with water to prevent drying and shrinkage.
be used simultaneously with PEG. After PEG treatment, the excess wax is wiped off and
the wool is allowed to cool. When cooled, any excess that was on the surface is removed
with a hot-air gun or with hot water. Alcohols may finally be used to clean the surface to
remove dark color and to regenerate woodlike color. When the Swedish warship Wasrl
was resurrected in 1961 after 333 years at the bottom of Stockholm Harbor, PEG was
selected to treat the entire vessel [IO].
Even though PEG treatment is a popular method, it has several significant drawbacks.
It is a relatively expensive and time-consuming process. Perhaps most significantly, PEG-
treated wood is sensitive to heat and high humidity. These unfavorable characteristics
require that treated wood be curated in climate-controlled facilities.
D. SucroseTreatment
To overcome the PEG-related limitations, sucrose has been developed as an alternative
method [ 11. In practice, the procedure is identical to that described for PEG. where sucrose
is applied by aqueous diffusion as a bulking agent. Itis less expensive and penetrates
wood quite well. For better results, refined white sugar (white sucrose) is recommended.
Unbleached, brown-colored sugar should be avoided because of its high hygroscopic
property.
For less degraded waterlogged wood, a sucrose solution with a sufficiently low con-
centration ( 1 - 5 % ) may be used to avoid dehydration of sound wood. For highly degraded
wood, a higher concentration of sucrose may be used. It is recommended to start with low
percentage increases, for example, 1-596, until a concentration of 50% is reached. Then
the solution can be increased in 10% increments. The treatment continues until sucrose
concentration reaches 70% and the wood has equalized at this concentration. During the
treatment, an antimicrobial agent may be added to the first batch of solution. When the
wood has reached equilibrium with the highest solution concentration, the treated wood
is air-dried slowly under conditions of controlled high humidity. The properly dried wood
should be, if possible, stored under conditions of less than 70% humidity. Sucrose-treated
wood has the appearance, density, and much of the strength of nondegraded wood.
E. Acetone-RosinTreatment
Acetone-rosin treatment involves the exchange of free water in the waterlogged wood
with a natural rosin, such as pine rosin. In this method, the wood is dehydrated completely
with acetone or a mixture of acetone and ether [ I 1 , l 21. It is important to remove all the
water because it is not compatible with rosin. The well-dehydrated wood is placed in a
saturated solution (67%) of rosin dissolved in acetone at about 50-55°C. After treatment,
treated wood is removed from the solution and the excess rosin is wiped off, This method
can give excellent results, but it involves heating volatile flammable solvents and is dif-
ficult to control. Acetone is also an expensive solvent. It is recommended only for treating
small artifacts.
F. Alcohol-EtherTreatment
In alcohol-ether treatment, waterlogged wood isfirst immersed in successive baths of
alcohol until all the free water is replaced by either isopropyl alcohol or ethanol. This is
followed by successive baths of acetone. The treated wood is then dried under vacuum to
remove ether. Since ether has a very low surface tension (0.7 dynekm), when it evaporates,
Preservation of Waterlogged Wood 813
the surface tension forces are so low that there is no appreciable collapse of the weakened
cell wall. Like acetone, alcohols and ether are highly flammable solvents, and they must
be handled very carefully. After the water is replaced by alcohols, camphor can be used
instead of ether. After the camphor fills the cavities and cell walls of the waterlogged
wood, the camphor then slowly sublimates without exerting any surface tension on the
cell walls. Thus the wood does not collapse, shrink, or distort.
G. In-SituPolymerization
PEG is a polymer. It takes time for the polymer to diffuse or penetrate into wood cells.
To overcome this problem, considerable work has been done in an attempt to develop
monomers which can be polymerized in situ after infusion [ 13,141. The polymerization
process maybe initiated using either heat or high-energy radiation to convert monomer
into polymer. The polymer functions as a consolidant, providing strength and dimensional
stability for the degraded wood cells. Many monomers can be used for this purpose.
Styrene, vinyl acetate, acrylonitrile, acrylates, and methacrylates are among the most com-
monly used monomers. Melamine formaldehyde resin has been used quite successfully
for this purpose. After infusion, copolymerization of unsaturated polyester oligomeric res-
ins with styrene by high-energy radiation has been employed successfully. The treated
products are mechanically strong. durable, and stable to a wide range of environmental
conditions [ 131.
H.Freeze-DryingTreatment
Freeze-drying is not actually a conservation process but rather only a dehydration process.
It is a physical process which sublimes ice. Water in the wood is frozen and leaves in the
form of gas without actually passing through the liquid state. There are several variations
of this process, including directly freeze-drying the wood, freeze-drying after impregnation
with a consolidant, freeze-drying after exchanging water in the wood with another solvent,
and consolidating the wood before freeze-drying and then freeze-drying in a natural en-
vironment [ 15,161. The technique is not as straightforward as it might seem to be, as there
is a considerable reduction in density when water freezes, and hence a physical expansion
of the material within the specimen. Accordingly, for freeze-drying to be effective in
treating waterlogged wood, a cryoprotectant must be added. The purpose of a cryoprotec-
tant is to reduce the volume change by continuing to increase in density as the mixture
cools through the water’s freezing point. The most commonly used cryoprotectant is PEG
with a MWof 400. I t is usual for waterlogged wood to be i n equilibrium with a 20%
solution of PEG prior to commencing the freeze-drying process.
Vacuum freeze-drying to remove water from waterlogged wood has been done at
several conservation laboratories. Of the many methods tested, pretreatment with PEG 200
and 400 worked well with undeteriorated wood. Highly deteriorated woods responded best
to pretreatment with a combination of higher-molecular-weight PEG or PEG dissolved in
solvents such as f-butyl alcohol [ 171. Directly freeze-drying wood was the process chosen
in the conservation of some parts of the Mary Rose [ 5 ] ,which was a ship built as part of
King Henry VIII’s naval expansion program.
1. SupercriticalDrying
The supercritical drying method was invented in the 1950s by Kistler 1181. This method
uses a high-density or supercritical fluid to replace the water in the wood. The supercritical
814 Hon
fluid is then removed from the wood by decompression, without forming a liquid phase.
As shrinkage is due to surface tension forces at a liquid surface, supercritical drying does
not damage the artifact. Carbon dioxide is a suitable supercritical fluid that can be used.
For treatment, water in the wood is first replaced with an organic solvent or alcohol. At
high pressure the carbon dioxide’s density is similar to thatof common liquids, and it
readily dissolves the alcohols.
FIGURE 1 Specific locations of wood specimens collected from the U.S.S. Cairo, Vicksburg Mil-
itary Park, Vicksburg, MS.
816 Hon
C. Wood Identification
The wood used in the construction of the U.S.S. Cairo was identified as a species of the
white oak group (QuCrcu.7 sp.). Some of the gundecks of the ship were made of southern
yellow pine (Pirzus sp.). It is not possible to identify the specific species of these group
of woods on the basis of their wood anatomyalone; fruit and flowers are needed for
positive species identification.
1. Chemical Analysis
Chemical analysis of specimens from the gunboat revealed that the wood fabric is severely
degraded, particularly at the surface layers. Chemical composition of the wood is tabulated
Preservation of Waterlogged Wood 817
in Table I . It is apparent that cellulose and hemicelluloses content have decreased signif-
icantly. Lignin appeared to suffer less degradation than cellulose. (The increase in lignin
content was due to the loss of cellulose and hemicelluloses.) Reduction of degree of
polymerization of cellulose is shown in Table 2. It also revealed that low-molecular-weight
fragments of detcriorated products can be extracted with co-solvents of ether and benzene
and with 1 % sodium hydroxide solution.
Degree of
Sample no.' polymerization
~~
Control
1 1471
4 467
5(c) 934
6(c) 1634
770
3 9(c)
9(s)
17(c)
17(s)
19(c)
19(s) 957
20(c) 1447.
20(s) 1027
'See Fig. 1 for location of speclmens; c, core;
S, surface.
FIGURE 3 Cross section of libriform fiber tissue showing extreme deterioration of the secondary
walls of the fiber. Normally these cells would have extremely thick walls with a very small lumen.
Magnification 300X.
readily from wood in a high-humidity climate. The addition of polyethylene glycol and
borate would only increase the risk of further damage of the ruined fabric. One hundred
and two years of submersion in the Yazoo River as well as a decade-long weathering in
Pascagoula and Vicksburg have brought the gunboat to a state of extreme fragility. As
discussed earlier, the gunboat has lost most of its polysaccharides(i.e., cellulose and
hemicellulose) and some parts of its lignin. Loss of these chemical components and phys-
ical stress have resulted in the formation of cracks, checks, crevices, and cavities in the
wood structure at the macro- and microscopic levels. Given these conditions, the buildup
of polymer consolidants within the cell walls appears to be the most acceptable approach
to restoring fabric solidity.
Six different types of polymer resins were selected to consolidate the deteriorated
wood samples. The resins were phenol formaldehyde, resorcinol-phenol-formaldehyde,
epoxy, polyacrylic, polyurethane, and wood rosin. With the exception of wood rosin, all
the polymer consolidants favorably restored the strength of the wood fabric. Of these
polymers, only resorcinol-phenol-formaldehyde resin exhibited acceptable appearance,
i.e., color and texture of wood after impregnation; wood treated withother resins displayed
a glossy surface which is not acceptable for restoration and conservation purposes. The
resorcinol-phenol-formaldehyde-treated wood specimens exhibited significant improve-
ment of hardness, reduced water pick-up, and good dimensional stability.
1. Resorcinol-Phenol-Formaldehyde Treatments
Various concentrations of resorcinol-phenol-formaldehyde resins can be prepared by us-
ing water and alcohol as the diluents, as shown in Tables 3 and 4. Several solutions were
820 Hon
100 16 0 80 120
100 30 0 80 236
I 00 30 0 120 76
100 30 0 I60 36
I00 20 20 I20 52
"K301 is a rcsorclnol-phenol-formaldehyde rcsin manufactured hy Koppcrs Company. Inc.
"cp = centipoise.
used in tests in which wood fabric was treated-each with a given solution and cured at
room temperature. Allof the formulations worked well with the wood fabric to achieve
good consolidation. Strength was improved. Original color and texture were retained. It
was found that the more alcohol used, the darker the color of wood surf'ace resulted. The
use of filler (peanut flour) could fill the large gaps and cavities properly. Scanning electron
micrographs showed that the resorcinol-phenol-formaldehyde resin deposited on the sur-
face of cell lumens properly to achieve good consolidation.
N. Chamcteristics qf Resorcinolic Resins. Resorcinolic resins are condensation
products of resorcinol with formaldehyde, or with various phenol-formaldehyde resoles.
The latter were used for this study due to their much lower cost. Although the resorcinolic
adhesives were initially found to be capable of being efficiently cured under neutral con-
ditions, it has generally been found to be advantageous for commercial purposes to cure
them under mildly alkaline conditions. Basic catalysts may provide bonds of strength at
low curing temperature, such as that used in this work. Resins are prepared from resorcinol
by reaction with efficient amounts of formaldehyde. The theoretical amounts necessary to
produce a cured resin are slightly in excess of 1 molof formaldehyde per mole of res-
orcinol. However, even 80% of this amount would be sufficient to give an unstable or
gelatinous product, since the molecules react disproportionately in bulk situations, so that
1 100 20 0 30 280
2 100 20 0 60 70
3 100 20 S0 0 200
4 100 20 IS 18 280
S I 00 20 25 2s 120
6 1 00 20 35 35 72
7 IO0 20 40 40 64
8 100 20 h0 20 S2
9 100 20 80 IS 64
20 80 0 -
10 100
Preservationof Waterlogged Wood 821
OH
OH
OH
I 7
CH2OH
some of the resorcinol molecules are left unreacted, while some of the first-formed olig-
omers acquire a superior share of the formaldehyde to form polymers of higher molecular
weights. Resorcinolic resins of between 0.5 and 0.7 mole of formaldehyde per mole of
resorcinol are made having infinite stability. At the point of eventual use, some additional
formaldehyde is provided and the resin is converted, within a short period of time, to a
very highly cross-linked resin. This product is characteristically insoluble, infusible, and
physically strong when properly aged.
b. Chemistry of' Resorcinol- Fornlaldehyde Resin Formation. Resorcinol readily
combines with formaldehyde to form methylol derivatives, with the methylol groups oc-
cupying either the positions ortho to both hydroxyl groups, or ortho to one and para to
the other. The meta position is not ordinarily reacted (see Fig. 4).
The reactivity of these methylol derivativesis so high that they cannot easily be
isolated in pure stable form. They continue to react under ambient, uncatalyzed conditions
with formaldehyde,resorcinol,phenol, or othermethylol-containingmolecules to form
polymer chains of higher molecular weight, with branched, as well as linear, configurations
of great complexity. These reactions continue until spatial considerations prevent further
interaction. In these polymers the resorcinol nuclei are joined together through methylene
linkages to give complex molecules as shown schematically in Fig. 5 .
c. Resorcinol-Pherzol- Fr~rnluldehyne Resin. In orderto produce resorcinol-phe-
nol-co-polymer resin, the phenol is combined with the formaldehyde before the resorcinol
is introduced. If the resorcinol were added to the initial charge, it would preempt most of
&
CH,---
CH,---
OH OH OH CH2---
FIGURE 5 A typical polymcric structurc of resorcinol-formaldehyde rcsin
822 Hon
the formaldehyde and form a gel, because it is many times as reactive as phenol. As a
result, most of the phenol would remain unreacted.
In order to obtain a block co-polymer, phenol would be combined withformaldehyde
in one reactor to form a resole, and resorcinol would be combined with formaldehyde in
a second reactor. The two resins would then be mixed, and the methylol groups on the
resole would combine with available nuclear positions of the resorcinol to form mixed
polymers. The other way to do it is to let phenol react with an excess of formaldehyde to
form a low-condensed resol. Resorcinol would then be added in sufficient proportion to
combine with all of the methylol groups, thus forming a resinous co-polymer.
2. ScanningElectronMicroscopyExamination
Since the resorcinol-phenol-formaldehyde resin consolidates wood fabric successfully, it
is assumed that the penetration and solidification of the polymer in the cell walls is effi-
cient. SEM was used to studythe characteristics of resin treated wood fabric. Typical SEM
pictures are shown in Figs. 6-9.
Figures 6 and 7 showed that the cross section of the red oak was coated with res-
orcinol resin. Some of the small cells still can be seen. Figure 6 shows that most of the
cells in the cross section were filled with resins. A large vessel on the left was also filled.
Vessels on the right were still left empty.
Figure 8 shows the tangential surface aspect. Some of the procumbent cells in the
body of the ray still can be seen, but are heavily coated with the resin. The surface of the
radial section is also heavily coated with resorcinol resin, in which some of the pit cavity
was filled with resin too (Fig. 9).
FIGURE 6 SEM micrograph of the cross section of red oak fabric of the gunboat, treated with
resorcinol-phenol-formaldehyde resin. The vessel at the left was filled with resin. The surface of
the lumen was also coated with the resin. Magnification 250X.
Preservation of Waterlogged Wood 823
FIGURE 7 SEM micrograph of the cross section of red oak fabric of the gunboat, treated with
resorcinol-formaldehyde resin. The surface was heavily coated with the resin. Many small cells
still can be seen. Magnlfication 250X.
FIGURE 8 SEM picture of the tangential section of red oak fabric of the gunboat, treated with
resorcinol-formaldehyde resin. One of the ray cells at the left was filled with resin. Magnification
250X.
824 Hon
FIGURE 9 SEM micrograph of radial section of red oak of the gunboat, treated with resorcinol-
formaldehyde resin. The surface was heavily coated with the resin. Some of the pits were entirely
coated. Magnification 250X.
It is obvious that resorcinol resin can penetrate, wet, and consolidate fabric favorably
to increase the strength of the deteriorated fabric. It may serve as a useful resin to preserve
the U.S.S. Cairo gunboat.
REFERENCES
1. D. W. Grattan and R. W. Clarke, in Conservation of Marine Archaeological Objects (C. Pear-
son, ed.). Butterworths, London, p. 164 (1987).
2. J. I. Hedges, in Archaeological Wood (R. M. Rowel1 and R. J. Barbour, eds.), American
Chemical Society, Washington, DC, p. 111 (1990).
Preservation of Waterlogged Wood 825
1. INTRODUCTION
During the past 50 years, synthetic polymers utilizing petroleum as their raw material have
been advanced. A wide variety of plastic materials is now being used commercially that
supportour daily life with considerablecomfort. Even plastics which have properties
similar to those of metals have appeared as engineering plastics.
Before the start of the synthetic polymer industry, there were a number of attempts
to obtain moldable materials from natural polymers, mainly from cellulose. Trials of de-
veloping cellulose derivatives into industrially acceptable materials [ I ] , as well as efforts
to identify excellent plasticizers for cellulose acetate (CA) [2,3], provide good examples.
However, these effortsbecame unpopular with thestart-up of theindustrialization of
petrochemistry.
Actually, many synthetic polymers exhibit properties that polymers of natural origin
do not possess, especially in relation to melt processability. Many opinions expressed i n
textbooks claim that cellulose has such a rigid backbone that it cannot be converted to
plastic materials. Because of these circumstances, it becomes understandable that attempts
to develop plastics of natural origin have not been emphasized during the past half-century.
Recently, however, several changes have occurred and provided motivation for the
authors to start and continue studies on the conversion of biomass into plastics. The first
change concerns gradually growing demands for circulating materials which are desirably
originated from biomass. One of the works of the authors’ group must be includable, in
which wood could be converted into thermally flowable material by chemical modification,
such as esterification and etherification [4-71. Thatis, plastics couldbeobtainedfrom
suchlow-cost materials as wood wastes. At the present, there are actually almost no
sophisticated methods or technologies generally available that can make use of biomass
wastes for the purpose of adding satisfactory value.Thus, wood plasticization canbe
considered as one attempt to pursue recycling technology.
The second motivation resulted from a recently occurring requirement for developing
biodegradable plastics. To meet this need,development of biodegradable plastics from
natural polymers becomes attractive, together with that of bacterial polyesters and synthetic
827
828 Yoshioka and Shiraishi
polymers (aliphatic polyesters and water-soluble polymers). Many reviews concerning the
biodegradable plastics have appeared recently [S- I O ] .
Among these biodegradable plastics, investigations in the field of bacterially pro-
duced polymers and synthetic polymers are more actively and extensively pursued than
polymers from natural origin, especially those from lignocellulosic and cellulosic materials.
Actually, even in recently published books on biodegradable polymers, it is difficult to
find descriptions of lignocellulosic biomass wastes being converted to meaningful biode-
gradable plastics [8-10]. These biodegradable polymers must not only be cost-effective;
they must also have performance characteristics that are comparable to common synthetic
polymers and they must be degradable in the cnvironment. These requirements, however,
are often mutually exclusive, and practical biodegradable polymers have not yet been
realized. It has been pointed out that it will take S - 10 more years before the development
of biodegradable polymers reaches a practical level [ l l].
It is known that the biodegradable chemical intermonomer bonds include glycosides,
peptides, and aliphatic esters [ 121. Thus, some of the most attractive materials with greatest
potential i n terms of cost, material applications, and environmental compatibility include
cellulose derivatives, especially cellulose esters. Among the cellulose esters, cellulose ac-
etate (CA) has been produced industrially in the largest amount. Thus, the big interest has
recently focused on the potential biodegradability of CAS.
Until the end of the 1960s, it was accepted as axiomatic that cellulose acetates having
DS 2 1.0 are resistant to hydrolysis by enzymes [ 131. In 1969, however, Cantor and
Mechalas [ 141 found that even cellulose diacetates (CDAs) having a degree of substitution
up to 2.5 could be degraded by microbial attack. Their investigation was carried out with
the objective of relating losses i n semipermeability of cellulose acetate (DS 2.5) reverse-
osmosis membranes to microbiological degradation. They were discussing the durability
of cellulose diacetate used for the reverse-osmosis membranes; that is, they did not have
any interest in biodegradable plastics. In that sense it can be said that the first people to
find the microbiological degradation of CDA in relation to biodegradable polymers were
the research groups of Eastman Chemical Company and the University of Massachusetts.
Buchanan et al. [IS] and Komarek et al. (161 demonstrated that CA with a degree
of substitution (DS) up to 2.5 can be degraded microbially. In the former study [IS], they
used two separate assay systems to evaluate the biodegradability of CA: an in-vitro cn-
richment cultivation technique (closed batch system), and a system in which CDA films
were suspended in a water treatment system (open continuous-feed system). The in-vitro
assay employed a stable enrichment culture. which was initiated by activated sludge into
a basal salts medium containing CA with 5 % (v/v). Extensive degradation of CDA (DS =
2.5) fibers was found after 2-3 weeks of incubation. In-vitro enrichments with CMA (DS
= 1.7) films were able to degrade 80% of the films in 4-5 days. Films prepared from
cellulose triacetate remained essentially unchanged after 28 days in the in-vitro assay. The
wastewater treatment assay was less active than the in-vitro enrichment system. For ex-
ample, approximately 27 days were required for 70% degradation of CMA (DS = 1.7)
fill11s to occur, while CDA (DS = 2.5) films required approximately 10 weeks before
significant degradation was obtained.Evidence for the biodegradation ofCA was also
obtained through the conversion of cellulose~l-“C]-acetateto “CO, in the in-VitI-0 assay
[ 151.
The last viewpoint was conclusively established in their successive study by use of
naturally derived mixed microbial culture derived from activated sludge and “C-labeled
CA and “C-labeled cellulose propionate [ 161. Biodegradation was measured in an in-VitrO
aerobic culture system that was designed to capture “CO, produced by the aerobic mi-
Biodegradable Plastics from Lignocellulosics 829
crobial metabolism of cellulose esters. More than 80% and 60% of the original lJc-poly-
meric carbon was biodegraded to ]'CO; for CA substrates with a DS of 1.85 and those of
2.07 and 2.57, respectively, over periods of 14-31 days. The amount of biodegradation
that was observed with cellulose[ I-'"C] propionate with DS of 2.1 I , 2.44, and 2.64 was
lower than that of the corresponding acetyl ester and ranged from 0.09% to 1.1 %. How-
ever, cellulose[I-"C] propionate with a DS of 1.77 and 1.84 underwent very rapid deg-
radation in the mixed culture system, with from 70% to over 80% conversion of labeled
polymeric carbon metabolized to "CO2 in 29 days. The high level of microbial utilization
of carbon from both cellulose esters and its conversion to CO, confirms the biodegrad-
ability of these polymers and the potential they have for total mineralization in natural
microbiologically active environments [ 16).
Gu et al. [ 171 studied the cellulose acetate biodegradation upon exposure to simulated
aerobic composting and anaerobic bioreactor environments. CA films with DS of 1.7 and
2.5 were exposed to biologically active in-laboratory aerobic composting test vessels main-
tained at approximately 53°C. The CA 1.7- and 2.5-DS films (thickness values of 1.3 X
10"-2.5 X 10 ' and 5.1 X 10" mm,respectively) had completelydisappeared by the
end of 7- and 18-day exposure time periods in the biologically active bioreactors, respec-
tively. On the other hand, almost no change in CA film weight losses was observed when
the samples were exposed in the poisoned control vessels(aqueous KCN was added),
showing the conclusion that CA film erosion during the composting exposures resulted
from, at least in part, biologically mediated processes.Treatments of CA 1.7-DS film
samples (1.3 X 10 '"2.5 X IO-' and 5. l X 10 mm thickness) in anaerobic serum bottles
with or without KCN poison gave the same natures of results mentioned above. Therefore,
it was concluded that degradation of the CA 1.7-DS films upon exposure to the anaerobic
bioreactors was due, also at least in part, to biologically mediated processes. Gu et al. [ 181
also reported the degradation and mineralization of CA in simulated thermophilic compost
environments. They studied the aerobic degradation of CA (DS 1.7 and 2.5) films exposed
for up to 7 and 18 days, respectively. The number- and weight-average molecular weight
(Mn and MW) values for both 1.7- and 2.5-DS CAS decreased significantly for extended
composting exposure times. For example, Mn of residual polymers (CA 1.7 and 2.5 DS)
decreased by 30.4% by day 5 and 20.3% by day 16, respectively. Furthermore, a decrease
i n the DS from 1.69 to 1.27 (4-day exposure) and from 2.5 1 to 2.18 (12-day exposure)
was observed for the respective CA samples. In contrast, CA films (DS 1.7 and 2.5) which
were exposed to poisoned control vessels for identical time periods showed no significant
changes in Mn and DS. Scanning electron microscopy (SEM) photographs of CA (DS 1.7
and 2.5) film surfaces after compost exposures revealed severe erosion and corresponding
microbial colonization. Similar exposure times for CA films in poisoned control vessels
resulted in only minor changes in surface characterization by SEM observations.
Theconversion of CAS (DS 1.7 and 2.5) to CO, was monitored by respirometry.
A lag phase of IO- and 25-day duration for CA DS 1.7 and DS 2.5, respectively, was
observed, after which the rate of degradation increased rapidly. Mineralization of 1.7- and
2.5-DS CA powders, reported as the percentage of theoretical CO, recovered, reached
72.4% and 77.6% in 24 and 60 days, respectively. The results of this study demonstrated
that microbial degradation of CA films exposed to aerobic thermophilic compost reactors
not only results in film weight loss but also causes severe film pitting and a corresponding
decrease in chain Mn and DS for the residual material. Furthermore, a high degree of CA
mineralization was observed by the significant attainments of CO, conversion.
Buchanan et al. [ 191 studied the influence of DS on blend miscibility and biodegra-
dation of cellulose acetate blends. They reported their findings on blends of CA having a
830 Yoshioka and Shiraishi
DS of 2.49 with those having a DS of 2.06. This blend system was examined over the
composition range of 0- 100% 2.06-DS CA employing both solvent casting of films (no
plasticizer) and thern-ral processing (melt-compressed films and injection molding) using
poly(ethy1ene glycol) as a common plasticizer. Thermal analysis and measurement of phys-
ical properties indicate that blends in the middle composition range are partially miscible.
while those at the end of the composition range are fully miscible. The miscibility of these
cellulose acetate blends is suggested as being influenced primarily by the monomer com-
position of the co-polymers. Bench-scale simulated municipal composting confirmed the
biodestructability of these blends and indicated that incorporation of a plasticizer accel-
erated the composting rate of the blends. In-vitro aerobic biodegradation testing involving
radiochemical labeling demonstrated conclusively that both the lower DS CA (DS 2.06)
and plasticizer significantly enhanced the biodegradation of the more highly substituted
CA (DS 2.49).
This last point, that the presence of the lower DSCA significantly enriches the
biodegradability of the more highly substituted CA, was already reported by Itoh et al.
[20,21]. They reported that, when the former, more degradable CA, exists in an amount
of more than 10%. the more highly substituted, thus less degradable CA can be signifi-
cantly enhanced in its biodegradability. It was estimated that because of the presence of
the lower-DS CA, microorganisms. usually not workable for destroying the higher-DS CA,
can attain the degradation ability through so-called acclimatization.
On the other hand, Sakai et al. [22] have searched for fungi that decompose CAS.
They demonstrated that Neisseritr SI'CCN can degrade CA with a DS at least up to 2.3. The
isolated strains, identified as Neisseria s i c m , degraded CA membrane filter (DS mixture
of 2.8 and 2.0) and textiles (DS 2.34) in a cultivating medium. Biodegradation of 1.81-
and 2.34-DS CAS on the basis of biochemical oxygen demand reached S 1-60 and 40-
45%, respectively, in the culture of N . sicccr within 20 days. It also has been suggested
that CA would undergo an enzymatic splitting by acetyl esterase in a first stage, down to
a DS of 1 .O, before the degradation would continue by the action of cellulase enzymes
[ l6.17,22].
Since cellulose diacetate (CDA) became thus recognized as a biodegradable polymer,
various trials have been undertaken to impart sufficient thermoplasticity to CAS in order
to render them melt-processable. This is because CDA, which has the greatest thermo-
plasticity among all kinds of CAS, fails to show adequate melting behavior without de-
composition or discoloring. Thus, lowering the flow temperature of CAS is necessary, and
it requires the addition of plasticizers and/or flow promoters.
Traditional plasticization of CAS has been accomplished by using conventional p k -
ticizers with low molecular weights, such a s phthalates. glycerol derivatives, phosphates.
etc. At present. phthalates and phosphates are used industrially i n procedures that are often
very time-c(~nsuming (i.e.,4-5 h per batch). These plasticizers are usLlally not suitable for
the prep:1ration of biodegradable polymers because of the harmful natures of their decom-
position products. In this connection, there have been several attempts to utilize aliphatic
polyesters of bacterial origins a s well as synthetic ones as plasticizers for CAS 123,341.
Of these experimental studies, the following are includable as typical examples.
Scandola et a l . reported miscibility of bacterial poly(3-hydroxybutyrate) (P(3HB))
with cellL1lose esters i n 1992 1231. They prepared blends of P(3HB) with cellulose acetate
butyrate (CAB) and cellulose acetate propionate (CAP) by melt compounding. It is known
that P(3HB)/CAB blendscontaining S-SOQI P(3HB) and P(.?HB)/CAP blends with 5-
h()(%p(3HB) arc transparent. stable homogeneous amorphous glasses, while blends with
highel- p(3HB) content are partially crystalline. When i n the amorphous state. both P(3HB)/
Biodegradable Plastics from LignOCellUlOSiCS 83 1
CAB and P(3HB)/CAP blends show aglass transition which decreases regularly with
increasing P(3HB) content, in excellent agreement with the behavior predicted for totally
miscible blends. Both dynamic mechanical thermal analysis (DMTA) and differential scan-
ning calorimetric (DSc) show that P(3HB) and CAB can crystallize from the blends only
at temperatures higher than the composition-dependent T,. When crystallization is induced
by thermal treatments. the melting temperature of the crystalline phase obtained depends
on composition, a s expected for miscible blends of crystallizable polymers. Besides the
strongly composition-dependent glass transition, another relaxation is observed, located in
proximity to the T , of P(3HB) and slightly shifting to higher temperature with increasing
CAB or CAP content. That is, another relaxation associated with mobilization of the IOW-
T, component is observed at alowertemperature. It was suggested that the two glass
transitions are the manifestation of two mobilization processes coexisting in blends which
appear in all respects to be single-phase, homogenous mixtures. After getting this infor-
mation, Scandola’s group studied the effect of a low-molecular-weight plasticizer on the
thermal and viscoelastic properties of the miscible blends of P(3HB) with CAB [24]. The
low-molecular-weight plasticizer selected was di-n-butyl phthalate (DBP). The plasicizer
DBP is miscible i n all proportions with both CAB and PHB. It is known that, analogous
to the polymeric CAB/P(3HB) blends, the two polymer/diluent systems [CAB/DBP and
P(3HB)/DBP] show a dual dependence on T, in composition. Thus, it can be said that i n
binary mixtures such behavior appears to be independent on the macromolecular or low-
molecular-weight nature of the Iow-T, component. On the other hand, addition of a fixed
amount of DBP plasticizer to CAB/P(3HB)blends with varyingcomposition [P(3HB)
content from 0 to 100%] causes a significant decrease of T, of the binary polymer blends;
the higher the amount of DBP in the ternary blend, the greater the T, depression. Con-
comitant with the expected plasticizing effect on T,, the presence of DBP also induces a
decrease in the characteristic temperature of the additional low-temperature transition ob-
served in CAB/P(3HB) blends. In the ternary blends, the temperature of such a transition
is a function of DBP content only, being independent of the relative amount of the two
polymers [CAB and P(3HB)I.
Buchanan et al. (251 reported on their work on CAB and bacterial poly-
(hydroxybutyrateco-valerate)(PHBV)co-polymerblends.They prepared blends in the
composition range 20-80 wt% of CAB and aco-polymer of PHBV by thermal com-
pounding. Measurement by I3C-NMR and gel permeation chromatography (GPC) showed
that no transesterification occurred during thermal mixing and that little change in molec-
ular weight occurred.Blendscontaining20-50% PHBV were found to be amorphous,
optically clear, miscible blends, while the blends containing 60-80% PHBV were semi-
crystalline, partially miscible blends. Both thermal and DMTA revealed the presence of a
high-temperature transition that was sensitive to blend composition and a low-temperature
transition whose position was uninfluenced by the blend composition. The high-tempera-
ture transitions of the 20-50% PHBV blendsclosely match calculated T V ’ s fora fully
miscible blend. It was proposed that the dual transitions in the blends containing 20-50%
PHBV arise from dynamic heterogeneity and not from a classical miscibility gap. Blend
morphology was found to strongly influence physical properties such as tensile strength
and tangent modulus. Blends containing 70% and 80% PHBV were found to exhibit tear
strengths that were superior to either of the blend components.
Buchanan et al. [26] also studied CAP and synthetic poly(tetramethy1ene glutarate)
(PTC) blends. They prepared blends of synthetic polyester, PTC, and CAP, in the range
of 50-90 wt% of the latter, by thermal compounding. During the compounding, no trans-
esterification and little loss in molecular weight occurred. PTC was found to be a low-
832 Yoshioka and Shiraishi
idly upon composting, while the injection-molded bars exhibited weight loss of 10- 12%.
For the CAP/polyester blends. the type of polyester (PEG versus PTC) in the blends made
no difference in cornposting rate. In general, a s the DS of the CAP decreased and the
amount of polyester in the blend increased, the rate of composting and the weight loss
due to colnposting increased. When the CAP was highly substituted, almost all the weight
loss was ascribed to loss to polyester. When the DS of CAP was below approximately
2.0, both components degraded.
Buchanan et al. also reported their work on composting of miscible CAPhliphatic
polyester blends in another journal 1301. In the article, they evaluated a series of miscible
blends consisting of CAP and PEG or PTC in a static bench-scale simulated nlunicipal
compost environment. Samples were removed from the compost at different intervals, and
the weight loss was determined before evaluation by GPC, SEM, and 'H-NMR. The type
of polyester (PEG versus PTC) in the blend made no difference in composting rate. At
fixed CAP DS. when thecontent of polyester in the blend was increased, the rate of
composting and the weight loss due to composting increased. When the CAP was highly
substituted, little degradation was observed within 30 days and almost all of the weight
loss wasascribedto loss of polyesters. Although the polyesters were still observed to
degrade faster, when the CAP DS was below approximately 2.0, both components were
observed to degrade. The data obtainedin this experiment suggested that initial degradation
of the polyester is by chemical hydrolysis and the rate of this hydrolysis is very dependent
on the temperature profile of the compost and on the DS of CAP.
Buchanan's group also developed and reported a bench-scale compost methodology
that emulates a high-efficiency municipal windrow composting operation [ 31 1. A series of
CA films, differing in DS,wasevaluated in this bench-scalesystem. In addition,com-
mercially availablebiodegradablepolymers such as PHBV and PCLwere included as
points of reference. Based on film disintegration and on film weight loss, CA having DS
less than approximately 2.2 composts at rates comparable to that ofPHBV. NMR and
GPC analyses of composted films indicate that low-molecular-weightfractionsare re-
moved preferentially from the more highly substituted and more slowly degrading CA.
Buchanan et al. submitted a patent application consisting of various data included
in their above-mentioned publications 1321.
Vazquez-Torres and Cruz-Ramos [33] studied binary blends of PCL with cellulose
esters(CDA,CAB. and CTA) by using D S c , DMTA, and wide-angle X-ray scattering
(WAXS)techniques.Aqualitativecomparisonwasmade with the results obtained by
polarizing optical microscopy. PCL having MW of 35,000-45,000 was used. The PCL/
CAB system was proved to be partially miscible, whereas PCL/CDA and PCL/CTA ap-
peared to be immiscible. A double-melting behavior was showed for PCL/CAB and PCL/
CTA blends. As these peaks did not shift by varying the heating rate of D S c run, this
behavior can be due to melting of two populations of crystals of PCL, which maybe
different in size. On the other hand, blends of PCL containing a small amount of CAB or
CDA seem to develop more crystallinity for the PCL than this polymer alone. The solvent
seems to have a certain influence on the thermal and morphological behaviors of the as-
cast blends of these three systems, affecting the extent of crystallinity of PCL, as well as
its T,,, and AH,. This finding is discussed in the light ofWAXS and polarizing optical
microscopy results.
Zhang et al. [34] studied the melting and crystallization behavior and phase mor-
phology of bacterial P(3HB) and hydroxyethyl cellulose acetate (HECA) blends prepared
by casting films by D S c , Fourier-transform infrared spectroscopy (FT-R), SEM, and PO-
larizing optical microscopy. The melting temperatures of P(3HB) in the blends were in-
Shiraishi
834 and Yoshioka
dependent of the blend composition with P(3HB) contents above 20%. The melting en-
thalpy of the blends decreased with increase in the HECA component and was close to
the additive value of the enthalpy of the two components.The glass transition temperatures
of P(3HB) in the blend were constant at about 8°C. No specific interaction between the
two components was found by FT-IR. The crystallization of P(3HB) in the blend was
affected by the HECA component, especially in the P(3HB)/HECA (20/80) blend. During
the DSC cooling run at a lower cooling rate, two separate transitions werefoundfor
P(3HB)IHECA (80/20), (60/40), and (40/60) blends, which corresponded to the crystalli-
zation of P(3HB) and the phase transformation of HECA from an isotropic phase to me-
sophase in the blends, respectively. The phase transformation of HECA from an isotropic
phase to a mesophase was almost independent of the P(3HB) component.
As has been shown in the above literatures [23-341, thermoplastic miscibilities could
be found between bacterial or synthetic polyesters and cellulose esters, and amorphous
optically clear miscible blends could be formed in certain ranges of combinations. It can
be pointed out that novel thermoplastic materials moldable to give transparent homoge-
neous sheets could be prepared by these blendings. The problems with these are the main
uses of cellulose esterssuch as CAB, the biodegradabilities of which have not been studied
to a sufficient level, and higher cost performance of CAB compared with CA.
At any rate, these studies have been continued with the intent to enhance the bio-
degradabilities of the materials obtained and furthermore to increase their thermoplastici-
ties by blending the polymeric materials with high biodegradabilities.
On the other hand, the addition of low-molecular-weight plasticizers has been taken
as the second type of plasticization of cellulose acetates. Buchanan et al. [29] reported
that CA can be effectively plasticized by thermal compounding with TEC.Thecom-
pounded resins were converted to compression-molded film and injection-molded bars.
Their biodegradabilities were evaluated and confirmed by composting.
Concerning this plasticization of CA with low-molecular-weight plasticizers, there
appeared in newspapers several announcements of commercialization of biodegradable
plasticized CA. One was an announcement from Planet Polymer Technologies. Inc. (Cal-
ifornia, USA) that CA plasticized with triacetine was entering the Japanese market with
the trade name of Lunare. Theother was from Daicel Chemical Industries Ltd., using
polycaprolactone oligomer having a molecular weight of 500 as plasticizer for CA.
As for methods for plasticization of CA, in addition to the two kinds of external
plasticizations mentioned above, which are usually adopted, it is possible to make use of
internal plasticization-that is, chemical modification or grafting methods.
In view of this situation, we have been attempting in the past 5 years to find novel
plasticizers and plasticizing procedures by which biodegradable thermoplastic polymers
can be obtained from CAS. In the early stage, attempts have been made to introduce
oligoester side chains into CA molecules by reaction of CA with dicarboxylic acid an-
hydrides such as maleic anhydride (MA) and succinic anhydride (SA) together with mono-
epoxides such as phenyl glycidyl ether (PGE), styrene oxide (SO), and allyl glycidyl ether
(AGE). Theoligoesterifications of CAS have been carried out by the use of a compounding
machine at high temperatures with constant kneading speed. The results of these attempts
are shown later,
With the advancement of this study, it became clear that CA must be sufficiently
graft co-polymerized to prevent the bleeding of co-produced homo-oligomers or homo-
polymers. From these findings came the understanding that the more effective the grafting
attained, the more ideal plasticization of CA can be effected. With this idea in mind, the
Biodegradable Plasticsfrom Lignocellulosics 835
authors have come to grafting work, in which CAS are plasticized by cyclic ester grafting
using tin(II)2-ethylhexanoate (SnEht,) as catalyst.
0 0
II II
OH
FIGURE 2 Effects of oligoesterification on L-40: upper left, untreated L-40; upper right, L-40
hot-pressed at 190°C under 15 MPa for 30 S; lower left, L-40/SA/PGE (100;11.0/33.8) just after
being kneaded at 120°C for 20 min without pretreatment; lower right, a sheet from the kneaded
sample (lower left) prepared by hot-pressing at 190°C under 15 MPa for 30 S.
FIGURE 3 Stress-strain curves of representative CAS plasticized with SA and PGE. SP-l0 (40):
LL-l0 (L-40)/SA/PGE = 100/11.0/33.8; prepared without pretreatment; kneading condition 120"C,
30 rpm (5 min)-90 rpm (20 min); hot-pressing condition 190"C,15 MPa, 30 S ; tensile test, span
40 mm, crosshead speed 5 mmlmin.
rom Plastics
Biodegradable Llgnoce~~u~osics 837
50 100
150 200 250 300
Temperature ('C)
FIGURE 4 Thermal softening curves for L-40 and its plasticized materials. L-40: untreated CDA,
SP-40: L-40ISNPGE = 100/11.0/33.8, MP-40: L40IMAPGE = 100/10.8/33.8, SS-40: U O I S A I S O
= 100/11.0/27.0. Prepared without pretreatment; kneading condition 120°C, 30 rpm (5 min)-90 rpm
(20 min); flow test condition, load 10 kgf; heating rate lO"C/min; T,, flow temperature.
FIGURE 5 Bleeding of plasticizer from sheets of plasticized L-40 and LL-10. SP-10: LL-IOISAI
PGE = 100/11.0/33.8; prepared on April 21, 1993; SP-40: L40/SA/PGE = 100/11.0/33.8; prepared
onApril19,1993: kneading condition 120°C, 30 rpm (5 min)-90 rpm (20 min); hot-pressing
condition 190°C, 15 MPa, 30 S; photographed on June 26, 1995.
838 Yoshioka and Shiraishi
A B C D E
FIGURE 6 Effect of catalyst (Na2COS) on weight gain of cellulose acetate. Kneading, 12OoC, 90
rpm, 15 min; flow test, die diameter 1 mm, length 2 mm; plunger, 1 cm2;load 5 MPa; heating rate
1O0C/min. A, L-4O/MA/PGE = 100/16.9/25.9 (30%); B, L-40/SAPGE = 100/17.3/25.9 (30%); C,
L-40OISAPGE = 100/21.7/32.6 (35%); D, LL-lO/SA/PGE = 100/21.7/32.6 (35%); E, L-40/SA/GMA
= 100/17.7/25.1 (30%). Each value in parentheses shows reactive plasticizers content (wt%) in the
startingmaterial. 0,withoutcatalyst(Na,CO,);withcatalyst(Na,CO,).
fore, a large amount of oligoester side chain attached to the CA molecule can beexpected
to enhance the affinity betweenthe modified CAS andthe homo-oligomer. Thatis, grafting
can effectively suppress or prevent the bleeding of nongrafting homo-oligomers. CMA can
be proceeded by grafting to a higher level compared with the case for CDA.
Based on these considerations, methods for enhancing the amount of grafting were
pursued by varying the combination of dibasic acid anhydride and monoepoxide, by ex-
tending the kneading reaction period, by using grafting catalyst (Fig. 6), and so forth [36].
By these trials, the grafting could be enhanced, which actually resulted in the suppression
of the plasticizer bleeding.
The biodegradabilities of the representative samples obtained were examined by a
soil burial test in a controlled environment (30°C, 80% RH) and by the measurement of
oxygen consumption in a closed system where test samples were exposed to standard
activated sludge [36]. Concerning the former, it was found in a few examples, as in Fig.
7, that the plasticized CA samples were degraded and disappeared within relatively short
periods (i.e., within 3-12 months). Results of the biodegradation test measured by means
of the oxygen consumption within a closed activated sludge suspension are shown in Fig.
8. It is apparent that all samples are subjected to significant biodegradation. The CMA
control sample, of which biodegradation had been literarily confirmed, was degraded more
slowly than any of the oligoesterified samples.
Samples disappeared
l *er l2 months
FIGURE 7 Changes SA-l0 and SP-IO specimens during the soil burial test in the incubator.
839
840 Shiraishi Yoshioka and
conducted in the presence of CDA. On the contrary, Kricheldorf and his co-workers studied
the polymerization of L-lactide catalyzed with SnEht, i n the presence or absence of benzyl
alcohol [42]. When SnEhtz and benzyl alcohol are used as a catalyst and a co-initiator,
respectively, NMR spectroscopicexamination of all polylactidesobtained by the ring-
opening polymerization revealed the presence of benzyl ester end groups but the absence
of 2-ethylhexanoate end groups [30]. This result means that the hydroxyl group of alcohols
plays an essential and direct role in initiating the ring-opening polymerization of cyclic
esters. In this sense, graft polymerization can be expected to occur selectively when the
cyclic esters are polymerized in the presence of CA and SnEht?.
Thus, the authors started a co-grafting study of s-caprolactone (CL) and I,-lactide
(LACD) onto CDA using SnEhtz as a catalystin order to realize grafting with considerably
2011 , , , , , ,
O O 20 40 60
Reaction time (min)
0 20 40 60
Reaction time (min)
high graft efficiency. In these cases, the reaction temperature was kept constant at 140°C
and the reaction time was changed from 0 to 60 min. The other reaction conditions are
shown in the footnotes of the relevant figures (Figs. 9-12).
One of the characteristics of this grafting is that the grafting reaction can proceed
rapidly and can be completed within 10-30 min (Fig. 9), and in correspondence with this,
the flow temperature decreases to about halfof that of CDA(Fig. 10). The products
obtained are moldable to films or sheets without using any plasticizer. Furthermore, it is
known from Fig. 1 1 that LACD is grafted more rapidly than CL, producing relatively
rigid and brittle products in the earlier stages and elastomer-like ones in the latter stages
of the grafting. From Fig. I I , it can also be said that the total molar substitution reached
its theoretical maximum value after 10 min of grafting. This result confirms that the graft
reaction proceeds at a high rate and shows that selective grafting is achieved completely
without significant production of homo-polymers or homo-oligomers. This supports the
reaction mechanism proposed by Kricheldorf et al. (421. Transparent sheets were obtained
depending on the reaction conditions, showing their amorphous nature. In this regard, the
triad structure of the grafted side chain was analyzed by means of high-resolution NMR
spectroscopy. An example of the NMR data is shown in Fig. 12, in which the E-oxycaproyl
unit is denoted by C and the lactidyl unit by LL. Each of the splitting spectral lines for
a, /3, y, &methylene carbons of c-oxycaproyl units and the methyl carbon of lactidyl unit
LL-1""
20 40 GO
Rcaction timc (min)
FIGURE 11 Effects of thereaction time on the molar substitution. Reaction temperature 140°C;
L40/(LACD + CL)/catalyst, 100/600/15 (by weight); LACD/CL, I.O/I.O(by mole). 0 , LACD; A,
CL: 0 , LACD + CL; ~ , theoretical maximum value of (LACD + CL).
842 Yoshioka and Shiraishi
Y 38 3‘ 31 0 1I 26
PP*
14 B 10 I, l8 11 I1
in a grafted product was assigned by reference to Kasperczyk and Bero’s work 1431. The
spectrum of Fig. 12 means that even for a grafted CDA (H-l) prepared by using a liquid
ratio of 2, LACD/CL = 2/5 (by mole), reaction time of 30 min, and reaction temperature
of 140°C, the introduced graft side chains are long enough to reveal triad sensitivity.
Appearance of LLCC, CCLL, and LLCLL sequences, total signal strength of which
are comparable or larger than that due to the CCC sequence, reveals meaningful occur-
rences of random polymerization of CL and LACD within the graft side chains. This fact
confers irregularity in graft chains and is considered to be related to the appearance of the
high thermoplasticity and amorphous nature found and discussed above. In other words,
it can be said that the analysis of the structure of the grafting side chain by means of
NMR spectroscopy showed that, although the side chains are composed of large amounts
of c-oxycaproyl or lactidyl block polymer portions, considerable amounts of randomly
polymerized parts coexist in the grafting chains, which confers high thermoplasticity and
amorphous nature on the grafted product obtained.
ical modification does not necessarily require special techniques; conventional and simple
methods work satisfactorily for this purpose [6,7].
The phenomenon of thermal flow can be explained in terms of internal plasticization
of wood. The introduction of large nonpolar substituent groups into wood can result in a
chemically modified material with high thermoplasticity. When small substituent groups
and/or polar groups are introduced, such thermal fluidity cannot be achieved; therefore,
the latter modification alone cannot produce plastic properties. However, this lack of plas-
ticity can be solved by external plasticization.
All of the above reveals that chemically modified wood derivatives can be considered
as novel biobased plastic materials, and experimental studies have targeted the develop-
ment of composites with enhanced physical properties.
There are various types of thermoplasticized wood derivatives. Among them, BzW
is known as a moldable material giving excellent mechanical properties. Indeed, a molded
sample of BzW showed a tensile strength of 42.7 MPa, which is fairly high when com-
pared, for example, with that of polystyrene (PS) (Styron 666, a product of Asahi Chemical
Industry Co. Ltd.). The tensile strength of Styron is 29.4 MPa under the same conditions.
Differential scanning calorimetric measurements have revealed that both polymers are
amorphous. Thus, the thermal softening behavior is similar, qualitatively, when measured
with a thermomechanical analyzer.
Quantitatively, there aresomedifferences. Experimental observations with a flow
tester revealed that PS undergoes flow at 153"C, while BzW starts to flow at 175°C. Almost
identical results are found when BzW is compared with polypropylene (PP) (J700G; MI
= 7, a product of Idemitsu Petrochemical Co. Ltd.), which is also widely used as a ther-
moplastic polymer.
However, even though the flow temperature of BzW is higher than that of PS or PP,
the temperature difference can be reduced to almost zero by blending polycaprolactone
(PCL) (PLACCEL H-4; Mn = 40,000, a product of Daicel Chemical Industries, Ltd.) in
an amount of 20% with BzW.
Figure 13 shows the dependence of melt viscosity (p)and shear stress (TU) on the
shear rate ( 7 ) of BzW and PP, measured with a Capirograph. Itis seen that both the
n
W
.-0
v)
W
c 5t
FIGURE 13 Melt viscosity and shearstress versus shear rate for BzW and PP.
844 Yoshioka and Shiraishi
viscosity and the shear stress of BzW reveal similar shear rate dependences as those for
PP. However, the values for BzW are one order larger than those for PP. This means that
there exists a greater resistance to thermal processing for BzW.
In order to reduce this resistance, blends of BzW with PCL were studied. since the
latter has considerably greater thermoplasticity. Since PCL itself has a low melting point
(ca. 60°C) and low tensile strength (19 MPa) at room temperature, its compounding with
BzW would be of great significance. Results obtained are shown in Fig. 14. As BzW
content increases. both the viscosity and shear stress are seen to decrease.
At a PCL content of 30%, the characteristic melt flow values converge with those
for PP, which are shown as dotted lines. Thus, the blended composite consisting of 70%
BzW and 30% PCL has practically the same fluidity as PP.
However, the mechanical properties of the blends were reduced by this blending as
shown in Fig. 15. This is a widely observed phenomenon in polymer blends. The me-
chanical properties of the latter. can, however, be improved by the useof appropriate
compatibilizers, which enhance the adhesion of interfaces between or anlong the domains
or phases of blended polymers.
As one of such attempts, the styrene-maleic acid anhydride co-polymer (SMA) was
used and its effect as a compatibilizer was investigated. The results obtained are shown
in Fig. 16. With an increase in the amount of added SMA, the tensile strength of the BzW/
PCL sheet recovered; and when the SMA content became 596, the tensile strength rose to
200 kgflcm' ( 19.6 MPa).
The composite also showed practical moldability as shown in Fig. 17. The photo-
graph displays trays that were vacuum-formed from sheets of the blends. This kind of tray
is commonly used in grocery stores. These trays possess not only reasonable strength but
they also have functional properties as biodegradability and photodegradability, as shown
later.
n
W
.-m0
U
c
a
M
-
0
M
-
0
Biodegradable Plastics from Lignocellulosics 845
GOO - -600 - 6
0 : Tensile strength
0:Break elongation
A ' Tensile modulus
-
0
o
400 -A - 4 0 0 .c2 . 4
-5M
C
F
-U
Y
z Y
m
U
.-- 200 -
U
0'
U
200 G-2
-U
.-
U
YI c
C
c
0
c
0.8
Q ? A
o'o', ,:, I I I c I .
0 -0
10/0 5/5 0/10
FIGURE 15 Relationship between mechanical properties of the BzW sheets and PCL content
SMA (part)
Furthermore, the BzWrPCL sheets were found to undergo faster biodegradation than
those of each individual sheet component, as revealed by tests carried out under the same
conditions (not shown). This fact is remarkable inasmuch as PCL is well known for its
high biodegradability.
W. CONCLUSION
In this review, recent studies on biodegradable plastics from cellulose and lignocellulose
were elucidated.It is known that thesestudies have just been developing. Even elastomeric
biodegradable polymers could be introduced from cellulose, which had been considered a
rigid polymer. This new development anticipates a wide and prospective future for that
kind of development.
REFERENCES
1. H. Marusawa.and K. Uda, Plrrstic Mrttc~riczlsSeries 1/71. Ct~llulosic~
Re.sirt.s, Nikkan Kogyo
Shinbun. Tokyo ( 1970).
2. M. Kita. Htctrdlm)k of N m d Wootklw Mtrteritrls, (T. Haragtlchi et a l . . eds.). Gihodo, Tokyo. p.
S94 (1996).
3. C. R. Fordyce, and L. W. A. Meyer, Irrtl. Errs. Cherrr., 32: 1053 ( 1940).
4. N. Shiraishi, T. Matsunaga, and T. Yokota. J. A p p / . Polyrm‘r Sri.. 24:2361 (1979).
S. N. Shiraishi, in Cllrrnistry of Wootl U t i l i x t i o n (H. Imamura, H. Okamoto, T. Goto. Y. Ynsue,
T. Yokota. and T. Yoshinloto, eds.). Kyoritsu. Tokyo, p. 294 (1983).
6. N. Shiraishi, in Ct~ll~tlosc Utj/j:cctiorr, Kt~stwrchrrrrd Kmwrds in Ce//lrlosic~.s (H. lnagaki and
G. 0 . Phillips, eds.). Elsevier Applied Science, London & New York. p. 97 (1989).
7. N. Shiraishi, in Wood t r r d Cc4/rrlo.sic. Chmrisrry (D. N. S. Hon and N. Shiraishi. eds.). Marcel
Dekker. New York & Basel. p. 861 ( 1991).
8. G. J. L. Griffin (ed.), C h ~ ~ r ~ ~nrtdi s t rTyw h r l o l o g y t~f’Biotlc~~~rcrtla/~le
P o / y r w r s , Blackie Academic
&L Professional, London ( 1994).
Takanori Arima
The University of Tokyo, Tokyo, Japan
Wood has been the most important construction and decorative material since ancient
times. The importance of wood will certainly increase in future years, since among all the
construction materials, it is the only major renewable resource. Further, in view of growing
energy consciousness and the resulting emissions of carbon dioxide to the earth's atmo-
sphere, greater wood utilization is to be encouraged, since wood processing requires rel-
atively smallamounts of energy. Since wood is a typical organic compound which is
constituted by absorbing carbon dioxide from the atmosphere, biodegradation and burning
leads to the release of carbon dioxide to the atmosphere. The effective use of wood waste
produced from industries that are processing or consuming wood and of timber recovered
from demolished houses has become a serious issue. Solving this problem could be an
effective means for saving wood resources and reducing the emission of carbon dioxide,
even though the processing requires a reasonable amount of energy.
For wood-based resources, the breakdown ofraw material basically follows a se-
quence of steps from logs to lumber, to chips, to fibers, to charcoal, and tinally to fuel.
After wood-based products such as building products, furniture. paper, etc., leave the
factory, they are utilized and eventually disposed of. With disposal, they again become
available as raw materials to various wood-utilizing industries if they are collected,
shipped, and reprocessed. In such a case, it can be said that there is relatively little negative
impact associated with this type of recycling. For example, there are many cases in Japan
where, after removing foreign matter, wood recovered from demolished timber structures
is reused in chip form for particleboard production or as a fuel source. If the material is
prepared in this way, it presents no technical difficulties and can therefore be recycled.
However, when the recycling conditions are not met, the wood material will simply be
thrown away or incinerated along with other garbage. The essential point is to address
adequately the problems of (1) usable life spans of wooden materials such as furniture
and building materials; ( 2 ) their subsequent quality considerations asa recyclable raw
material, including the removal of foreign debris; and ( 3 ) the collection ofthewood
material. It is vital that this should be coordinated in order to address effectively the issues
of environmental protection and energy concerns.
In this chapter, an overview is given of the current conditions and issues related to
the recycling of wood materials both from the wood industries and from the demolition
of wooden structures in Japan.
849
850 Arirna
Wood waste
from forest
l
l
l-lCharcoal
Ir
Wood particle cement board
Fiberboard
around the urban areas. Most of the hardwood chips are produced from logs of miscel-
laneous species. As most of the lumber mills in Japan are minor enterprises and are widely
dispersed, transportation costs and the low prices of imported chips often restrict the
effective use of wood waste. The wood integration plan as described could make a definite
contribution to better utilization of wood waste.
D. ParticleboardandFiberboard
The particleboard and fiberboard industries have made a considerable contribution to the
utilization ofwood waste produced from lumber and plywood mills, so-called ordinary
chips. Figure 2 shows the annual particleboard production in 11 Japanese industries, which
produce about 80% of the total amount of Japanese production. The shaded portion in-
dicates the ratio of recycled wood chips. The sum of the shaded portion reached about
20% of the total production. According to a survey 10 years ago, only a few mills used
Wastes
from mill Wastes
. . Wastes
from from
Grouping Total Log Total Own mill Other forest demolition
~ _ _ __ ~~ ~~
0
4.’
0
0 Recycledwoodchip
0
0
10
A B C D E F G H I J K
PB Maker
FIGURE 2 Particleboard production by 1 1 makers in 1992, and the proportion of recycled wood
used.
recycled chips. Changes in conditions for procuring raw material could have elevated the
importance of recycled wood in particleboard production over the last decade, as described
in Section 111.
E.Wood-ParticleCementBoard
Wood-particlecementboard of Japanese Industrial Standard is used as sheathing and
fireproof material for buildings. The wood chips are produced from lumber, plywood, and
from the demolition lumber of wooden structures.
InJapan, large quantities of plywood are used as formworkpanels for concrete
construction in buildings and civil engineering works. The waste panels from this work
are contaminated withcementand sand, and are usuallydumped or burned.Ways of
utilizing this material by crushing the waste panels and producinga type of wood-particle
cementboardhavebeen investigated, butwhetherthis is technically or economically
feasible depends on collecting wood waste constantly and expanding the market for this
type of board.
According to the Ministry of Health and Welfare’s “Change in Waste by Industry” report,
in 1994 approximately 8 million tons of wood debris were transported from demolition
sites. When this is converted to a volume basis, the figure exceeds 11 million cubic meters
of wood waste. The ratio of reuse is approximately 36%. Additional waste such as fur-
niture, pallets, leftover waste materials from new construction sites, etc., are thrown to-
gether, making up the difference. The actual volume of wood which can be reused varies
considerably according to the demolition method employed. Below, the desired situation
and problems of recycling or cascade use of wood demolition material are stated.
(1) The nature of wood waste has changed, the wood particles have become smaller,
more mixed with foreign matter, and are of many varieties. The use of a magnet can easily
remove the steel objects such as nails, but nonferrous objects such as aluminum, plastic,
cement, gypsum board, etc., present difficulttechnical problems inthe sorting and selection
of chips, which need to be solved.
Recycling of Wood and Fiber Products 853
( 2 ) The condition and amount of the chips are determined by the manner in which
the building is demolished-primarily, whether it was done by machine or by hand. AS
shown in Table 2, this determines whether the chips are of suitable quality for pulping,
board production, fuel chips, or just plain garbage, and has a tremendous effect on the
price and volume of chipsobtained.Therefore, it is necessary to establish a means of
sorting the chips according to quality so that they can be properly utilized to their maxi-
mum potential. Already, there is a slowly growing trend at the chip collection sites for a
minimum acceptable level of chip quality. On a regional basis, there are instances where
one can see thought has been put into the recycling of demolition material as well as cases
where there appears to be no consideration at all as to the reuse of the recovered wood
waste. The important deciding factor seems to be whether there is a facility to handle the
wood waste in the locality, or sufficient volume to justify such a facility.
(3) Ideally, the demolition and site preparation would be done carefully, but in many
instances these steps are rushed, exacerbated by a shortage of labor that does not allow
for hand demolition andwood waste segregation. As a result, the reliance on machine
demolition damages muchof the material and introduces foreign matter into thewood
waste. This also adversely affects the efficiency of the waste transportation, as it becomes
impossible to load the machine-demolished and -loaded material in a way which maxi-
mizes the use of the truck. This adds further cost to a low-value waste.
By wayof comparison, let us lookat an example of a machine-demolished but
material-segregated case and a totally machine-demolished case where no effort is made
to recover any of the material. In the sorted case, where the segregated demolition material
can be more neatly loaded onto a truck, the load will weigh about 0.72 ton per cubic
meter. compared to 0.38 ton per cubic meter for the unsorted case. Thus, a truck in the
sorted case will carry almost twice the load of one in the unsorted case. Also, among the
mixed waste to dump, the volume of wood in the unsorted case is about five times that
of the sorted case. If the demolition material were to be separated, it would be possible
to recycle a greater portion of it; however, when this is not done, the material will be
disposed of as mixed waste and will not be recycled. The reliance on mechanical demo-
lition arises from a limited amount of time available to devote to the job, the ease with
which demolition may be conducted by machine compared to hand demolition,and a
general shortage of labor. As a result, a larger volume of garbage is generated which is
difficult to reuse or recycle.
TABLE 2 Wood Waste from Timber Structures Demolished by Machine or by Hand 171
(4) The generation and collection of waste material varies widely across regions and
also by seasons, making it very difficult to obtain a stable supply and quality. Therefore,
individual companies that must consider profit implications of using a new material will
be reluctant to do so unless the supply is stable and the costs are low. Presently, the price
of chips from demolition material is determined by their quality and suitability for fuel,
board, or pulp production. At the same time, chips of pulping quality must compete with
chips of a virgin nature, while fuel chips must compete with fossil fuels. Foreign currencies
have a huge impact on the competition of these materials.
(S) With greater urbanization, it has become very difficult to build recycling facilities,
due to site restrictions. operating hours, and noise-prevention problems. For more distantly
located sites, working hours and transportation considerations create a large burden, mak-
ing it unprofitable and unfeasible to collect and process waste wood material. Because
efforts on behalf of the government and individual companies are being conducted piece-
meal and according to individual agendas, there is little active effort at coordinating the
policies of recycling and garbage disposal.
(6) The wood products industry has accepted the inevitability of using demolition
from only a cost-and-supply viewpoint, but the usage will be determined by a balance of
technological risk associated with manufacturing and the cost of raw materials. Accord-
ingly. if a new raw material were to have stable price and supply, the wood products
industry would naturally tend to use the new material. The recovered wood waste would
tend to become a reduction i n trash. In this scenario, competition between new raw ma-
terials and the demolition wood waste on these simple economic terms would potcntially
negate environmental conservation efforts in regard to the utilization of wood waste from
demolition sites.
From a technological standpoint it has become possible to utilize wood resources, includ-
ing recovered demolition material. in a systematic cascade-type approach described earlier,
where finally the wood is converted to carbon dioxide (CO,) by incineration. However,
the great volumc of material generated creates difficulties in recycling and incineration
with respect to social and economic efficiencies. According to the different methods of
demolition, primarily mechanical and hand. the quality of the wood waste is changed
where the material may be suitable raw material for manufactured wood products, chips,
or unusable trash. I n order to prevent the recovered wood from becoming trash to be
thrown away, it has become clear that an active and aggressive effort is needed. In spite
of the fact that wood waste recycling efforts have shown positive results compared to the
past, the recycling of demolition wood waste (demolition chips) still finds itself i n a very
difficult situation.
Wood chips arc generally segregated for utilization i n pulp, wood-based composite
boards, or fuel production. In the regions around Tokyo, Chubu, and Osaka which are
close to the typical urban areas, the composition of the chip material in the region’s 35
chip facilities is shown i n Table 3 . Approximately half of the chip materials are collected
from the waste of house demolition. The waste from concrete-form. pallet, and packaging
demolition also is considerably used because it is easy to collect and also to remove foreign
debris.
I n the casc of the board industrics, the large proportion of chips originating from the
demolition of houses can easily be seen, a s shown i n Table 4.
Recycling of Wood and Fiber Products 855
Table 5 shows the breakdown of industries which utilize wood waste chips, where
the users are forest products-related industries.Particleboard and wood-particlecement
board industries use recycled chip materials mostly as raw material. In fiberboard and pulp
industries, the ratio of fuel use tends to increase because the quality level of raw material
is required to be high. In the plywood, gypsum board, and dye industries, wood chips are
used as fuel for drying.
Table 6 shows, in order of relative importance, the primary considerations in chip
quality as a raw material for production of pulp and boards, and for use asfuel.Pulp
producers presently accept very little demolition wood chips for pulp production and are
not actively pursuing furtherexpansion i n demolition wood chipacquisition.Thepulp
sectors which do not accept demolition chips are involved primarily in the production of
paperboard products such as cardboard. Demolition chips have great difficulty competing
with virgin fiber sources due to the presence of foreign particles in the chips, cost, and
other factors which make demolition chips unattractive to pulp producers. At present the
supply and price of imported chips and waste from other wood industry sectors is both
stable and economical, which is probably the main reason for the lack of interest and
effort regarding greater demolition wood chip utilization in the pulp industry. In Japan,
the recycling of old paper such as newspaper is very high; nevertheless, the industry is in
a difficult situation and the incineration of large volumes of paper presents a considerable
burden when considering society's economic and social efficiencies and priorities.
If the chips from demolition wood waste are sorted, the negative aspects of these
chips become negligible or nonexistent and the wood-based boards produced from such
material can receive recognition by means of the Ecomark seal. Using demolition chips
alsosaves on dryingexpenses,as they are usually drier than material from alternative
sources. However, in recent years the percentage of nonferrous waste material mixed with
Grouping of material
Fuel,
Raw tons Other, tons Total, tons
industry tons (%) (%) (%l (%Io)
the wood chips has created difficulties, as the simple use of a magnetic sorting system is
no longer sufficient. Accordingly, it becomesquiteimportant to what degreesortingis
carried out at the demolition site. Moreover, the nonselection of materials, especially ones
which will create waste disposal problems when demolition ultimately occurs, becomes
vital from theconception of the building process. Along theselines, material selection
based simply on functionality should be discarded in favor of greater consideration of the
complete life cycle of the materials up to the ultimate conversion to CO, with incineration.
Because the scale of wood-based material factories, such a particleboard, is limited due
to site restrictions among other factors, geographic factors become important when con-
sidering the use of demolition chips for board production.
In the wood industries, the waste raw material which does not find its way into pulp
or board production is often burned to supplement energy requirements of the mills or
sent for incineration elsewhere. Demolition wood chips used for fuel have comeinto
greater use as an economic alternative to fossil fuels, and are often used as fuel for public
baths and burned in boilers of various industries. However, in recent years the use of fuel
chips has come into difficult times due to sagging prices and volumes. The economic costs
of conveyor maintenance, ash disposal, and personnel for wood waste boilers have become
increasingly burdensome, and many facilities are increasingly switching to oil and, espe-
cially, gas.
The oil shocks of the past sounded the alarm of too much reliance on limited oil
resources across all industries and sectors, including those related to building materials,
where too much energy was consumed in the production, assembly, transportation, and
demolition relatcd to those materials, at the same time came increased awareness of waste-
ful resource consumption. The use of wood and wood materials as a supplemental and
alternative energy source was examined; however, burning wood has a lower heat effi-
ciency, and in ordertoobtain the sameamount of energyas by burningoil, a greater
volume of CO2 would be generated and released to the atmosphere, contributing a greater
amount to the global warmingphenomenon. In otherwords, the move toward burning
wood as an alternative to oil, though possible, provides an inferior choice concerning the
environmental problems associated with excessive COz generation. When viewed from the
point of efficiency and economics, the only likely trend is a move to oil and gas. However,
at the end of its life cycle, wood will eventually be burned, and wood waste destined for
incineration (versus fuel generation) combined with the burning of oil or gas used to fuel
the incinerators themselves results in essentially a double release of carbon to the atmo-
sphere. As a consequence, the burning of wood waste should be the last step i n a series
of steps which utilizes the wood through a cascade-type recycling process. In this vein, it
is necessary to give due consideration to effective use of waste materials, including the
safety of the waste and any by-products from utilizing that waste. In other words, simply
incinerating wood waste materials should be avoided if possible, and consideration should
be given to systems which can capture and utilize the energy contained in the wood.
The city of Sapporo,Japan, has wastedisposalresourcefacilities which produce
solid-waste fuel, fuel chips, and chips for particleboard production. The uses of this waste
include heat generation for greenhouses, electricity generation, and supplemental energy
for industrial facilities. In order to carry out the proper sorting and material selection and
to adjust to the seasonal variations in the waste types and quality, the use of silos becomes
a vital factor in enabling operations to proceed smoothly.
For Japan’s construction industry, and the housing industry in particular, the era of em-
phasis on volume of construction has yielded to a new era where emphasis is placed on
quality. From this has arisen the “scrap and build” phenomenon, where large numbers of
structures built i n previous years when volume was emphasized are demolished and new,
high-quality buildings take their place. Because of the large numbers of buildings being
scrapped after a relatively short life span, the problem of waste disposal and recycling has
become an unavoidable issue.
It is possible to reuse and recycle the waste from residential demolition sites as well
asother public works and construction/demolitionsitesforvarious uses such as pulp,
board, and fuel chips in the cascade-type manner describedearlicr. However, the huge
volumes of waste being generated and the economic efficiency problems associated with
utilizing waste chips have created a very difficult situation regarding the full utilization of
the demolition chips. To meet the demand for reducing the volume of waste, i n the worst
case, it could simply be incinerated or decomposed in a manner with comparatively few
pollution problems, but disposing of the huge volumes of waste i n such a manner would
emit large quantities of CO, into the atmosphere, which should be avoided to the degree
possible.
858 Arima
From this background, the goal is to develop further the reuse and recycling alter-
natives while simultaneously reducing the volume of wood waste and keeping the CO, in
a stored form (i.e., not released into the atmosphere until as late a time as possible) by
converting waste wood resources into charcoal products with sequential carbonizing heat
treatment steps.
In the past the production of the charcoals, from ordinary general-purpose charcoal
to special types such as for filters, has relied on relatively clean wood resources such as
virgin material, and not material such as demolition waste. Charcoal is a very stable
material compared to other biodegradable materials.
Because of this, it is possible to change the method of disposal of charcoal, since it
may be added to just about any soil anywhere without any adverse effects, thus alleviating
some pressure on landfills.
Because the physical properties differ according to the different levels of heat treat-
ment, proper heat treatment is aimed at producing a charcoal suitable for prescribed uses.
Charcoal is used for soil improvement, absorption of odors, moisture, and for preservatives.
Charcoal was mixed with soil in ordinary house planters and plant growth and soil con-
ditions were shown to be improved.
REFERENCES
I. T. Arima, Rev. Forest Culture, 13:IO9 ( 1992).
2. Research Group on Forest Products Administration, Wood Supply and Wood Industry 1997,
Wood Industry Integration Plan Areas 27 1-283 ( 1997).
3. Research Group on Forest Products Administration, Wood Supply and Wood Industry 1997,
Wood Chip Industry, 189-196 (1997).
4. S. Suzuki, Proc. Int. RlLEM Workxllop, Tokyo, p. 223 (1995).
S. T. Arima, Proc. Preserlt arrd Future of‘ Wood Wctstc~,Japan Wood Research Society Hokkaido
Branch. p. I14 (1993).
6. T. Arima, J. J p 7 . Agric. Sys. Soc.. 8:69 ( 1992).
7. Osaka Industrial Waste Corporation. Report on Generating and Recycling of Wood Waste i n
the Construction Site ( 1988).
8. Japan Housing Wood Technology Center, Report on Recycling of Wood Waste, p. 36 ( 1992).
9. Association Housing Demolition, Composition Analysis on Waste from Wooden House Dem-
olition ( 1993).
IO. Japan Housing Wood Technology Center, Report on Recycling of Wood Waste (I). p. 6 (1994).
11. Japan Wood Research Society, Wood Scirncc~m t d T ~ ~ . k r ~ o /IV o g l~~t/rt.sfrirrl
y r r r d 1Ir~il.vW ~ t c . .
Wood Chip, p. 7 ( 1996).
12. Japan Wood Research Society. Wood Scierrcc. ( I d Techrrology IV Irltlustrirrl rrrrd Drti1.v W~ISIL’,
Roctrd, p. 104 (1996).
13. T. Arirna. Rcsectrck Ptcp”:s o r 1 H o u s i n g trntl L ~ I I I ~211255
, ( 1997).
14. K. Fujike, Pro(.. Prt~serrrt r r r d Future qf’ Wood Wrl.stC.Japan Wood Research Society Hokkaido
Branch, p.22 (1993).
15. T. Arima, Reports on Grant-in-Aid for Developmental Scientific Research Development of
Fundamental Technics for Recycling Wood Resources, p. 42 ( 1994).
16. T. Arima. S. Kobayashi, F. Socda, and M. Tokuda, RC): Forest Culfttrr. IS: I78 ( 1 997).
17. Japan Wood Research Society. Wood Scicwcc t u r d T d ~ ~ r o l o II! g y Irtthrstritrl rrrlrl L)tti!\. Wtr.stt>.
Chcrrt.otr/. p. 78 (1996).
1 8. Japan Housing Wood Technology Center. Report on Recycling 01’ Wood Waste, Charcoal. p.
66 ( 1997).
25
Pulping Chemistry
Goran Gellerstedt
Royal Institute of Technology, Stockholm, Sweden
1. INTRODUCTION
The pulping of wood for the production of paper, board, and other cellulosic products has
a large economic impact in several countries in the world. North America, Scandinavia
and, more recently, SouthAmerica are importantareasforpulpproduction,sincethe
availability of suitable wood species is high. Japan has a large pulp industry which, to a
large extent, is based on imported wood, whereas the pulp industry in Western Europe is
of amoderatesize. In otherareas of theworld,suchasChina, annual plants play an
important role as fiber supply, but here, as well as in other developing areas, new pulp
mills utilizing wood are rapidly increasing in number. Secondary fibers constitute an im-
portant fiber resource that has been used for paper production in countries with no or little
domestic wood supply. In the future, the importance of secondary fibers will increase
considerably, and legislation in several countries has been used to force the paper industry
to further increase the content of secondary fibers in a variety of paper products.
The chemistry of pulping has developed in parallel with the development in pulping
technologies. In several pulp-producingcountries, central researchinstitutes have been
created to support the industry with fundamental research on selected problem areas of
technical significance. Today, the chemicalknowledgeaboutpulping and bleaching of
wood is comprehensive, and numerous articles, reviews, and book chapters are available
on the subject. Nevertheless, due to the complexity of wood as a composite of polymers,
our knowledge about the chemistry and chemical reactions involved when wood is con-
verted to pulp of different types is still far from being complete.
In this chapter, the chemistry involved in the production of unbleached and bleached
mechanical and chemical pulps is discussed. with an emphasis on reactions which are
thought to play a significant role in technical pulping systems.
II. (CHEM1)MECHANICALPULPING
A. TechnicalOverview
In mechanical pulping, wood is disintegrated into fibers by grinding or refining, using a
rotating grindstonc or a disk refiner, respcctively. In the former case, wood logs are pressed
against the rough stone surface with simultaneous water spraying. In the latter case, refin-
859
860 Gellerstedt
ing is done by forcing wood chipsto pass from the center and outward between two
circular disks, with at least one rotating at a high speed. The inner surfaces of the disks
are covered with bars and grooves of successively smaller dimensions. The disintegration
of wood to mechanical pulp results in a broad distribution of fiber sizes, including course
fibers and fiber bundles, whole fibers, fiber fragments, and small particles. To some extent
the relative distribution of the various fractions is determined by the process and the
conditions needed to meet various end uses.
The resulting pulps are termed stone groundwood (SGW) and thermomechanical pulp
(TMP), respectively. Refining is also used to make refiner mechanical pulp (RMP), which
differs from TMP by the way the chips are pretreated in the process. Fresh spruce is the
preferred wood species for making mechanical pulp, although other species can be used
as well. The majority of the mechanical pulp produced is used for making newsprint and
other “wood-containing” printing papers in large integrated mills. The pulp yield in me-
chanical pulping is usually of the order of 95-989’0; i.e., a l l the wood constituents are
retained i n the finished product.
Brightness isan important quality parameter for spruce mechanical pulps. Values
around 60-63% IS0 are generally encountered. For several paper products this is too low,
and bleaching is frequently used to adjust the brightness to different levels in the range
of around 80% ISO, which, at present, constitutes the upper technical limit. The bleaching
of mechanical pulps is done in such a way that a l l wood constituents are retained, using
either dithionite under neutral conditions or alkaline hydrogen peroxide. This“lignin-
retaining bleaching” results in bright pulps which, however, suffer from brightness rever-
sion caused by either heat or light. For this reason, bleached mechanical pulps cannot
replace bleached chemical pulps in paper products where brightness permanence is
important.
Mechanical pulping can also be performed on chips which have been pretreated with
small amounts of chemicals, usually sodium sulfite, either alone (softwoods) or in com-
bination with sodium hydroxide (hardwoods). A somewhat lower pulp yield is obtained,
but strength and brightness properties are improved. These pulps are referred to as chemi-
thermomechanical (CTMP) or chemimechanical pulps (CMP), and they are often further
bleached with alkaline hydrogen peroxide.
0 +o 0
0 0
OCH,
a, - -
390 410 nm amax - 360 nm
-
(420 440 nm) (410 - 420 nm)
a,, - -
335 350 nm a,, - 310 nm
-
(340 350 nm)
FIGURE 1 Suggested chromophores in wood and their approximate light absorption maxima (val-
ues in parentheses are for the solid state).
of redox reactions and incomplete methylation, respectively, in the growing tree. Under
oxidative conditions, as in the presence of air, such structures are easily oxidized, with
formation of the corresponding quinone structure [6].
End groups of the cinnamyl alcohol type do not contribute to the wood color directly.
They constitute, however, structures of high reactivity in, e.g., oxidative processes. Since
the frequency of these structures is rather high in the native lignin (approximately 3-4
per 100 phenylpropaneunits) 171, evena reaction proceeding to aminuteextent may
contribute to color changes.
Among the major structural units in lignin, the p-1 structures are of particular in-
terest. In mild acidic or alkaline conditions such structures are easily converted into the
corresponding stilbene structure according to the reaction shown in Fig. 2 [ & S ] . The stil-
bene, in turn, is easily oxidized to form deeply colored quinones of various types.
H'or HO-
P
OH OH
FIGURE 2 Formation of a stilbene from p- 1 structures in wood.
swelling effect which directs the fiber-fiber fracture to this cell wall layer. This results in
a pulp having more intact and homogeneous fibers as compared to a mechanical pulp.
The reactions between sulfite and wood are thought to involve sulfonation of reactive
lignin structures as the predominant reaction mode. Thus, coniferyl alcohol structures,
which are abundant in the primary wall 1121, react readily with the strongly nucleophilic
sulfite ions to form a series of sulfonated coniferyl alcohol structures [ 13,141. The related
structure, coniferaldehyde, is even more reactive and easily converted into sulfonatesunder
CTMP conditions [ 151. Other aldehydes, such as vanillin structures, can be assumed to
form a-hydroxy-sulfonates in the well-known aldehyde-bisulfite equilibrium reaction.
Conjugated carbonyl structures, such as orrho- and para-quinones, will also react with the
sulfite ion to form the corresponding aromatic sulfonates [ 161. The various reactions taking
place between sulfite and wood chromophores result in a certain bleaching effect, leading
to CTMP being somewhat brighter as compared to TMP.
5 10 15
Time, min
color formation which, together with the native wood chromophores. determinesthe bright-
ness of the final pulp. Several factors can influence the degree of color formation in the
fibers, including wood quality, content ofbark in the wood, content of transition metal
ions in the wood, pH during fiber liberation, presence of sulfite, and presence of oxygen.
The importance of these parameters indicates that autoxidation reactions play a role in the
discoloration reactions. This is further supported by experiments both with wood [ 17,l S]
and with various model compounds [6]. The major contribution is likely to come from
the oxidation of reactive phenols in ligninlike catechols and hydroquinones. Such reactions
occur rapidly even at room temperature and result in the formation of the corresponding
quinone structure. In the presence of metal ions, such as copper or manganese, the oxi-
dation rates are further enhanced. Catechol structures in lignin are also able to form strong
chelates with ferrous/ferric ions; these are colored and, once formed,difficult or impossible
to remove [19]. Thus, the presence of high amounts of iron in the wood results in very
dark-colored mechanical pulp [20].
The presence of bark in defibration is detrimental, since reactive phenolic compounds
are easily leached out in the surrounding aqueous solution. Subsequently, these may par-
ticipate in lignin autoxidation reactions, and new, and more intensely colored addition
products linked to the fiber, may be formed [6].
The mechanical action on wood in grinding or refining can alsoalter the lignin
structure in such a way that more reactive structures are created. In a series of investi-
gations, several different lignin model compounds have been subjected to either ball mill-
ing at room temperature or added to coarse mechanical pulp which subsequently was
subjected to further refining [21]. The products obtained from the model compounds reveal
that mechanical action on lignin structures may result in both homolytic and heterolytic
reactions. Important products include structures with an a-carbonyl group in the side chain,
as well as stilbenes; the later originate from p-5 and p-l structures. One example is shown
in Fig. 4.
2. CTMP
The addition of sulfite to wood prior to defibration results in a certain brightening effect,
since some wood chromophores are eliminated due to sulfonation. The effect is limited,
CH0
I
Hi: HC
II It
CH CH
I
uo
FIGURE 4 Chemicalchange in a p-5 structure as the result of mechanical treatment of wood.
864 Gellerstedt
however, since new chromophores are created in the defibration process. Therefore, the
net result on brightness of the simultaneous creation and elimination of chromophores can
vary widely due to the exact reaction conditions before and during refining. In particular,
the presence of transition metal ions in the wood and/or the process water will affect the
extent of autoxidation and other homolytic processes.
The simultaneous presence of sulfite and a strong chelating agent such as EDTA or
DTPAin the defibration process has been shown to give a synergistic bleaching effect,
since the color-forming reactions are minimized while the bleaching effect of sulfite is
still operating. Thus, a brightness gain of the order of 8-10% I S 0 is attainable using this
mixture [22].
E. Bleaching Chemistry
1. Dithionite
Sodium dithionite (sodium hydrosulfite) is a mild reductive bleaching agent which is used
to obtain small or moderate brightness gains in mechanical pulps with a maximum increase
of around 8 I S 0 units. Although the reducing power in alkaline solution is much higher
(eo = 1.12 V versus e,, = 0.66 V), such conditions result in discoloration of the pulp due
to rapid alkali-induced discoloration reactions. Consequently, dithionite isusedat a pH
around 5-6 and at temperatures around 60°C. Under aerobic conditions, dithionite easily
decomposes into sulfite and sulfate according to reaction ( l ) , resulting in a loss of bleach-
ing power, even though the sulfite formed may participate in reactions of the type discussed
above. A slower decomposition is encountered under anaerobic conditions, resulting in the
formation of sulfite and thiosulfate [reaction (2)]. Thus, the bleaching solution mustbe
freshly prepared, which is done either by simple dissolution of the sodium salt in water
or by reduction of sulfur dioxide by sodium borohydride.
S,Oi- + H,O + 0, = HSO, + HSO,
2S,O:- + H,O = 2HSO; + S @ -
The bleaching chemistry of dithionite is a simple reduction of easily reduced chro-
mophores, such as quinones present in lignin [23]. Accordingly. the maximum color re-
duction is found in the wavelength region of 440-480 nm.
2. Hydrogen Peroxide
The bleaching of mechanical pulp with hydrogen peroxide in alkaline media requires a
prior elimination of transition metal ions from the pulp since, otherwise, a rapid decom-
position of hydrogen peroxide to oxygen and water takes place according to reaction (3).
The metal ion concentration is reduced by treatment of the pulp with DTPA or EDTA at
a neutral or slightly alkaline pH, followed by dewatering. The bleaching itself is carried
out in the presence of silicate, which acts as both a buffer agent and a stabilizer for the
peroxide. The brightness gain in peroxide bleaching is dependent on the charge of peroxide
and alkali; however, the bleaching liquor cannot be allowed to contain a large excess of
alkali, sincesuchconditions will result in darkening reactions in competition with the
bleaching reactions [24].
H,O, + H 0 2 = O2 + H,O + HO- (3)
The chemistry of peroxide bleaching has been thoroughly studied both with lignin
model compounds and with pulps. Analysis of pulp reflectance spectra before and after a
Pulping Chemistry 865
t A
1.6
1.4
1.2
FIGURE 5 Relative reflectance spectra of mechanical pulp beforehfter a peroxide stage [25].
peroxide bleaching reveals that two major areas of the spectrum are subject to changes,
viz., 320-400 nm and 400-600 nm, as shown in Fig. 5 [25]. The former of these is
attributed mainly to coniferaldehyde structures, which are known to react rapidly with the
peroxy anions to form the corresponding aromatic aldehyde(Fig. 6) [26]. In phenolic
units, the aldehyde may react further in a Dakin reaction, with formation of a hydroquinone
structure [27,28].
CH0
I
-
CH
CH0
I
H202 I HO- + 2 HCOOH
/ OCH, / OCH,
OR OR
PH
+ HCOOH
0 OCH,
OH
4 0
OCHj
H*~dHO” COOH
COOH
+ others
The elimination of coniferaldehyde structures with peroxide bleaching has been sup-
ported both by analysis of wood tissue using UV microscopy [29] and by chemical analysis
of pulp samples [30].However, this elimination does not seem to be quantitativeand,
even in highly bleached mechanical pulp, coniferaldehyde structures can still be found,
thus indicating that the bleaching process is not optimal.
Ortho- and yaru-quinones are also attacked by the strongly nucleophilic peroxide
anion and degraded to a mixture of carboxylic acids (Fig. 7) [31]. In a second reaction
pathway, quinones may also react with peroxide to form the corresponding hydroxy-qui-
none. Although occurring to only a minute extent, this reaction creates a new chromophore
which, despite being a quinone, has a low reactivity toward alkaline peroxide, due to the
acidic hydroxyl group.
The stoichiometry in peroxide bleaching has been found to vary depending on the
charge of peroxide. Thus, at high charges of peroxide, more peroxide is consumed to
obtain a certain brightening effect (Fig. 8) [32]. A high charge of peroxide is, however,
necessary to reach the highest brightness levels and, consequently, such bleaching systems
are quite expensive. The reason for the increase in peroxide consumption seems to be a
formation of radicals through a homolytical decomposition reaction of hydrogen peroxide
according to reaction (4) [33-351. This reaction is catalyzed by certain transition metal
ions, such as manganese, and results in the formation of hydroxyl radicals and superoxide
ions. In the absence of other substrates, these radicals will combine to give oxygen and
water [reaction (5)]; but the presence of pulp may alter the reaction pathways and a large
Relative Relative
bleaching effect bleaching effect
5- 5-
10.7 11.1 11.5 pH;
L- L-
P
3- 3-
2- 2-
1-
‘1 I
0.2 0.6 1.0 0.2 0.6 1.0
Consumption of b o 2, mol/kg pulp
FIGURE 8 Stoichiometry in the alkaline peroxide bleaching of mechanical pulp [ 3 2 ] .PH, denotes
the initial pH in the pure bleaching liquor.
Pulping Chemistry 867
number of both lignin and carbohydrate reactions are possible. The effect of these radicals
on wood chromophores is unknown; however, it has been shown that bleaching mechanical
pulp in the presence of a radical scavenger results i n asomewhat reduced bleaching
response under otherwise identical conditions [36].
F. Chemistry of Yellowing
Both unbleached and bleached mechanical pulps undergo brightness reversion when ex-
posed to daylight or heat treatment. For a given set of aging conditions, the extent of color
formation is highest when the brightness of the original pulp is high [37]. For bleached
mechanical pulps, these yellowing reactions prohibit a wide use of the pulp in high-quality
paper products and thus constitute a severe technical problem.
The reflectance spectra of bleached mechanical pulp before and after accelerated
light-induced yellowing reveal that two wavelength areas change as the result of irradia-
tion, viz., 310-350 nrn and 380-470 nm (Fig. 9) [38]. The former of these can be attrib-
uted to the formation of carbonyl groups in conjugation with aromatic rings; the latter is
from quinone structures.
The possibility of reducing or eliminating the yellowing of mechanical pulps is small,
due to the fact that several functional groups present in the lignin absorb daylight and
therefore act as sensitizers for photochemical oxidation reactions. The addition of ultra-
-
AAbs
0.3
t ~ 0 5
0.2
0.1
0 3
250 350 L50 550 Xnm
FIGURE 9 Absorption spectra beforehfter light-induced yellowing of mechanical pulp [ X ] .
868 Gellerstedt
0 1 2 3 Accelerated (h)
Irradiation time
&
1 2 3 Daylight (months)
0 OH
?
FIGURE 11 Suggested mechanism for the light-induced yellowing of mechanical pulps.
Pulping Chemistry 869
dithionite or peroxide bleaching, as discussed above. The first step in the reaction, the
absorption of light and formation of a phenoxy radical, can be accomplished by the phenol
itself [42,43] or by a reaction involving an excited conjugated carbonyl structure [44-461.
The latter, in turn, can be either native or formed in the peroxide bleaching stage.
At a later stage of yellowing, unspecific monohydric phenols in lignin can be con-
verted to phenoxy radicals and further oxidized to quinones by the action of light and
oxygen. Again, conjugated carbonyl structures can act as photosensitizers. In addition to
available carbonyl structures, lignin end groups containing conjugated double bonds are
themselves photo-active. Forexample, coniferyl alcohol structures can be oxidatively
cleaved to yield further aromatic aldehydes [47]. Photo-induced oxidation of benzyl al-
cohol groups to the corresponding ketyl radical may also constitute an important pathway
for discoloration, since such structures, when present in P-aryl ether structures, induce a
cleavage of the P-aryl ether linkage and direct formation of a phenoxy radical (Fig. 12)
[46, cf. 451.
The heat-induced (thermal) aging of mechanical pulps can take place i n the manu-
facturing process itself, during pulp storage, and in the finished paper product. The color
formation seems to be due to autoxidation processes of catechols and hydroquinones,
resulting in the formation of the correspondingquinones in reactions similar to those
discussed above [6]. In the absence of light, these reactions proceed through phenolate
anion intermediates, making them dependent on pH with a minimum around pH 5 . The
presence of certain transition metal ions, such as copper or manganese, accelerates the
reactions. A simple way of reducing the autooxidation of mechanical pulps is by addition
of sulfite at a pH of around 5-6. The simultaneous presence of a chelating agent such as
DTPA or EDTA prevents autooxidation of the sulfite and leads to long-term stabilization
[37].
CHzOH
-
I
Hi:
I: a
hv +
CH30
OCH, OCH,
/O
further reactions
FIGURE 12 Direct photo-induced cleavage of an oxidized p-0-4 structure in lignin with forma-
tion of a phenoxy radical.
870 Gellerstedt
The sulfite system contains two equilibrium reactions [reactions (6) and (7)], and the
pulping liquor is usually prepared by dissolution of sulfur dioxide in an aqueous solution
or a slurry of the appropriate hydroxide. The proportions are chosen such that a certain
excess of sulfur dioxide is obtained, resulting in a pH in the range of 1.5-4.0, depending
on process and product requirements. In NSSC (neutral sulfite semi-chemical) pulping, a
solution of sodium sulfite is used at a slightly alkaline pH.
SOz H,O + H,O = HSO; + H30' pK,, = 1.9 (6)
HSO, + H,O = SO:- + H30' pK,, = 7.0 (7)
The presence of a base is essential in sulfite pulping, and pulping with sulfur dioxide
alone cannot be done since such conditions would result in a dark wood residue with very
little lignin dissolution.This is due to the fact that the strongly acidic aqueoussulfur
dioxide promotes condensation reactions in the lignin, e.g., between aromatic and benzyl
alcoholic carbon atoms, and does not lead to any appreciable amount of sulfonation re-
action. The relationship between the total amount of sulfurdioxide and the required
amount of base in the cooking liquor has been determined experimentally. Thus, if a certain
relationship between the combined and total sulfur dioxide is maintained in the system,
the proportions between sulfonation and condensation are such that sulfonation is favored
[W.
The desired chemical reaction types in sulfite pulping are sulfonation and acid hy-
drolysis. Acid catalysed condensation is a nondesirable side reaction. In these reactions,
the lignin is solubilized through the introduction of a large number of sulfonate groups in
the lignin side chains. This solubilization is further facilitated by the acid-catalyzed hy-
drolysis of alkyl aryl ether linkages in lignin and of benzyl alkyl ether linkages between
lignin and carbohydrates. Further acid hydrolysis takes place in the polysaccharides, re-
sulting in a certain loss of pulp yield and in the formation of low-molecular-weight neutral
sugars suitable for fermentation to ethanol.
( I . Lignin Chemistry. The rates for sulfonation and dissolution of lignin in sulfite
pulping have been shown to be dependent on the pH of the liquor. Thus, sulfonation is
always the fastest reaction type, and at a low pH a complete sulfonation of all lignin units
takes place within a few hours of pulping (Fig. 13) [50].The hydrolytic reactions resulting
in cleavage of ether linkages between lignin and polysaccharides, as well as between lignin
units, seem to be somewhat slower. Together with possible restrictions in the mobility of
dissolved lignin fragments through the fiber wall, the resulting delignification thus shows
an apparent retardation as compared to sulfonation.
Around neutral pH values, the sulfonation of lignin is much more selective and only
around 20% of the phenylpropane units react, although at a comparably fast rate. There-
fore, the dissolution of lignin is limited. The initial sulfonation reactions in the pH region
are the same as those described for CTMP. Under more strongly alkaline conditions, sulfite
pulping gives chemical pulps with characteristics similar to kraft pulps [51,52].
The reactions of lignin in acid sulfite pulping have been studied using a variety of
model compounds representing the different structural units [53,54].The major reaction
mode is a sulfonation of benzylic carbon atoms through the intermediate formationof
carbonium ions, as outlined in Fig. 14. In principle, both phenolic and etherified lignin
units can react, although the rates for phenolic structures seem to be somewhat higher
[ S ] . A stronginfluence on the rate of sulfonation is also exerted by the pH under otherwise
identical conditions.
At higher pH values, i.e., around neutral and above, the sulfonation of lignin becomes
more selective and only phenolic structures react. The reaction involves addition of sulfite
872 Gellerstedt
CHZOH
CHzOH I CHzOH
I I
HC- HC-
I
HC- HC-0-R I o
SOz/ H @ HSO,0 0
-e + R-OH + H
/O /O
2. Organosolv Pulping
Pulping with ethanol at a high temperature can be done with wood species such as aspen
and poplar, as well as with annual plants [62,63]. Due to the facile liberation of low-
molecular-weight organic acids, such as acetic acid, the pulping process is carried out in
a weakly acidic environment. Hydrolytic reactions, resulting in a liberation of lignin from
the lignin-carbohydrate matrix, is assumed to play an important role [64]. In poplar spe-
cies this is particularly facilitated since the lignin contains para-hydroxybenzoic acid end
groups that are linked to carbohydrates through ester linkages 165,661. In grass species,
ferulic acid plays a similar role as “spacer” 1671. Homolytic cleavage of linkages in lignin
may, however, also occur and contribute to a lignin fragmentation and subsequent disso-
lution in the solvent 168,691. The carbohydrates in organosolv pulping seem to undergo
reactions similar to those occurring in sulfite pulping; neutral sugars, as well as furfural,
can be found in the spent liquor 1621.
finding that the addition of sulfide to a soda process gave a pulp of superior strength as
compared to soda alone. In addition, the pulping time could be considerably shortened.
Although the pulp was dark in color, the possibility of obtaining strong pulps from a
variety ofraw materials made the process attractive. The poor bleachability prevented
major expansion, however, until chlorine dioxide was developed as a bleaching agent in
the 1940s, thus permitting the production of fully bleached kraft pulps.
The kraft pulping liquor is a mixture of sodium hydroxide and sodium sulfide; the
charge of each chemical is usually expressed on a wood basis as effective alkali and
sulfidity, respectively. The sulfide system contains two equilibria, as shown in reactions
(8) and (9). The equilibrium constants are such that, under all conditions prevailing in the
digester, the sulfur is present as hydrosulfide ions [70]. These ions accelerate the rate of
delignification, the rate increase being proportional to the charge of sodium sulfide. No
similar influence on the rate of carbohydrate dissolution is found according to Fig. 16
[71], but from the figure it is obvious that the carbohydrate losses in kraft pulping are
substantial.
H,S + H,O = HS- + H,O+ pK,, = 7.1 (8)
HS- + HO- = S” + H 2 0 pK,, = -0.7 (9)
gOl
80
20
FIGURE 16 Dissolved amounts of lignin and carbohydrates in kraft and soda pulping [71].
Pulping Chemistry a75
If the dissolutions of lignin and carbohydrates in kraft pulping are plotted against
each other, the resulting curve can be divided into three distinct phases. These are termed
initial, bulk, and residual delignification, respectively; each has a different selectivity with
respect to carbohydrate losses. The kinetics and activation energies for the initial and bulk
phases have been determined from laboratory kraft cooks and are given in reactions ( 1 0)
and ( 1 1) [72]. Obviously, there is no apparent influence of either alkali or sulfide on the
delignification in the initial phase. The activation energy here is low. The bulk phase, on
the other hand, shows a strong dependency on alkali and has an activation energy typical
of chemical reactions. A certain influence of sulfide is also noticed.
dL
-- = k . L - [HO-I0[HS-]” E ,= 40 kJ/mol
dr
The apparent slight influence of sulfide on the delignification kinetics has, however,
a great impact on the resulting pulp characteristics. Both the transition point from bulk to
residual delignification and the viscosity of the pulp at a given kappa number will be
greatly changed if the sulfide concentration is so low that a sulfide deficiency occurs
toward the end of the initial phase [73,74]. Thus, a low availability of sulfide in this part
of the cook will result in a high amount of lignin when the cook reaches the residual
(slow) delignification phase. Continued cooking results in a pulp of inferior quality with
a low viscosity.
In the traditional kraft cook, the whole charge of chemicals present in the “white
liquor” is mixed directly with the wood, resulting in a very high concentration of alkali
at the beginning of pulping. Simultaneously, the charge of sulfide is such that after a short
period of pulping, the sulfide concentration may decrease to very low levels before it again
increases [70]. These concentration profiles are highly detrimental for pulping selectivity
and, in modern pulping, split charge of white liquor and recirculation of the used pulping
liquor (black liquor) have been introduced. These changes result in a more even alkali
profile during the cook and in a high initial concentration of sulfide, thus permitting a
more selective delignification [75].
2. Lignin Reactions
a. @Aryl EtherStructures. The dissolution of lignin in kraft pulping is mainly a
consequence of the cleavage of p-aryl ether linkagesin p-0-4 structures, resulting in lignin
fragmentation, together with the liberation of free phenolic hydroxyl groups. The latter,
together with the originally present phenolic lignin end groups, are ionized in the alkaline
solution. Phenolate salts are more soluble in water than un-ionized phenols, and this results
in dissolution of lignin fragments.The reactions of p-aryletherstructures have been
studied extensively, both in experiments with low-molecular-weight model compounds and
through analyses of black liquor and pulp lignin samples [76-811. In phenolic structures
of the guaiacyl type, the rate of cleavage has been found to follow the expression given
in reaction (12), which, in accordance with the relationship found for wood, shows no
influence of alkali or sulfide. In nonphenolic units [reaction (13)1, the corresponding re-
lationship shows a first-order dependency on alkali [82,83].
876 Gellerstedt
The published pseudo-first-order rate constants for these reactions [84] suggest that
the kinetic half-life for a phenolic P-aryl ether structure under realistic pulping conditions
is around 1.3 min at 170°C, whereas the corresponding value for a nonphenolic structure
is 45 min. These values are based on model compound studies and can be regarded as
ideal, since complete accessibility between model and pulping liquor exists.
The reaction mechanism for cleavage of a phenolic P-0-4 structure is shown in Fig.
17. The first reaction step, the formation of a quinone methide, is reversible, but once it
is formed this intermediate may react further with nucleophiles present in the pulping
liquor. The most reactive nucleophiles are hydrosulfide and/or polysulfide ions. Under
conditions where sulfide is absent or present in very low concentration, carbanions from
Reduction
Condensation
Further reactions
lignin or carbohydrates may compete successfully [85-881. The latter reactions would
result in condensation products, the presence of which is discussed further below.
By addition of hydrosulfide or polysulfide to the quinone methide, a thiol (polythio)
structure is formed (Fig.17) [76,89-911. In the alkaline solution this may undergo an
intramolecular attack to form a cyclic intermediate with simultaneous release of the p-
substituent. In further reaction steps, the cyclic intermediate loses elemental sulfur, re-
sulting in a temporary loss of sulfide ions in the pulping liquor. In the alkaline solution,
elemental sulfur reacts with sulfide to form polysulfide, which finally disproportionates
into hydrosulfide and thiosulfate (Fig. 18). Although this reaction sequence is not fully
supported by experimental data under conditions prevailing in a digester [92], it offers an
explanation of the formation of thiosulfate and the re-formation of hydrosulfide in kraft
pulping.
In competition with the addition of a hydrosulfide ion or a carbanion to the inter-
mediate quinone methide, two further reaction routes are possible on this intermediate.
The first of these is elimination of either the P-hydrogen or the y-hydroxymethyl group,
resulting in the formation of an enol ether structure 1761. Once formed, such a structure
is relatively stable under alkaline conditions and may thus survive the kraft cook [79]. As
a consequence, the lignin fragmentation will not proceed to an optimum extent. Thesecond
reaction type is less established but involves a reduction of quinone methides to the cor-
responding aromatic a-methylene structure, a reaction which has been indicated by anal-
ysis of lignin samples with "C-NMR [93]. Further support for this reaction type has been
obtained by alkaline treatment of quinone methide model compounds with alkali in the
presence of either glucose, anthrahydroquinone, or both [94]. Again, such reactions would
constitute a limitation of the lignin fragmentation, since p-0-4 structures are withdrawn
from the desirable cleavage reaction.
The cleavage of nonphenolic p-0-4 structures is dependent on the concentration of
hydroxy ions and, as shown above, the reaction is comparably slow. The mechanism
involves a formation ofan epoxide between the a - and the p-carbon atoms as shown in
Fig. 19 1951. Subsequent hydrolysis would give a glycerol structure, together with the
released @-substituent. Again, a competing reaction iwolving anions from lignin or car-
bohydrates may lead to condensation products [96].
The chemistry of p-0-4 structures in kraft pulping has been supported by analyses
of the remaining fiber lignin, as well as the corresponding dissolved lignin, as a function
of delignification. Analysis by acidolysis or thioacidolysis provides semiquantitative data
of the amount of remaining p-0-4 structures in lignin. Figure 20shows the expected
decrease in p-0-4 structures as pulping progresses but also that there is an amount re-
maining at the end of the cook (in softwood lignins). Furthermore, the dissolved lignin
contains an appreciable amount of this structure throughout the cook [78,81]. Obviously,
the rapid cleavage reaction indicated by the kinetic data above is not fully supported by
the analytical data. The reason for this is not known, but factors such as accessibility of
hydrosulfide, both in the fiber wall and in the lignin after dissolution, must be assumed
CH2OH
I
HC-OR
I
HC-OH
CHIOH CH2OH
I
H L - O e
I -
HC-OH
+ "Q-
/ \
CH30
OCH,
/o o
/
CHZOH
I
HC-OH
I
HC-OH
I
500 -
-
300 -
100 -
I 1 1 I I 1
0 20 40 60 80 100
Degree of delignificotion. D/'' on wood
FIGURE 20 The presence of noncondensed phenolic p-0-4structures in pulp and dissolved lignin
as a function of delignification (analysis by acidolysis).
to be important. The consumption of hydrosulfide and formation of sulfur in the fiber wall
may, for example, result in a certain deficit of sulfide at some of the reactive sites during
pulping. Even if the charge of sulfide is high and added to the wood through an efficient
impregnation stage, this type of deficit seems to be present during pulping. The presence
of p - 0 - 4 structures in the dissolved lignin can, for similar reasons, be assumed to be due
to the formation of molecules or molecular aggregates having negatively charged shells
with an interior that is inaccessible to the hydrosulfide ions.
b. Other Lignin Substructure.s. Phenolicphenylcoumaran (p-S) and 1,2-diarylpro-
pane-l ,3-diol (p-1) structures in lignin are reactive under alkaline conditions, since they
contain an a-hydroxy or a-ether function that can be lost in the reversible formation of a
quinone methide. In analogy with the reactions of p-aryl ether structures discussed above,
addition of a nucleophile or elimination of a hydrogen or formaldehyde may take place.
Thepresence of acarbon-carbonlinkage between the phenylpropane units, however,
prevents any fragmentation of the lignin macromolecule around this linkage. Model studies
indicate that phenolic p-5 or p-l structures preferentially undergo elimination reactions,
resulting in the formation of stilbene structures (cf. Fig. 2) [97]. The nonphenolic coun-
terparts, on the other hand, are essentially stable under alkaline pulping conditions.
Other major lignin linkages, such as those in biphenyl (S-S) and biphenyl ether (4-
0 - S ) structures, are completely stable in alkaline pulping, whether they are phenolic or
not. In phenolic units, the side chains in such structures can, however, react via formation
of a quinone methide according to the reactions outlined above. Recently, the behavior of
a phenolic dibenzodioxocin (S-S-0-4) structure has been studied by the use of a model
compound 1981. Under kraft cookingconditions, the aryl etherlinkageiscleaved to a
large extent, leaving the S-S structure as a major reaction product.
c. Aronzrrtic Metl1o.ry Groups. In softwood species, the lignin contains one aromatic
methoxyl group per phenylpropane unit, with the exception of compression wood lignin,
where some pcrrrr-hydroxy-phenylpropane units are present. In hardwoods, a mixture of
880 Gellerstedt
guaiacyl and syringyl structures are found, resulting in a methoxyl value per C-9 unit of
between 1 and 2, depending on species. In kraft pulping, these methoxyl groups are at-
tacked by the nucleophilic hydrosulfide ions, resulting in a partial demethylation and for-
mation of methyl mercaptan. The formation of a new phenolic hydroxyl group in lignin
adds to the hydrophilicity, whereas the other product, methyl mercaptan, is able to react
further as a nucleophile to form dimethyl sulfide by further attack on aromatic methoxyl
groups (Fig. 21) [99]. The two low-molecular-weight sulfur compounds survive the kraft
cook and constitute major components in the mixture of malodorous gases leaving the
digester.
From both softwoods and hardwoods, the amount of these products corresponds to
a degree of demethylation in the lignin of around 1-2%. In modern kraft mills, methyl
mercaptan, dimethyl sulfide, and small amounts of dimethyl disulfide, formed through air
(oxygen) oxidation of methyl mercaptan, and some hydrogen sulfide, are usually collected
and burned to eliminate almost all of the odor.
d. Condensation Reuctions. The presence of condensation reactions in lignin has
been used frequently to explain the slow rate of delignification toward the end of kraft
pulping, as well as the high demand of chemicals in a subsequent bleaching operation.
This assumption is further supported by a variety of lignin model studies demonstrating
that condensation reactions may easily occur, e.g., between quinone methides and different
types of carbanions or between carbanions and liberated formaldehyde (Fig. 22) [85].The
possibility of establishingequilibria between quinone methides and (ionized) hydroxyl
groups, as well as reactions between quinone methides and enol structures, have also been
suggested as modes of formation of new linkages between lignin and carbohydrates
[87,88].
However, conclusive evidence for the formation of condensed structures in the wood
residue during pulping, as well as their presence in the unbleached pulp, is not available.
Although the existence of comprehensive condensation reactions giving rise to diaryl-
methane structures during pulping has been suggested [ 100,101], the analyses used in that
work cannot be used unequivocally for making such an interpretation [ 1021. The presence
of lignin-carbohydrate linkages in unbleached pulps. however, is strongly supported by
indirect means; but, as discussed further below, it is not known whether these are native
or not. The presence of condensed structures and products in dissolved lignin have been
identified, although the extent of such reactions seems to be small [1031. Recently, it has
also been suggested that cellulose and glucomannan. when subjected to a kraft cook in
the presence of coniferyl alcohol, will result in a “grafting” reaction [ 1041. In summary,
(R = H or CH,)
CHS-S-S-CH,
however, the presence of condensation reactions in draft pulping is still a matter of un-
certainty which requires further attention.
3. Carbohydrate Reactions
In alkaline pulping, the initial consumption of alkali is caused mainly by the rapid hy-
drolysis of acetyl groups from hemicelluloses: galactoglucomannans in softwood and glu-
curonoxylan in hardwoods. Essentially all acetyl groupsare eliminated as acetic acid,
requiring an equivalent molar amount of sodium hydroxide. Thus, in softwood pulping,
this reaction consumes approximately 0.35 mol of alkali per kilogram of wood, whereas
the corresponding figure for hardwoods is in the range of 0.9 mol.
The yield loss in kraft pulping due to degradation and dissolution of polysaccharides
is substantial and constitutes a serious drawback of the process. Typical values for the
amount of wood (pulp) components after kraft pulping of pine and birch are shown in
Fig. 23. Thus, in addition to the desirable dissolution of lignin, all types of polysaccharides
are degraded to some extent. In softwoods, substantial amounts of the xylan and around
75% of the dominating hemicellulose, glucomannan, are lost. In hardwoods, the xylan loss
is almost 50%.
The predominant reaction responsible for the degradation and dissolution of poly-
saccharides is an alkali-induced stepwise elimination of monomeric sugar units, starting
from the reducing end; this is the so-called peeling reaction. Even at low temperatures,
around 100°C, the rate of reaction is substantial and a decrease in the degree of polymer-
ization (DP) of the polysaccharide is encountered [ 1051. For hemicelluloses having a short
chain length, such as in glucomannans, the change in DP facilitates nearly acomplete
dissolution already in the early part of the kraft cook. Xylans, being somewhat larger
polymers and having a less efficient peeling reaction, survive the cook to a larger extent,
but the remaining xylan chains can be assumed to have suffered a substantial loss i n DP.
The mechanism for the peeling reaction is shown in Fig. 24. The major polysaccha-
rides in wood are linked through 1 + 4 glucosidic linkages; the overall reaction results
in a loss of the terminal sugar moiety as an isosaccarinic acid and the formation of a new
reducing end group capable of undergoing the same reaction. Intermediate reaction steps
involve a reversible opening of the hemiacetal ring, reversible isomerization to a fructose
unit, a p-elimination of the 4-substituent, a keto-enol equilibrium, and finally a benzilic
acid rearrangement. As a side reaction, the p-elimination reaction may occur already in
the first-formed intermediate, the open aldehyde structure, with formation of a mdcr-~ac-
Pine Birch
-
Cellulose 35 34
Glycomannan 4 1
Xylan 5 16
Other carbohydrates and various
components
Sum of carbohydrates 44 51
Lignin 3 2
Pitch 0.5 0.5
Sum of components (yield) 47 53
FIGURE 23 Yield loss in kraft pulping for pine and birch. (All figures based on wood.)
Pulping Chemistry 883
cell-0
H QH
0
(OH
OH - cell-oH 0
(OH
/y-"oz, OH
Metasaccharinic
acid
-
("stopping" reaction)
H0
& CHZOH
p-elirn cell-0
(Fructose unit)
I
H OH
HoYCooHCHIOH
lsosaccharinic acid
carinic acid. For cellulose, the latter reaction, termed the stopping reaction, occurs when
an average of 65-70 sugar units have been peeled off the chain.
In softwood xylans, the presence of arabinose substituents in the 3-position favors a
direct @-elimination from the corresponding open aldehyde form of the end group. This
results in a somewhat higher stability of this xylan as compared to a hardwood xylan,
since the stopping reaction is facilitated. Both softwood and hardwood xylans also contain
4-0-methyl-glucuronic acid as substituent in the 2-positions, thereby making the isomer-
ization to a terminal keto sugar more difficult and decreasing the extent of the peeling
reaction.
In alkaline pulping, the 4-0-methyl-glucuronic acid substituents in xylans are to a
large extent converted to the corresponding unsaturated structure by loss of methanol (Fig.
25) [ 1061. The formed "hexenuronic acid" groups, which are still attached to the xylan
backbone, are rclativcly stable under alkaline conditions and contributc a substantial pro-
portion of the acidic groups in kraft pulps (Fig. 26) [ 1071.
884 Gellerstedt
HO' ~
- CHaOH
cH30*1H0
H0
"Xyl-xyl-xyl-
H0
"Xyl-xyl-xyl-
H0
The extent of peeling reactions is high in the beginning of a kraft cook, resulting in
a rapid loss of hemicelluloses from the wood. In later parts of the cook, a further loss of
carbohydrates can be initiated through alkaline hydrolysis of polysaccharide chains. This
reaction occurs randomly along the chain and results in the formation of shorter chains or
alkali-soluble chain fragments(Fig. 27). Since each cleavage reaction alsoleads to the
formation of a new reducing end group, this in turn may rapidly undergo further degra-
dation by the peeling reaction ("secondary peeling").
20 I " 175
18 - l70
16 - HexA
W -8-
4 165
Y
3;
J
14-
12- "
MeGlcA
Ara
"
160
155
g
3
5a
0 10-
S -El-
0 W
o a- 150 t
l
remperature
i0 5F
E
6' 4 145
4 '
140
2'
n - 135
"0 50 100 150 200 250
TIME, min
FIGURE 26 The presence of hexenuronic acid groups i n pulp xylan as a function of cooking time
in a kraft cook [1071.
Pulping Chemistry 885
+ X
0
9+
+
C
.*
'0
Xt
886 Gellerstedt
The reaction mechanism in the oxidation of a carbohydrate end group can be envis-
aged as involving the enediol intermediate which is oxidized to the corresponding diketo
structure (Fig. 28). A benzilic acid rearrangement then results in the formation of aldonic
acid structures [cf. 11 I]. The reductive reactions give rise to the corresponding sugar
alcohol or thioalcohol end groups with borohydride or hydrogen sulfide, respectively.
The oxidation of carbohydrates with either polysulfide or AQ results in the formation
of hydrosulfide ions or anthrahydroquinone (AHQ). In the alkaline pulping liquor, the
latter is ionized and may add to quinone methides in lignin analogously to the hydrosulfide
ion. The adduct formed, when present in a P-0-4 structure, has been shown to decompose
with cleavage of the P-aryl ether linkage and re-formation of AQ, thus resulting in an
efficient AQ redox cycle (Fig. 29) [ 112,113]. An alternative mechanism, which has been
supported by several experimental studies, indicates that the reaction of AHQ with lignin
involves single-electron-transfer reactions, leading to efficient P-aryl ether cleavage [ 1 141.
With either mechanism, the cleavage reactions result in the formation of a cinnamyl al-
coholstructure. In analogy to kraft pulping, a mixture of sodium hydroxide and AQ
therefore acts as a powerful pulping system able to give “kraftlike” chemical pulps. In
practice, however, AQ is used mostly as an additive in kraft pulping and, like polysulfide
addition, this will result in a yield increase.
The oxidizing power of polysulfide is not restricted to the carbohydrates; certain
lignin structures can be oxidized as well. Some examples are shown in Fig. 30 [89,115].
Since oxidized lignin structures generally are more easily degraded in alkaline media, it
can be assumed that the presence of polysulfide ions in kraft pulping will facilitate the
delignification. Further support for such reactions was recently obtained by impregnating
wood with polysulfide prior to kraft pulping and observing a better selectivity in the cook
[ 1161.
-0
CHIOH
= HOQ
-0
C C H O 7
-0
GOH
OH OH
CHIOH
I
0 OCH3 0
0
OH
HC-0
OH
OCH,
- PS 0. CH,
I
c=o
+ &H3
OH
OCH,
CHZOH CH0
I I
CH CH CH3
II I It
PS
OCH3
OH
FIGURE 30 Alkaline polysulfide (PS) oxidation of an enol ether and a coniferyl alcohol structure.
at best semiquantitative, indicate a remaining amount, around 15% in spruce kraft pulp
lignin [78]; this figure probably constitutes a lower limit. In birch pulp, on the other hand,
the figure seems to be much lower [SO]. In addition to intact p-0-4 structures in the
remaining kraft pulp lignin, some enol ether structures can also be found [79]. Obviously,
the availability of hydrogen sulfide ions at the reacting sites in the fiber wall is not suf-
ficient to suppress the elimination reactions completely throughout the cook, as discussed
above.
The presence of phenolic hydroxyl groups in lignin determines the solubility in
alkaline solution. To some extent, such groups are present in the native lignin, but the
liberation ofnew such groupsduring pulping is essential for the lignin dissolution to
proceed efficiently. In dissolved lignins from a spruce kraft cook, it has been demonstrated
that the content of phenols should be in the range of 3-4 mmol/g of lignin for attaining
solubility in the pulping liquor; higher values are needed for the higher-molecular-weight
molecules. In addition,a kraft lignin contains around 0.7 mmol/g of carboxyl groups
1I17,l IS]. In the residual fiber lignin, the values are much lower and the content of
phenolic hydroxyl groups changes from approximately 0.7 in wood to 1.5 mmol/g of lignin
in the unbleached pulp with a kappa number around 30 [ 1 17,1191.
The desirable fragmentation of lignin in kraft pulping, brought about by the cleavage
of p-0-4 structures, can be restricted due to a low availability of hydrosulfide and by
nondesirable side reactions. The reduction of quinone methides mentioned above is one
such reaction, but condensation reactions between lignin and polysaccharides would also
Pulping Chemistry 889
act in a restrictive way, since the resulting product would be of much higher molecular
weight and have the possibility of being firmly attached to the polysaccharides. The for-
mation of such linkages during pulping has been suggested [ 1201, but lignin-carbohydrate
linkages may also be native. It has been found that the residual delignification phase is
accompanied by a slow dissolution of lignin, which successively contains higher amounts
of polysaccharides, most notably xylan [ 1181. Extraction of unbleached kraft pulp with
strong alkali at an elevated temperature will also result in the dissolution of a material
which contains lignin linked to xylan as the predominant polysaccharide [ 12 1 1.
It is reasonable to believe that the lignin reactions in kraft pulping proceed toa
different extent in different parts of the fiber wall and in different types of fibers. Therefore,
all analytical data on residual lignin must be regarded as averages, and the true values for
phenolic hydroxyl groups, etc., must, in fact, include a range of values. The lignin mac-
romolecules also include a range of structural features from rather intact “native” mole-
cules to those which have been severely degraded and chemically changed as a result of
the pulping conditions. In line with this hypothesis, it has been shown that a substantial
portion of the residual lignin can be leached out of the fibers in a process which is governed
by a high temperature and the presence of alkali [ 122,1231.
b. C~rr1~okyclmte.s.The peeling reactions occurring in the polysaccharides during
pulping result in the formation of acidic end groups. In addition, some side groups in the
hemicelluloses are either eliminated, such as galactose from galactoglucomannan, or chem-
ically modified, like 4-0-methyl-glucuronic acid from xylan. The latter, together with the
formed unsaturated hexenuronic acid groups, add a substantial portion of the acid groups
i n unbleached kraft pulps, the total of which can be around 60-80 mmolkg of (softwood)
pulp [ 1241. The presence of hexenuronic acid groups also add to the kappa number of the
pulp (Fig. 3 1 ) [ l 25,1261. In recent work, it has been found that this structure alone cor-
responds to around 2 kappa units inan unbleached softwood kraft pulp (kappa no. 18)
whereas the corresponding figure for a hardwood pulp (kappa no. 1.5) is around 5 units
11271.
Pulp Kappa
Hexenuronic
number acid, Kappa number
pMollg pulp equivalence
Birch, unbleached
11.3 3.6
14.0 64 5.5
14.5 4.9
15.9 85 5.6
18.8 5.8
Birch, bleached
6.0 53 4.6
4.5 39 3.4
Pine, unbleached
18.2 1.2
18.6 22 1.9
18.0
FIGURE 31 Contribution to kappa nulnbcr of hcxcnuronic acid in solnc unbleached and bleached
kraft pulps.
890 Gellerstedt
W. PULP BLEACHING
A. TechnicalOverview
The poor selectivity and the slow delignification encountered in the final delignification
phase in kraft (as well as in sulfite) pulping cannot be tolerated from a pulp quality point
of view. Consequently, the cook is interrupted and, when required, the pulp is further
delignified and bleached using oxidation as the principal reaction mode. The residual lignin
in chemical pulps, notably kraft pulps, is difficult to remove, and several oxidative steps
are required with intermittent extraction of the lignin which was oxidized in the previous
acidic stage. Contrary to mechanical pulp bleaching, the complete bleaching of chemical
pulps to full brightness must be accompanied by the almost total removal of all residual
lignin from the fibers. Obviously, the chromophoric structures introduced in the pulping
process are such that no bleaching agent can be used for the selective oxidation of the
colored groups without a simultaneous dissolution of the lignin.
The most efficient bleaching agent for chemical pulps is elemental chlorine, which
has been used for several decades (together with hypochlorite) as the predominant oxidant
in chemical pulp bleaching. With the development of chlorine dioxide bleaching, the pos-
sibility of making fully bleached kraft pulp in sequences such as C E D E D* was realized
and introduced in the 1940s.
The growing awareness that elemental chlorine might produce chlorinated products
with a negative impact on the environment and the large leakage of organic matter oc-
curring from bleaching plants paved the way for further changes in the bleaching tech-
nology. Today, these changes are aimed at strongly decreased levels of chlorinated organic
products. As a consequence, new possibilities arise for reducing the use of fresh water and
closing up the pulp mills.
The discovery that magnesium compounds excerted a positive influence on the se-
lectivity in oxygen (0)bleaching opened up the possibility of starting the bleaching se-
quence with an 0 stage. Here, around 50-6570 of the residual lignin in a softwood kraft
pulp can be eliminated without detrimentally affecting the pulp strength characteristics.
The dissolved material can be introduced into the chemical recovery system, thus consid-
erably reducing the content of organic matter in the bleaching effluent.
In modern bleaching technology, further changes have been introduced together with
new bleaching agents. The use of elemental chlorinehas been reduced or completely
eliminated, while the use of chlorinedioxide has increased. In elemental chlorine-free
(ECF) bleaching, only chlorinedioxide is used. Alternatives, not involving the useof
chlorine-containing bleaching agents, have been developed (totally chlorine-free, TCF,
bleaching) and include oxygen, hydrogen peroxide (P), and ozone ( Z ) as major oxidants.
This development has resulted in large improvements in the effluent quality and in in-
creasing possibilities for reducing the discharge of both water and organic material from
pulp mills.
Bleaching reactions with oxygen, hydrogen peroxide, or ozone involves, in addition
to ionic reactions, hydroxyl radical reactions. Oxygen is itself a radical, and therefore the
initial reaction with lignin is a one-electron reaction giving rise to the formation of SU-
peroxide ions, hydrogen peroxide, and organic peroxides, which in turn can give hydroxyl
radicals in secondary reactions. Hydroxyl radicals can also be formed in hydrogen peroxide
and in ozone bleaching stages. The formation of the highly reactive and unselective hy-
droxyl radical in these bleaching systems will result in a certain oxidation and degradation
of the cellulose unless special precautions are taken. Even so, certain cellulose degradation
cannot be completely avoided.
B. Oxygen
1. Lignin Reactions
In oxygen &lignification, carried out in an alkaline medium, a partial oxidation of phenolic
structures in lignin takes place. In structures with conjugated double bonds, such as in
enol ethers and stilbenes, the oxidation may also take place in the side chain, resulting in
a fragmentation of the lignin. The reaction starts with the formation of a phenoxy radical
by electron transfer to the oxygen. In subsequent reaction steps, the phenoxy radical is
converted into a hydroperoxide, which in turn reacts further, with formation of oxidized
products. The reaction sequence is outlined in Fig. 32 [ 128,1291.
The kinetics for the oxidation of a series of ligninlike phenols has been studied and
found to be rather slow [ 1301. Thus, the kinetic half-life for phenolic structures having a
guaiacyl structure is of the order of 14 min under technically relevant conditions. For
structures like stilbenes and diphenols, such as catechols, much higher reaction rates are
encountered. The importance of these structures for the lignin degradation is, however,
limited due to their low abundance.
Since the solubility of oxygen in water is low, it must be assumed that the extent of
oxygen oxidation of lignin is highly dependent on the efficiency of mixing the pulp and
the oxygen gas. Once the concentration of oxygen in the aqueous phase has ceased, only a
very limited further oxidation of phenols can take place and secondary reactions, such as
alkaline hydrolysis and lignin extraction from the fibers, become important. These reactions
are accompanied by a certain brightening action caused by the liberated hydrogen peroxide
in the system. The latter, together with organic peroxides, may also give rise to hydroxyl
radicals by metal ion-induced decomposition and, thus, degradation of the cellulose.
The consumption of phenolic hydroxyl groups in oxygen delignification has been
found to be rather limited [ 131,l 321. In accordance with the kinetics discussed above, the
residual lignin after an oxygen stage contains a considerable amount of remaining phenolic
groups, whereas the solubilized lignin has a higher amount of phenolic groups as compared
to the average for the residual lignin in the unbleached pulp. Thus, it seems likely that
the oxygen oxidation occurs predominantly with those parts of the residual lignin which
are the most phenolic and the most accessible, i.e., a lignin which already is rather soluble
but trapped in the fiber wall for reasons of molecular size or a few remaining linkages to
other pulp constituents. In addition to phenolic groups,thedissolved lignin from the
oxygen stage contains an increased number of carboxyl groups. The total number of acidic
groups is of the order of 5 mmol/g of lignin.
After an oxygen stage, the remaining lignin contains a higher amount of carboxyl
and a lower amount of phenolic groups as compared to the unbleached fiber lignin. Other
differences are small, but the residual lignin has a somewhat higher amount of condensed
structures [ 1321 and the content of p-0-4 structures has decreased in comparison with the
residual lignin after the kraft cook. A second oxygen stage results in further &lignification
of the pulp, but again the structural differences in the residual lignin are minor [ 1331.
2. Carbohydrate Reactions
The carbohydrate reactions in oxygen delignification result in a loss of viscosity due to
attack by hydroxyl radicals along the cellulose chains. The reaction, which is shown in
892 Gellerstedt
9
\ \ c
0
Pulping Chemistry a93
Fig. 33, involves the formation of a carbonyl group in the sugar moiety, followed by a p-
elimination reaction that results in a chain scission [134]. Unless magnesium ions are
present in the bleaching liquor, the viscosity loss isof such a magnitude that technical
utilization of oxygen delignification is prohibited 11351. Addition of magnesium ions,
however, will stabilize the formed hydrogen peroxide to some extent [136], thereby re-
ducing the formation of hydroxyl radicals. It has also been shown that the presence of
magnesium ions gives a certain reduction in the rate of reaction between hydroxyl radicals
and carbohydrate structures [ 1371.
The yield loss in oxygen delignification is limited, since the carbohydrate peeling
reactions (see above) are suppressed in favor of an oxidative stabilization reaction. Thus,
the reducing end groups in the polysaccharide chains are converted to aldonic acid groups
through oxygen oxidation of the open enediol structure as shown in Fig. 28 [ 1381.
After a kraft cook, the pulp contains a xylan in which a major fraction of the
4-OMe-glucuronic acid groups has been converted into the corresponding unsaturated
structure, "hexenuronic acid," by loss of methanol, as described in Fig. 25. A subsequent
oxygen delignification step is not able to degrade this structure, and consequently the
relative contribution to the resulting pulp kappa number from the hexenuronic acid groups
will increase after the 0 stage.
CHpOH
H0
H0
o ' o m0
H o '0 HH
aOQow 00
CH20H
- Hop
0 CHpOH
H0 t"
0
0
H0
'0
1 - @-OH
H0
CHZOH
0
-
rearrangement
HeLow
H CHpOH
coo0
FIGURE 33 Mechanism for the oxidative degradation of cellulose by hydroxyl radicals.
894 Gellerstedt
C. HydrogenPeroxide
Alkaline hydrogen peroxide in the presence of sodium silicate is the traditional bleaching
system for mechanical pulps, as discussed earlier. An efficient bleaching action requires
that all transition metal ions be removed as completely as possible prior to bleaching, an
operation usually done by pretreatment of the pulp with a strong chelant such as EDTA
or DTPA. The same criteria must be fulfilled in the peroxide bleaching of chemical pulps.
Unlike mechanical pulps, however, chemical pulps cannot be bleached to high brightness
values with hydrogen peroxide unless the residual lignin is removed, since the chromo-
phoric systems are much less reactive [ 1391. Efficient delignification and bleaching of
chemical (kraft) pulps can be obtained if the bleaching stage is preceded by a chelating
stage with EDTA, carried out under controlled conditions [140,141] in order to eliminate
manganese, iron, etc., as completely as possible while keeping the concentration of ions
such as magnesium and calcium high. The bleaching stage itself can be done with alkaline
hydrogen peroxide (without silicate) at temperatures around or slightly above 100°C for
2-4 h.In the latter case, an oxygen pressure can be applied (POstage). Addition of
magnesium ions to the bleaching liquor is beneficial.
The ability of alkaline hydrogen peroxide to attack and degrade chromophoric struc-
tures, such as conjugated carbonyl structures, is well known. In the bleaching of chemical
pulp, however, this mode of reaction must be accompanied by reactions resulting in a
more comprehensive lignin oxidation and dissolution. One possibility for achieving such
a reaction would be a radical-induced oxidation of phenolic structures in the residual
lignin. Thus, the (controlled) decomposition of hydrogen peroxide leading to hydroxyl
radicals and superoxide ions should result in the degradation of aromatic rings through
reactions similar to those occurring in oxygen delignification. Analyses of the structure of
the dissolved and residual lignin from a peroxide bleaching stage do not, however, un-
equivocally support such a mechanism since, i n contrast to the corresponding lignins from
an oxygen stage, the content of aromatic rings in these lignins does not change to any
large extent [ 1321. On the other hand, the content of carboxyl groups i n lignins from a
peroxide stage is high, demonstrating the capability of hydrogen peroxide as an oxidant
for lignin. Based on I3C-NMR data, the majority of the formed carboxyl groups seem to
be aliphatic rather than aromatic, thus indicating that side-chain oxidation in lignin may
be a predominant reaction mode in peroxide bleaching [142]. Recently, this view was
further supported by a model compound study in which a phenolic lignin structure was
shown to be easily degraded by the action of alkaline hydrogen peroxide at an elevated
temperature (Fig. 34) [ 1431.
Analysis by "C-NMR of isolated residual lignin after a peroxide stage (in the se-
quence OQP) demonstrates that this lignin, although coming from a pulp with less than
1 % lignin, still has most of the features of a lignin structure; the major differences are a
low amount of p-0-4 structures, the presence of carboxyl groups, and the presence of
aliphatic methylene and methine groups [142]. Therefore, the successive delignification
that takes place in the kraft cook and in the subsequent oxygen delignification and hydro-
gen peroxide bleaching stage does not seem to alter (degrade) the remaining lignin struc-
ture in a profound way.
The reactions of carbohydrates in a peroxide bleaching stage seem to be very similar
to an oxygen stage. Thus, any decomposition of hydrogen peroxide, resulting in the for-
mation of hydroxyl radicals, is probably inducing an oxidation of hydroxyl groups to
carbonyl groups along the polysaccharide chains followed by a chain scission. In analogy
with oxygen delignification, the hexenuronic acid groups present in the xylan will survive
an alkaline peroxide stage, thus resulting in a comprehensive contribution to the kappa
nlltnher :d'ter the stape (Fig. 31).
Pulping Chemistry 895
I
I HC-0-OH
-
/ OCH,
0
OH 0
CHZOH
I I
HC-O* HC
-
-
I OCH3
coo@
HC-O*
OCH, I
+ ox.
+ HO'
f"
OH
I
OCH,
0 OCH, OH
30 6
-i
l§
10
51k:n*
0
0 10
0 CE
0 CRECE
A C~ECRECE
cR: 2 min,
,
20
5.74% cl2 on 0.d. pulp
C : t mln, 5.74% Cl2 on 0.d. pulp
om*
30
Time (t), min
,
40
o% ,
50
as chloride ions, most of which are formed in the E stage. In addition, chlorate is formed
in the D stages.The remaining chlorine is found as organically bound chlorine in the
effluent and, to a small extent, in the bleached pulp.
The reactions between lignin and aqueous chlorine have been thoroughly studied,
and four reaction principles have been identified I 1341. Chlorination of aromatic rings may
Cl (<0,1 %)
to air
IC1021
ICl,I
g act. CllL
0.50
0.25-\
"" CIO,
,'C-
0 I I I
0 10 20 30
Time, min
phase 1 phase 3
sequent extraction, the remaining fiber lignin still has the general feature of a lignin poly-
mer [ 1571, indicating that oxidative reactions in the fiber wall cease before a complete
lignin oxidation has been achieved. After the extraction stage, the remaining fiber lignin
can be subjected to further oxidation, again involving more or less the same structures as
before.
E. Ozone
In recent years, ozone has gained considerable interest as a bleaching agent for chemical
pulps, and during the 1990s, several mill installations have occurred. The one important
objective being met was the reduction or elimination of chlorine-containing bleaching
effluents. Bleaching with ozone is always carried out in acidic media in order to minimize
the self-decomposition of ozone in water according to reaction (14) [ 1581.
1 0-0
c” 0?
J
O
CH
3
0 0’
I
FIGURE 39 Mechanism for the ozonolysis reaction of a double bond and formation of carboxyl
groups in the presence of water.
FIGURE 40 Formation of a hydroxyl radical from the reaction between ozone and a phenol.
Pulping Chemistry 901
H0
0 HOC+
/ 0
0 0
0
FIGURE 41 Suggested elimination of chloride ions from chlorinated aromatic structures in lignin
upon treatment with alkali.
group will increase this yield for such structures. The degree of ionization of the phenol
(i.e., the pH value) will determine the extent of radical formation.
F. Alkaline Extraction
The role of alkaline extraction (E) following an acidic bleaching stage is to remove the
oxidized lignin from the cellulosic fiber by neutralization of the formed carboxyl and other
acidicgroups. In modern bleaching,oxygen, hydrogen peroxide,or both areadded in
small amounts to furtherenhance the degree of oxidation and,thus, the dissolution of
lignin. In addition to being dissolved, chlorinated lignin from a previous C or D stage will
also undergo a substantial dechlorination by nucleophilic attack of hydroxyl and hydro-
peroxy ions [ 1661. This type of substitution may proceed through reactions between
chlorinatedquinonestructures in lignin and a nucleophile,asoutlinedschematically in
Fig. 4 I .
Under acidic (as well as alkaline) bleaching conditions, carbohydrates may become
oxidized along the polysaccharide chain, resulting in the formation of carbonyl groups in
the sugar units. This type of reaction takes place particularly in ozonebleaching. In a
subsequent extraction or alkaline bleaching stage, the introduced carbonyl group will in-
duce a p-elimination reaction that results in a cleavage of the polysaccharide chain and,
thus, a loss of viscosity.
REFERENCES
1. E. Adler and J. Marton, ActLI Chern. Scnrzd., 13:75 (1959).
2. J. C. Pew and W. J. Connors, Z ~ p p i 54r245
, (1971).
3. W. J. Connors, J. S. Ayers. K. V. Sarkanen, and J. S. Gratzl, Tappi, 54:1284 ( l 9 7 I).
4. S. Larsson and G. E. Miksche, Acto C/lem. Scrrnd., 26:2031 (1972).
5. G.Gellerstedtand E-L. Lindfors, Hol$o,sclz., 38:151 (1984).
902 Gellerstedt
907
908 Index
One-electron oxidation mechanism, 547, 549, Primary wall, 12, 13, 15, 18-20,25,28.30, 58,
550, 559. 562 60, 62, 63, 176, 185, 861. 862
Oxalic acid, 413, 415, 418, 420, 421, 425, 435, Proanthocyanidin, 218, 229, 246, 255, 256, 258,
456, 694, 695 260, 26 1, 263, 265
Oxidative degradation, 161, 288, 320, 321, 344, Probe tack test, 749
346,445, 456, 493, 5 13, 530, 566, 894 Property enhancement of wood, 784, 789, 793
Property enhancer, 787, 788, 792, 793
p-quinone (see Quinones) Propiconazole, 803
Pallet, 802, 850 Propylene oxide treatment, 575, 577. 580, 582,
Particleboard, 264, 266, 280, 405, 577, 580, 59 I , 592
849, 850, 852, 855, 857
Pectin, 26, 52, 56, 59, 175,180,185,189,190, Quinone methide, 698, 871, 876-879, 880, 886,
203, 205, 245, 266, 353, 809 888
Peel strength, 709. 735, 743, 748, 749, 757, Quinones, 123,125,128,130, 226, 388. 539,
772, 77.5 861, 863, 864, 866, 867, 869, 897, 901
Peeling reaction, 470, 471, 473, 475, 476, 478, o-quinone, 344, 499
493, 496, 882, 884, 889, 893 p-quinone, 860, 862, 866
Pentachlorophenol, 263, 797, 799, 815
Pentosan, 55, 59, 302, 310, 452, 809 Radial compression, 584
Peracetic acid, 282, 498, 499, 565, 664, 666 Radicals (see Free radicals)
Periderm, 242, 244 Ray parenchyma, 2, 4. 19, 20, 21, 66, 67. 70,
Periodate, 124-126, 137, 3 10. 31 I , 320, 493, 72, 74, 369, 817
496, 619, 620 Reaction kinetics, 482
Permanent fixation, 586 Reaction wood (see Wood)
Permethylation, 184, 286, 613, 616 Regenerated cellulose, 85, 444, 601, 606, 613,
Peroxyacetic acid, 284 647
Phanerochaete chrysosporium, 547-550, 554. Regiospecific control, 628, 629
557 Relaxation process, 592. 593, 595, 675, 721,
Phenol-formaldehyde resin, 264, 265, 439, 688. 832
819-821 Resorcinol-phenol-formaldehyde,8 19-822, 824
Phenolic acid, 206, 228, 246, 258, 557, 561 Rhamnogalacturonan, 185, 189,190.203
Phenolic hydroxyl (see Hydroxyl group) Rhytidome, 244
Phenolic units, 128, 137, 229, 344, 478, 483, Ribbonassembly,96-98.100-104
486, 865, 875. 879 Ring opening, 156. 498, 552, 562, 564
Phenolics, 175. 188. 232, 246, 25 l , 3 15 Ring-opening polymerization, 627, 628, 639,
Phenylcarbanilation,609 641, 644. 645, 648, 650, 65 I , 838, 840
Phenylpropanoid. 214, 226, 232, 251 Rolling friction coefficient. 752, 758, 770, 775
Phloem, 4. 26, 28, 242, 244. 252 Rosin, 222, 364, 768, 769, 778, 812, 819
Photo-induced discoloration (see Discoloration)
Photodegradability, 525. 845 Saccharitication, 267, 444, 466, 469
Plasticizer, 580, 582, 583, 593, 595. 656, 667. Sapwood (see Wood)
668. 670, 68 I. 682, 830, 834. 841 Sawdust. 278. 358, 369, 432, 850
external, 667. 668, 673. 681 Secondary wall, 4. 5 , 12, 13, 16, 18. 24, 28. 30,
Polycaprolactone. 832, 834, 842. 843 35. 37. 39. 40, 57, 58, 61. 66. 70. 72.
Polyethylene glycol. 406. 575, 579, 694, 696, 74, 75. 201, 203, 346. 369. 445. 500,
811. 814. 815, 817, 819 534. 536, 574, 817
Polyflavanoid, 258, 264-266 Semicarbazide. 405, 436
Polysulfide, 473. 870, 876, 884, 886, 888 Shipwreck, 807, 808
Polyterpene, 222. 768, 769, 778 Shrinkage, 575, 808, 810, 814
Polyurethane, 263, 267. 405, 409, 420, 432, Sizing. 97. 222. 610, 61 I
437, 688. 689, 696. 819 Sodium borate. 302. 433. 8 I0
Primarycellwalls,175.176.178. 192. 201. Sodium dihydrogen phosphate. 4 I9
207, 2 10 Softwood lignin ( s c v Lignin)
index 913
Water repellency, 610, 805 Wood chips, 251, 325, 362, 432. 548, 694, 696,
Waterlogged wood, 807 850, 855, 860, 869, 870
preservation, 8 I O Wood chromophores, 860, 862, 863, 867 (see
properties, 809, 810 also Chromophores)
Wattle, 232, 263, 264, 266. 267 Wood industry integration, 850, 858
Wax, 232, 246-248, 353, 362,409,526, 607, 81 1 Wood preservation, 795, 799, 804
Weathering, 512, 528, 535, 541, 766, 805, 807, pressure processes, 795
816, 819 supercritical-fluid process, 796
Weight percent gain (WPG), 577, 784 vapor-phase process, 796
Whiteness, 429, 547-549, 557, 566, 788 Wood preservatives, 795
Wood copper naphthenate, 802.
compression, 2, 22, 32, 36, 37, 39, 41, 42, oil-borne, 797
44, 45, 54, 55, 57, 58, 63, 68, 116, 134, creosote, 797
142,168, 879 pentachlorophenol, 799
hardwood, 1, 7, 25, 31, SI, 69, 92, 108, 112, tributylin oxide, 799
119. 123, 1 4 4 , 150, 175, 214, 246, 251, oxide copper, 802
265, 282, 300, 350, 434, 587, 604, 490, Timbor, 802
810, 870, 879, 880 water-borne, 799
heartwood, 2, 4, 20, 54, 114, 213, 217, 218, ammoniacal copper zinc arsenate, 800
221, 232, 235, 264, 281, 366, 385, 388, chromated copper arsenate, 800
391-393, 400, 402, 410-413, 435, 438, Wood-inorganic composites, 782, 784, 788, 789,
796, 805, 873 792
normal, 2, 12, 22, 40, 43, 5 1, 55, 61, 116, Wood-methyl methacrylate composite, 575
142, 168, 658
reaction, 2, 22, 36, 51, 52, 63
sapwood, 2, 4, 20, 32, 54, 114, 213, 232, Xylan, 56-58, 201, 207, 293, 300, 303, 305,
244, 28 1, 341, 385, 388, 391, 393, 400, 314, 325, 447, 450, 456. 466, 476,
402, 410, 430-432 477, 490, 497, 71 I , 873, 882, 883,
softwood, 1-4, 17-20, 26, 28, 3 I , 5 I , 56, 59, 889. 894
63, 65, 68, 71, 75, 92, 108,116, 144, Xylem, 1-4. 20, 26, 28, 3 1, 46, 66, 72, 175,
156,165,175, 214, 246, 252, 274, 284, 176, 242, 244
300, 319. 337, 358. 394, 425, 467, 490, Xyloglucan, 98, 192, 195, 199, 201, 205, 207,
532, 573, 601, 850, 860, 880 314
tension, 2 , 9, 12, 22-24, 36, 40, 44, 51, 54,
55, 60, 445 Yellowing, 332, 402, 540, 589, 867-869