Bioresource Technology 101 (2010) 1361–1368

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Application of liquefied wood as a new particle board adhesive system

Matjaž Kunaver a,*, Sergej Medved b, Nataša Čuk a, Edita Jasiukaitytė a, Ida Poljanšek a, Tatjana Strnad c
a
National Institute of Chemistry, Hajdrihova 19, SI-1000 Ljubljana, Slovenia
b
University of Ljubljana, Biotechnical Faculty, Rožna dolina C VIII/34, SI-1000 Ljubljana, Slovenia
c
Melamin d.d. Kočevje, Tomšičeva cesta 9, SI-1330 Kočevje, Slovenia

a r t i c l e i n f o a b s t r a c t

Article history: Different types of southern European hardwoods and softwoods were subjected to a liquefaction process
Received 1 May 2009 with glycerol/diethylene glycol. The liquefied spruce wood was reacted in a condensation reaction in the
Received in revised form 22 September hot press with different melamine–formaldehyde and melamine–urea–formaldehyde resin precursors
2009
and used as adhesives for wood particle boards. The mechanical properties of these particle boards
Accepted 23 September 2009
and the determination of formaldehyde release, proved that addition of 50% of the liquefied wood to such
resin precursors caused the product to meet the European standard quality demands for particle boards.
Up to 40% reduction of the formaldehyde emission was achieved. The temperature of the press unit was
Keywords:
Wood liguefaction
lowered from 180 °C to 160 °C with no significant influence on the mechanical properties. On the basis of
Melamine–formaldehyde adhesives the presented results it was possible to conclude that liquefied wood can be used as substitute for syn-
Melamine–urea–formaldehyde adhesives thetic resin precursors in adhesives that are used for the particle board production.
Particle board Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction
Biomass based materials and wood in particular are among the
more abundant renewable resources. Recently, considerable atten-
tion has been given to the preparation of environmentally friendly
polymeric products from liquefied biomass materials and their
derivatives.
The liquefaction of wood and other lignocellulosic materials in
the presence of phenol or polyhydroxy alcohols has been inten-
sively studied by several authors (Lin et al., 1994; Alma et al.,
1998; Yamazaki et al., 2006; Kržan and Kunaver, 2006; Qiuhui
et al., 2006; Hassan and Shukry, 2008a,b). Such a liquefaction pro-
cess is usually carried out at elevated temperature and in the
presence of an acid catalyst. The liquefied wood contains depoly-
merized products from the b 1–4 glucosidic bond cleavage of the
cellulose and hemicelluloses molecules as well as from low molar
mass oligomers. The complex structure of the lignin molecule is
broken to smaller fragments. However, the associated reaction
pathways and products are yet to be defined completely (Kurimoto
et al., 1999; Lin et al., 2001a,b; Yamada and Ono, 2001; Kobayashi
et al., 2004).
With respect to the liquefaction of wood and wood wastes and
their use as a liquid fuel, Wu et al. (2009) liquefied poplar leaves,
wood and bark at high temperatures and pressures to produce al-
kanes. Rezzoug and Capart (2002) used solvolysis in ethylene–gly-
* Corresponding author. Tel.: +386 1 476 0363; fax: +386 1 476 0300.
E-mail address: matjaz.kunaver@ki.si (M. Kunaver).
col and catalytic hydrogenation as the second stage, for the
production of oils that possessed a high heating value.
Liquefied wood has a high reactivity due to a large amount of
phenolic and alcoholic hydroxyl groups that are present. The hy-
droxyl value of the liquefied wood has been determined by several
authors and is generally determined to be higher than 200 mg of
KOH/g (Hassan and Shukry 2008a,b). The value depends on the liq-
uefaction time and on the ratio between the wood content and the
liquefying reagents. These functional groups can be used in the cre-
ation of polyurethane foams, polyurethane resin precursors
(Kurimoto et al., 2000, 2001; Wei et al., 2004) and in the recently
developed wood–polyalcohol based urethane adhesives (Tohmura
et al., 2005). Hydroxyl groups react with the isocyanate unit form-
ing an urethane bond. Lee et al. (2002) have studied the thermal
stability, biodegradability and genotoxicity of a range of polyure-
thane foams that were produced from polyols made from waste
paper. They found that the foams possessed the same thermal sta-
bility as those made from liquefied starch or wood. The foams were
biodegradable to some extent.
The hydroxyl groups of the liquefied wood can react with epi-
chlorohydrin, thus introducing the epoxy functionality. Kobayashi
et al. (2000, 2001) prepared such epoxy compounds from the
heartwood meal of Japanese cedar. Kishi et al. (2006) prepared a
wood based epoxy resin from wood meal of German spruce. The
authors cured their epoxy resin precursors with suitable amino
precursors. Such two component systems were used as adhesives
in plywood preparation, good mechanical and physical properties
were obtained.

0960-8524/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biortech.2009.09.066
1362 M. Kunaver et al. / Bioresource Technology 101 (2010) 1361–1368
Hydroxyl groups in liquefied lignocellulosics can also react with
different reactive sites in thermosetting systems as well as in two
component systems. Lee and Liu (2002) prepared a resol type of re-
sin from liquefied bark that was used in particle board preparation.
Similarly, a phenol formaldehyde adhesive was prepared by Sheny-
uan et al. (2006) from liquefied bamboo. Zhang et al. (2007) used
Chinese fir and poplar in this context. They tested the bonding
strength of such adhesive in plywood showing the product to meet
the requirements successfully. Gagnon et al. (2004) have under-
taken research into the softwood bark pyrolysis oils and their use
as adhesives for particle boards in combination with isocyanates.
Amen-Chen et al. (2002) have used bark residues from pulp and
paper industry to produce pyrolysis oil. They replaced different
levels of phenol in resol synthesis with pyrolysis oil. Strand boards
were prepared with these resols with comparable mechanical
properties to panels made with commercial resols.
The first objective of this study was to synthesize a liquefied
wood that possesses high hydroxyl group content, with a good
yield in the liquefaction reaction. The second goal was to establish
the criteria for creating a melamine–formaldehyde or a melamine–
urea–formaldehyde resin precursor that would react at elevated
temperature with liquefied wood and could be used as an adhesive.
The third objective was to optimize the formulation and to test
such systems, to confirm their potential use in the furniture
industry.
2. Methods
Wood meals (flours) of poplar (Populus spp.), oak (Quercus spp.),
spruce (Picea spp.) and beech (Fagus sylvatica spp.) were sieved
through a 2 mm screens and dried at room temperature to a con-
stant water content.
All chemicals were of synthesis grade (Merck) and were used
without further purification.
Methylated melamine resins, urea–formaldehyde, melamine–
formaldehyde and melamine–urea–formaldehyde resins were
used as delivered from the producers. The type and the producer
of each resin are shown in Table 1.
2.1. Liquefaction of wood
The liquefaction of wood was carried out in a 1000 mL three-
neck glass reactor, equipped with the mechanical stirrer and con-
denser. The reactor was charged with 100 g of wood and 300 g of
glycerol – diethylene glycol mixture (4:1). 9 g of p-toluenesulfonic
acid was added. The proportion of the constituents in the reaction
mixture was chosen after the initial trials, with the goal to maxi-
mize the glycerol content as a by-product of the biodiesel produc-
tion. A certain proportion of glycerol was replaced with diethylene
glycol to provide the required degree of viscosity reduction for the
final product. This step was taken because the use of 100% glycerol
gave a pasty, highly viscous final product that did not meet appli-
cation requirements. The mixture was heated for 3 h at 180 °C
while being constantly stirred. A sample was withdrawn from
Table 1
Resin precursors used in our experiments.
Name Type
Cymel 323 Methylated melamine resin, 80% solution in i-butanol,
Cymel 350 Methylated melamine resin, 100%
Meldur SIS-BM Urea–formaldehyde resin, 50% water solution, free formaldehyde: less than 1%
Meldur MS/1 Melamine–formaldehyde resin, 58% water solution, free formaldehyde: less than 0.2%
Meldur H-97 Melamine–urea–formaldehyde resin, 63% water solution, free formaldehyde: less than 0.5%
Melapret NF 70M Methylated melamine–formaldehyde resin, 70% water solution, free formaldehyde: less than 0.3%
Meldur IV Melamine–urea–formaldehyde resin, 66% water solution, free formaldehyde: less than 0.2%
the reaction system periodically and immersed in cold water to
quench the reaction.
The extent of liquefaction was evaluated by determining the
residue after the washing out the sample with dioxane and water
(4:1 v/v). The residue was dried in an oven at 105 °C to constant
weight. The conversion yield was calculated as the weight percent-
age based on the starting wood material.
2.2. Measurement of hydroxyl number of liquefied wood
The hydroxyl numbers (ASTM 2005) of the liquid woods (LW)
were determined as follows: 0.5 g to 1.0 g of sample was dissolved
in 25 cm3 of a phthalation reagent and heated at 115 °C, for 1 h, un-
der reflux. This was followed by an addition of 50 cm3 of pyridine
through the condenser. The mixture was backtitrated with a 0.5 M
sodium hydroxide solution. The indicator was 1% phenolphthalein
solution in pyridine. The phthalation reagent consisted of 115 g of
phthalic anhydride dissolved in 700 cm3 of pyridine.
The hydroxyl number, defined as mg KOH/g of sample, was cal-
culated as follows:
ðB  AÞ  M  56:1
Hydroxyl number ¼ þ acid number
w
Here, A is the volume of the 0.5 M sodium hydroxide solution
required for the titration of sample (cm3). B is the volume of the so-
dium hydroxide solution required for the titration of the blank
solution (cm3). M is the molarity of the sodium hydroxide solution
and w is the amount of the sample (g) to be analysed.
If the sample is acidic, the acid uses the phthalation reagent
during the analysis and the hydroxyl number must be corrected
accordingly.
The acid number was determined as follows. 0.4 g of sample
was weighed into a 400 cm3 Erlenmeyer flask and dissolved in
50 cm3 of solvent mixture. The solvent mixture consisted of diox-
ane and water (4:1 v/v). 0.5 cm3 of phenolphthalein indicator solu-
tion (1% in ethanol) was added and titrated with 0.1 M KOH
solution, in ethanol, to the equivalent point. The acid number
(mg KOH/g of sample) was calculated using the following
equation:
ðC  BÞ  M  56:1
Acid number ¼
w
Here, C is the titration volume of the potassium hydroxide solu-
tion (cm3). B is the titration volume of blank solution (cm3). M is
the molarity of the potassium hydroxide solution and w is the
amount of the sample (g) being analysed.
2.3. GC/MS measurements
The concentration of the residual diethylene glycol and glycerol
in the reaction mixture was determined by use of an AGILENT
6890N/5973N gas chromatographic system, equipped with mass
spectrometer detector. Prior to the analysis, the hydroxyl groups
of reaction components were transformed into the corresponding
trimethylsilyl (TMS) derivatives, as described by Rojas-Escudero
Producer
Cytec Industries Inc.
Cytec Industries Inc.
Melamin Slovenia
Melamin Slovenia
Melamin Slovenia
Melamin Slovenia
Melamin Slovenia
 M. Kunaver et al. / Bioresource Technology 101 (2010) 1361–1368
et al. (2004). The chromatographic conditions were as follows. A
30 m HP MS-5 capillary column with 0.25 mm i.d. was used. He,
with flow rate of 0.9 cm3, was the carrier gas. The initial tempera-
ture of the oven was 120 °C for 4 min, then increased to 300 °C at
20 °C min1 and held for 15 min. The split ratio was 100:1. Mass
spectral data were collected in the range between 30 and
350 Amu. Neopentyl glycol was used as the internal standard.
2.4. GPC measurements
Gel permeation chromatographic (GPC) measurements were
carried out on an HP – AGILENT system, consisting of isocratic
pump HP 1100 and a Perkin Elmer LC235 diode array detector.
Two PLgel columns, connected in series (PLgel 5 lm MIXED D
2.5  300 mm and Polar Gel L 8 lm 7.5  300 mm), were used.
Poly(styrene) standards were used for the calibration curve. The
mobile phase was prepared from DMA containing 0.01 M LiBr.
The latter was added in order to prevent interaction between the
sample molecules and stationary phase. The mobile phase flow
rate was 0.7 cm3/min. The chromatographic data were processed
with PSS (polymer standards service) WinGPC unity software.
2.5. Measurements of the gel time
In order to measure the reactivity of a system consisting of liq-
uefied wood and several different melamine–formaldehyde or
melamine–urea–formaldehyde resin precursors approximately
3 cm3 of such a mixture was placed in a test tube and heated in
boiling water or in silicone oil bath at temperatures above 100 °C
under continuous mixing. The reactivity of a mixture is defined
as the time measured from the beginning of heating until gelation
occurred.
2.6. Particle board preparation
The viscosity of the liquefied wood was first reduced with the
addition of 15–20% of distilled water. A reduction in viscosity is
needed for applications that require the use of nozzles as well as
for improved mixing between the adhesive and hardener. The pro-
cess of preparing the adhesive was carried out by mixing urea–
formaldehyde resin precursor or melamine–formaldehyde resin
precursor with liquefied wood. The obtained mixture was stable
Fig. 1. Reaction yields: liquefaction of oak, spruce, beech and poplar wood meals. ANNEX: SEM micrographs of spruce wood meal (left) and residue after liquefaction (right).
1363
for 72 h at room temperature. This approach gave rise to a formu-
lation that was stable for 72 h at room temperature. After 72 h, the
viscosity started to rise due to the occurrence of step-growth pro-
cesses. The prepared adhesive mixture was applied through noz-
zles onto wood particles on a wood particle blending unit.
The blending was carried out with different adhesive mixture
loadings: 7.5% of adhesive (dry mass) in the particleboard core
layer and with 11.5 % of the adhesive in the surface layer.
The adhesive mixture was composed of the resin precursor (Ta-
ble 1) – 0–50%, liquefied wood – 0–50%, ammonium chloride as the
catalyst – 0.3% and water 15–20%, respectively.
The blended particles were pressed into a 500  500 mm board
formate. The particle mat was pressed at 180 °C, 3 N/mm2, for
5 min. The nominal thickness was 16 mm and the nominal density
was 0.7 g/cm3. The boards were later cooled and conditioned at
20 °C and 65% relative humidity.
3. Particleboard testing
3.1. Mechanical and physical properties
All of the evaluations of the resulting particle boards were car-
ried out according to the appropriate European standards (EN).
These concern thickness (EN 324), density (EN 323), swelling in
thickness after immersion in water (EN 317), bending strength
and modulus of elasticity (EN 310), internal bond strength (EN
319) and surface soundness (EN 311).
3.2. Formaldehyde emission tests
The total formaldehyde content was determined using the per-
forator extraction method, according to the European standard (EN
120). The perforator extraction method allows one to determine
the total formaldehyde content in the board samples by extraction.
4. Results and discussion
4.1. Characterization of the incompletely liquefied/derivatized residues
of the wood meal (flour)
The liquefaction process gave yields of the liquefied wood that
were in excess of 99%. However, in all cases, a small amount of res-
1364 M. Kunaver et al. / Bioresource Technology 101 (2010) 1361–1368
idue was present as a dispersive phase in the liquefied wood solu-
tion. This material was made the subject of a characterization
study. The typical reaction yields are shown in Fig. 1 for the differ-
ent wood meals. It is clear that a reaction time of at least 105 min,
under the specified heating regime, was needed if unwanted resi-
dues were to be minimized. For oak and beech a relatively pro-
nounced condensation occurs with time, increasing the amount
of the residue formed. On the other hand, in the case of poplar
and spruce, there is little or no increase in residue formation with
time of reaction, over the time period that was mentioned.
Spruce wood meal was used in all further experiments due to
high reaction yield and very low condensation with time. More-
over, spruce is the dominant softwood tree in Slovenia (28% of total
forest area) and therefore, the most available wood for the lique-
faction process.
4.2. SEM analysis of residues from the wood meal liquefaction process
The wood particles that remained after the liquefaction process
were analysed by SEM. The micrographs of the various residues
were very similar in their appearance. For this reason, the residues
of the spruce wood flour system are given as a representative
example. (For SEM micrographs see Supplementary Electronic An-
nex 1 in the online version of this article.) It is obvious that due to
liquefaction, the wood cells became completely disintegrated. The
Fig. 2a. GPC evaluation of a water soluble residue of the liquefied spruce wood
meal.
Fig. 2b. GPC profile for a water insoluble residue of the liquefied spruce wood meal.
residue appears to contain only irregular, powdery particles mixed
with some of the products of the polycondensation process that oc-
curs towards the end of the liquefaction process.
4.3. GPC analysis of liquefied wood meal based materials
The liquefied spruce wood meal was subjected to GPC analysis.
The sample of the liquefied wood was taken after 120 min of the
reaction. The sample was filtered. Both the water soluble part
and water insoluble part were analysed. The water insoluble part
represented 12% of the sample. Fig. 2a gives the chromatogram de-
rived from the water soluble part of the liquefied spruce wood
meal.
The chromatogram shows the presence of the residual diethyl-
ene glycol and a mixture of low molar mass oligomers. The average
molecular mass is 670 Dalton. All molar masses are expressed as
average molar masses, calculated from a calibration curve that
was based on poly(styrene) standards.
Fig. 2b gives information for the water insoluble part of the liq-
uefied spruce wood meal.
Here, the strong peak represents the average molar mass of
9700 Dalton, originating from liquefied wood components that
are insoluble in water due to their higher molar masses and con-
densation products from reactions between lignin and cellulose
low molar mass derivatives that are created during the liquefaction
process (Kurimoto et al., 1999).
4.4. Hydroxyl group content in liquefied wood
The hydroxyl group content is one of the more important fac-
tors of relevance to the use of such materials in the preparation
of polyurethane foams as well as for calculations of the reactant ra-
tios in other condensation reactions (Kurimoto et al., 1999). As
shown in Fig. 3, the OH number of each sample, taken at different
time intervals, decreased almost in a linear manner. Here, the
residual glycols were of the major influence. Since the amount of
each glycol in the sample was determined by gas chromatographic
analysis, it was possible to calculate the theoretical OH value for
the glycols that were present in the reaction mixture. The OH value
of products, originating from the wood meal, was then calculated
as the difference between the OH value of the sample and the
OH value of the glycols, respectively.
The decrease in OH value is thought to be the dehydration and
thermal oxidation of the glycols as well as due to the condensation
reactions between the glycols and wood components, such as cel-
lulose, hemicelluloses and lignin. Yao et al. (1996) subjected gly-
cols alone to the same liquefaction conditions and measured the
Fig. 3. Hydroxyl content in the liquefied spruce wood meal samples, taken at
different time intervals.
 M. Kunaver et al. / Bioresource Technology 101 (2010) 1361–1368 1365

OH values. They found that the OH values of the glycols did not Table 2
change significantly – indicating that the major decrease in the Gel times of mixtures of liquefied spruce wood with different melamine–formalde-
hyde and melamine–urea–formaldehyde resin precursors, without catalyst addition.
OH values of the reaction mixture is due to the reactions between
the wood components and the glycols that are present. Gel times
Liquefied wood 30% 20%
Cymel 323 974 s 1205 s
4.5. Reactivity of different melamine–formaldehyde and melamine– Cymel 350 186 s 493 s
Meldur H-97 320 s 325 s
urea–formaldehyde resin precursors in combination with liquefied Meldur IV 570 s 720 s
wood Melapret NF-70M 335 s 365 s
Meldur SIS-BM 520 s 580 s
The preliminary testing of mixtures of liquefied wood with mel- Meldur MS-1 185 s 160 s
amine–formaldehyde resin precursors or melamine–urea–formal-
dehyde resin precursors consisted of measuring the time
required for the formation of gel like crosslinked polymer, at
Table 3
100 °C. The gel time is the initial criterion for assessing the useful- Gel times of Meldur MS-1 and Meldur H-97 + LW with different loadings of the
ness of such an adhesive in the particle board production. liquefied spruce wood (LW) in the mixture and with different catalysts added.
Two different loadings of the liquefied wood in these mixtures
Catalyst Meldur H-97 + LW Meldur MS-1 + LW
were used to evaluate the influence of the wood type on the reac-
tion time. The reactivity was determined for mixtures with 20% 50% LW 30% LW 0% LW 50% LW 30% LW 0% LW
(s) (s) (s) (s) (s) (s)
and 30% loadings of different liquefied wood with the addition of
1% AF3,3 (ammonium formate) catalyst. Variations in such behav- 1% PTSA 230 165 121 165 120 91
1% HCOOH 153 105 74 124 80 67
ior were observed, depending on the type of wood flour that was
2% AF3,3 165 100 76 143 85 54
used. In a moist atmosphere, oak wood generates the lowest pH 2% NH4CI 193 125 74 150 80 58
among the chosen wood meals, thus more greatly accelerating 2% NH4NO2 180 120 81 140 85 53
the above mentioned step-growth process. The spruce wood sam- 2% AlCl3 107 75 60 86 55 53
ples were exposed to a moist atmosphere for a longer period of 0% 338 320 223 185

time. Such exposure results in lower pH values being generated

and the subsequent release of the extractive substances. Shorter
reaction times were achieved with the spruce wood meal. How-
ever, the differences were relatively small. It was for such reasons
that spruce wood meal was used in subsequent experiments.
Formaldehyde is a known carcinogen. It can be released from
products containing cured urea–formaldehyde, melamine–urea–
formaldehyde and melamine–formaldehyde adhesives. Recent
findings that the hardening of melamine–urea–formaldehyde resin
precursors occurs mostly due to the reactivity of the melamine, has
led to the conclusion that pure melamine–formaldehyde resin pre-
cursors should give better adhesive performance and lower form-
aldehyde emissions than those that arise with current urea–
formaldehyde resins or melamine–urea–formaldehyde resins
(Kim and Kim, 2005). In order to establish a basis for creating
superior adhesive systems that have lower formaldehyde emis-
sions, seven different melamine–formaldehyde, melamine–urea–
formaldehyde resin precursors were tested, in combination with
different catalysts that are currently used in particle board
production.
The reactivity of liquefied spruce wood with different mela-
mine–formaldehyde and melamine–urea–formaldehyde resin pre-
cursors was determined with 20% and 30% loadings of liquefied
wood without the addition of a catalyst. The results are shown in
Table 2.
The shorter reaction times were achieved using Meldur MS-1
and Meldur H97. Although Melapret NF-70M reacted at a similar
time to that of Meldur H97, the mixing with liquefied spruce wood
was hindered by the low compatibility of the resin precursor with
the liquefied spruce wood. At 30% loading, Cymel 350 performed
equally as the Meldur MS-1. Since both, Cymel 350 and Meldur
MS-1 are highly methylated melamine–formaldehyde resin pre-
cursors with similar characteristics but from different producers,
Meldur MS-1 was used in further experiments.
Different catalysts are usually added to melamine and mela-
mine–urea–formaldehyde resin precursors in order to speed up
the curing reaction. The catalysts used were p-toluenesulfonic acid
(PTSA), formic acid (HCOOH), ammonium formate (AF3,3), ammo-
nium chloride (NH4Cl), ammonium nitrate (NH4NO3) and alumin-
ium chloride (AlCl3).
The gel times of mixtures of liquefied spruce wood with Meldur
MS-1 and Meldur H-97, catalysed with one of the acid catalysts, are
shown in Table 3.
The initial addition of liquid wood slows the reactivity of the
blend due to the buffer effect of the liquid wood during the selfcon-
densation of the Meldur resin. However, at the concentrations be-
tween 20% and 40% the reactivity remains relatively constant as
shown in Fig. 4. Condensation reactions between liquid wood
and Meldur resins prevail. The pH of liquid wood is 2 or less due
to the acid catalyst and the acidic depolymerisation products of
wood that are present in the product. Higher loadings of liquid
wood dilute the system and the reaction becomes slower.
Since the goal of this research was to optimize the adhesive sys-
tem for the potential use in industry different catalysts were used.
Among them, ammonium chloride gives the shortest reaction time
and the highest level of the released formaldehyde while ammo-
nium nitrate gives the opposite effect in industrial applications.
The results in Table 3 show that the chloride salts are efficient
catalysts but are avoided in particle board production. Formic
acid and ammonium formate were then chosen for further
experiments.
The influence of the catalyst loading on the gel time of the mix-
ture of Meldur H-97 and Meldur MS-1 with liquefied spruce wood
was monitored. The results are shown in Table 4.
The influence of the amount of formic acid on the gel time and
thus, on the speed of the condensation reaction, was studied for
systems consisting of 30% and 50% of liquefied spruce wood in
combination with Meldur MS-1 and Meldur H-97. The catalyst, for-
mic acid, was added at room temperature. The measured gel times
show the positive effect of the catalyst on the speed of the reaction.
The addition of 2% of the catalyst to the mixture reduced the gel
time significantly. Greater loadings had a less pronounced effect.
The gel time of the mixture of liquefied spruce wood and the
Meldur resin precursors depends on the loading of the liquefied
wood added, the process temperature and the type and quantity
of the catalyst. Optimization of these parameters is essential prior
the use of such mixtures in particle board production.
1366 M. Kunaver et al. / Bioresource Technology 101 (2010) 1361–1368

Fig. 4. Relationship between gel time and loading of liquefied spruce wood mixed with Meldur MS-1 and Meldur H-97.

Table 4 stant. At higher loadings of the liquefied spruce wood, the system
Gel time of the mixture of Meldur H-97 and Meldur MS-1 with liquefied spruce wood becomes diluted by the excess of liquefied wood and the reaction
(LW) with different concentrations of formic acid. becomes slower. Meldur MS-1, a melamine–formaldehyde resin
Catalyst Meldur H-97 + LW Meldur MS-1 + LW with higher melamine content is, in general, the more reactive sys-
0% (s) 30% (s) 50% (s) 0% (s) 30% (s) 50% (s)
tem. Consequently, the gel times are shorter relative to those given
by the Meldur H-97 mixtures. The pH value of the liquefied wood is
0% 320 338 185 223
1% HCOOH 74 105 153 67 80 124
2 or less and it accelerates the condensation reaction even at room
2% HCOOH 72 80 110 71 93 117 temperature after a several days.
3% HCOOH 60 78 102 63 86 119 The relationships between the gel time and the temperature of
4% HCOOH 44 50 95 48 84 117 the mixtures (50/50 ratio with 1% catalyst ammonium chloride)
5% HCOOH 38 44 84 44 86 116
are shown in Fig. 5, for the same formulations of mixtures.
Fig. 5 shows that the condensation reaction is faster at high
temperatures. Above 120 °C, the influence of the temperature is
Mixtures of liquefied spruce wood with Meldur MS-1 and with less pronounced and the differences in the gel time is smaller. Con-
Meldur H97 were prepared with different liquefied spruce wood sidering the particle board production process, temperatures be-
loadings. All mixtures were catalysed with the addition of a 1% low 180 °C could be applied. Since many industrial applications
aqueous solution of ammonium formate (AF3,3). The gel times of use a press heat of 180 °C, such a reduction in the temperature
both systems are shown in Fig. 4. would reduce the energy costs. The difference in the reactivity be-
It is clear that at 10% liquefied spruce wood loading, the reactiv- tween the more reactive Meldur MS-1 and Meldur H-97 can again
ity decreases till 20% loading. The initial addition of liquefied wood be observed. Meldur MS-1 has shorter gel times at all of the mea-
slows the reactivity of the blend due to the buffer effect on the self- sured temperatures.
condensation of the Meldur resin precursors. After that the gel
time of both mixtures remains almost constant or slightly de-
creases until 40%. Then the gel time increases rapidly, it is consid- 5. Properties of particle boards
ered that by loadings between 20% and 40% there are enough
hydroxyl groups available originating from the liquefied spruce Table 5 shows the properties of particle boards made with the
wood so that the condensation reaction between the liquefied mixture of the liquefied spruce wood and the resin precursors Mel-
spruce wood and the methylol groups on melamine–formaldehyde dur H97 and MS-1. The required values, according to the European
and melamine–urea–formaldehyde resin precursor remains con- standard EN 312 (2003), type P2, are also shown. The formalde-
hyde content was measured according to the European standard
(EN 120), using the perforator extraction method that determines
the formaldehyde concentration by extraction, giving the formal-
dehyde emission potential for the long term use of particle boards.
The limiting value for the formaldehyde Class E1 is 68 mg/100 g of
particle board, measured using the perforator method.
The addition of 50% of the liquefied spruce wood shows no ma-
jor influence on the mechanical properties of particle boards. How-
ever there is a significant decrease in the formaldehyde release of
the board. This may be due to that the depolymerization products
of lignin maintain their aromatic character and remain highly reac-
tive; they act as radical scavengers and can therefore readily react
with free formaldehyde in the mixture, during crosslinking pro-
cess. The extent of swelling is marginally greater. However, since
the particle boards are meant to be used in dry places, such boards
Fig. 5. Gel time of the mixtures of LW with Meldur MS-1 and Meldur H-97 in can still be used. The controlling sample with no liquefied wood
relation to the reaction temperature. added has a similar value. The swelling could be reduced through
 M. Kunaver et al. / Bioresource Technology 101 (2010) 1361–1368 1367

Table 5
Mechanical and physical properties of particle boards prepared using an adhesive that contained the addition of 50% (w/w) liquefied spruce wood and controlling samples
without liquefied wood. Numbers in brackets represent standard deviation.

Required values Achieved values Meldur H-97 Achieved values Meldur MS-1
0% LW 50% LW 0% LW 50% LW
Board thickness EN 324 (mm) 16.0 ± 0.3 15.8 16.1 16.2 16.2
(0.101) (0.112) (0.140) (0.201)
3
Density EN 323 (g/cm ) – 0.69 0.68 0.67 0.67
(0.033) (0.021) (0.041) (0.031)
Internal bond strength Above 0.35 0.42 0.40 0.49 0.43
EN 319 (N/mm2) (0.081) (0.076) (0.114) (0.098)
Surface soundness Above 0.8 0.9 1.0 0.9 1.0
EN 311 (N/mm2) (0.096) (0.101) (0.194) (0.213)
Bending strength Above 13 17.5 15.9 18.4 16,5
EN 310 (N/mm2) (0.121) (0.134) (0.075) (0.106)
Swelling in thickness Below 15 12.4 14.6 10.7 12.9
EN 317 (%) (1.272) (1.397) (0.902) (0.998)
Formaldehyde release Below 8 7.1 5.7 5.9 3.6
Perforator method (0.0472) (0.076) (0.100) (0.117)
EN 120 (mg/100 g board)

Table 6
Properties of laboratory prepared particle boards with regard to the pressing conditions (Meldur H-97 resin precursor and liquefied spruce wood 50/50). Numbers in brackets
represent standard deviation.

Required values Pressing temperature

180 °C 160 °C
Board thickness EN 324 (mm) 16.0 ± 0.3 16.1 (0.112) 16.1 (0.201)
Density EN 323 (g/cm3) – 0.68 (0.021) 0.60 (0.483)
Internal bond strength EN 319 (N/mm2) Above 0.35 0.40 (0.076) 0.39 (0.279)
Surface soundness EN 311 (N/mm2) Above 0.8 1.0 (0.101) 1.1 (0.531)
Bending strength EN 310 (N/mm2) Above 13 15.9 (0.134) 13.6 (0.516)
Swelling in thickness EN 317 (%) Below 15 14.6 (1.397) 15.8 (1.734)
Formaldehyde release perforator method EN 120 (mg/100 g board) Below 8 5.7 (0.076) 5.4 (0.0.91)

modification of the technology of hot pressing and the use of a
greater catalyst loading, thus intensifying the crosslinking reaction.
The manufacturing process efficiency gain can be achieved by
reducing the temperature of the press or by shortening the press-
ing time. Thus, the temperature of the press was reduced from the
standard 180–160 °C. Few changes in the mechanical properties of
particle boards resulted. The only difference was observed in the
thickness swelling (higher at the lower pressing temperature).
The results are presented in Table 6. Since the hot boards from
the press are cooled and conditioned to equilibrate the moisture
content lower press temperature would result in reduced cooling
time and thus, faster production process.
6. Conclusions
Different types of South European hardwood and softwood have
been used in the liquefaction process with yields greater than 99%.
The optimum reaction time needed for the liquefaction was
105 min. A condensation reaction occurred with beech and oak
wood samples after a prolonged reaction time.
A significant difference in average molar mass was determined
between the water insoluble part of the liquefied spruce wood
(9700 Dalton) and water soluble part (670 Dalton) indicating the
presence of a diverse solvolysis pathway for the cellulose, hemicel-
luloses and lignin molecules in the reaction mixture.
The OH value for all liquefied wood samples was determined at
different time intervals. It was possible to distinguish between the
remaining glycols contribution to the total OH value and the lique-
faction products of wood components, ranging from 400 mg KOH/g
at the beginning of the liquefaction process to 200 mg KOH/g at the
end of the reaction.
These OH groups, including those of the remaining unreacted
glycols were available for the condensation reactions with different
melamine–formaldehyde and melamine–urea–formaldehyde resin
precursors. The initial tests involved gel time measurements at dif-
ferent concentrations of the liquefied wood in the mixture with
these resins. The loading and the type of the catalyst were opti-
mized and resin precursors chosen for particle board preparation
on a laboratory scale. By measuring the mechanical properties of se-
lected particle boards and by measuring the formaldehyde release,
it was found, that a 50% addition of the liquefied wood met the
European standard quality demands for particle boards. It has been
shown that the temperature of the press could be lowered by 20 °C
with no significant influence on the mechanical properties and
formaldehyde emission characteristics. Formaldehyde release was
lower than 8 mg/100 g in all experiments due to the positive influ-
ence of the liquefied wood components. It can be concluded that the
products of the liquefied lignin with their aromatic character be-
haved as a formaldehyde scavenger. Lower formaldehyde emissions
from particle boards due to the use of the liquefied wood, are extre-
mely important in the provision of better quality of life.
On the basis of the presented values one can conclude that liq-
uefied woods can be used as a substitute for synthetic resin precur-
sors in adhesives that are used for particle board production.
Acknowledgements
The authors wish to gratefully acknowledge the support for the
presented work received from the Ministry of Higher Education,
Science and Technology of the Republic of Slovenia within the Pro-
gram P2-0145-0104, GGP Postojna for their generous financial sup-
1368 M. Kunaver et al. / Bioresource Technology 101 (2010) 1361–1368
port and Melamin Kočevje who kindly provided resins used in this
study.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.biortech.2009.09.066.
References
Alma, M.H., Maldas, D., Shiraishi, N., 1998. Liquefaction of several biomass wastes
into phenol in the presence of various alkalis and metallic salts as catalysts.
Journal of Polymer Engineering 18, 162–177.
Amen-Chen, C., Riedl, B., Wang, X.M., Holzforschung, Roy C., 2002. Softwood bark
pyrolysis oil-PF resols part 1. Resin Synthesis and OSB Mechanical Properties 56,
167–175.
ASTM Standard D4274-05, 2005. Standard Test Method for Testing Polyurethane
Raw Materials: Determination of Hydroxyl Numbers of Polyols. ASTM
International, Pennsylvania, United States of America.
European Standard EN 120, 1992. Wood-based Panels – Determination of
Formaldehyde Content – Extraction Method Called the Perforator Method.
European Committee for Standardization, Brussels, Belgium.
European Standard EN 310, 1993. Wood-based Panels – Determination of Modulus
of Elasticity in Bending and of Bending Strength. European Committee for
Standardization, Brussels, Belgium.
European Standard EN 311, 2002. Wood-based Panels – Surface Soundness – Test
Method. European Committee for Standardization, Brussels, Belgium.
European Standard EN 312, 2003. Particleboards - Specifications. European
Committee for Standardization, Brussels, Belgium.
European Standard EN 317, 1993. Particleboards and Fibreboards – Determination
of Swelling in Thickness After Immersion in Water. European Committee for
Standardization, Brussels, Belgium.
European Standard EN 319, 1993. Particleboards and Fibreboards – Determination
of Tensile Strength Perpendicular to the Plane of the Board. European
Committee for Standardization, Brussels, Belgium.
European Standard EN 323, 1993. Wood-based Panels – Determination of Density.
European Committee for Standardization, Brussels, Belgium.
Gagnon, M., Roy, C., Riedl, B., 2004. Adhesives made from isocyanates and pyrolysis
oils for wood composites. Holzforschung 58, 400–407.
Hassan, E., Shukry, N., 2008a. Polyhydric alcohol liquefaction of some lignocellulosic
agricultural residues. Industrial Crops and Products 27, 33–38.
Hassan, E., Shukry, N., 2008b. Polyhidric alcohol liquefaction of some lignocellulosic
agricultural residues. Industrial Crops and Production 27, 33–38.
Kim, S., Kim, H.J., 2005. Comparison of standard methods and gas chromatography
method in determination of formaldehyde emission from MDF bonded with
formaldehyde-based resins. Bioresource Technology 96, 1457–1464.
Kishi, H., Fujita, A., Miyazaki, H., Matsuda, S., Murakami, A., 2006. Synthesis of wood
based epoxy resins and their mechanical and adhesive properties. Journal of
Applied Polymer Science 202, 2285–2292.
Kobayashi, M., Tukamoto, K., Tomita, B., 2000. Application of liquefied wood to a
new resin system- synthesis and properties of liquefied wood/epoxy resins.
Holzforschung 54, 93–97.
Kobayashi, M., Hatano, Y., Tomita, B., 2001. Viscoelastic properties of liquefied
wood/epoxy resin and its bond strength. Holzforschung 55, 667–671.
Kobayashi, M., Asano, T., Kajiyama, M., Tomito, B., 2004. Analysis on residue
formation during wood liquefaction with polyhydric alcohol. Journal of Wood
Science 50, 407–414.
Kržan, A., Kunaver, M., 2006. Microwave heating in wood liquefaction. Journal of
Applied Polymer Science 101 (2), 1051–1056.
Kurimoto, Y., Doi, S., Tamura, Y., 1999. Species effect on wood liquefaction in
polyhydric alcohols. Holzforschung 53, 617–622.
Kurimoto, Y., Takeda, M., Koizumi, S., Yamauchi, S., Doi, S., Tamura, Y., 2000.
Mechanical properties of polyurethane films prepared from liquefied wood
with polymeric MDI. Bioresource Technology 74, 151–157.
Kurimoto, Y., Koizimi, A., Doi, S., Tamura, Y., Ono, H., 2001. Wood species effect on
the characteristics of liquefied wood and the properties of polyurethane films
prepared from the liquefied wood. Biomass and Bioenergy 21, 381–390.
Lee, W., Liu, C., 2002. Preparation of liquefied bark based reso resin and its
application to particle board. Journal of Applied Polymer Science 87, 1837–
1841.
Lee, S.H., Teramoto, Y., Shiraishi, N., 2002. Biodegradable polyurethane foam from
liquefied waste paper and its thermal stability, biodegradability and
genotoxicity. Journal of Applied Polymer Science 83, 1482–1489.
Lin, L., Yoshioka, M., Yao, Y., Shirashi, N., 1994. Liquefaction of wood in the presence
of phenol using phosphoric acid as a catalyst and the flow properties of the
liquefied wood. Journal of Applied Polymer Science 52, 1629–1636.
Lin, L., Yao, Y., Shiraishi, N., 2001a. Liquefaction mechanism of b-O-4 lignin model
compound in the presence of phenol under acid catalysis, part 1. Identification
of the Reaction Products. Holzforschung 55, 617–624.
Lin, L., Nakagame, S., Yao, Y., Yoshioka, M., Shiraishi, N., 2001b. Liquefaction
mechanism of b-O-4 lignin model compound in the presence of phenol under
acid catalysis, part 2. Reaction behavior and pathways. Holzforschung 55, 625–
630.
Qiuhui, Z., Guangjie, Z., Jinpeng, C., 2006. Effect of inorganic acid catalysts on
liquefaction of wood in phenol. Frontier Forestry in China 2, 214–218.
Rezzoug, S.A., Capart, R., 2002. Liquefaction of wood in two succesive steps:
solvolysis in ethylene–glycol and catalytic hydrotreatment. Applied Energy 72,
631–644.
Rojas-Escudero, E., Alarcon-Jimenez, A.L., Elizalde-Galvan, P., Rojo-Callejas, F., 2004.
Optimization of carbohydrate sylation for gas chromatography. Journal of
Chromatography A 1027, 117–120.
Shenyuan, F., Lingfei, M., Wnzhou, L., Shuna, C., 2006. Liquefaction of bamboo,
preparation of liquefied bamboo adhesives, and properties of the adhesives.
Frontier Forestry in China 2, 219–224.
Tohmura, S., Li, G., Qin, E., 2005. Preparation and characterization of wood
polyalcohol-based isocyanate adhesives. Journal of Applied Polymer Science
98, 791–795.
Wei, Y., Cheng, F., Li, H., Yu, J., 2004. Synthesis and properties of polyurethane resins
based on liquefied wood. Journal of Applied Polymer Science 92, 351–356.
Wu, L., Guo, S., Wang, C., Yang, Z., 2009. Production of alkanes (C7–C29) from
different part of poplar tree via direct deoxy-liquefaction. Bioresource
Technology 100, 2069–2076.
Yamada, T., Ono, H., 2001. Characterization of the products resulting from ethylene
glycol liquefaction of cellulose. Journal of Wood Science 47, 458–464.
Yamazaki, J., Minami, E., Saka, S., 2006. Liquefaction of beech wood in various
supercritical alcohols. Journal of Wood Science 52, 527–532.
Yao, Y., Yoshioka, M., Shiraishi, N., 1996. Water-absorbing polyurethane foams from
liquefied starch. Journal of Applied Polymer Science 60, 1939–1949.
Zhang, Q., Zhao, G., Yu, L., Jie, S., 2007. Preparation of liquefied wood-based resins
and their application in molding material. Frontier Forestry in China 9 (1), 51–
56.