Beruflich Dokumente
Kultur Dokumente
Fuli Wang
The Fracture
Mechanics of Plant
Materials
Wood and Bamboo
The Fracture Mechanics of Plant Materials
Zhuoping Shao Fuli Wang
•
123
Zhuoping Shao Fuli Wang
School of Forestry and Landscape School of Forestry and Landscape
Anhui Agricultural University Anhui Agricultural University
Hefei, Anhui Hefei, Anhui
China China
This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
part of Springer Nature
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Preface
Wood and bamboo are the plant materials that can be directly used as structural
materials, meanwhile, they are also the oldest and still the most widely used natural
structure materials. Many of them are used as structural materials, such as beam,
frame, floor, and support. And bamboo is also a fine engineering structure material
with high strength, good stiffness, and high wear resistance, which is used to build
bamboo house, as construction scaffolding, and as bamboo ladder. Thus, to study
the strength, toughness, and the failure behaviors of wood and bamboo is very
important for the safety assessment and structural design of wood and bamboo.
Wood and bamboo are natural composites that possess obvious meso-structure
and can be studied in multiscale. Because of inhomogeneous and anisotropic
structure and microscopic or macroscopic natural defects or damages, when loaded,
the macroscopic mechanical behavior of wood and bamboo would be determined
by the irregular evolution behaviors of the defects or damages. Although wood and
bamboo both are cell body plant materials, the differences in macroscopic and
microscopic structure bring wood and bamboo different failure mechanisms cor-
responding to different study methods. Thus, it is significant for the design and
safety analysis of wood and bamboo components to understand how to use fracture
mechanics and meso-mechanics to analyze the fracture behaviors of wood and
bamboo, what the changes of inner microstructure are when loaded, and what the
relation between the changes and macro-mechanical response is. Meanwhile, it will
provide guiding function for the development of new biocomposites that possess
special strength and toughness properties and overcome the defects of biomaterials.
This is a book on the fracture behaviors and toughness mechanism of bamboo
and wood, which reflects the research work of authors in the past decade. In the
sections on wood, varieties of trees are selected, for example, softwoods: China fir
(Cunninghamia Lanceolata), Mongolian pine (Pinus sylvestris var. mongolica
Litv.), Picea jezoensis (Picea asperata), Larch (Larix gmelinii), and so on are
chosen, which possess growth rings in different clarity because of the changes of
early wood and late wood; hardwoods: Populus spp I-69, Castanopsis hystrix,
Koompassia spp, Melia azedarach, and so on are chosen considering their differ-
ence in construction such as diffuse-porous wood, ring-porous wood, wood ray,
v
vi Preface
grain, etc. And moso bamboo (Phyllostachys pubescens) is the study object in the
sections on bamboo. Theoretical analysis is combined with experiments assisted
with various mathematical tools and experiment means.
There are nine chapters in total. The content involves the mechanical charac-
teristics and stress–strain relationship of wood structure, the fracture of wood along
grain, the transverse fracture of wood, the finite element analysis of wood crack tip
stress field and prediction of the crack propagation direction, acoustic emission
characteristics and Felicity effect of wood fracture perpendicular to the grains, the
mechanical characteristics of bamboo structure and its components, the interlaminar
fracture properties of bamboo, and the toughness fracture model and energy
absorbing mechanism of bamboo. Many of the researches are studied for the first
time. Most of the chapters are written by Shao Zhuoping, and Sects. 8.4 and 8.5 are
written by Wang Fuli. In addition, Chap. 9 is written by the corporation of Shao
Zhuoping and Wang Fuli. There might be defects and mistakes inevitably for the
limited specialized knowledge of authors, so please put out the mistakes so that they
can be corrected.
vii
Contents
ix
x Contents
Abstract In this chapter, the history of fracture mechanics is introduced, and the
difference between fracture mechanics and mechanics of materials is discussed.
Then, the application history and research progress of fracture mechanics applied to
wood are reviewed. And it is described briefly that although wood and bamboo are
both cell body plant materials, the differences in macroscopic and microscopic
structure bring wood and bamboo different failure mechanisms corresponding to
different study methods. When studied the fracture properties, wood can be treated
as orthotropic macro-homogeneous body, while bamboo is composite reinforced by
fiber distributed non-uniformly, so meso-mechanics should be used to build
mesoscale model to study the fracture of bamboo.
In 1920, Griffith studied the fracture problem of ideal brittle materials such as glass
and ceramics and obtained some new concepts on material strength [1]. However,
Griffith’s theory has not attracted extensive attention, because the brittle materials
above were not used as engineering structure material in that time, and other
structure materials that showed brittle fracture were few.
Since World War II, many large welding elements were manufactured with
high-strength steel; meanwhile, major accidents increased obviously. Although the
welding fabrications all met the traditional design requirements, brittle fracture that
always gave rise to catastrophic devastation often happened under low stress level
without any forewarning. For example, 40 welded steel bridges collapsed contin-
uously with no omen from 1938 to 1940 [2]. During World War II, for about five
thousand welding ships, more than one thousand fracture faults happened and 238
ships had to be scrapped, some of which fractured into two suddenly in a calm sea
[3]. In 1965, Britain, the offshore drilling platform sank because of the brittle
fracture of the pull rod [4]. In 1969, America, because of the brittle fracture of wing
spindles, warplane F-111 and C-56 crashed [5], and so on. The frequent occur-
rences of the series major accidents shocked the engineering community, because
the structures all met the traditional design requirements when failed. After realized
that there must be something neglected in traditional design idea, people found that
crack in structure was the immediate cause of failure for most accidents by
extensive investigation and research.
In the past 50 years, a lot of theoretical and experimental researches have been
performed on the failure of structure with crack, thus fracture mechanics comes into
being. Fracture mechanics provides the safety design of structure new idea and
method, which is extensively applied to metal, inorganic nonmetal, polymer, tim-
bers, and composites. Hence, fracture mechanics is a subject to study the strength of
crack body, and its generation and development has a close relationship with the
occurrence of engineering major accidents.
rStructure ½r
where for plastic materials, ½r ¼ rns ; for brittle materials, ½r ¼ rnb . n is safety factor
and n > 1. And for the structure bearing impact load and pressure vessels, toughness
must be checked.
However, the traditional design idea above is based on the hypothesis of
material’s continuity and uniformity with no crack and defect. In fact, crack and
defect of engineering material are inevitable, and the fracture of material with crack
is not only determined by crack length and external force but also by the sensitivity
of material to crack, namely, fracture toughness [6].
Taken elasticity and plastic mechanics as theoretical approach, fracture
mechanics abandon the hypothesis of material’s continuity and uniformity, by
which the stress field and displacement field of material with crack can be deter-
mined. And the physical quantity that determines the propagation ability of crack
can be obtained; thus, a new strength design idea is proposed. The new strength
criterion could be developed by measuring the ability of material to resist crack
1.2 Mechanics of Materials and Fracture Mechanics 3
propagation, namely, fracture toughness KC and GC, for example, the fracture
criterion of brittle material is shown as follows:
K criterion: K ¼ KC
G criterion: G ¼ GC
Fracture mechanics was first used in wood by Porter [8]. Based on energy balance
principle, linear elastic fracture mechanics (LEFM) was successfully applied by
Porter to the opening fracture in longitudinal and tangential (LT) and longitudinal
and radial (LR) plan of Western white pine (Pinus monticola). The results showed
that the strain energy release rate GIC, which represented the ability of materials to
resist the propagation of crack along grain, had no relation with the geometric
dimensions of sample and crack length. And the comprehensive value of stress and
crack length could be predicted by the knowledge on GIC, which would cause the
rapid propagation of crack and fracture of sample. After that, researchers from
various countries have done tremendous work on wood fracture, and various wood
fracture modes, fracture toughness test methods, and fracture criterions have been
proposed. Preliminary results have been obtained when fracture mechanics was
applied to solve some practical problems of wood and wood structure.
So far, there are two fracture criterions for the macro fracture of wood: one has
thermodynamics meaning, based on the energy criterion in crack tip process zone,
such as energy release rate criterion (G criterion) [9, 10]; the other mechanics
meaning, based on the strength criterion in crack tip process zone, such as stress
intensity factor criterion (K criterion) [11, 12]. However, wood possesses many
characteristics which are obviously different from those of other orthotropic
materials. And the major characteristic of wood is the high anisotropy among three
principal directions from the perspective of the composition and structure of wood,
which makes it difficult to apply LEFM to wood. The complexity of orthotropic
case is that crack may not propagate along its initial direction in plan. Until now,
regarding the problem of wood fracture, most discussions are on the case of
4 1 Introduction to the Application of the Fracture Mechanics …
structures. Fracture mechanics has been used by French Institute of Wood Science
to predict the crack direction and crack depth in wood caused by the accumulation
and release of growth stress, which is a breakthrough and innovation research idea.
It is also a new development tendency that fracture of wood is combined with its
microscopic structure to explore the toughness mechanism. For example, Ren and
Jiang [25] tested the fracture toughness of Chinese Fir and Masson Pine by LEFM,
and microscopic structure of wood had an influence on wood fracture due to the
fracture surface observed by SEM. Stefanie and Stanzl [26] studied the evolution of
wood microscopic structure under load and the fracture mechanical response in the
chamber of environment scanning electron microscope (ESEM).
Acoustic emission (AE) was first applied in the fracture of wood by Prof. Porter [8],
and then Knuffel proposed that the fracture of wood included three stages: crack
initiation, crack propagation, and failure, and due to the viewpoints, the failure of
wood was treated as a continuous process with different AE characteristics in
different stages but not a single event [27].
Later, AE was used to measure the elastic modulus of wood due to the sound
transmission feature in wood. Ansell [28] found that the shape of AE strain curve
was affected by earlywood or latewood percent, when studied the tensile properties
of three softwoods. Sato et al. [29] found a negative correlation between the number
of AE events and wood strength in tensile test, and differentiated between slow AE
response when microcrack propagated cross growth rings and fast AE response
when macrocrack propagated cross growth rings.
The research of Ogino et al. [30] showed that AE energy would increase rapidly
soon after the first microcrack occurred. Spectral analysis indicated that
high-frequency component would occur before cracking began, and low frequency
only occurred after cracking (once crack began, the frequency would decrease).
Ogino also found that the AE signal from wood drying could be divided into four
patterns and by observation, the appearance of any two patterns could be used as an
early warning signal of cracking.
Suzuki and Schniewind [31] found that there was a linear relation between
fracture toughness and the number of AE event in unit crack zone, when studied the
AE characteristic of the fracture of timber composites with different adhesive
contents. Rice and Skarr [32] studied the AE characteristic of steam beech (Zelkova
schneideriana) panel under transverse bending force and found that the AE char-
acteristic of dry wood was different with that of green wood. Ando et al. [33]
studied the relation between fracture toughness and AE signal of sample with single
edge crack and the influence of different grain angles.
6 1 Introduction to the Application of the Fracture Mechanics …
Schniewind et al. [34] collected AE signals during mode I and mixed mode tests
at different moisture contents and temperatures and found that the AE activity in
mixed mode tests was much higher than that for mode I. Dill-Langer et al. [35] used
AE technique to monitor the fracture of clear spruce wood under tensile loading and
found that there was an onset of AE prior to the first visible crack growth
step. Aicher et al. [36] used AE to localize crack nucleation in glulam loaded in
tension perpendicular to the grains. Reiterer et al. [37, 38] used AE to monitor mode
I fracture of softwoods (spruce and pine) and hardwoods (alder, oak, and ash) and
stated that the AE counts up to maximum force are much higher for the softwoods.
Chen et al. [39] used AE to monitor the failure process of hardwood and softwood
test-pieces under static and fatigue torsion loading and found it was possible to
monitor and analyze the failure process in wood by AE techniques. Choi et al. [40]
studied the fracture processes of typical fiber-reinforced plastic composites lami-
nates with continuous fiber reinforcement and the results showed that the AE
characteristics might represent the process of fiber breakages according to the
various loading stages, which expressed characteristic fracture processes for indi-
vidual fiber-reinforced composite laminates. The feature of the AE hit-event rate, in
combination with AE amplitude classifications, could be utilized for nondestructive
identification of different fracture mechanisms.
Bamboo, “The Second Forest”, is an important part of forest resource in China and
has the advantages of short growth cycle, easy renewal, high production, strong
regeneration capacity, extensive use, and high economic value. There are a wide
variety of bamboo species, among which moso bamboo (Phyllostachys pubescens)
is the most widely distributed one, mostly in Yangtze River Basin. It has a straight
stem, high stiffness, strong strength, and some other fine mechanical properties,
such as high static bending strength and tensile strength, high elastic modulus, and
hardness.
Through millions of years’ evolution, bamboo has formed a particular structure
in order to bear the bending load mainly caused by snow or wind. This structure
contributes to its high transverse bending strength and toughness. In contrast, the
anti-cleaving and anti-shearing strength along the grain of bamboo are relatively
low. Therefore, bamboo often cracks along the direction of grain while drying.
Many studies have been reported on the general mechanical properties of bamboo
[41–43]. But researches on fracture characteristics of bamboo are few [44, 45], and
the transverse fracture toughness KIC of bamboo measured by fracture toughness
test method of metal materials was just a nominal toughness that had limited
physical significance, because stress intensity factor method is not feasible to deal
with the fracture problem of fiber composites.
In view of macro-mechanical behavior, bamboo is a typical unidirectional
long-fiber-reinforced bio-composite. It exhibits significant anisotropy in strength
1.3 Brief Review of the Fracture Mechanics of Wood 7
To study the fracture behavior and fracture mechanism of plant materials com-
prehensively, varieties of trees are selected as experimental subjects, for example,
softwood: China fir (Cunninghamia lanceolata), Mongolian pine (Pinus sylvestnis
var. mongolica Litv.), Picea jezoensis (Picea asperata), Larch (Larix gmelinii), and
so on, which possess growth rings in different clarity because of the changes of
early wood and late wood; hardwood: I-69 poplar (Populus spp.), Castanopsis
hystrix, Koompassia spp., Melia azedarach, and so on are chosen considering their
difference in construction such as diffuse-porus wood, ring-porus wood, wood ray,
and grain. And moso bamboo (Phyllostachys pubescens) is the study object in the
sections on bamboo.
Theoretical analysis is combined with experiments assisted with various math-
ematical tools and experiment means. This book includes six principal aspects,
many of which are studied for the first time, involving:
(1) The fracture properties of wood along grain and test method of wood fracture
toughness; the influence of sample size on test result; the relationship between
critical stress intensity factor KIC and critical energy release rate GIC and the
8 1 Introduction to the Application of the Fracture Mechanics …
influence factors; fractal feature of the fracture surface of wood along grain and
the internal relationship between the fracture toughness and meso-structure of
wood due to the relation between fractal dimension of fracture surface and
fracture toughness; Mode III interlaminar fracture property of wood and the test
method.
(2) The transverse fracture properties of wood and the analysis on stress filed at
crack tip by finite element method, prediction of the crack propagation direc-
tion, the transverse fracture mechanism; the influence of transverse crack on the
tensile strength, bending strength, and impact strength; the suggestion of a net
stress criterion for the safety assessment and strength design of wood compo-
nent with crack perpendicular to grain; analytic solution of energy release rate
when crack propagates laterally based on energy theory.
(3) The raise and preliminary definition of four basic damage structure elements;
the evolution character of microstructure during the fracture of wood; the
application of AE technique to the identification of wood fracture pattern;
research on the felicity effect of wood structure bearing load.
(4) The mechanical characteristics of bamboo structure; analysis on the strength
and elastic modulus of bamboo fiber and ground tissue by Mixture law and
Shear lag theory of meso-mechanics; the difference between bamboo internodes
and node in construction and strength.
(5) Mode I, II, III interlaminar fracture properties of bamboo and the fracture
mechanism due to fracture morphology observed by SEM.
(6) The transverse fracture model of bamboo and the toughness contribution of
each damage pattern; the energy absorbing mechanism of bamboo transverse
fracture.
References
1. Griffith AA (1920) The phenomenon of rupture and flow in solid. Philos Trans R Soc Lond A
221:163–198
2. Xiang H (2001) Advanced theory of bridge structure. China Communication Press, Beijing
(in Chinese)
3. Gao Q (1986) Engineering fracture mechanics. Chongqing University Press, Chongqing
(in Chinese)
4. Yang G (1995) The disasters of offshore engineering and the environment load. China
Offshore Platform 10(5):202–203
5. Li H, Zhou C (1990) Engineering fracture mechanics. Dalian University of Technology Press,
Dalian (in Chinese)
6. Kuang Z, Ma F (2002) Crack tip fields. Xi’an Jiaotong University Press, Xi’an (in Chinese)
7. Gordon JE (1968) The new science of strong materials or why you don’t fall through the
floor. Penguin Books Limited, Haromondsworth
8. Porter AW (1964) On the mechanics of fracture in wood. Forest Prod J 8:325–331
9. Larsen HJ, Gustafsson PJ (1990) The fracture energy of wood in tension perpendicular to the
grain. In: 23th CIB-W18 meeting, Lisbon Portugal, p 23-19-2
References 9
10. Stanzl-Tschegg SE, Tschegg EK, Teischinger A (1994) Fracture energy of spruce wood after
different drying procedures. Wood Fiber Sci 26:467–478
11. Ewing PD, Williams JG (1979) Thickness and moisture content effect in the fracture
toughness of Scots Pine. J Mater Sci 14(12):2959–2966
12. Barrett JD, Foschi RO (1977) Model II stress-intensity factors for cracked wood beams. Eng
Fract Mech 9(2):371–378
13. Wu EM (1967) Application of fracture mechanics to anisotropic plates. J Appl Mech 34:967–
974
14. Wu EM (1968) Fracture mechanics of anisotropic plates. In: Tsai SW (ed) Composite material
workshop, Technomic Publishing Company, Lancaster, p 23
15. Mindness S, Nadeau JS, Barrett JD (1976) Stow crack growth in Douglas-fir. Wood Sci
1:389–396
16. Schnewind AP (1977) Fracture toughness and duration of load factor. Duration factor for
cracks propagating perpendicular to grain. Wood Fiber 9(3):216–226
17. Smith TW, Penny DT (1980) Fracture mechanics of butt joints in laminated wood beams.
Wood Sci 12(4):227–235
18. White MS, Green DW (1980) Effect of substrate on the fracture toughness of wood-adhesive
bonds. Wood Sci 12(3):149–153
19. Triboulot P, Pluvinage G (1984) Validity of fracture mechanics concepts applied to wood by
finite element calculation. Wood Sci Technol 18(1):51–58
20. Boatright SWJ, Garrentt GG (1983) The effect of microstructure and stress state on the
fracture behaviour of wood. J Mater Sci 18:2181–2199
21. Murphy JF (1979) Strength of wood beams with end splits. Research Paper. FPL 347. USDA
Forest Service Products Laboratory. Madison, WI, p 12
22. Foschi RO, Barrett JD (1976) Stress intensity factors in anisotropic plates using singular
isoparametric elements. Int J Numer Meth Eng 10(6):1281–1287
23. Canadaian Stanards Association (1984) Engineering design in wood (working stress design).
CAN3-086-M84
24. Sun Y, Lu Z (1999) Calculation of stress strength factor at crack tip of Fraxinus mandshurica
Rupr. by finite element. J Beijing Forest Univ 21(3):53–57 (in Chinese)
25. Ren H, Jiang Z (2001) Morphology of wood fracture of Chinese Fir and Masson Pine.
Scientia Silvae Sinicae 37(3):118–121(in Chinese)
26. Stefanie E, Stanzl T (2006) Microstructure and fracture mechanical response of wood. Int J
Fract 139:495–508
27. Knuffel WE (1988) Acoustic emission as Strength predictor in structural timber.
Holzforschung 42:336–348
28. Ansell MP (1982) Acoustic Emission from softwood in tension. Wood Sci Technol 16(1):35–
58
29. Sato K, Okano T, Asano I, Fushitani M (1985) Application of AE to mechanical testing of
wood. In: 2nd internatioal conference on acoustic emission, Lake Tahoe, pp 240–243
30. Ogino S, Kaino K, Suzuki M (1986) Prediction of lumber checking during drying by means
of acoustic emission technique. J Acoustic Emission 5(2):61–65
31. Suzuki M, Schniewind AP (1987) Relationship between fracture toughness and acoustic
emission during cleavage failure in adhesive joints. Wood Sci Technol 21:121–130
32. Rice RW, Skarr C (1990) Acoustic emission patterns from the surfaces of red oak wafers
under transverse bending stress. Wood Sci Technol 24:123–129
33. Ando K, Sato K, Fushitani M (1992) Fracturetoughness and acoustic emission Characteristics
of Wood II. J Japan Wood Res Soc 38:342–349
34. Schniewind AP, Quarles SL, Lee H (1996) Wood fracture, acoustic emission, and the drying
process Part 1. Acoustic emission associated with fracture. Wood Sci Technol 30:273–281
35. Dill-Langer G, Aicher S (2000) Monitoring of microfracture by microscopy and acoustic
emission. In: International conference on wood and wood fiber composites, Stuttgart, pp 93–
104
10 1 Introduction to the Application of the Fracture Mechanics …
36. Aicher S, Höfflin L, Dill-Langer G (2001) Damage evolution and acoustic emission of wood
at tension perpendicular to fiber. Holz als Roh Werkst 59:104–116
37. Reiter A, Stanzl-Tschegg SE, Tschegg EK (2000) Mode I fracture and acoustic emission of
softwood and hardwood. Wood Sci Technol 34(5):417–430
38. Reiter A, Stanzl-Tschegg SE, Tschegg EK (2002) Fracture characteristics of different wood
species under Mode I loading perpendicular to the grain. Mater Sci Eng A 332:29–36
39. Chen Z, Gabbitas B, Hunt D (2006) Monitoring the fracture of wood in torsion using acoustic
emission. J Mater Sci 41:3645–3655
40. Choi N-S, Woo S-C, Rhee K-Y (2007) Effects of fiber orientation on the acoustic emission
and fracture characteristics of composite laminates. J Mater Sci 42:1162–1168
41. Zeng QY, Li SH, Bao XR (1992) Effect of bamboo nodal on mechanical properties of
bamboo wood. Sci Silvae Sinica 28(3):247–252 (in Chinese)
42. Ahmad M, Kamke FA (2005) Analysis of calculate bamboo for structural composite
materials: physical and mechanical properties. Wood Sci Technol 39:448–459
43. Obataya E, Kitin P, Yamauchi H (2007) Bending characteristics of bamboo (Phyllostachys
pubescens) with respect to its fiber-foam composite structure. Wood Sci Technol 41:385–400
44. Xi X, Xi D (1991) Fracture behaviour of bamboo. Mater Sci Progr 5(4):336–341(in Chinese)
45. Amada S, Untao S (2001) Fracture properties of bamboo. Compos B 32:451–459
Chapter 2
Mechanical Characteristics
and Stress–Strain Relationship
of Wood Structure
Fibril
Amorphous
substance
Microfibril
Crystal-like
arrangement
Hydrogen bond
Cellulose chain
molecules
OH H CH2OH
H H OH O
H OH H H
OH
OH H H H OH OH
O
CH2OH OH H
R(y)
2.2 The Stress–Strain Relation of Solid Material 13
According to elasticity theory [6], the stress state or strain state of an arbitrary point
in solid, as seen in Fig. 2.3, could be expressed by six stress components or strain
components, namely, normal stress: rx ; ry ; rz , shear stress: syz ; szx ; sxy ; normal
strain: ex ; ey ; ez , shear strain: cyz ; czx ; cxy .
In the elastic range, stress component is a linear function of strain component
due to generalization Hooke law, as follows:
2 3 2 32 3
rx C11 C12 C13 C14 C15 C16 ex
6 ry 7 6 C21 C22 C23 C24 C25 C26 76 ey 7
7 6
6 7 6 7
6 rz 7 6 C31 C36 7 6 7
6 7¼6 C32 C33 C34 C35 7 6 ez 7 ð2:1Þ
6 syz 7 6 C41 C46 7 6 7
6 7 6 C42 C43 C44 C45 76 cyz 7
4 szx 5 4 C51 C52 C53 C54 C55 C56 4 czx 5
5
sxy C61 C62 C63 C64 C65 C66 cxy
elastic body, the elementary work increment per unit volume dW has no relation
with order, namely:
@W @W
¼ Cji ej ) ¼ Cji ð2:4Þ
@ej @ej @ei
@W @W
¼ Cij ej ) ¼ Cij ð2:5Þ
@ej @ei @ej
For anisotropic material, there is no interaction between normal stress and shear
strain, as well as shear stress and normal strain, and shear stress and shear strain, so
there are only nine independent constants in stiffness matrix. The stress–strain
relation in coordinate system along principal direction of anisotropic material can
be expressed as
2 3 2 32 3
rx C11 C12 C13 0 0 0 ex
6 ry 7 6 C12 C22 C23 0 0 0 7 6 ey 7
7 6
6 7 6 7
6 rz 7 6 C13 0 7 6 7
6 7¼6 C23 C33 0 0 7 6 ez 7 ð2:7Þ
6 syz 7 6 0 0 7 6 7
6 7 6 0 0 C44 0 76 cyz 7
4 szx 5 4 0 0 0 0 C55 0 4 czx 5
5
sxy 0 0 0 0 0 C66 cxy
Engineering elastic constants are the elastic constants of material including gen-
eralized elastic modulus, Poisson’s ratio, and shear modulus, which can be mea-
sured by simple tests, such as axial tensile test and pure shear test. Engineering
constants are more intuitive compared with stiffness coefficients and flexibility
coefficients in stiffness matrix and flexibility matrix, so engineers are used to
express generalized Hooke’s law in the form of engineering constants Ei, Gi, and
lij. The flexibility matrix of anisotropic material can be expressed with engineering
constants as follows:
2 l l
3
1
Ex Eyxy Ezxz 0 0 0
6 lxy l 7
6 Ey 1
Ezyz 0 0 0 7
6 l Ey 7
6 xz l
Eyzy 1
0 7
6 Ex 0 0 7
Sij ¼ 6 Ez
7 ð2:9Þ
6 0 0 0 1
0 0 7
6 Gyz 7
6 0 1
0 7
4 0 0 0 Gzx 5
1
0 0 0 0 Gxy
where Ex, Ey, and Ez are the elastic modulus along three elastic axes x, y, and z,
respectively. lij is transverse sensitivity ratio of the strain in direction i caused by
the stress in direction j, called Poisson’s ratio.
ej
lij ¼ ð2:10Þ
ei
Gyz, Gzx, and Gxy are the shear modulus in x–y, z–x, and x–y planes, respectively.
For anisotropic material, the flexibility matrix has a property of symmetry, and it is
lij lij
¼ i; j ¼ x; y; z ð2:11Þ
Ei Ej
E
G¼ ð2:12Þ
2ð1 þ lÞ
16 2 Mechanical Characteristics and Stress–Strain Relationship …
Due to symmetry principle, the flexibility matrix above has the property of
symmetry [6], so the following relationship expressions can be obtained:
lLR lRL lLT lTL lRT lRT
¼ ; ¼ ¼ ð2:14Þ
ER EL ET EL ET ER
2. Stress characteristics
Due to description ③, for the upper surface and lower surface, there is
rz ¼ szx ¼ szy ¼ 0;
Due to description ①, at any point, there is rz ¼ szx ¼ szy ¼ sxz ¼ syz ¼ 0;
Due to description ②, there are only three stress components at any point, rx , ry
sxy ¼ syx , which are parallel to xy plane, and they are only the function of x and
y for the uniform thickness, as seen in Fig. 2.5.
The situation above is called as plane stress problem.
2. Stress characteristics
Cut a thin cross section at where both far from the two end, then:
➀ the stress on both sides are symmetric, szx = szy = sxz= syz = 0;
x
18 2 Mechanical Characteristics and Stress–Strain Relationship …
8 8 s
< ex ¼ E1 rx lðry þ rz Þ < cyz ¼ Gyz
e ¼ 1 r lðrz þ rx Þ : c ¼ szx : ð2:15Þ
: y E1 y : zx sGxy
ez ¼ E rz lðrx þ ry Þ cxy ¼ G
In plane stress state, as rz ¼ 0, szx ¼ 0 and szy ¼ 0, Hooke’s law can be sim-
plified as follows:
8
< ex ¼ E ðrx lry Þ
1
l
E 0 and l0 are used to substitute 1l
E
2 and 1l, and then, the relations above can be
simplified as follows: 8 0
< ex ¼ E= ðrx l ry Þ
1
0
e ¼ ðr l rx Þ :
1 ð2:18Þ
: y E10 y
cxy ¼ G sxy
It can be seen that the form of Hook’s law in plane stress state is same with that
in plane strain state. Thus, if the geometrical shape of xy plane and the load
condition in plane stress state are same with that in plane strain state, once the
solutions of plane stress problem are obtained, the solutions of plane strain problem
can be obtained by substituting E 0 and l0 for E and l.
There are many methods to measure the elastic coefficients of wood, such as
extensometer method, laser speckle interferometry method, ultrasonic method, and
electrometric method. Electrometric method was used to measure the elastic
coefficients of wood since 1960s. Digital speckle correlation method (DSCM) was
applied on wood by Xu et al. [7], Viotti et al. [8], and Wang et al. [9]. The elastic
coefficients of wood are important parameters for the design of wood structures and
20 2 Mechanical Characteristics and Stress–Strain Relationship …
F F F
4# 5# 6#
wood composite, while the experimental data of wood elastic coefficients are not
sufficient to be used for design.
Among the 12 elastic coefficients, three elastic modulus (EL, ER, and ET) and six
Poisson’s ratio (lRT, lTR, lTL, lLT, lLR, and lRL) can be tested by electrometric
method and DSCM with the rectangle samples as shown in Fig. 2.8. The three
shear elastic modulus (GRT, GTL, and GLR) can be tested by three rectangle samples
with 45° off-axis due to the relationship between elastic coefficients of normal axis
sample and those of off-axis sample, as seen in Eq. (2.19).
EX45
Gij ¼ ð2:19Þ
2ð1 þ lXY Þ
where EX45 is the elastic modulus of sample with 45° off-axis.
Resistance strain gage is the sensor in electrometric method. It is that the strain of
sample can be converted into the resistance variation of resistance strain gage, and
the resistance variation can be detected and converted into the strain of sample by
resistance strain gage with an accuracy of 1 le. The testing bridge is shown in
Fig. 2.9, where R1 is working gage, R2 is temperature compensating gage, and R3
and R4 are the fixed internal resistances.
As wood is inhomogeneous body, the resistance strain gage should have large
gage length to extend test range. In this book, BQ12O-IOAA foil paper-based strain
gages are used, and they are attached on sample by No.502 glue, room temperature
2.6 Tests of Wood Elastic Coefficients 21
R1
R2
A C U
R3
R4
curing. Before this, as wood is a porous material, the surfaces of sample should be
sanded to ensure the pores are filled with wood flour. The arrangement of resistance
strain gages has been shown in Fig. 2.8. And the coordinate lines on resistance
strain gage should overlap with the longitudinal line and transverse line drawn on
the surfaces of sample in advance.
The tensile, compressive, and bending elastic modulus of wood are about equal
[5], so uniaxial compression test can be performed to measure the elastic coeffi-
cients of wood. Before that, failure test should be performed on study subject to
determine the lower limit load and upper limit load, and they are 0.3 and 0.7 time of
proportional limit. Stepwise loading method is often adopted, and it is that test is
performed on a computer-controlled testing machine, and after setting up step
number k and step length DF, loading and unloading will be performed by test
machine automatically. The stepwise loading curve of spruce sample 1# (as seen in
Fig. 2.8) is shown in Fig. 2.10 with a load speed of 20 N/s. The initial load is
0.5 kN keeping 5 s, DF = 500 N, until the fifth step with a load of 2.5 kN keeping
5 s, then unloaded with a speed of 100 N/s, and the process should be repeated six
times. However, the maximum load and step length should be decreased corre-
spondingly for samples 2# to 6#.
The strain data can be read on test machine directly or automatically collected by
the assistance of strain indicator. Creep is inevitable for wood during loading, but if
test time is not long and increment formula is adopted, the influence of viscous
deformation on elastic strain rate can be ignored. The increment formula of wood
elastic coefficients is shown as follows:
DFi REi
Ei ¼ ; E¼
A Dei n
0 ð2:20Þ
Dei Rl
li ¼ ; l ¼ i
De i n
(1) Speckle
After the surfaces of the six samples were polished, black matted paint and white
matted paint were sprayed on the surfaces of sample by high-pressure spray method
in sequence to form artificial speckle. The size of speckle would affect the accuracy
of associated processing, which should satisfy the requirement of CCD’s resolution
and sampling theorem. Too small speckle would cause a great error during the
2.6 Tests of Wood Elastic Coefficients 23
PC
CCD Image card
processing of different values, while too large speckle would affect the sensitivity of
the variation of correlation coefficient to the deformation of sample surface and
cause an error in related calculation process.
(3) Test
Stepwise loading method was applied in DSCM just like that of electrometric
method. Step number k, step length DF, and load maintaining time were set up in
advance. Loading and unloading will be performed by test machine automatically,
and speckle image would be collected during each load maintaining stage, as seen
in Fig. 2.12.
1X n
vi;j þ n vi;j 1 X2n
Dei ¼ ; De ¼ Dei
n j¼1 nm 2n i¼1
ð2:21Þ
0 1X n
ui þ n;j ui;j 0 1 X2n
Dej ¼ ; De ¼ Dej
n j¼1 nm 2n j¼1
(1,1) (2n,1)
(1,2n) (2n,2n)
PK
DFk Ek
Ek ¼ ; E ¼ k¼1
A Dek K
0 PK ð2:22Þ
Dek l
lk ¼ ; l ¼ k¼1 k
De k K
where DFk was the load increment, and A was the cross-sectional area of sample.
The elastic coefficients of orthotropic material should meet the following require-
ments: the product of stress component and corresponding strain component should
be the work done by stress; the sum of work done by all stress components should
be a positive value; thus, flexibility or stiffness matrix must be positive definite [12].
The requirements above present a thermodynamic limitation on the value of elastic
coefficients, and if described with engineering coefficients, the elastic coefficients of
orthotropic material should satisfy the following equations:
However, researches indicated that there were always some differences between
the measured value and theoretical value of wood elastic coefficients, and the
26 2 Mechanical Characteristics and Stress–Strain Relationship …
relative error was in the range of 15–25% [12–14]. The main reasons are as follows:
wood is approximately cylindrically symmetric, but rectangular coordinate system
is adopted in the orthotropic hypothesis of wood, which leads to a certain model
error especially to wood with small diameter and growth ring with large curvature;
wood properties vary greatly; the errors come from sample processing and the
attachment of strain gage, such as deviation longitudinal axis of sample and fiber
orientation, error on the angle of strain gage and the attachment quality of strain
gage, and so on.
References
1. Meyer KH, Misch L (1937) Posiondes atemes dans le nouveru modele spatial do la cellulose.
Helv Chim 20:232–244
2. Roelofsen PA (1959) The plant cell-wall. Gegruder Borntraeger, Berline-Nikolassee, pp 126–
189
3. Junqing Cheng (1985) Wood science. China Forestry Publishing House, Beijing, p 1379
4. Price AT (1928) A mathematical discussion on the structure of wood in relation to its elastic
properties. Philos Trans R Soc Lond 228:1–62
5. Kollmann FFP (1991) Wood science and principle of wood technology (translation). China
Forestry Publishing House, Beijing, pp 278–281
6. Jones RM (1972) Mechanics of composite materials. Scripta Book Company, New York
7. Xu M, Jin G, Lu Z (2003) Digital spackle correlation method (DSCM) for measurement of
wood compression elastic modulus. Scientia Silvae Sinicae 39(2):174–176
8. Viotti MR, Kaufmann GH, Galizzi GE (2006) Measurement of elastic moduli using spherical
indentation and digital speckle pattern interferometry with automated data processing. Opt
Laser Eng 44(6):495–508
9. Wang QH, Xie HM, Tang PF et al (2009) A study on the mechanical properties of beagle
femoral head using the digital speckle correlation method. Med Eng Phy 31(10):1228–1234
10. Yamaguchi IA (1981) Laser-speckle strain gage. J Phys E Sci Instrum 14:1270–1273
11. Ranson WF, Peters WH (1982) Digital image techniques in experimental stress analysis. Opt
Eng 21(3):427–431
12. Lempriere BM (1968) Poisson’s ratio in orthotropic materials. AIAA 6:2226–2227
13. Liu Z, Liu Y, Yu H et al (2004) Research progress of automatic measuring technology of
wood modulus of elasticity. Forest Sci Technol 29(1):45–48
14. Wang X, Ross RJ et al (2001) Several nondestructive evaluation techniques for assessing
stiffness and MOE of small-diameter logs. Research Paper, FPL-RP-600, USDA, Forest
Service, FPL, Madison, WI, p 12
Chapter 3
Fracture of Wood Along Grain
Abstract In this chapter, the theory of LEFM and the special application of LEFM
on wood are introduced. Then, different samples and methods are applied to
measure the Mode I fracture toughness of wood along grain, and the results show
that the fracture toughness of wood along grain is the basic attribute of wood and
has nothing to do with test method, geometrical shape, and size of crack. The
research indicates that LEFM based on isotropic body is applicable to the crack
propagation of wood along grain. And based on energy theory, double cantilever
inversion symmetry bending load method is applied to measure the Mode III
fracture toughness of spruce. The results show that the average Mode III inter-
laminar fracture toughness 1.05 kJ/m2 can be seen as a basic attribute of spruce that
represents the capacity of spruce to resist the propagation of Mode III crack.
3.1 Introduction
In recent years, with the further application of fracture mechanics on wood, the
application fields have developed from safety assessment to processing and uti-
lization of wood. Currently, there are many test methods to determine the fracture
parameters of wood, among which fracture toughness is one of the most important
indexes. Fracture toughness represents the capacity of wood to resist the unstable
propagation of opening crack, which has significant meaning for the quality
assessment, safety design, and optimization of processing method of wood.
However, there may be great difference in the results of different test methods,
because there is no universal test standard. What’s more, comparative studies
among different test methods are not sufficient as well. In this chapter, the partic-
ularity of the application of LEFM on wood is discussed, then the fracture property
of wood along grain and test method are studied, at last, fractal theory is applied to
study the fractal feature of the fracture surface of wood, and the relationship
between fractal dimension of fracture surface and fracture toughness is developed.
3.2.1 Crack
Crack and defects are inevitable in engineering materials, which might be generated
during manufacture, processing, or using, such as fatigue crack caused by alter-
nating loads, radial or ring shake caused by compression damage or drying. As seen
in Fig. 3.1, crack can be classified into three types due to the characters of load and
fracture.
Mode I (Opening mode) crack—external normal stress is perpendicular to crack;
Mode II (Sloping mode) crack—shearing stress is parallel to crack;
Mode III (Tearing mode) crack—stress staggers crack surfaces.
Among the three types of crack, Mode I and Mode II cracks are more common
and dangerous than Mode II crack. If there are normal stress and shearing stress or
tensile stress is not perpendicular to crack line on crack body, Mode I and Mode II
(Or Mode III) cracks will occur in material at the same time, which is called mixture
crack. And in fact, mixture crack is the most common situation, and it is treated as
Mode I crack sometimes from the angle of safety and convenience. Thus, Mode I
crack is the research focus of engineering fracture mechanics.
r x
To study the stress distribution at crack tip, Irwin illuminated the general form of
stress field at crack tip by Westergaard stress function [2, 3] as follows:
pffiffiffiffiffiffiffiffiffi
r1 p a
rij ¼ p ffiffiffiffiffiffiffiffiffiffiffi fij ðhÞ ð3:2Þ
2p r
where a is the size of crack; r1 is the working stress far away from crack tip; r and
h are polar coordinates with crack tip as origin; fij ðhÞ is azimuth function neglecting
the higher order term of r.
As seen in Fig. 3.2, there is a central through crack with a length of 2a in a plate.
It is assumed that the thickness of plate is one unit, and length and width are much
greater than 2a, so the crack can be treated as Mode I crack in an infinite plate under
tensile stress. The stress and strain fields in the small area at crack tip are
pffiffiffiffiffiffi
r pa h h 3h KI
rx ¼ pffiffiffiffiffiffiffiffi cos 1 sin sin ¼ pffiffiffiffiffiffiffiffi f ðhÞ
2pr 2 2 2 2pr
pffiffiffiffiffiffi
r pa h h 3h KI
ry ¼ pffiffiffiffiffiffiffiffi cos 1 þ sin sin ¼ pffiffiffiffiffiffiffiffi f ðhÞ
2pr 2 2 2 2pr
pffiffiffiffiffiffi
r pa h h 3h KI
sxy ¼ pffiffiffiffiffiffiffiffi sin cos cos ¼ pffiffiffiffiffiffiffiffi f ðhÞ ð3:3Þ
2pr 2 2 2 2pr
pffiffiffiffiffiffi rffiffiffiffiffi rffiffiffiffiffi
r pa 2r 2h h KI 2r
u¼ ð1 lÞ þ ð1 þ lÞ sin cos ¼ f ðhÞ
E p 2 2 E p
pffiffiffiffiffiffi r ffiffiffiffi
ffi r ffiffiffiffi
ffi
r pa 2r h h KI 2r
v¼ ð1 lÞ þ ð1 þ lÞ sin2 sin ¼ f ðhÞ
E p 2 2 E p
E
where for plane stress: E ¼ E, l ¼ l; for plane strain: E ¼ 1l 2 , l ¼ 1l.
l
y
30 3 Fracture of Wood Along Grain
It can be seen that for an arbitrary point in front of crack, due to Eq. 3.3, the
pffiffiffiffiffiffiffiffiffi
stress at the point is determined by r p a called as stress intensity factor K, which
is a new physical quantity. Generally, stress intensity factor can be written as
pffiffiffiffiffiffiffiffiffi
K ¼r paY ð3:4Þ
where r is the working stress assuming there was no crack in the center, and Y is
shape factor related to the shape and location of crack.
Thus, stress intensity factor K is the function of working stress, crack size, and
shape. And according to the comprehensive factors, when K reaches to a critical
value KC, crack will propagate unstably, so a basic criterion in LEFM is established:
K ¼ KC ð3:5Þ
dW ¼ dU þ dC þ dP ð3:6Þ
Or
3.2 Theory of LEFM 31
dW dU ¼ dC þ dP ð3:7Þ
In Eq. (3.7), dC þ dP is the energy consumed when crack area increases dA,
namely the energy needed to resist the propagation of crack, which is determined by
the toughness of material. For ideal brittle material, dP ¼ 0, while for metal
material, dP dC, where dC can be neglected.
dW dU in Eq. (3.7) represents the energy released by the system when crack
area increases dA, which is used to drive the propagation of crack. If П is total
potential energy, P ¼ U W, then dP ¼ dW dU represents the energy
released when crack increases da, which is the active force to drive the propagation
of crack. So there is:
dP ¼ dW dU ¼ dC þ dP ð3:8Þ
The energy release rate (G) of crack is defined as the energy released by elastic
system when crack propagates per unit area, so there is
@P @W @U
G¼ ¼ ð3:9Þ
@A @A @A
If the thickness of crack body is B, the length of crack is a, when crack increases
da, the increased area is dA ¼ Bda; so Eq. (3.9) can be written as
1 @P
G¼ ð3:10Þ
B @a
The energy consumed when crack propagates per unit area is defined as ratio of
crack propagation resistance (R or GC), so
@C @P
R ¼ GC ¼ þ ð3:11Þ
@A @A
To a certain material, the crack surface work and plastic work of crack propa-
gation are material constants that have nothing to do with external load and the
geometry of crack, so R or GC reflects the ability of material to resist fracture
damage, called as fracture toughness of material and is determined by test.
When energy release rate G reaches to GC, the crack system would loss balance
and crack begin to propagate unstably. So G criterion is
G ¼ GC ð3:12Þ
There are two views on the study of crack propagation law by linear elastic fracture
mechanics: K criterion and G criterion. Though the starting points of these two
criteria are different, they are actually the same, because there is a certain rela-
tionship between G and K under the linear elastic conditions. Taking the mode-I
crack of isotropic material as an example, due to Westergaard [3] stress function,
the calculation formulas of stress and displacement perpendicular to crack surface at
crack tip are
KI h h 3h
ry ¼ pffiffiffiffiffiffiffiffi cos 1 þ sin sin ð3:13Þ
2pr 2 2 2
rffiffiffiffiffi
ð1 þ lÞKI 2r h 3h
v¼ ð2v þ 1Þ sin sin ð3:14Þ
4E p 2 2
ðPlanestressÞ
3l
where E is elastic modulus, l is Poisson’s ratio, v ¼ 1þl .
3 4l ðPlanestrainÞ
When crack length increases Da, if the energy released equals to the work done
to close the opening crack to its original state, as seen in Fig. 3.3, the energy
released can be calculated as follows:
ZDa
1
U¼ ry 2v Bdn ð3:15Þ
2
0
ZDa
U 1
GI ¼ ¼ ry h¼0;r¼n vh¼p;r¼Dan dn
B Da Da
0
ðv þ 1Þð1 þ lÞ 2 ð3:16Þ
¼ KI
4E
1
¼ 0 KI2 ¼ S0 KI2
E
0
E ðPlane stressÞ 0
where E ¼ E
ðPlane strainÞ , S ¼ E0 is compliance.
1
ð1lÞ2
For Mode II and Mode III crack, there are similar relationships (Assuming crack
propagates along the original crack line):
1 2
GII ¼ K ð3:17Þ
E0 II
1l 2
GIII ¼ KIII ð3:18Þ
E0
The relationship between K and G indicates that K is the strength of elastic stress
field around crack tip, while the square of K could determine the energy release rate
of crack propagation, therefore, K criterion equals to G criterion when solving linear
elastic fracture problems. G criterion has a simple form and clear concept, but the
analytical expression of G should be given when a crack is introduced to a structure,
which will encounter great difficulty in mathematical and mechanical analysis.
Irwin pointed out that for a linear elastic crack body, G could be expressed by the
compliance of the system [4]. For a plate with a central through crack, G is dis-
cussed in constant displacement situation and constant load situation, respectively.
1. Constant displacement situation
As seen in Fig. 3.4, for an elastic crack body, the deformation energy is U ¼ 12 FC,
and the displacement is D ¼ CF, where C is compliance. And U is only in related
34 3 Fracture of Wood Along Grain
(a) (b)
F
A
FA
U
G=−
2a 2a +2da A
FB B
C
0
Fig. 3.4 The energy release rate of crack propagation in the situation of constant displacement
with the final state and has nothing to do with process, so when crack propagates
from a to a + da, DU = Ua+da − Ua.
Case A: Crack length 2a remained unchanged, when plate is stretched over D, the
deformation energy UA =FAD/2;
Case B: Crack length 2a +2da remained unchanged, when plate is stretched over D,
the deformation energy UB = FBD/2.
Thus, after elastic plate is stretched over D under load F, the up and down ends
of plate are fixed to form a energy closed system with constant displacement. When
crack propagates from 2a to 2a +2da, deformation energy of plate will be released
and decrease, namely @U @A ¼ ðSOBC SOAC Þ\0.
As the displacement of load point is unchanged, namely, dD = 0, the increment
of external work is dW = 0, thus Eq. (3.9) can written as:
@ðW UÞ @U 1 @U
GI ¼ ¼ ¼ ð3:19Þ
@A @A D B @a D
The equation above shows that in the constant displacement situation, the strain
energy released by the system is used to drive the propagation of crack, so the
energy rate of crack propagation is the energy release rate of elastic body. And the
compliance C of elastic body is the function of crack length, namely, C = C(a).
As
dD ¼ FdC þ CdF ¼ 0
1 1 1 1
dU ¼ FdD þ DdF ¼ FCdF ¼ F 2 dC
2 2 2 2
3.2 Theory of LEFM 35
1 1 1
dU ¼ FdD ¼ FðFdC þ CdFÞ ¼ F 2 dC ð3:21Þ
2 2 2
It can be seen from the equation above that the energy consumed by the prop-
agation of crack is the energy of external work minus the increment of elastic strain
energy in constant load condition.
The comparison between Eqs. (3.20) and (3.23) indicates that no matter the
situation is constant displacement or constant load, energy release rate GI has a
uniform expression:
(a) (b)
F
A B
F
2a 2a +2da
F
F 0 A B
Fig. 3.5 The energy release rate of crack propagation in the situation of constant load
36 3 Fracture of Wood Along Grain
F 2 @C
GI ¼ ð3:24Þ
2B @a
Or dimensionless crack length a/W can replace crack length a, then Eq. (3.24)
can be written as
P2 @C
GI ¼ ð3:25Þ
2BW @ða=WÞ
For the fracture problem of anisotropic material, the propagation of crack is com-
plex, and even for Mode I crack in anisotropic material, it may not propagate along
the extension of the original crack. The fracture toughness of anisotropic material is
in close relation to the direction of material property.
The analytical solution of stress, strain, and displacement field at crack tip has
been given in the early research work of Sih et al. [4] and Wu [5, 6]. The stress and
strain around the Mode I crack tip in the center of infinite anisotropic plate
(Fig. 3.2) can be expressed as
KI n1 n2 n2 n1
rx ¼ pffiffiffiffiffiffiffiffi Re
2pr n n2 u2 u1
1
KI 1 n1 n2
ry ¼ pffiffiffiffiffiffiffiffi Re
2pr n1 n2 u2 u1
KI n1 n2 1 1
sxy ¼ pffiffiffiffiffiffiffi
ffi Re ð3:26Þ
2pr n1 n2 u1 u2
rffiffiffiffiffi
2r 1
u ¼ KI Re ðn p2 u2 n2 p1 u1 Þ
p n1 n2 1
rffiffiffiffiffi
2r 1
v ¼ KI Re ðn1 q2 u2 n2 q1 u1 Þ
p n1 n2
where
3.3 Fracture Mechanics of Anisotropic Material 37
S11 n4 2S16 n3 þ ðS12 þ S66 Þn2 S26 n þ S22 ¼ 0, where, Sij is the coefficient of
flexibility matrix.
For Mode I crack problem, as load is acted on crack body, the energy released by
crack body still equals to the work done to close the opening crack to its original
state:
ZDa
1
U¼ ry 2vBdn
2
0
ZDa
ð3:28Þ
¼ ry h¼0;r¼n vh¼p;r¼Dan Bdn
0
1 n þ n2
¼ BKI2 S22 Re i 1
2 n1 n2
pffiffiffiffiffiffi
where K ¼ r pa.
So the energy release rate is
U K2 n þ n2
GI ¼ ¼ I S22 Re i 1 ð3:29Þ
B Da 2 n1 n2
Similarly,
KII2
GII ¼ S11 Re½iðn1 n2 Þ ð3:30Þ
2
For orthotropic material, when Mode I crack is parallel to one of the symmetry
planes, as S16 ¼ S26 ¼ 0, the relationship expression of K and G with four inde-
pendent elastic constants can be obtained. n1 and n2 are the roots of the following
equation:
38 3 Fracture of Wood Along Grain
S11S12
12 h S22
12 i12
where S* is equivalent flexibility, and S ¼ 2 S11 þ 2S12 þ S66
2S11 .
Sij can be appropriately replaced by engineering elastic parameters, thus the
equivalent flexibility of Mode I crack body can be expressed as
2 !1=2 31=2
1=2 1=2 lTL
1 E 2 þ 1
S ¼ 4 ET GLT 5
L
þ ð3:35Þ
2EL ET ET 2
EL
The small value term with Poisson’s ratio can be neglected, so the equivalent
flexibility can be written as
1=2 " 1=2 #1=2
1 EL EL 1=2
S ¼ þ ð3:36Þ
2EL ET ET 2GLT
Sih et al. [4] proved that the concept and relationship between elastic stress and
input energy of isotropic crack body can be generalized to the similar situation of
anisotropic material. And if self-balancing forces are acted on crack body, the form
of 2D stress intensity factor of anisotropic material is same with that of isotropic
material. Therefore, the related concepts in classical fracture theory are appropriate
for the crack problem of anisotropic material, for example,
P2 @C
GI ¼ ð3:37Þ
2BW @ða=WÞ
So
3.3 Fracture Mechanics of Anisotropic Material 39
P2 @C
KI2 ¼ ð3:38Þ
2BWS @ða=WÞ
RL RT LR
TL TR LT
2a
K
rij ¼ pffiffiffiffiffiffiffiffi Re fij h; a; aij ; u1 ; u2 ; u3
2pr
rffiffiffiffiffi ð3:39Þ
K 2r
vij ¼ Re fij h; a; aij ; u1 ; u2 ; u3
G p
where aij is elastic constant of material, Re is the real part of complex function fij.
u1, u2, and u3 are complex parameters depending on the anisotropic degree of
material and the angle a between crack and longitudinal fiber (Fig. 3.7).
From the equations above, it can be found that the fracture of wood is much
more complex than that of isotropic material. The difference between anisotropic
material and isotropic material has been discussed by Shahrokh Parhiagar et al. [8],
when LEFM was used in composite materials. However, for wood, the great dif-
ferences lie in the following aspects:
1. In general, crack in wood always propagates along the direction of fibers instead
of its original direction, while LEFM presupposes that crack propagates along
its original direction all the time.
2. Even under simple load, compound form displacement may occur at crack tip in
wood, which is different with Mode I, Mode II, and Mode III crack in LEFM.
3. The stress field at crack tip in wood is the function of complex parameters that
are the function of wood property and the angle a between crack and longitu-
dinal fiber, which is different with LEFM developed from the stress field at
crack tip and in no relation with material property and crack direction.
Generally, LEFM is not applicable to approximately orthotropic wood consid-
ering the three facts above. And there are no material constants KIC, KIIC, and KIIIC
for wood to represent its fracture toughness, unless all the cases that cracks are at
different angles to fiber direction are studied and the applicability of LEFM to each
case is verified, which is very inconvenient even impossible in practice. However, if
crack is along fiber direction and the orthotropic principal axes are superposed with
the direction of crack surface and the propagation direction of crack, respectively,
the three differences can be eliminated, and this particular situation has been proved
by many experiments that:
3.4 The Special Application of LEFM on Wood 41
1. Crack propagates along its original direction in the same direction with fibers;
2. Displacement is not in compound form;
3. To a fixed angle between crack and fiber (a = 0), the complex parameters of
material are constants, so the stress at crack tip is only the function of r and h.
The situation above indicates that LEFM is applicable to material when original
crack is along fiber direction. For wood, most of the crack and defects formed
during growth and processing are along fiber direction, while the resistance of wood
to the propagation of crack along fiber direction is minimum. In experiments, TR
crack and RT crack are similar to radial shake and ring shake of wood, respectively,
so LEFM is applicable to wood with a crack propagating along fiber direction,
namely grain. There will be significant meaning for the quality assessment, safety
design, and optimization of processing method of wood to study and measure the
fracture toughness along grain of wood.
Now there are test standards at home and abroad to measure KIC of materials, such
as ASTM-E399 [9] in America, BS5447 [10] in Britain and GB4161-84 [11] in
China. And there are various samples, such as compact tension (CT) sample, single
edge notched bending (SEB) sample, tensile (TS) sample and double cantilever
beam (DCB) sample, and so on (Fig. 3.8), and the improved forms of above
samples.
In theory, KIC is the natural property of material and has nothing to do with load
mode, sample form, and size of sample. From the view of measurement, the CT
method is simple and easy to perform on specimen with small size, which is
particularly suitable for wood with small diameter class; SEB sample often used to
measure transverse fracture toughness needs large diameter timber to make crack
along grain; TS sample has high requirements for tensile grips; DCB sample is
adopted by compliance method by which the energy release rate can be measured,
while nine elastic coefficients need to be measured to calculate KIC, so the workload
is heavy. Therefore, CT sample is always applied, and the validity of result can be
improved by increasing sample number.
There are three types of load–displacement (F-d) curves in the test of KIC, as
seen in Fig. 3.9. To a certain sample, it needs to measure crack size and geometrical
dimensions of sample before test and determine the critical load Fcr after test to
calculate KIC. It should follow the requirements below to determine Fcr. For brittle
sample or sample with large size, it will fracture unstably once crack initiates with
no obvious subcritical propagation stage, then the maximum load Fmax is the critical
load Fcr. In general, crack would propagate slowly before sample fractures, and
42 3 Fracture of Wood Along Grain
H
a
F B
a
e W
CT sample TS sample
F
e a W
F
DCB sample
B
L/2 L/2
SEB sample
there is no visible sign unstable of propagation, thus the maximum load is not the
critical load. For metal material without visible yield stage, the rule is that the stress
corresponding to 0.2% of strain is regarded as nominal yield limit. And in many
literature and national standard, the critical load is the corresponding load when the
propagated length Da equals to 2% of the initial crack length.
3.5 The Stress Intensity Factor … 43
However, in practical test, the curve obtained is F-d curve but not F-Da curve, so
the corresponding load (Da/a = 2%) should be determined on F-d curve. In theo-
retical analysis and approximation [12], it is found that the slope of secant line of
the corresponding load on the F-d curve decreases 5% compared with the slope of
the line segment of the F-d curve before the propagation begins, so graphing
method can be used to determine the critical load on the F-d curve.
The third type of curve will be obtained if sample has enough thickness. When
loaded, sample is in plane strain state except the surfaces of sample, and there is no
propagation in crack front zone. When load reaches its maximum, brittle fracture
will happen suddenly, and the maximum load is Fcr. For sample with smaller
thickness, the second type of curve can be obtained. There is an obvious “burst”
platform on the curve during loaded, the reason is that the central layer of sample is
in plane strain state, where crack propagates first, while there is no crack propa-
gation on surfaces of sample that are in plane stress state, so the propagation of
crack in central layer will be dragged by surfaces. During test, when the “burst”
platform appears, clear sound of crack can be heard, so the corresponding load is
Fcr.
In the two cases above, there is no need to record the F-d curve during test, but
to obtain the maximum load or load at the “burst” platform, while sample with large
dimensions is required, and correspondingly test machine with wide range.
To overcome the difficulties above, sample with minimal thickness is always
adopted, and the first type of curve can be obtained. In this case, the maximum load
cannot be used to calculate fracture toughness, because crack propagates slowly
before load reaches its maximum value, which is imperceptible. Thus, particular
engineering assumption is applied to determine Fcr, as seen in Fig. 3.9, and it is that
the load at the intersection point of the F-d curve and a secant line with a slope 5%
less than that of F-d curve is the Fcr.
To verify that the fracture toughness along grain is a basic attribute of wood and
has nothing to do with shape and dimensions of sample, experiments are performed
on Pinus sylvestris (Pinus sylvestris var. mongolica Litv.) with TL crack. CT sample
and its improved sample (WOL sample) are adopted considering wood diameter,
the manufacture of sample and load way, and the influence of sample size on KIC is
44 3 Fracture of Wood Along Grain
also studied. Then KIC values of CT samples with TL crack and GIC values of DCB
samples of four kinds of woods (China fir, spruce, poplar, and Castanopsis hystrix)
are tested, and the tested KIC values are compared with the critical stress intensity
factor calculated from GIC.
3.5.2 The K TL
IC of CT Samples with Different Thickness
F B
a
e W
3.5 The Stress Intensity Factor … 45
Fcr
a
KIC ¼ pffiffiffiffiffi f ð3:40Þ
B W W
where Fcr is the critical load, a/W is the dimensionless crack length, and f(a/W) is
the function of sample geometry given by the formula below:
h a
2
3
4 i
a 2þ a
W 0:886 þ 4:64 W 13:32 Wa þ 14:72 Wa 5:6 Wa
f ¼
3=2 :
W 1 Wa
It can be seen that the F-d curve of air-dried CT sample is kept straight basically
from the initiation of crack to the failure of sample under tensile force. And all the
F-d curves are the third type according to GB4161-84 [11], so Fmax was the critical
load Fcr. Substituting sample dimensions and Fmax into Eq. 3.40, the fracture
toughness of sample can be obtained.
The sizes and KIC values of CT samples are shown in Table 3.1, and statistical
description results and variance analysis results are shown in Tables 3.2 and 3.3. It
can be seen that the coefficients of variation (CV) of the KIC values are less than
20%, and the accuracy indexes are all less than 5%, which indicates that the average
KIC value of CT samples is effective, and it is 9.86 N mm−3/2. The results
demonstrate that the thickness of sample nearly has no influence on the fracture
toughness of wood, which is in accordance with the conclusions of Barrent [13] and
Boatright and Garrentt [14].
3.5.3 The K TL
IC of WOL Samples with Different Crack
Length
Fcr
a
KIC ¼ pffiffiffiffiffi f ð3:41Þ
B W W
a
a
1=2 a
3=2
5=2
7=2
9=2
where f W ¼ 30:96 W 195:8 W þ 730:6 Wa 1186:3 Wa þ 754:6 Wa .
3.5 The Stress Intensity Factor … 47
The sizes and KIC values of WOL samples are shown in Table 3.4, and the
average value of fracture toughness is 10.09 Nmm−3/2 that is close to the average
fracture toughness of CT samples. The statistical description results and variance
analysis results are shown in Tables 3.5 and 3.6. It can be seen that the CV of the
results of WOL sample is larger than that of CT sample, while the variance analysis
results in Table 3.6 show that the sizes and crack length (0.3 < a/W < 0.8) of CT
sample have no obvious influence on fracture toughness.
Table 3.4 Sizes of WOL samples and test value of Pinus sylvestris
No. W (mm) a (mm) a/W B (mm) H (mm) Fmax (N) f(a/W) KIC (Nmm−3/2)
W1-1 56 16.8 0.3 21.93 54.56 291.44 6.603 11.73
W1-2 56 16.8 0.3 21.96 54.56 290.48 6.603 11.67
W2-1 56 22.4 0.4 21.93 54.56 251.28 8.178 12.52
W2-2 56 22.4 0.4 21.94 54.56 266.36 8.178 13.27
W3-1 56 28 0.5 21.91 54.56 158.80 10.313 9.99
W3-2 56 28 0.5 22.01 54.56 174.88 10.313 10.95
W4-1 56 33.6 0.6 21.91 54.56 122.60 13.983 10.46
W4-2 56 33.6 0.6 21.94 54.56 97.48 13.983 8.30
W5-1 56 39.2 0.7 21.96 54.56 68.36 21.898 9.11
W5-2 56 39.2 0.7 21.91 54.56 48.24 21.898 6.44
W6-1 56 44.8 0.8 21.98 54.56 40.20 38.999 9.53
W6-2 56 44.8 0.8 21.93 54.56 30.16 38.999 7.17
DCB sample is always used to test pure Mode I critical strain energy release rate
GIC of wood. GIC can be calculated by the compliance of crack measured by the
symmetrical bending test method. Triboulot et al. [15] measured the fracture
toughness of wood with TL crack by DCB method, and the results are compared
with the solution of finite element analysis (FEA). The results tally well with each
other.
DCB samples of China fir, spruce, poplar, and C. hystrix are prepared, and a is the
length of initial crack, as seen in Fig. 3.11. There are 6–9 samples for each wood.
China fir samples are divided into two groups: one with TL crack and the other with
RL crack, and only TL crack is prepared on the samples of spruce, poplar, and C.
hystrix. The sizes of sample and crack are shown in Table 3.7. As thickness of
e a W
F
Table 3.7 Sizes of DCB sample and test value of Cunninghamia lanceolata
No. W (mm) H (mm) B (mm) a (mm) a/W Fmax (N) C (mm/N) GIC (J/m2)
1 260 84 21.84 49.5 0.190 366.284 1.692 87.37
2 260 84 21.84 59.5 0.229 341.268 2.053 89.27
3 260 84 21.84 71 0.273 313.612 2.598 90.93
4 260 84 21.84 82.5 0.317 302.066 3.014 101.75
5 260 84 21.84 95.5 0.367 273.310 3.641 102.96
6 260 84 21.84 107.5 0.413 247.798 4.516 102.93
7 260 84 21.84 120 0.462 221.682 5.611 100.99
8 260 84 21.84 137 0.527 188.144 7.225 95.97
9 260 84 21.84 155 0.596 156.750 9.727 89.33
3.6 The Fracture Toughness … 49
samples is larger than 20 mm, the fracture can be treated as plane strain problem.
The moisture content (MC) of specimens is about 12%, the temperature is about
20 °C and relative humidity is about 60% in laboratory.
0.25
a8
0.2 a9
0.15
0.1
0.05
0
0 0.5 1 1.5 2
(mm)
50 3 Fracture of Wood Along Grain
C ¼ qemða=WÞ ð3:42Þ
Fmax
2
@C
GIC ¼ ð3:43Þ
2BW @ða=WÞ
where q and m are the fitting coefficients of the compliance curve. And the coef-
ficients of determination (R2) of compliance and crack length are all above 0.98.
Figure 3.13 Relationship between corresponding compliance and a/W of
Cunninghamia lanceolata DCB sample
Mode I fracture toughness GIC can be calculated by Eq. (3.43). The arithmetic
mean values of toughness of one specimen x and one group of specimens X are
calculated, respectively, as
1X k
x¼ ðGIC Þi ð3:44Þ
k i¼1
1X n
X¼ xj ð3:45Þ
n j¼1
where k is the amount of measured points of one specimen and n is the amount of
the specimens in one group.
Figure 3.14 shows the distribution relationship between the fracture toughness
and crack length of DCB samples of China fir. The average energy release rate of
Cunninghamia lanceolata
DCB sample
a/W
3.6 The Fracture Toughness … 51
GIC (J/m 2)
104 J/m 2
DCB sampler
TL
a ( mm)
RL crack
GIC (J/m2)
101 J/m 2
RL
a ( mm )
China fir samples with TL crack and RL crack are 104.17 J/m2 (SD = 13.35 J/m2)
and 101.49 J/m2 (SD = 22.07 J/m2) respectively, and there is no significant
difference.
The fracture toughness of China fir samples with TL crack and RL crack are
about the same because of the fine wood ray of China fir. However, for many
woods, there may be difference between the results of samples with TL crack and
RL crack caused by the size and content of wood ray and the construction of wood.
For the opening crack of homogeneous linear anisotropic material, the relationship
expression between strain energy release rate and fracture toughness was given by
Sih, Paris, and Irwin [17], called S.P.I. relationship, as follows:
Substituting Sij by engineering elastic parameters, and neglecting the small value
term with Poisson’s ratio, the equation above can be written as
52 3 Fracture of Wood Along Grain
KI2
GI ¼ ðIsotropic materialÞ ð3:48Þ
E
KI2
GI ¼ ðAnisotropic materialÞ ð3:49Þ
E
1 ð2EL ET Þ1=2
E ¼ ¼
1=2
1=2 1=2 ð3:50Þ
S EL EL
ET þ 2GLT
Wood is anisotropic porous biomaterial, but when cutting a small rectangle piece
at a certain distance from the pith and making one of its symmetry plane perpen-
dicular to the growth rings, then it can be seen as an orthotropic body in macro-
scopic scale [18]. Triboulot et al. [15] had converted the measured energy release
rate to stress intensity factor according to Eq. (3.46) or Eq. (3.49) to demonstrate
that it was feasible to treat wood as orthotropic and elastic body, and fracture
mechanics was applicable to wood. However, although the converted values of
stress intensity factor of two type samples were in accordance with each other, no
experiment was performed to verify the S.P.I. relationship, so it is not stated that
whether the KIC values of one wood obtained by different test methods are in
agreement.
Ashby et al. [19] considered that when crack peeled or layered in opening mode
along grain, the fracture process was the same as the peeling of bonded points. As
the composition and structure of cell wall had almost no difference in different kind
of woods, the energy absorbed per unit area was approximately a constant when
peeled (under given moisture content) for all woods. In fracture process, the fracture
energy was provided by elastic energy released by the surrounding tissue and the
work done by the applied load. Therefore, by the principle of energy balance, the
energy release rate of Mode I delamination could be sufficiently approximated as
follows (Called Ashby relationship):
KIC
2
GIC ¼ ð3:51Þ
ER
3.6 The Fracture Toughness … 53
where ER was Young’s modulus in radial direction, and KIC was the stress intensity
factor of delaminating fracture.
In the same experimental environment, there are two main factors that affect the
equivalent relation between KIC and GIC, and they are the construction of wood and
test method, which both have relation with the occurrence of bridge. KIC is a
physical quantity in instantaneous state when wood fractures along grain unstably,
and the measurement is based on pre-crack, so there is nearly no fiber bridge.
However, GIC is an average physical quantity during the multiple continuous
propagation process of crack, thus fiber bridge in different levels would occur in
different woods. Generally, for softwood with straight grain, there is nearly no fiber
bridge during the propagation process of crack in DCB sample, while for hardwood
with interlocked grain, lots of fiber bridges will occur during crack propagation
process in DCB sample, which would increase the energy consumption, and con-
sequently, the equivalent relation between KIC and GIC is affected.
To illustrate this point, the KIC of the CT samples of China fir, spruce, poplar,
and C. hystrix are measured according to GB4161-84 [11]. And the GIC of DCB
samples of the four kinds of woods are measured by compliance method in two
ways, namely multi-sample with single-point method and single sample with
multipoint method. For multi-sample with single-point method, the pre-crack of
DCB sample satisfies a/W = 0.2–0.8, 6–7 samples in each group, and only one
point will be tested; for single sample with multipoint method, that is, for one single
DCB sample, multiple points will be tested during the multiple continuous prop-
agation process of crack. The process of multi-sample with single-point method is
the same as that of CT sample, during which there is nearly no fiber bridge, while,
fiber bridges may occur during the multiple continuous propagation process of
crack by single sample with multipoint method.
China fir possesses the characters of straight grain, uniform structure, little and
very thin wood ray, so no bridge occurred in DCB sample during the continuing
propagation of crack. Thus, there is no significant difference between the results
obtained by multi-sample with single-point method and single sample with multi-
point method. The relationship between the compliance of F-d curves obtained by
multi-sample with single-point method and the corresponding a/W is shown in
Fig. 3.15. Then, the energy release rate GIC can be calculated according to
Eq. (3.43), and the results are shown in Table 3.8. The average energy release rate
of China fir DCB samples is 99.03 J/m2 obtained by multi-sample with single-point
method, which is very close to the average value 101.49 J/m2 obtained by single
sample with multipoint method.
The statistical description results of energy release rate of the four kinds of
woods by multi-sample with single-point method and single sample with multipoint
method are shown in Table 3.9. It can be seen that there are significant differences
between the results obtained by the two methods for spruce, poplar, and C. hystrix.
As seen in Fig. 3.16, for poplar, fiber bridges occurred during the continuing
propagation of crack leading to the high fracture toughness.
The elastic coefficients of the four kinds of woods measured by electrometric
method are shown in Table 3.10, then the equivalent modulus E* can be calculated.
54 3 Fracture of Wood Along Grain
Compliance (mm/N)
a/W
Fig. 3.15 Relationship between corresponding compliance and a/W of Cunninghamia lanceolata
DCB sample (single sample with multipoint method)
Table 3.8 Sizes of DCB samples and test value of Cunninghamia lanceolata (single sample with
multipoint method)
No. W (mm) H (mm) B (mm) a (mm) a/W Fmax (N) C (mm/N) GIC (J/m2)
1 200 80 20.01 40 0.2 329.76 2.352 114.73
2 200 80 20.12 60 0.3 270.30 3.216 113.39
3 200 80 19.98 80 0.4 213.00 4.513 103.57
4 200 80 20.07 100 0.5 179.60 6.359 108.32
5 200 80 20.11 120 0.6 140.54 10.225 97.56
6 200 80 20.09 140 0.7 110.40 15.544 88.56
7 200 80 20.15 160 0.8 90.30 22.967 87.15
Table 3.9 Statistical analysis of GIC between many measure and single measure for DCB samples
Species Single sample with multipoint method Multi-sample with single-point
method
Crack Sample Test GIC SD Crack Sample GIC SD
point (N/m or (N/m) (N/m or (N/m)
J/m2) J/m2)
Cunninghamia TL 7 50 104.17 13.35 TL
lanceolata
Cunninghamia RL 6 48 101.49 22.7 RL 6 99.03 26.7
lanceolata
Picea asperata TL 9 69 306.07 63.33 TL 6 163.38 57.9
Populus sp TL 7 42 582.92 151.03 TL 6 300.3 51.13
Castanopsis TL 6 51 1222.09 333.39 TL 6 392.7 134.8
hystrix
3.6 The Fracture Toughness … 55
Populus sp.
The measured stress intensity factor by CT samples, energy release rate by DCB
samples, and the corresponding stress intensity factor converted by S.P.I. rela-
tionship or Ashby relationship are shown in Table 3.11. Conclusions can be drawn
that:
(1) The stress intensity factor of China fir CT samples is 7.98 Nmm−3/2. The stress
intensity factor converted by S.P.I. relationship from the energy release rate of
China fir DCB samples obtained by two methods are 8.09 and 8.19 Nmm−3/2,
respectively. The stress intensity factors converted by Ashby relationship from
the energy release rate of China fir DCB samples obtained by the two methods
are 7.18 and 7.26 Nmm−3/2, respectively. It can be seen that as there is no fiber
bridge during the propagation process of crack in China fir, there is fine
equivalent relation between stress intensity factor converted by S.P.I. rela-
tionship or Ashby relationship and energy release rate of China fir.
(2) The stress intensity factors of CT samples of spruce and poplar have good
equivalent relation with the energy release rate of DCB samples obtained by
multi-sample with single-point method, but no equivalent relation with the
energy release rate is obtained by single sample with multipoint method. And
the stress intensity factor converted from the energy release rate obtained by
single sample with multipoint method is larger than the measured value and the
value converted from the energy release rate is obtained by multi-sample with
single-point method, because fiber bridges occurred during the continuing
propagation of crack leading to the high fracture toughness.
(3) For C. hystrix, there is no equivalent relation between stress intensity factor and
energy release rate obtained by the two methods, because the orthotropic
property is affected by the interlocked grain and wide ray of C. hystrix. Fiber
bridges occurred, and the fracture surface is rough and uneven that much
rougher than that of other woods. And the relationship between the feature of
fracture surface and fracture toughness along grain will be discussed in Chap. 6.
Species CT GIC (N/m or J/m2) S.P.I. KIC = (GIC E*)1/2 (N/mm3/2) Ashby KIC = (GIC ER)1/2 (N/mm3/2)
sample relation relation
Crack KIC (N/ DCB DCB E* = 1/ DCB DCB E* = ER DCB DCB
mm3/2) (Single-point (Multipoint S*(N/ (Single-point (Multipoint (N/mm2) (Single-point (Multipoint
method) method) mm2) method) method) method) method)
Cunninghamia RL 7.98 99.03 101.49 662 8.09 8.19 520 7.18 7.26
lanceolata
Picea asperata TL 13.83 163.38 306.07 1229 14.17 19.39 1204 14.03 19.20
Populus sp TL 20.59 300.30 582.92 1544 21.53 30.00 1205 19.02 26.50
Castanopsis TL 21.64 392.70 1222.09 2815 33.25 59.66 3873 39.00 68.80
hystrix
For Cunninghamia lanceolata, GIC(TL) GIC(RL) of DCB sample by multipoint method
57
58 3 Fracture of Wood Along Grain
elastic coefficients, which will bring errors to the description and analysis on wood
mechanical properties.
3.7.1 Introduction
Mode I fracture along grain is the most common and dangerous failure way of
wood, so it is the focus of wood fracture mechanics. And there are various samples,
such as compact tension (CT) sample, single edge notched bending (SEB) sample,
tensile (TS) sample and double cantilever beam (DCB) sample, and the improved
forms of above samples. There are also many researches on the Mode II fracture
toughness of wood based on end notched flexure (ENF) sample. However, the
research on the Mode III fracture of wood is rarely reported. Ehart et al. [20]
performed the Mode III fracture test of wood sample by a twist machine, but the
complex experimental facility hindered the spread of the method.
Donaldson [21, 22] characterized the Mode III interlaminar fracture of com-
posites successively by using split cantilever beam (SCB) specimens and proposed
that torsion issue could be dealt well by applying the split cantilever beam with
initial crack bonded between two parallel aluminum bars. Gary Becht [23] tested
the Mode III interlaminar fracture toughness of continuous-fiber-reinforced com-
posite materials by using crack rail shear (CRS) specimens. Shaw Ming Lee [24]
measured the Mode III interlaminar fracture toughness of several carbon fiber/
epoxy composites by edge crack torsion (ECT) method. Xiao Jun [25] modified the
Donaldson SCB fixture device by using guide limited post to simplify the Mode III
interlaminar fracture toughness test method of laminates. Mohammad Reza
Khoshravan and Mohsen Moslemi [26] calculated and analyzed the Mode III
interlaminar fracture behavior of woven fabric-reinforced glass/epoxy composite
laminates by compliance method and virtual crack closure technique. Here, the
Mode III interlaminar fracture toughness of spruce (Picea jezoensis) is tested by
improved designed SCB fixture device with the compliance method.
Air-dried spruce (P. jezoensis) is selected to make sample with straight grain, and
the size of sample is 240 (L) 20 (T) 20 (R) (mm). To simulate a naturally
sharp crack, the pre-crack is cleaved along middle level parallel to grain by knife,
and the length of initial crack is controlled about 40 mm from the center of loading
points.
3.7 Mode III Fracture Property of Wood Along Grain 59
Double cantilever inversion symmetry bending load method was used to test
Mode III interlaminar fracture (Fig. 3.17). Sample was placed in the fixture device,
then loaded by a computer-controlled testing machine, and the load–load point
deflection (F-d) curve was recorded by computer automatically during the test
(Fig. 3.18). Single specimen with multipoint method was applied, that is, on one
specimen, repetitive loading-unloading-loading was performed and the loading
speed was 1 mm/min. Temperature was 25 °C and the humidity was 65–70% in
laboratory.
At the initial of loading, the F-d curve kept straight until the crack began, then
the slope of F-d curve increased suddenly, meanwhile, the crack propagated slowly
along the grain. After the F-d curve rose nonlinearly to the maximum load, it began
to decline. At this moment, stopped loading to save the F-d curve and mark the
crack tip. The critical load Fcr was the load when the crack began, namely the point
Rollers
Crack
a5 = 87 mm, a6 = 99 mm,
a7 = 105 mm, a8 = 117 mm,
a9 = 123 mm,
a10 = 129 mm
(mm)
60 3 Fracture of Wood Along Grain
the slope of the F-d curve increased. After the test is over, length between the crack
tip marked each time and the center of the load point was measured. The average
value of the crack length on to sides of sample was taken as the final crack length.
As seen in Fig. 3.18, it was the test curve of one spruce sample. It can be seen
that the slope of the F-d curve straight portion is inverse to crack length and the
reciprocal of the slope is the compliance Ci corresponding to different crack lengths
ai. And by curve fitting, the exponential relationship between the compliance and
crack length is obtained
C ¼ qema ð3:52Þ
where q and m are the fitting coefficients of the compliance curves and the deter-
mination coefficients (R2) were all above 0.95. As shown in Fig. 3.19, the fitting
curve of compliance C and crack length a of one spruce sample was shown.
So the propagation resistance of Mode III interlaminar crack along grain, namely
the interlaminar fracture toughness can be calculated by the following formula:
@U @ðFcr
2
C=2Þ Fcr
2
@C
GTL
IIIC ¼ ¼ ¼ ð3:53Þ
@A @ðBaÞ 2B @a
a (mm)
3.8 Conclusions 61
GIIIC (kJ/m2)
1.05
a (mm)
20mm
3.8 Conclusions
In this chapter, the theory of LEFM and the special application of LEFM on wood
are introduced. Then different samples and methods are applied to measure the
Mode I fracture toughness of wood along grain, and the results show that the
fracture toughness of wood along grain is the basic attribute of wood and has
nothing to do with test method and the geometrical shape and size of crack. The
research indicates that LEFM based on isotropic body is applicable to the crack
propagation of wood along grain.
Fractal theory is applied to study the fractal feature of the fracture surfaces of five
woods along grain. The relationship between fractal dimension of fracture surface
and fracture toughness of wood is established. The results show that fracture
toughness parallel to the grains of various woods is different because of their
textural diversity and such differences are also shown on the morphology of fracture
surfaces. Furthermore, there is evident and direct proportional relation between the
fractal dimension and fracture toughness along grains.
Double cantilever inversion symmetry bending load method is applied to mea-
sure the Mode III fracture toughness of spruce based on energy theory. The results
have shown that the average Mode III interlaminar fracture toughness 1.05 kJ/m2
can be seen as a basic characteristic of spruce that represents the capacity of spruce
to resist the propagation of Mode III crack.
62 3 Fracture of Wood Along Grain
References
1. Snedden IN (1946) The distribution of stress in the neighborhood of a crack in an elastic solid.
Proc R Soc Lond A 187:229–260
2. Irwin GR (1957) Analysis of stresses and strain near the end of a crack traversing a plate.
J Appl Mech 24:361–364
3. Westergaard HM (1939) Bearing pressures and cracks. J Appl Mech 6:49–53
4. Sih GC, Prais PC, Irwin GR (1965) On cracks in rectilinearly anisotropic bodies. J Fract Mech
1(3):189–203
5. Wu EM (1967) Application of fracture mechanics to anisotropic plates. J Appl Mech 34:967–
974
6. Wu EM (1968) Fracture Mechanics of anisotropic plate. In: Tsai SW (ed) Composite material
workshop. Technomic Publishing Company, p 23
7. Kanninen MF, Popelar CH (1985) Advanced fracture mechanics. Oxford University
8. Parhizgar Shahrokh, Zachary Loren W, Sun CT (1982) Application of the principles of linear
fracture mechanics to the composite materials. Int J Fract 20:3–15
9. American Society of Testing Materials (1981) Standard methods for plain-strain fracture
toughness of metallic materials. ASTM E399-74
10. British Standards Institution (1997) Methods of test for plane strain fracture toughness.
BS5447
11. National Standard (1984) Metallic materials-determination of plane-strain fracture toughness.
GB4161-84
12. Cui Z (1981) Theory and test method of fracture toughness. Shanghai Scientific and Technical
Publishers, Shanghai, pp 15–34
13. Barrent JD (1976) Effect of crack-front width on fracture toughness of Doouglas-fir. Eng Fract
Mech 8(4):711–717
14. Boatright SWJ, Garrentt GG (1983) The effect of microstructure and stress state on the
fracture behaviour of wood. J Mater Sci 18:2181–2199
15. Triboulot P, Jodin P, Pluvinage G (1984) Validity of fracture mechanics concept applied to
wood by finit element calculation. Wood Sci Technol 18(6):448–459
16. Hodgkinson JM (2000) Mechanical testing of advanced fiber composites. Woodhead
Publishing Limited, Niklewicz
17. Sih GC, Prais PC, Irwin GR (1965) On cracks in rectilinearly anisotropic bodies. J Fract Mech
1(3):189–203
18. Coleman (1991) Wood and wood process principle. China Forestry Publishing House,
Beijing, pp 227–235
19. Ashby MF, Easterling KE, Harrysson R et al (1985) The fracture and toughness of woods.
Proc R Soc Lond A 398:261–280
20. Ehart RJA, Stanzl-Tschegg SE, Tschegg EK (1999) Mode III fracture energy of wood
composites in comparison to solid wood. Wood Sci Technol 33:391–405
21. Donaldson SL (1988) Mode III interlaminar fracture characterization of composite material.
Compos Sci Technol 32:225–249
22. Donaldson SL, Mall Shankar, Lingg Cynthia (1991) The split cantilever beam test for
characterizing Mode III interlaminar fracture toughness. J Compos Tech Res 13(1):41–47
23. Becht Gary, Gillespie John W (1988) Design and analysis of the crack rail shear specimen for
Mode III interlaminar fracture. Compos Sci Technol 31:143–157
24. Lee SM (1993) An edge crack torsion method for Mode III delamination fracture testing.
J Compos Tech Res 15(3):193–201
25. Jun X (1996) Experimental study on Mode III fracture toughness of multi-directional interface
of composite materials. J Nanjing Univ Aeronaut Astronaut 28(2):267–270
26. Khoshravan MR, Moslemi M (2014) Investigation on Mode III interlaminar fracture of glass/
epoxy laminates using a modified split cantilever beam test. Eng Fract Mech 127:267-279
Chapter 4
Transverse Fracture of Wood
Abstract In this chapter, first, the stress field at crack tip of wood component with
crack perpendicular to grain is analyzed by finite element method to explore the
reason why the crack always initiates along grain. The critical stress intensity factor
of four kinds of woods is measured with acoustic emission (AE) technique, and the
results show that Mode I crack in wood component initiates along fiber, and then
ductile fracture happens along the cross section of the component. The whole
propagation process of crack involves three stages: linear stage, stable stage, and
unstable stage. Then the influence of crack perpendicular to grain on the strength of
wood component is studied. Bending test, impact test, and tensile test as experi-
mental verification are performed on standard samples and samples with crack
perpendicular to grain of two softwoods and two hardwoods. The results show that
wood possessed fine capacity to resist transverse fracture because of its specific
organization structure, so it would not fail under low work stress in spite of the
stress singularity at crack tip. So for the safety assessment and strength design of
wood component with crack perpendicular to grain, net stress criterion should be
adopted, namely the strength of the clear part of component apart from the size of
crack. At last, the analytical formula of energy release rate of the Mode I inter-
laminar fracture of wood and the bending delamination damage of plywood are
derived.
4.1 Introduction
on the crack body [1–4]. However, for the transverse fracture of wood, LEFM is not
applicable anymore because the crack does not propagate along its original direc-
tion. Jeronimidis [5] proposed that fracture energy or fracture work, namely the
ratio of energy absorbed during the failure process and double of cross-sectional
area, could be used to characterize the fracture property of wood component.
Although fracture energy could be used to assess the toughness of different
materials and could help to understand the fracture process, it could not be used as
design parameter directly.
Wood is anisotropic and inhomogeneous material, and for air-dried wood, the
strain–stress curve possessed linear characteristic, so the mechanical behavior of
wood fitted into the linear elastic behavior and wood could be approximately
assumed as an orthotropic material [6]. While wood possesses many characteristics
which are obviously different with those of other orthotropic materials, the major
characteristic of wood is the high anisotropy among the three principal directions
from the perspective of the composition and structure of wood. As known, most
wood cells arrange along axial direction and only a few ray cells arrange along
radial direction. What’s more, because of the great difference in chemical bond
energy between valence bonds of cellulose chain molecules in the axial directions
and in radial direction, the transverse tensile strength is only 1/50–1/24 of the
longitudinal tensile strength [7, 8]. Wood species, density, and load direction could
influence the fracture of wood. Meanwhile, the structure of wood was also an
important factor that determined the fracture of wood [9]. Bodner et al. [10, 11]
have studied the initiation and propagation of fracture behavior of wood clear
specimens by tensile along grain test and bending test. Bodner et al. [12] also
studied the bending and tensile fracture behavior of wood specimen with oblique
grain. Grekin [13] studied the perpendicular-to-grain tensile fracture behavior of
Scots pine and obtained that the TR crack propagation was in tortuous, stepwise
crack path. Galicki and Czech [14] studied the failure of pine with grain in different
angles to the longitudinal direction of specimen under uniaxial tensile stress. In
recent years, finite element technique is also extensively used to simulate the
fracture behaviors of composites as well as wood [15–18].
During the bending test of wood specimen with crack perpendicular to grain, we
found that crack propagated along grain to a certain length then stopped, at this
moment, specimen had the ability to bear load as well. It was found that the
maximum failure load was several times the critical load when crack initiated. It
was obvious that if the critical load when crack initiated was taken as the criterion
for the safety design of wood component, there must be a waste of strength. Thus, a
thought was inspired that for the safety design of wood component with crack
perpendicular to grain, only the strength of the clear part without crack needed to be
considered instead of considering the angle of fracture mechanics. To verify the
thought above, numerical simulation and experimental verification were performed.
ABAQUS finite element software was used to simulate the crack tip stress field of
Picea asperata SEB (Single edge notched bending) specimen with crack perpen-
dicular to grain to explore why the crack always initiated along grain. Then stan-
dard specimens and specimens with crack perpendicular to grain of two softwoods
4.1 Introduction 65
and two hardwoods were chosen as samples to perform bending test, impact test,
and tensile test. And a criterion applicable for the strength design and safety
assessment of wood component with crack perpendicular to grain was proposed.
x
r
w
y a
Fig. 4.3 The nephogram of stress (ry) perpendicular to crack around crack tip
4.2 Analysis on Stress Field at Crack Tip 67
Fig. 4.4 The nephogram of stress (rx) parallel to crack around crack tip
There are 113 nodes right above crack tip (the plane Y = 0) in total, and after the
nodes around load point and the nodes with negative stress in the bending com-
pression zone are neglected, stress analysis is performed on 70 nodes taken from the
bottom up, as seen in Table 4.2.
For the 70 nodes, the variation tendencies of tensile stress perpendicular to crack
(rY) and tensile stress parallel to crack (rX) are shown in Fig. 4.7.
It can be seen that there are both rY and rX within a large zone in front of the
crack tip, and rY/rX is almost a constant in the range of 4–5 within a large zone
around the crack tip apart from singular points around the crack tip (Table 4.2). The
stresses declined quickly with the increase of Distance from crack-tip and after
reached a certain value, stresses changed gently.
The fracture of wood always begins at the micro-area in front of a crack tip. The
micro-area called the fracture process region is not only the high-stress area but also
where the microstructure has decisive influence. The stress analysis has shown that
there are both rY and rX in the fracture process region, and rY/rX is almost a
constant ranging from 4 to 5 (The values are different for different wood). Thus,
4.3 The Cracking Direction of Transverse Crack 69
Table 4.2 The stress of 70 nodes taken from the bottom up right ahead of crack tip
Node ry rx ry/rx Node ry rx ry/rx
(Mpa) (Mpa) (Mpa) (Mpa)
1 95.78 13.94 6.87 36 3.48 0.75 4.62
2 21.31 4.87 4.37 37 3.38 0.73 4.63
3 19.46 4.62 4.21 38 3.28 0.70 4.65
4 16.82 3.78 4.45 39 3.17 0.68 4.66
5 14.56 3.34 4.36 40 3.07 0.66 4.68
6 13.14 2.96 4.43 41 2.98 0.63 4.70
7 11.95 2.69 4.44 42 2.89 0.61 4.72
8 10.97 2.47 4.43 43 2.80 0.59 4.74
9 10.21 2.30 4.45 44 2.71 0.57 4.77
10 9.57 2.15 4.46 45 2.62 0.55 4.79
11 9.00 2.02 4.46 46 2.54 0.53 4.82
12 8.50 1.90 4.47 47 2.46 0.51 4.85
13 8.06 1.80 4.47 48 2.38 0.49 4.89
14 7.67 1.72 4.47 49 2.30 0.47 4.93
15 7.31 1.64 4.47 50 2.22 0.45 4.97
16 6.99 1.56 4.48 51 2.15 0.43 5.01
17 6.69 1.49 4.48 52 2.07 0.41 5.07
18 6.42 1.43 4.49 53 2.00 0.39 5.12
19 6.17 1.37 4.49 54 1.93 0.37 5.19
20 5.93 1.32 4.49 55 1.86 0.35 5.26
21 5.71 1.27 4.50 56 1.79 0.33 5.34
22 5.51 1.22 4.50 57 1.72 0.32 5.43
23 5.31 1.18 4.51 58 1.65 0.30 5.54
24 5.13 1.14 4.51 59 1.58 0.28 5.66
25 4.96 1.10 4.52 60 1.52 0.26 5.79
26 4.79 1.06 4.52 61 1.45 0.24 5.96
27 4.63 1.02 4.53 62 1.37 0.23 6.14
28 4.48 0.99 4.54 63 1.32 0.21 6.37
29 4.34 0.96 4.54 64 1.26 0.19 6.64
30 4.20 0.92 4.55 65 1.20 0.17 6.98
31 4.07 0.89 4.56 66 1.13 0.15 7.39
32 3.95 0.86 4.57 67 1.07 0.14 7.94
33 3.82 0.84 4.58 68 1.01 0.12 8.65
34 3.71 0.81 4.59 69 0.95 0.10 9.66
35 3.59 0.78 4.60 70 0.89 0.08 11.15
70 4 Transverse Fracture of Wood
160
20
120 15
80 10
40 σY 5 σX
σX
0 0
0 1 σY 2 3 4 5 6 0 1 2 3 4 5 6
Distance from crack-tip (mm) Distance from crack-tip (mm)
Fig. 4.7 The variation tendency of rY and rX along the crack initial direction within 5 mm from
the crack tip
when the tensile stress region in front of crack tip developed to the interface, if the
interfacial strength of the wood (namely tensile strength perpendicular to grain) was
higher than 1/5 of longitudinal strength, the interface would not crack. Then the
crack would propagate through the interface. At this moment, the mechanical
behavior of wood is just like that of common brittle materials, and brittle failure
would happen. On the contrary, if the transverse strength was lower than 1/5 of
longitudinal strength, the interface would be pulled open by the tensile stress rX
leading to the formation of a new crack perpendicular to the original crack. The
passivated crack could eliminate the stress concentration and the propagation of
original crack was prevented, then the unstable fracture was avoided. In fact, the
tensile strength perpendicular to grain is only about 1/40–1/20 of the tensile
strength parallel to grain (Here, the rT/rL of the P. asperata is only 1/25). As a
result, the crack in timber component always initiates along the grain.
Then for the timber beams with cracks in different angle to grain, when subjected
to transverse bending force, what are the directions that cracks propagate along?
The question will be discussed in Chap. 5. The next question is if the load when
crack initiates along the grain is treated as the critical load to calculate the fracture
toughness, whether the fracture toughness is still the basic attribution of wood. This
question will be discussed in next section.
The SEB sample (Fig. 4.1) is adopted to measure the critical stress intensity
factor (KIC) of wood transverse fracture according to national standard “Metallic
materials—Determination of plane-strain fracture toughness” (GB4161-84) [23].
As the accuracy of KIC depends on the load when crack initiates, acoustic emission
(AE) technique is applied to monitor the load when crack initiates along the grain.
According to GB4161-84 [23], there are three groups of samples for each wood.
In group A, samples are standard with a size of W/B = 2, S = 4 W; in group B,
samples are nonstandard with a size of S/W = 3–5, W/B = 2; in group C, samples
are nonstandard with a size of S/W = 4, 1 < W/B < 4. The crack is single edge
notched LT crack with a length of a = W/2, and the sizes of sample are the same for
different woods, as seen in Table 4.3.
Tests are performed on the computer-controlled mechanics of material testing
machine. Thin steel sheets are placed on the supports to minimize the friction
between sample and supports. AE sensor is attached on sample to monitor the
whole fracture process of the sample (Fig. 4.1). Loading speed is 2 mm/min, and
the load–deflection (F-d) curve is drawn automatically by the computer during
testing. Temperature was 15 °C and the humidity was 50–55% in laboratory.
From the F-d curve as seen in Fig. 4.8, it can be seen that the whole transverse
fracture process of wood includes three stages. (1) In the initial stage (OA), the F-d
curve is linear, and there is no crack along grain, while at point A, crack initiates
along grain, and the slope of F-d curve changes slightly. Also at that moment, AE
signals appear and slight sound of cracking can be heard by a medical stethoscope,
1.6 B Load
1.2
F (kN) 0.8
A
0.4
C
00 2 4 6 8 10 12
(mm)
so the load at point A is the critical load Fcr. (2) In the stable propagation stage of
crack (AB), crack propagates slowly with the increase of load, then ceases grad-
ually, during which more cracks along grain occur accompanied with slight sound
of cracking corresponding to the curved section on the F-d curve. (3) In the unstable
fracture stage (BC), when load increases to its maximum Fmax, the tissues around
the load point collapse on the upside of sample by pressure, and the fibers on the
downside of sample fracture by tensile force, accompanied with high sound of
cracking. Meanwhile load decreases stepwise, and the crack opens quickly until the
sample fractures.
Figure 4.9 shows the curve of force versus time and AE cumulative counts
versus time compared with effective voltage of discrete events versus time. As the
effective voltage (mv) is low when crack initiates along grain, decibel (dB) is
adopted to identify the critical load easily, as seen in Fig. 4.10.
B
AE cumulative (counts)
A C
Time (s)
Fig. 4.9 Load/time/cumulative events/relative energy (mV) curve for SEB specimen of wood
4.4 Test of Critical Stress Intensity Factor 73
AE cumulative (counts)
A C
Time (s)
Fig. 4.10 Load/time/cumulative events/relative energy (dB) curve for SEB specimen of wood
Due to GB4161-84 [23] and substituting the critical load and the sizes of sample
into the following equation, the nominal fracture toughness of wood transverse
fracture can be obtained
PC S a
KIC ¼ f ð4:1Þ
BW 3=2 W
LT
Table 4.4 The statistics description of KIC for four kinds of woods
Species Group Number F0 Fmax Fmax/ KIC SD CV
(N) (N) F0 (Nmm−3/2) (Nmm−3/2) (%)
Picea A 10 715 2062 2.9 57.71 4.6 7.98
asperata B 11 518 1601 3.15 52.55 6.98 13.3
C 11 501 1345 2.72 59.47 8.28 13.93
Total 56.54 7.29 12.9
Castanopsis A 12 1136 2192 1.93 147.45 29.44 19.97
hystrix B 13 901 1758 1.95 146.09 27.18 18.61
C 15 1323 2825 2.21 128.75 28.58 22.2
Total 134 29.04 20.74
Koompassia A 16 1640 5748 3.55 152.11 24.94 16.4
spp. B 16 1271 5675 4.55 147.44 30.25 20.51
C 18 895 3388 3.87 147.55 53.84 36.49
Total 150.7 31.86 21.14
Melia A 16 936 2190 2.34 89.2 15.33 17.19
azedarach B 15 936 2190 2.34 78.17 25.1 32.11
C 15 703 1415 2.27 78.17 25.1 32.11
Total 83.06 20.92 25.19
Two softwoods and two hardwoods were chosen as samples and they were spruce
(P. asperata), larch (Larix gmelinii), I-69 poplar (Populus spp.), and Castanopsis
hystrix. Standard specimens and specimens with crack perpendicular to grain were
used to perform bending test, impact test, and tensile test. The aim was to inves-
tigate the influence of crack perpendicular to grain on the bending strength, impact
toughness, and tensile strength of wood component.
LT
Table 4.5 The analysis of variance (ANOVA) of KIC for four kinds of woods
Species Source of Sum of df Mean F F-crit Sig.
variance Squares Square
Picea Between 283.45 2 141.72 3 3.33 No
asperata groups
Within 1365.84 29 47.1
groups
Total 1649.29 31
Castanopsis Between 3045.22 2 1522.61 1.88 3.25 No
hystrix groups
Within 29,836.46 37 806.39
groups
Total 36,230.94 39
Koompassia Between 263.89 2 131.94 0.12 3.21 No
spp. groups
Within 45,420.5 43 1056.29
groups
Total 45,684.38 45
Melia Between 1003.71 2 501.86 1.15 3.21 No
azedarach groups
Within 18,687.98 43 434.6
groups
Total 19,691.69 45
20
20
20
20
10
30
Fig. 4.11 Schematic diagrams of standard specimen and specimen with crack for bending test
76 4 Transverse Fracture of Wood
Fig. 4.12 Bending specimen with crack under load- crack initiated along grain
As seen in Fig. 4.12, for specimen with crack, crack initiated along grain. As
described in Sect. 4.2, the failure process of bending specimen with crack was that
the crack initiated along grain, and after propagated to a certain length, the speci-
men fractured with a fracture surface similar to that of standard specimen.
The statistical analysis results of the MOR of both groups of the four wood
species were shown in Table 4.6. It could be seen that the numerical values of the
two groups were similar.
Impact toughness is the energy absorbed or consumed by per area when wood
fractures because of impact load and it is the inherent ability of wood to resist
impact damage. To investigate the influence of crack perpendicular to grain on the
Table 4.6 Statistical analysis of MOR of standard specimen and specimen with crack
Species Specimen Number Mean (MPa) SD (MPa) CV (%)
Picea asperata Standard 22 74.95 11.06 14.76
Crack 22 78.23 9.62 12.3
Larix gmelinii Standard 27 129.84 12.98 10
Crack 26 125.28 16.35 13.05
I-69 Populus spp. Standard 30 65.28 10.47 16.03
Crack 42 66.12 8.28 12.52
Castanopsis hystrix Standard 16 122.97 13.26 10.78
Crack 16 118.08 10.88 9.21
4.5 The Influence of Transverse Crack on the Normal Strength of Wood 77
Table 4.7 Statistical analysis of impact strength of standard specimen and specimen with crack
Species Specimen Number Mean (MPa) SD (MPa) CV (%)
Picea asperata Standard 30 55.4 10.51 18.97
Crack 30 55.87 5.38 9.62
Larix gmelinii Standard 29 68.01 14.63 21.51
Crack 31 68.33 11.23 16.43
I-69 Populus spp. Standard 30 139.16 46.46 33.38
Crack 30 159.11 31.07 19.53
Castanopsis hystrix Standard 31 83.51 18.38 22.01
Crack 30 88.87 15.3 17.22
impact toughness of wood, there were two groups of specimens for each wood
species: one was standard specimen with a size of 300 mm (L) 20 mm
(T) 20 mm (R) according to GB1927-1943-91 [24]; the other one was specimen
with crack perpendicular to grain with a size of 300 mm (L) 30 mm
(T) 20 mm (R), and the crack was cut at the middle of specimen with a depth of
10 mm.
Impact tests were performed in the mechanical testing machine. The failure
surface of specimen with crack was similar to that of standard specimen. While it
could be observed on the fracture surface that the crack did not initiate along its
original direction but to both sides along grain just like that of bending specimen
with crack. When calculated the impact toughness of the specimen with crack, the
area parameter was 20 20 (mm2) theoretically.
The statistical analysis results of the impact toughness of both groups of the four
kinds of wood species were shown in Table 4.7. It could be seen that the impact
toughness of specimen with crack was all equal to or higher than that of standard
specimen. The main reason was that during impact test, the specimen with crack
absorbed more energy than the standard specimen because of the big size and
quality of specimen with crack.
Tensile strength parallel to grain represents the ability of wood to bear tensile load
along grain. To investigate the influence of crack perpendicular to grain on the
tensile strength parallel to grain of wood, there were three groups of specimens as
shown in Fig. 4.13. The specimens in first group were standard specimens
according to GB1927-1943-91 [24] (1991) and the cross-sectional dimensions in
the test section were 15 mm (R) 4 mm (T). The cross-sectional dimensions in
the test section of specimens in group 2 were 15 mm (R) 4 mm (T), while a
crack was cut at the middle of specimen with a depth of 3 mm, then the
78 4 Transverse Fracture of Wood
4
15
3 15
5
20
5
Fig. 4.13 Standard specimen and specimen with crack for tensile test
4.5.4 Discussion
The significant analysis results of the MOR, impact toughness, and tensile strength
of the four wood species were shown in Tables 4.9, 4.10, and 4.11. Variance
Table 4.8 Statistical analysis of tensile strength of standard specimen and specimen with crack
Species Specimen Number Mean (MPa) SD (MPa) CV (%)
Picea asperata Standard 20 95.54 9.77 10.23
Crack A 16 102.23 10.11 9.89
Crack B 22 107.32 10.36 9.65
Larix gmelinii Standard 20 109.79 26.93 24.53
Crack A 18 95.60 33.25 34.78
Crack B 18 87.20 23.13 26.53
I-69 Populus pp Standard 19 85.08 9.90 11.64
Crack A 20 80.85 17.47 21.60
Crack B 15 86.07 16.29 18.93
Castanopsis hystrix Standard 17 106.43 17.78 16.70
Crack A 15 92.58 17.46 18.86
Crack B 19 99.27 16.02 16.14
4.5 The Influence of Transverse Crack on the Normal Strength of Wood 79
Table 4.9 Significant analysis of bending test results of standard specimen and specimen with
crack
Species Source of Sum of df Mean F Fcrit Sig.
variance squares square
Picea Between 118.41 1 118.41 1.102 4.073 No
asperata groups
Within 4514.2 42 107.48
groups
Total 4632.6 43
Larix Between 275.63 1 275.63 1.27 4.03 No
gmelinii groups
Within 11,065.58 51 216.97
groups
Total 11,341.21 52
I-69 Populus Between 12.51 1 12.51 0.146 3.978 No
spp. groups
Within 5987.81 70 85.54
groups
Total 6000.32 71
Castanopsis Between 191.07 1 191.07 1.299 4.171 No
hystrix groups
Within 4411.43 30 147.05
groups
Total 4602.5 31
analysis showed that there was no significant difference between the results of
standard specimen and specimen with crack.
The tensile strength parallel to grain of wood is mainly determined by cellulose,
while the tensile failure of wood is not caused by the fracture of cellulose molecular
chain but the slide of cellulose molecular chains [25]. When subjected to tensile
stress, failure or slipping would first happen where the molecular chains at a dis-
advantage place born most of the stress. With the increase of tensile load, molecule
chains began to slide and flow, which caused the decrease of cross section and
resulted in the failure of wood. The fracture morphology of tensile specimen
embodied the fracture process above. It also could be seen that for standard
specimens, slide initiated randomly, so the initial point was the weakest place of
material (Fig. 4.14b); for specimen with crack, slide initiated from the pre-crack tip,
what’s more, crack propagated along the grain instead of along its original direction
during tensile process (Fig. 4.14a). And it could be obviously observed that the
distance between the cut pre-crack surfaces increased with the slide of material at
crack tip.
The results of bending test, impact test and tensile test demonstrated that wood
component with crack perpendicular to grain would not fail under low work stress
in spite of the stress singularity at crack tip. Wood possessed fine capacity to resist
transverse fracture because of its specific organization structure, so for the safety
80 4 Transverse Fracture of Wood
Table 4.10 Significant analysis of impact strength of standard specimen and specimen with crack
Species Source of Sum of df Mean F Fcrit Sig.
variance squares square
Picea Between 3.27 1 3.27 0.047 4.007 No
asperata groups
Within 4040.53 58 69.66
groups
Total 4043.8 59
Larix Between 1.5 1 1.5 0.01 4.01 No
gmelinii groups
Within 9773.9 58 168.52
groups
Total 9775.4 59
I-69 Populus Between 5968.28 1 5968.28 3.821 4.007 No
spp. groups
Within 90,584.6 58 1561.8
groups
Total 96,552.88 59
Castanopsis Between 436.89 1 436.89 1.523 4.004 No
hystrix groups
Within 16,924.03 59 286.85
groups
Total 17,360.92 60
Table 4.11 Variance analysis of tensile strength of standard specimen and specimen with crack
Species Source of Sum of df Mean F Fcrit Sig.
variance squares square
Picea Between 1457.98 2 728.99 2.552 3.165 No
asperata groups
Within 15,711.3 55 285.66
groups
Total 17,169.28 57
Larix Between 4981.37 2 2490.69 3.168 3.172 No
gmelinii groups
Within 41,671.2 53 786.25
groups
Total 46,652.57 55
I-69 Populus Between 282.49 2 141.24 0.639 3.179 No
spp. groups
Within 11,277.12 51 221.12
groups
Total 11,559.61 53
Castanopsis Between 1534.93 2 767.47 2.642 3.191 No
hystrix groups
Within 13,941.9 48 290.46
groups
Total 15,476.83 50
4.5 The Influence of Transverse Crack on the Normal Strength of Wood 81
Fig. 4.14 Tensile specimens: a for specimen with crack, crack initiated along grain; b failed
standard tensile specimen
For wood, cracking along grain is another damage form that has non-negligible
influence on the stiffness and stability of wood beam, and it is common in plywood,
so it is necessary to study the Mode I delamination and debonding damage of wood
beam and plywood further. Plentiful research work has been done by Williams [26,
27], Wang and Williams [28], Reeder [29, 30], and Du and Wang [31] to analyze
82 4 Transverse Fracture of Wood
and calculate the strain energy release rate of Mode I delamination and debonding
damage of laminate composites.
As seen, the schematic diagram of delamination crack tip under load and corner
analysis of wood beam is shown in Fig. 4.15. It is supposed that the thickness of
wood beam is H, the width is B, and there is a penetrative lamination fracture zone
and load distributed uniformly in width of the beam. The thickness above the
delamination is h1 and the thickness below is h2, and the applied moments are M1
and M2. When crack tip spreads from A-A to B-B by da, the energy release rate of
the whole system is
1 dUe dUs
G¼ ð4:4Þ
B da da
where Ue is the work done by external load to the system; Us is the strain energy of
the system.
When there is no delamination crack, the corners of section A-A and B-B under
load are u0 and u0 þ ddua0 da respectively, as seen in Fig. 4.15a.
When there is delamination crack, the corner of section A-A at crack tip was still
u0 under load (Fig. 4.15a). When crack tip spreads from section A-A to section
B-B, the corners of the section above and below section A-A are (Fig. 4.15b)
du1 du0
da ð4:5Þ
da da
du2 du0
ð Þ da ð4:6Þ
da da
Due to the flexure beam theory, the relationships between the corners and the
applied moments are
du0 M1 þ M2
¼ ð4:7Þ
da EI 0
(a) (b)
Fig. 4.15 Sketch of wood beam interlaminar fracture: a the corner of sectional A-A at
interlaminar crack tip. b While sectional A-A expanding to B-B with da, the corner of sectional
A-A and corner between superstratum and lower layer
4.6 Energy Release Rate of the Mode I Interlaminar Fracture … 83
du1 M1
¼ ð4:8Þ
da EI 1
du2 M2
¼ ð4:9Þ
da EI 2
3
where the inertia moment I0 ¼ BH
12 , and let n ¼ H , so:
h1
Bh31 BH 3 h1 3
I1 ¼ ¼ ¼ n3 I 0
12 12 H
Bh3 BH 3 H h1 3
I2 ¼ 2 ¼ ¼ ð1 nÞ3 I0
12 12 H
So:
du1 M1 M1
¼ ¼
da EI 1 En3 I0
ð4:10Þ
du2 M2 M2
¼ ¼
da EI 2 Eð1 nÞ3 I0
When delamination crack existed, the corners of the beam above and below the
lamination crack are different from the corners of the beam with no delamination
crack. The variation of the corners leads to the decrease of system energy and it is
also the driving force of the appearance and spread of delamination. Thus, during
the propagation of crack from section A-A to section B-B, the variation of the work
done by external load is
du1 du0 du2 du0
@Ue ¼ M1 da þ M2 da ð4:11Þ
da da da da
Then, substituting Eq. (4.10) into Eq. (4.11), the variation rate of the work done
by external load in the propagation process of delamination crack is
dUe du1 du0 du2 du0
¼ M1 þ M2
da da da da da
" #
M1 M1 þ M2 M2 M1 þ M2
¼ M1 þ M2 ð4:12Þ
En3 I0 EI0 Eð1 nÞ3 I0 EI0
" #
1 M12 M22
¼ 3
þ ðM1 þ M2 Þ2
EI0 n ð1 nÞ3
In the same way, when the delamination crack existed, the strain energy of the
beam above and below delamination crack were different from the strain energy of
84 4 Transverse Fracture of Wood
the system of those when there is no delamination crack. Due to the flexure beam
theory, the variation of system strain energy is
Thus, the variation rate of the strain energy of the system in the propagation
process of delamination crack is
So, the energy release rate of the system during the propagation of crack from
section A-A to section B-B is
" #
1 dUe dUs 1 M12 M22 2
Gi ¼ ¼ þ ðM1 þ M2 Þ ð4:15Þ
B da da 2BEI0 n3 ð1 nÞ3
c H
e
F/2 S F/2
4.6 Energy Release Rate of the Mode I Interlaminar Fracture … 85
The strain energy release rate will decrease with the increase of C and when it is
less than the critical strain energy release rate, the delamination crack will cease.
Especially, when C = 0, and the transverse crack converted to side crack along the
grain of wood beam, the rate of critical strain energy released can be calculated by
the following equation:
" #
3F 2 S2 1
GC ¼ 1 ð4:17Þ
8EB2 H 3 ð1 e=HÞ3
where F0 was the load when the crack along the grain began.
The analysis above can also be applied to the debonding behavior of plywood.
The energy consumed by single crack in the process of transverse bending fracture
of wood can be calculated by the following equation:
ZC Z " #
C
3 1
Uii ¼ 2 Gi Bdd ¼ 1 F 2 ðS 2dÞ2 dd ð4:18Þ
4EBH 3 0 ð1 a=HÞ3
0
The energy consumed by the cracking of a single layer is not high, thus no
significant change can be found on the F-d curve. However, multiple cracks will
happen in the whole process, so much energy will be consumed.
References
1. Schniewind AP, Pozniak RA (1971) On the fracture toughness of Douglas fir wood. Eng Fract
Mech 2:223–230
2. Prokopski G (1993) The application of fracture mechanics to the testing of wood. J Mater Sci
28:5995–5999
3. Reiterer A, Sinn G, Stanzl-Tschegg SE (2002) Fracture characteristics of different wood
species under mode I loading perpendicular to the grain. Mater Sci Eng A Struct 332:29–36
4. Keunecke D, Stanzl-Tschegg S, Niemz Peter (2007) Fracture characterisation of yew (Taxus
baccata L.) and spruce (Picea abies wL.x Karst.) in the radial-tangential and tangential-radial
crack propagation system by a micro wedge splitting test. Holzforschung 61:582–588
5. Jeronimidis G (1976) The fracture of wood in the relations to its structure. Leiden Bot Ser
3:253–265
6. Kollmann FFP, Côté WA (1968) Principles of wood science and technology I solid wood.
Springer-Verlag, New York
7. Xu YM (2006) Wood science. China Forestry Publishing, Beijing, p 181
8. Danielsson H, Gustafsson PJ (2013) A three dimensional plasticity model for perpendicular to
grain cohesive fracture in wood. Eng Fract Mech 98:137–152
9. Stanzl-Tschegg SE (2006) Microstructure and fracture mechanical response of wood. Int J
Fract 139:495–508
10. Bodner J, Schlag MG, Grüll G (1997) Fracture initiation and progress in wood specimens
stressed in tension. Part I. Clear wood specimens stressed parallel to the grain. Holzforschung
51:479–484
86 4 Transverse Fracture of Wood
11. Bodner J, Schlag MG, Grüll G (1997) Fracture Initiation and progress in wood specimens
stressed in tension part II compression wood specimens stressed parallel to the grain.
Holzforschung 51:571–576
12. Bodner J, Schlag MG, Grüll G (1998) Fracture initiation and progress in wood specimens
stressed in tension part III. Clear wood specimens with various slopes of grain. Holzforschung
52:95–101
13. Grekin M, Surini T (2008) Shear strength and perpendicular-to-grain tensile strength of
defect-free scots pine wood from mature stands in Finland and Sweden. Wood Sci Technol
42:75–91
14. Galicki J, Czech Michal (2015) Tensile strength of softwood in LR orthotropy plane. Mech
Mater 37:677–686
15. Vasic S, Stanzl-Tschegg S (2007) Experimental and numerical investigation of wood fracture
mechanisms at different humidity levels. Holzforschung 61:367–374
16. Schoenmakers JCM, Jorissen AJM, Leijten AJM (2010) Evaluation and modelling of
perpendicular to grain embedment strength. Wood Sci Technol 44:579–595
17. Qiu LP, Zhu EC, van de Kuilen JWG (2014) Modeling crack propagation in wood by
extended finite element method. Eur J Wood Prod 72:273–283
18. Cao J, Li FG, Wang QR et al (2016) Analysis of fracture criteria for 7050 aluminum alloy
with different geometries based on the elastic strain energy density. Theor Appl Fract Mect
81:50–66
19. Viotti MR, Kaufmann GH, Galizzi GE (2006) Measurement of elastic moduli using spherical
indentation and digital speckle pattern interferometry with automated data processing. Opt
Laser Eng 44:495–508
20. Wang QH, Xie HM, Tang P et al (2009) A study on the mechanical properties of beagle
femoral head using the digital speckle correlation method. Med Eng Phys 31:1228–1234
21. Method of testing in tensile strength parallel to grain of wood (GB 1938-91) (1991). China
Standards Press, Beijing
22. Method of testing in tensile strength perpendicular to grain of wood (GB/T 14017-92) (1991).
China Standards Press, Beijing
23. National standard Metallic materials-Determination of plane-strain fracture toughness
(GB4161-84) (1984)
24. Testing methods for physical and mechanical properties of wood (GB1927-1943-91) (1991).
China Standards Press, Beijing
25. Naoto W (1986) Basic of wood application. Shanghai Science and Technology Press,
Shanghai, pp 401–407
26. Williams JG (1988) On the calculation of energy release rates for crack laminates. Int J Fract
Mech 36:101
27. Williams JG (1989) End corrections for orthouropic DCB specimens. Compos Sci Technol
35:367
28. Wang Y, Williams JG (1992) Corrections for mode II feacture toughness specimens of
composite materials. Compos Sci Technol 43:251
29. Reeder JR, Crews JH (1990) Mixed-mode bending methed for delamination testing. AIAA 28
(7):1270
30. Reeder JR, Crews JH (1991) Redesigm of mixed-mode bending test for delamination
toughness. In: Proceedings of international conference on cognitive modeling, vol 8,
section 13-36
31. Du S, Biao W (1998) Micromechanics of composite materials. Science Press, Beijing,
pp 161–237
Chapter 5
Finite Element Analysis of Wood Crack
Tip Stress Field and Prediction
of the Crack Propagation Direction
Abstract The crack in timber components was always initiated along the grain no
matter what the original crack direction was. To investigate the fracture mechanism,
three-point bending (SEB) specimens of spruce (Picea jezoensis) are selected for
tests. And finite element method is applied to analyze the stress fields around crack
tips of 4 different SEB specimens with initial cracks orientated 0°, 30°, 60°, and 90°
to grain. The “tangential normal stress intensity factor ratio criterion” is adopted to
predict the propagation direction of crack. The results show that (1) the radial lines
of the maximum tensile stress of the 4 crack bodies are all along grain when a radial
plane is made at the crack tip; (2) The ratio of the tensile stress perpendicular to
crack surface (rY) and tensile stress parallel to the crack surface (rX) is almost a
constant ranging from 1 to 5 in a large fracture process region of the crack tip apart
from several singular points; (3) The “tangential normal stress intensity factor ratio”
gets maximum along the grain at the crack tip no matter how large the angle
between the crack and grain is. Both theoretical analysis and experimental results
prove that the propagation direction of crack will turn along the grain. Also, the
relationship between the interfacial strength and toughness of wood is discussed.
5.1 Introduction
Wood is one of the few plant materials that can be directly used as structural
materials, meanwhile, it is also the oldest and still the most widely used natural
structure material. The annual total production of wood is in the range of 500
million tons, among which about 220 million tons are used for material applications
[1]. What’s more, the demand quantity of wood increases day by day. And trans-
verse bending is the main load carrying way of timber components. So, it is very
important to predict the propagation direction of the cracks in different angles to the
longitudinal direction of wood under transverse force, which is also very important
for the design and safety evaluation of timber.
For air-dried wood, the strain–stress curve had a linear characteristic, so the
mechanical behavior of wood fitted into the linear elastic behavior and wood could
Picea asperata was selected as a sample with a moisture content of 13% by air
dried. Wood stem could be seen as cylindrical symmetry body. When a cube is cut
along three orthogonal principal axes: longitudinal (L), radial (R), and tangential
(T) axes at a certain distance from the pith, it could be regarded as the orthogonal
anisotropic body. So the stress–strain relationship could be expressed as follows by
the engineering elastic constants [16]:
8 9 2 38 9
>
> eL >
> 1=EL lLR =ER lLT =ET 0 0 0 >
> rL >>
>
> > >
> rR >
>
> eR >
>
>
6 lRL =EL
6 1=ER lRT =ET 0 0 0 7 7>>
>
>
>
< = 6 < =
eT lTL =EL lTR =ER 1=ET 0 0 0 7 r
¼6
6
7
7
T
> c
> RT >>
> 6 0 0 0 1=GRT 0 0 7> > s >
>
> > RT >>
>
> cTL >
>
>
4 0 0 0 0 1=GTL 0 5> >
> s > >
: ; : TL >;
cLR 0 0 0 0 0 1=GLR sLR
The elastic parameters and the tensile strength parallel to grain (rL) and per-
pendicular to grain (rT) should be measured before going to analyze the stress field
around crack tip and predict the crack direction. Three rectangle samples with the
positive axis and three with 45° off-axis (Fig. 5.1) were used to measure 12 elastic
parameters by the digital speckle correlation method (DSCM) [17, 18]. The 12
elastic parameters was EL = 12 GPa, ER = 1.2 GPa, ET = 0.6 GPa,
GRT = 85 Mpa, GTL = 209 Mpa, GLR = 495 Mpa, lLR = 0.4, lLT = 0.4,
lRL = 0.04, lRT = 0.45, lTL = 0.02, lTR = 0.23, respectively.
The tensile strength parallel to grain (rL) and tensile strength perpendicular to
grain (rT) were tested according to Chinese national standards GB1938-91 [19] and
GB/T14017-92 [20] respectively. The results were shown in Table 5.1. The ratio of
rL/rT was about 25. All the experiments were conducted at 23 ± 2 °C and relative
humidity of about 66 ± 3%.
F F F F F F
Fig. 5.1 Six kinds of speckle specimens for testing elastic modulus
90 5 Finite Element Analysis of Wood Crack Tip Stress Field …
The specimen was standard single-edge-notched bending (SEB) specimen. The size
of the specimen was S = 160 mm, W = 40 mm, B = 20 mm, a = 20 mm,
respectively, where S was the span length, W was the height of specimen, B was the
thickness of the specimen, and a was the length of the initial crack. The initial crack
was located in the middle of the specimen. The initial crack orientation (b) was 0°
(a crack parallel to the grain, i.e., TL crack), 30°, 60°, and 90° (a crack perpen-
dicular to the grain, i.e., LT crack) respectively (Fig. 5.2).
The numerical simulations of the three-point bending tests were performed by
ABAQUS finite element software. CPS8R (eight-node plane strain element,
reduced integration) was adopted to mesh the model. And degraded 1/4 singularity
element was introduced around the crack tip. The radius of the singularity element
was 0.05 mm and the crack tip was divided into 36 parts (10° per unit). The mesh
was dense near crack tip and sparse away from the crack tip (Fig. 5.3). The applied
load was 100 N.
F F
T
L X 60
X L
T 0
Y Y
a a
R R
F F
L
T X T 90
30 X
L W=40
Y Y
a a=20
R R
S=160 B=20
Fig. 5.2 Schematic diagram of cracks and fiber orientation of SEB samples
5.3 Results and Discussion 91
Fig. 5.3 Arrangement of the quarter point elements at the crack tip Mesh generation around crack
tip
The field cloud maps of the stress (rY) perpendicular to the initial crack surface and
field cloud maps of the stress (rX) parallel to the initial crack surface were shown in
Fig. 5.4. It could be seen that the maximum stress rY of the four crack bodies were
all along grain when a radial plane was made at the crack tip, while the maximum
stress rX tended to perpendicular to the grain. The value of rY, rX, and rY/rX around
the crack tips of the four crack bodies were shown in Table 5.2. Variation ten-
dencies of rY and rX within 5 mm from the crack tip were shown in Fig. 5.5. It
indicated that with the distance from the crack tip increased the stress declined
quickly, then tended to be stable and very low.
The fracture of materials always began at a micro-area in front of a crack tip. The
micro-area was called the fracture process region, which was not only the area with
high stress but also the area where microstructure played a determined role. When
predicted the crack direction of composite, Gregory and Herakovich [21] found that
no matter on any periphery within a certain distance from the crack tip (region
where critical stress intensity factor K played a leading role), crack would propagate
along the direction where the “tangential normal stress intensity factor ratio” got its
maximum. According to the hypothesis of Gregory and Herakovich [21], the
strength of material on the direction tangent to the periphery of arbitrary angle h
was given by
2.16
2.46
3.08
5.15 1.99
2.52
3.09
5.04
o
β =0
o
3mm β =0 1mm
σY σX
2.06
2.53 2.01
3.06
2.50
4.89
3.03
4.83
o o
β =30 β =30
σY 5mm σX 1mm
1.45
2.08 1.47
2.98
5.14 1.97
2.89
4.52
o o
β =60 5mm β =60
1mm
σY σX
0.47
Fig. 5.4 The stress field cloud maps of the 4 kinds of specimens (b was the angle between crack
and grain)
Table 5.2 The rY and rX around crack tip of the four crack bodies (F = 100 N)
D/mm 0° 30° 60° 90°
rY/MPa rX/MPa rY/rX rY/MPa rX/MPa rY/rX rY/MPa rX/MPa rY/rX rY/MPa rX/MPa rY/rX
0 227.5 83.47 2.726 256.5 71.64 3.58 217.09 42.45 5.114 142.1 14.48 9.814
0.0073 147.97 40.933 3.615 84.249 35.444 2.377 59.56 21.62 2.755 38.07 8.7734 4.339
0.05 63.346 14.431 4.39 31.362 12.898 2.432 21.625 8.1487 2.654 13.476 3.2714 4.119
0.1 45.428 10.09 4.502 21.23 8.9743 2.336 14.614 5.6156 2.602 9.3721 2.2142 4.233
5.3 Results and Discussion
0.15 37.415 8.3198 4.497 17.356 7.4037 2.344 11.978 4.6401 2.581 7.7176 1.8134 4.256
0.2 32.486 7.1922 4.517 14.969 6.4087 2.336 10.311 4.0237 2.563 6.6682 1.5646 4.262
0.25 29.048 6.4147 4.528 13.289 5.7222 2.322 9.1467 3.5994 2.541 5.9443 1.3933 4.266
0.3 26.492 5.833 4.542 12.046 5.21 2.312 8.2728 3.284 2.519 5.4034 1.2662 4.268
0.35 24.494 5.3775 4.555 11.069 4.8091 2.302 7.5904 3.0372 2.499 4.9798 1.1665 4.269
0.4 22.879 5.0077 4.568 10.272 4.4839 2.291 7.035 2.8369 2.48 4.6358 1.0856 4.27
0.45 21.531 4.6995 4.582 9.6054 4.2118 2.28 6.5703 2.6699 2.461 4.3491 1.0181 4.272
0.5 20.388 4.4372 4.595 9.0363 3.9821 2.269 6.1737 2.5278 2.442 4.1052 0.9606 4.274
0.55 19.4 4.2103 4.608 8.5421 3.7826 2.258 5.8295 2.4049 2.424 3.8942 0.9108 4.276
0.6 18.535 4.0113 4.621 8.1072 3.6076 2.247 5.5266 2.2971 2.406 3.7092 0.8671 4.278
0.65 17.769 3.8348 4.634 7.7201 3.4524 2.236 5.257 2.2015 2.388 3.5451 0.8283 4.28
0.7 17.083 3.6768 4.646 7.3725 3.3134 2.225 5.0149 2.1159 2.37 3.3982 0.7935 4.283
0.75 16.465 3.5341 4.659 7.0576 3.1879 2.214 4.7956 2.0386 2.352 3.2656 0.762 4.285
0.8 15.904 3.4044 4.672 6.7705 3.0738 2.203 4.5955 1.9683 2.335 3.145 0.7334 4.288
0.85 15.391 3.2857 4.684 6.5071 2.9694 2.191 4.4119 1.904 2.317 3.0347 0.7072 4.219
0.9 14.92 3.1764 4.697 6.2641 2.8733 2.18 4.2425 1.8449 2.3 2.9333 0.6831 4.294
(continued)
93
Table 5.2 (continued)
94
Fig. 5.5 The distribution of rY and rX in front of crack tips of the four crack bodies (b was the
angle between crack and grain)
rX þ rY r X rY
rhh ¼ þ cos 2h sXY sin 2h ð5:2Þ
2 2
where X and Y were the strength on the first and second principal directions of
material, respectively, sXY the shear strength, b the angle between the initial crack
and the first principal direction of material (Fig. 5.6). So the “tangential normal
stress intensity factor ratio” was defined as the ratio of tangential normal stress to
the strength of material in the same direction as follows:
96 5 Finite Element Analysis of Wood Crack Tip Stress Field …
rhh
R¼ ð5:3Þ
Thh
Thus, the crack would propagate along the direction of h when R obtains the
maximum value.
The “tangential normal stress intensity factor ratio” hypothesis has been applied
successfully to predict the crack propagation direction of composite materials,
which was also called “tangential normal stress intensity factor ratio criterion”. It
was assumed that the criterion was appropriate for the orthotropic material––wood.
To verify the assumption, ABAQUS was applied to analyze the stress of each node
on the singularity element at the crack tip of P. asperata SEB specimens, as seen in
Fig. 5.3. The applied load was 100 N. The stress of each node and Thh, rhh and R
calculated by Eqs. (5.1)–(5.3) were shown in Table 5.3.
It was obvious from Table 5.3 that no matter how large the angle between crack
and grain was, the R (The load was 100 N) got maximum on the radial direction
along the grain. Therefore, according to the “tangential normal stress intensity
factor ratio criterion”, it could be determined that for timber beam with initial cracks
in arbitrary angle to grain, cracks would always initiate along the grain under
transverse load. As seen in Fig. 5.7, the crack propagation paths of the four crack
bodies were shown.
After millions of years’ evolution, as a response to the bending load caused by wind
and snow, tree has evolved into particular structure, which enables tree with high
resistance ability to transverse bending failure. As known, most wood cells are
oriented in the longitudinal axis. The first microstructure level is a tubular structure,
and the second microstructure level is cell wall with multilayer structure reinforced
by microfibrils which are bonded together with various non-cellulose components.
Thus, the transverse interfacial strength is much lower than the axial tensile strength
Table 5.3 Stress and tangential normal stress intensity factor ratio of singularity element around crack tip (F = 100 N)
b h 0° 10° 20° 30° 40° 50° 60° 70° 80° 90° 100° 110° 120° 130° 140° 150° 160° 170° 180°
0° rX / 63.35 43.91 30.54 24.67 22.07 20.51 18.9 16.82 14.78 13.45 12.65 12.04 12.1 13.39 15.85 19.11 22.31 23.15 3.6
MPa
r Y/ 14.43 14.88 15.44 16.02 16.63 17.21 17.66 17.75 17.13 15.51 13.1 10.54 8.34 6.61 5.16 3.81 2.45 1.05 0.86
MPa
sXY/ 0 −2.77 −2.4 −1.54 −0.62 0.32 1.39 2.69 4.21 5.7 6.77 7.28 7.38 7.26 6.99 6.48 5.57 3.85 1.25
MPa
Thh 3.8 6.57 14.53 26.74 41.7 57.64 72.61 84.81 92.77 95.54 92.77 84.81 72.61 57.64 41.7 26.74 14.53 6.57 3.8
rhh 63.35 43.99 30.32 23.84 20.44 18.26 16.77 15.92 15.62 15.51 15.4 15.39 15.67 16.57 18.32 20.9 23.57 23.8 3.6
R 16.67 6.7 2.09 0.89 0.49 0.32 0.23 0.19 0.17 0.16 0.17 0.18 0.22 0.29 0.44 0.78 1.62 3.62 0.95
30° rX / 31.36 38.05 49.61 57.57 29.93 17.17 13.29 12.75 12.69 11.74 9.38 6.45 4.04 1.94 −0.18 −1.5 −1.21 0.23 4.12
MPa
rY/ 12.9 12.42 11.99 11.8 12.42 12.96 13.46 13.93 14.35 14.66 14.68 14.11 12.71 10.68 8.57 6.83 5.45 4.24 4.05
MPa
sXY/ 0 0.22 −0.32 −3.9 −5.43 −4.25 −2.56 −1.86 −0.94 −0.03 1.02 2.25 3.46 4.33 4.75 4.87 4.83 4.61 3.76
MPa
Thh 26.77 14.56 6.58 3.8 6.55 14.51 26.7 41.67 57.6 72.57 84.78 92.76 95.54 92.79 84.83 72.64 57.67 41.74 26.77
rhh 31.36 37.2 45.41 49.51 28.04 18.89 15.63 14.98 14.62 14.66 14.87 14.66 13.54 11.34 8.11 4.8 2.68 1.93 4.12
R 1.17 2.56 6.69 13.03 4.28 1.3 0.59 0.36 0.25 0.2 0.18 0.16 0.14 0.12 0.1 0.07 0.05 0.05 0.15
60° rX / 21.63 23.29 25.66 29.13 34.64 43.82 47.55 20.96 10.47 8.37 9.11 9.84 9.31 7.07 4.05 1.36 −0.96 −2.72 0.74
MPa
r Y/ 8.15 7.86 7.52 7.18 6.87 6.58 6.49 7.05 7.43 7.73 7.96 8.14 8.25 8.2 7.8 6.97 5.87 4.88 4.78
5.4 The Relationship Between Interfacial Strength and Toughness …
MPa
sXY/ −1.16 −0.71 −0.41 −0.25 −0.32 −1.06 −4.44 −5.44 −4.18 −2.99 −2.12 −1.53 −1.05 −0.52 0.12 0.74 1.19 1.48 1.63
MPa
Thh 72.59 57.62 41.69 26.72 14.52 6.56 3.8 6.57 14.54 26.75 41.72 57.65 72.62 84.82 92.78 95.54 92.77 84.8 72.59
rhh 21.63 23.07 23.8 23.86 23.48 23.01 20.6 12.17 8.96 7.73 7.27 7.35 7.6 7.22 5.71 3.41 0.61 −1.98 0.74
R 0.3 0.4 0.57 0.89 1.62 3.51 5.42 1.85 0.62 0.29 0.17 0.13 0.1 0.09 0.06 0.04 0.01 −0.02 0.01
(continued)
97
Table 5.3 (continued)
98
b h 0° 10° 20° 30° 40° 50° 60° 70° 80° 90° 100° 110° 120° 130° 140° 150° 160° 170° 180°
90° rX / 13.48 13.64 14.11 14.89 16.01 17.59 19.83 23.27 28.8 29.62 11.41 5.18 4.64 5.72 6.49 6.16 4.53 2.32 −0.43
MPa
rY/ 3.27 3.18 2.97 2.7 2.44 2.19 1.95 1.72 1.45 1.38 1.54 1.6 1.57 1.49 1.36 1.2 0.96 0.64 0.46
MPa
sXY/ 0 0.27 0.49 0.62 0.7 0.71 0.66 0.47 −0.16 −2.48 −2.96 −2.07 −1.33 −0.89 −0.69 −0.6 −0.53 −0.45 −0.38
MPa
Thh 95.54 92.78 84.82 72.62 57.65 41.72 26.75 14.54 6.57 3.8 6.56 14.52 26.72 41.69 57.62 72.59 84.8 92.77 95.54
rhh 13.48 13.23 12.5 11.3 9.72 7.85 5.85 3.93 2.33 1.38 0.82 0.69 1.19 2.36 3.69 4.39 3.77 2.12 −0.43
R 0.14 0.14 0.15 0.16 0.17 0.19 0.22 0.27 0.35 0.36 0.13 0.05 0.04 0.06 0.06 0.06 0.04 0.02 0
5 Finite Element Analysis of Wood Crack Tip Stress Field …
5.4 The Relationship Between Interfacial Strength and Toughness … 99
Fig. 5.7 The crack propagation direction of the four crack bodies in the SEB tests
of wood. And it is just the weakened interfacial strength that enables tree stem to
possess excellent toughness.
The relationship curves of rY/rX and distance from crack tips of the four crack
bodies were shown in Fig. 5.8. It indicated that no matter how large the angle
between the crack and grain was, rY/rX was almost a constant ranged from 1 to 5
within 4 mm from the crack tip apart from singular points (Table 5.4). Although
with increase in the distance, the ratio increased, and the stress was much lower
there.
In order to explain the relationship between interfacial strength and toughness of
wood better, the specimen with the crack perpendicular to the grain was taken as an
example as shown in Fig. 5.8. The stresses declined quickly with the distance from
the crack tip increased and then changed gently after reaching a certain value.
While, the ratio of rY to rX was almost a constant ranged from 4 to 5 (Fig. 5.5)
Table 5.4 The values of rY/rX within 4 mm from the crack tip
Distance 0 0.0073 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
from crack
tip (mm)
b (°) 90 9.81 4.34 4.12 4.23 4.26 4.26 4.27 4.27 4.27 4.27
60 5.11 2.75 2.65 2.6 2.58 2.56 0.54 2.52 2.5 2.48
30 3.58 2.38 2.43 2.37 2.34 2.34 2.32 2.31 2.3 2.29
0 2.73 3.61 4.39 4.5 4.5 4.52 4.53 4.54 4.55 4.57
Distance 0.45 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9
from crack
tip (mm)
b (°) 90 4.27 4.27 4.28 4.28 4.28 4.28 4.29 4.29 4.29 4.29
60 2.46 2.44 2.42 2.41 2.39 2.37 2.35 2.33 2.32 2.3
30 2.28 2.27 2.26 2.25 2.24 2.23 2.21 2.2 2.19 2.18
0 4.58 4.59 4.61 4.62 4.63 4.65 4.66 4.67 4.68 4.7
Distance 0.95 1 1.2 1.4 1.6 1.8 2 2.5 3 4
from crack
tip (mm)
b (°) 90 4.3 4.3 4.32 4.34 4.36 4.38 4.41 4.5 4.62 5.04
60 2.28 2.26 2.19 2.12 2.05 1.98 1.91 1.72 1.52 1.09
30 2.17 2.16 2.11 2.06 2.02 1.97 1.92 1.79 1.66 1.39
0 4.71 4.72 4.77 4.83 4.88 4.94 5 5.17 5.37 5.89
apart from some singular points. Thus, when the tensile stress region in front of
crack tip developed to the interface, the interface would not crack if the interfacial
strength (namely tensile strength perpendicular to grain) was higher than 1/5 of
longitudinal strength of the wood. Then, the crack would propagate through the
interface. At this moment, wood would have the same failure behavior with
common brittle materials, and brittle failure would happen. On the contrary, if the
interfacial strength was lower than 1/5 of longitudinal strength, the interface would
be pulled open by the tensile stress rX leading to the formation of a new crack
perpendicular to the original crack, as seen in Fig. 5.9. The passivated crack could
eliminate the stress concentration and prevent the propagation of original crack to
avoid unstable fracture. In fact, the tensile strength perpendicular to grain was much
lower than the tensile strength parallel to grain, for example, the ratio of rT/rL of
the P. asperata was only 1/25 in this section, so cracks in timber beam always
propagate along the grain. The bonding force between interfaces of cells in a living
tree is lower than that in dried wood because of water permeation, thus the inter-
faces are easy to slip. This is why living trees have fine flexibility and the ability to
resist bending fracture even after bitten by beasts or chopped by knife or axe.
The theory to improve the toughness of material by means of interface control
has been successfully applied in the design of artificial composites. The structure
and property of interface would influence the overall performance of composite
directly. Studies [14, 22–25] indicated that the interface with low bonding force and
5.4 The Relationship Between Interfacial Strength and Toughness … 101
good plasticity and easy to dissociate should be adopted to improve the strength and
toughness of composites. That was because low bonding force was beneficial for
the interfaces to slip, which would lead to the good plasticity of composites. And a
good plastic interface was helpful to eliminate the stress concentration. Therefore,
the key to produce high-performance composite lies in the study on the effect of the
formation, properties, and stress transfer behavior of interface on macroscopic
properties of the composite. Consequently, the study on interfacial debonding and
the numerical simulation of interface failure is the key research contents in
materials.
References
1. Mantau U et al (2010) Real potential for changes in growth and use of EU forests. Euwood
Final Report, Project: Call for tenders No. TREN/D2/491-2008. Hamburg
2. Kollmann FFP, Cate WA (1968) Principles of wood science and technology. I: solid wood.
Springer, New York
3. Wu EM (1967) Application of fracture mechanics to anisotropic plates. J Appl Mech 34
(4):967–974
4. Boatright SWJ, Ggrrette GG (1983) The effect of microstructure and stress state on the
fracture behaviour of wood. J Mater Sci 18:2181–2199
5. Stanzl-Tschegg SE, Tan DM, Tschegg EK (1996) Fracture resistance to the crack propagation
in wood. Int J Fract 75:347–356
6. Stanzl-Tschegg SE (2006) Microstructure and fracture mechanical response of wood. Int J
Fract 139:495–508
7. Keunecke D, Stanzl-Tschegg SE, Niemz P (2007) Fracture characterization of yew (Taxus
baccata L.) and spruce (Picea abies wL.x Karst.) in the radial-tangential and tangential-radial
crack propagation system by a microwedge splitting test. Holzforschung 61:582–588
8. Nairn JA (2007) Material point method simulations of transverse fracture in wood with
realistic morphologies. Holzforschung 61:375–381
102 5 Finite Element Analysis of Wood Crack Tip Stress Field …
Abstract In this chapter, the fractal theory is applied to study the fractal feature of
the fracture surfaces of five types of woods along the grain and the relationship
between the fractal dimension of fracture surface and fracture toughness of wood is
established. The results show that fracture toughness parallel to the grains of various
woods is different because of their textural diversity and such differences are also
shown in the morphology of fracture surfaces. Furthermore, there is an evident and
direct proportional relation between the fractal dimension and fracture toughness
along grains. Then the evolution characteristics of microstructure during the bending
failure of the clear samples and samples with the transverse crack of four types of
woods are studied by acoustic emission (AE) technique, and the initiation and
expansion of different types of damage of wood component are identified with AE
characteristic parameters. The results showed that: (1) AE event counts developed
slowly and most were the low-amplitude AE events at the low strains and a large
number of high-amplitude AE events appeared in peak load or fracture stage for the
standard sample. (2) The initiation and expansion of crack tip could be monitored
efficiently by AE technique in the whole process of wood three-point bending test for
the notched sample. (3) The AE signals were related to different damage patterns/
modes. The AE characteristics of cell wall fracture were high amplitude, high
energy, and long duration time AE events but the AE characteristics of cell wall
damage and spallation, cell wall buckling, and collapse were low-amplitude,
low-energy and short duration time AE events. (4) Kaiser effect appeared at low
loading and Felicity effect at high loading under repeated wood bending loading.
The Felicity ratio could better indicate the damage degree of wood structure.
6.1.1 Introduction
Observations have shown that fracture surfaces for different materials are statisti-
cally very well described by self-affine fractal [1], and fractal dimension has been
considered as a measure of fracture surface morphology of materials. It also reflects
6.1.2 Theories
The nature of fractals is reflected in the word itself, coined by Mandelbrot [5] from
the Latin verb frangere, ‘to break’, and the related adjective fractus, ‘irregular and
fragmented’. Fractal geometry indicates the property of self-similar or self-affine on
mathematics, and the fractal dimension is used to describe quantificationally the
degree of irregularity of geometric construction [15]. As a common phenomenon in
nature, observations have shown that material fracture surfaces are self-affine. The
application of fractal geometry in fracture studies is helpful to understand the
fracture itself and builds the possible relationship between fracture toughness and
fractal dimension.
The method to measure the fractal dimension of the material fracture surface is
that choose a kind of area element small enough to cover the fracture surface.
6.1 The Fractal Features of Wood Fracture 105
The size of the area element is e, and the number of the area elements needed to
cover the rupture surface is N(e), So
LðeL Þ ¼ L0 e1D
L
L
ð6:3Þ
where DL is the fractal dimension of contour line L of cross section. Actually, the
roughness of contour line is RL = L/L0. Similarly, we can see the surface roughness
as RS = S/S0. The relation between RS and RL is deduced by Underwood [16] as
4
RS ¼ þ1 ð6:4Þ
pðRL 1Þ
4
e2DS ¼ þ1 ð6:5Þ
pðe1D
L
L
1Þ
Since the fractal structure of S and contour line of section L exists in the same
scale.
Therefore, let e = eL, we get
pe4L
ln 1DL
4ðeL 1Þ þ p
DS ¼ ð6:6Þ
2 ln eL
The equation above is the analytical formula to calculate the fractal dimension
DS of the fracture surface. DL can be calculated by the method of box-counting
which is from the curve that is obtained by electronic scanning, and eL is a fixed
linear element yardstick.
Three kinds of air-dried softwood and two kinds of air-dried hardwood were chosen
for this experiment. They are Cunninghamia lanceolata, Picea jezoensis var.
microsperma, Larix gmelinii, Populus spp, and Castanopsis hystrix respectively.
The moisture content of the samples is about 13%.
According to the standard established by ASTM E399-74 [17], the size of the
CT sample containing TL crack is W = 50 mm, e = 12.5 mm, a = 25 mm,
H = 30 mm (Fig. 6.2). The number of samples of each kind of wood is 30, and the
total number is 150. In order to get such a CT sample, first, we cut a straight flute
about 30 mm with a band saw along the grain, then cut the straight flute forward 1–
2 mm with a sharp blade and press a knife with relatively thick back (or wedge)
into the straight flute to expand the crack port forward about 2–5 mm. After the
process above, the natural incisive crack port is obtained. Then, we saw the CT
samples according to the crack length e + a. The obtained natural incisive crack
samples would correspond well to the real situation in structural lumber. All there
experiments were taken under 23 ± 2 °C and the relative humidity is about
66 ± 3%.
Force B
a
e W
grains is quasi-brittle, except for a gentle curve in the initial loading stage because
of the gap between the sample and U hook, and then the F-d curve is nearly linear.
Once the crack began to craze along the grains, i.e., the unstable fracture expansion
happened. So the critical load FQ = Fmax. Then, taking the values of critical load
Fmax of each sample into the following equation recommended by ASTM E399-74
[17], we got average fracture toughness KIC parallel to the grains of the five kinds of
wood.
FQ a
KIC ¼ pffiffiffiffiffi f ð6:7Þ
B W W
where
a rffiffiffiffiffi
a
f ¼ 29:6
W W
a 3=2 a 5=2 a 7=2 a 9=2
185:5 þ 655:7 1017:0 þ 638:9 :
W W W W
After the experiment, the CT sample was not separated absolutely because of fiber
bridge, it was needed to be separated artificially. In order to measure DS, we put the
fracture surface at the digital microscope with 3CCD and its pixel is as high as
5,400,000. The digital microscope can show the 3D structure of the fracture surface
for its function of restructure and it can also measure the length of contour line
between any two points automatically. To collect a complete information of the
fracture surface, 4 5 = 20 testing surface elements (Fig. 6.4) were chosen from
the natural incisive crack surface (about 20 12.5 mm2). We obtained the 3D
image of the testing surface element under 60 times of microscopy magnification,
and drew 5 contour lines along longitudinal and tangential direction by electronic
scanning on the surface element (Fig. 6.5). Figure 6.6 shows an in log–log repre-
sentation curve of one C. lanceolata sample of the box-counting method which the
liner regression is used to measure the fractal dimension. Then DS of the fracture
108 6 Fractal Features and Acoustic Emission Characteristics of Wood Fracture
F/kN
F/kN
/mm /mm
Cunninghamia lanceolata Picea jezoensis var. microsperma
F/kN
F/kN
/mm /mm
Populus sp Larix gmelinii
F/kN
/mm
Castanopsis hystrix
surface in meso-field could be gained. The average values of DS of five kinds of tree
are shown in Table 6.1.
Fig. 6.5 The 3D image of the testing surface element and electron scanning microscope of
various woods fracture surface parallel to grains. a Cunninghamia lanceolata, b Picea jezoensis,
c Larix gmelinii, d Populus spp, e Castanopsis hystrix
2
1.5
y = -1.2291x + 4.1402
1
0.5
0
0 0.5 1 1.5 2 2.5
log ε
6.1 The Fractal Features of Wood Fracture 111
TL
Table 6.1 The fracture toughness KIC and the corresponding DS of the five woods
pffiffiffiffiffiffiffiffi
Species Sample TL
KIC =MPa mm Fractal dimension
number DS
Cunninghamia lanceolata 30 8.0 2.19
Picea jezoensis var. 30 13.8 2.21
microsperma
Larix gmelinii 30 19.5 2.22
Populus spp 30 20.6 2.23
Castanopsis hystrix 30 21.6 2.24
differences are also shown on the morphology of fracture surface parallel to grains.
So, the fracture toughness parallel to grains is the correlation to the fractal
dimension.
The fracture surface of CT sample with TL crack is the radial section in this
work. The morphological characteristics of fracture surface are mainly determined
by radiation speckle on the radial section, growth ring pattern, and wood grains.
The former three kinds of softwood (C. lanceolata, P. jezoensis, L. gmelinii) have
fine to extremely fine xylem ray, and belong to straight texture wood. So, the
fracture surface should be more smooth then that of hardwood with broad xylem
ray. While, in terms of solftwood, C. lanceolata is the gradual transition,
P. jezoensis is a little abrupt transition, and L. gmelinii abrupt transition. The growth
ring pattern of the above three are distinct successively. So, the fracture surface
roughness of the three increases successively. Populus spp is diffuse-porous wood,
and its structure is far more exquisite than that of C. hystrix with the xylem ray of
oak type, meanwhile, C. hystrix has cross grains, so the fracture surface of Populus
spp is flatter than that of C. hystrix. Because of the difference of microstructure the
five types of woods, the fractal dimension of the fracture surface ranges from small
to big as shown in Table 6.1 and it has a significant direct proportion (Fig. 6.7) to
the fracture toughness parallel to the grain:
TL
KIC ¼ 269:77DS 582:65 ðR2 ¼ 0:98Þ ð6:8Þ
20
15
10
IC
0
2 2.1 2.2 2.3 2.4 2.5 2.6
DS
112 6 Fractal Features and Acoustic Emission Characteristics of Wood Fracture
6.2.1 Introduction
technique to monitor the fracture of clear spruce wood under tensile loading and
found that there was an onset of AE prior to the first visible crack growth
step. Reiterer et al. [32, 33] used AE to monitor Mode I fracture of softwoods
(spruce and pine) and hardwoods (alder, oak, and ash) and stated that the AE counts
up to maximum force are much higher for the softwoods. Chen et al. [34] used AE
to monitor the failure process of hardwood and softwood test pieces under static and
fatigue torsion loading and found it was possible to monitor and analyze the failure
process in wood by AE techniques. Choi et al. [35] studied the fracture processes of
typical fiber-reinforced plastic composites laminates with continuous fiber rein-
forcement and the results showed that the AE characteristics might represent the
process of fiber breakages according to the various loading stages, which expressed
characteristic fracture processes for individual fiber-reinforced composite laminates.
The feature of the AE hit-event rate, in combination with AE amplitude classifi-
cations, could be utilized for nondestructive identification of different fracture
mechanisms.
But there are few studies on how to identify the different fracture modes with AE
on wood bending fracture. In this work, we study the AE characteristics during
wood three-point bending testing and try to identify or distinguish different modes
of wood fracture processes, and discuss the fracture evolution and mechanisms in
light of its microscopic structure with the help of double cantilever beam
(DCB) and compression tests.
6.2.2.1 Materials
10mm
F
Load
AE
AE Sensor
Sensor
240mm
300mm
6.2.2.2 Methods
When the wood sample was loaded at bending test, the load–deflection curve is
presented in three stages, i.e., (1) linear elastic deformation stage, (2) nonlinear
deformation stage, and (3) toughness fracture stage (Fig. 6.9a). Wood fracture is a
complex multilevel and multistage process. We will observe the molecular chain
rearrangement, slip, orientation, and fracture in the microscopic, at the same time,
the cracks of wood will growth and propagation and finally, fracture. All the
microscopic and macroscopic material structures changes come along with energy
saving and releasing. Therefore, apart from the AE cumulative counts and AE
6.2 Acoustic Emission Characteristics and Felicity Effect … 115
events peak amplitude Amax, the event energy and energy rate were adopted in this
analysis. The event energy was quantified by RMS2 where RMS was the root mean
square of the signal voltage calculated from
2 31=2
ZT
1
RMS ¼ 4 V 2 ðtÞdt5 ð6:9Þ
T
0
With t time, T duration of the AE event, and V the signal voltage value at
moment t.
Figures 6.9 and 6.10 show curves of force versus deflection and AE cumulative
counts versus Deflection compared with events RMS versus deflection and curves
of force versus deflection and AE counts versus deflection compared with energy
rate versus deflection for P. jezoensis and C. hystrix for the standard sample under
three-point bending tests. There was no AE count in the linear elastic deformation
stage for a standard sample, i.e., there was no damage and it could come back.
When it entered into the nonlinear deformation stage and because the compressive
strength was less than the tensile strength of wood, the compressed area started to
yield and the neutral area was offset to drawing area to maintain the overall balance
(Fig. 6.11), the curve of force versus deflection was nonlinear. There was only a
small number of AE signal in the front half of this stage (the difference depends on
the different tree species, the texture orientation, and the threshold value because the
helically wound cellulose cell wall reinforcement extends elastically within the
matrix of hemicellulose and lignin) and in the second half of the stage, the rapid
increases in AE signal due to the cell interface and interlaminar shear (and layer
resulted from the slip and shear within the molecular chain of cellulose). Then, it
entered into the toughness fracture stage and it came along with the high-energy
elastic wave produced by the fiber fracture and pull-out. Usually, the wood beam
sample kept on the integrality to certain load ability in the post-fracture period.
Along with the sample convex bending of fiber to the continuous tension, the
tension stress micro-fracture damage zone expanded and saved the energy. External
(a) (b)
AE cumulative counts
AE cumulative counts
Events RMS (mv)
Force (kN)
Energy rate
0 2 4 6 8 10
0 2 4 6 8 10
Deflection (mm)
Deflection (mm)
Fig. 6.9 a The curves of force versus deflection and AE cumulative counts versus deflection
compared with events RMS versus deflection and b the curves of force versus deflection and AE
cumulative counts versus deflection compared with energy rate versus deflection for Picea
jezoensis for standard sample under three-point bending tests
116 6 Fractal Features and Acoustic Emission Characteristics of Wood Fracture
(a) (b)
AE cumulative counts
AE cumulative counts
Events RMS (mv)
Force (kN)
Energy rate
Force (kN)
Fig. 6.10 a The curves of force versus deflection and AE cumulative counts versus deflection
compared with events RMS versus deflection and b the curves of force versus 0 deflection and AE
cumulative counts versus deflection compared with energy rate versus deflection for Castanopsis
hystrix for standard sample under three-point bending tests
power was absorbed before pull-out of the fiber cluster fracture, and which would
be released as the high-energy wave with the pull-out of the fiber cluster fracture.
We would find the interlocking form from the fracture surface in the tension stress
zone, but it was almost flush in the compression stress zone, similar to brittle
fracture (Fig. 6.12), which resulted from the reduction of the anti-break strength
due to the crushing loss of the wood cells in this zone.
Figures 6.13 and 6.14 show curves of force versus deflection and AE cumulative
counts versus deflection compared with events RMS versus deflection and curves of
force versus deflection and AE cumulative counts versus deflection compared with
energy rate versus deflection for P. jezoensis and C. hystrix for notched sample
under three-point bending tests. The notched wood sample was also presented as
three stages in bending failure process. But a large quantity of low-amplitude
low-energy AE was generated when the load increases to about 30–50% of the
maximum force Fmax (Figs. 6.13a and 6.14a), at the same time, lateral cracks would
be found around the visible crack tip on the sample surface and the force–deflection
curve would present a salient point due to the stiffness changes. Figure 6.15 shows
the cell interface and interlaminar shear of lateral cracks for P. jezoensis. The lateral
cracks expanded parallel to the wood grain in the interlayer and with the increase of
load the cracks expanded slowly and eventually stopped. And about a 20 mm high
new beam section was formed behind the original transverse crack, similar to the
standard sample (Fig. 6.16). And then the AE characteristics of the notched sample
fracture behavior were similar to the standard sample. It showed good toughness.
There were no significant differences of the bending strength in the statistical sense
between the standard sample and the notched sample after deducting the prefab-
ricated sharp crack (Table 6.2), and once again showed that sample containing
crack perpendicular to the wood grain would not produce low-stress rupture phe-
nomenon because of the crack tip stress singularity.
6.2 Acoustic Emission Characteristics and Felicity Effect … 117
σ− σ b− σ b−
Original
h/2 Compression neutral layer
stress Zone
σ+ σ+ σ b+
Fig. 6.11 Sketch of center axial of beam was moved to pulling side on the bending process
10mm 350μm
Fig. 6.12 Scanning electron microscopy observation of the fracture surface of the standard
sample of Picea jezoensis: a the toughness fracture zone and b brittle fracture zone
(a) (b)
AE cumulative counts
AE cumulative counts
Events RMS (mv)
Force (kN)
Force (kN)
Energy rate
0 2.5 5.0 7.5 10.0 12.5
0 2.5 5.0 7.5 10 12.5
Deflection (mm) Deflection (mm)
Fig. 6.13 a The curves of force versus deflection and AE cumulative counts versus deflection
compared with events RMS versus deflection and b the curves of force versus deflection and AE
cumulative counts versus deflection compared with energy rate versus deflection for Picea
jezoensis for notched sample under three-point bending tests
(a) (b)
AE cumulative counts
AE cumulative counts
Energy rate
0 1.5 3.0 4.5 6.0 7.5 0 1.5 3.0 4.5 6.0 7.5
Deflection (mm) Deflection (mm)
Fig. 6.14 a The curves of force versus deflection and AE cumulative counts versus deflection
compared with events RMS versus deflection and b the curves of force versus deflection and AE
cumulative counts versus deflection compared with energy rate versus deflection for Castanopsis
hystrix for notched sample under three-point bending tests
Fig. 6.16 Sketch of bending process on notched sample: a the formation of lateral cracks at crack
tip; b the formation and expansion of collapse area; c the fiber fracture layer by layer of the tensile
zone brittle fracture of the collapsed zone
Table 6.2 Bending strength in the statistical sense between the standard sample and the notched
sample
Species Sample type Number of Average value SE CV
samples (MPa) (MPa) (%)
Picea asperata Standard 22 74.95 11.06 14.76
sample
Notched 22 78.23 9.62 12.30
sample
Castanopsis Standard 16 122.97 13.26 10.78
hystrix sample
Notched 16 118.08 10.88 9.21
sample
SE standard error; CV coefficient of variation
(a) (b)
Amplitude (mV)
Amplitude (mV)
Fig. 6.17 The time-domain curve of different damage models: a interlaminar fracture; b cell wall
fracture
120 6 Fractal Features and Acoustic Emission Characteristics of Wood Fracture
expansion of micro-fracture damage area, and cell wall fracture. Figure 6.17 shows
the typical time-domain curve of interlaminar fracture and transverse fracture of the
notched sample for P. jezoensis combined with Figs. 6.13a and 6.14a, the char-
acteristics of AE signal in different damage and fracture models are summarized as
follows:
1. The characteristics of AE signal for interlaminar fracture stage (i.e.,: cell wall
damage evolution and spallation) is low amplitude, short duration, low-count,
and low energy;
2. The AE amplitude of wood cell compressive yields stage (i.e.,: cell wall
buckling and collapse) is lower or shows lower energy. And signal usually does
not count when a signal of the probe is lower threshold value because of the
attenuation of AE wave propagation;
3. The AE signal of wood macroscopic fracture stage (i.e.,: fiber bundle fracture
and pull-out) is high amplitude, long duration, high-energy, and high AE counts;
4. The AE signal characteristic with the correspondence of the formation and
expansion of the micro-fracture damage area is more complicated and it is the
prelude to damage fracture and also exists in the whole period. What makes it
differ from an interlaminar fracture is that the former mainly occurs in type I or
peeling cracking in the cell or the cell layer, while the latter is caused due to the
whole material damage by formation, expansion, and connection of the
micro-fracture caused by the cell wall tear or rupture. The AE characteristic,
which lies between interlaminar fracture and transverse fracture, is affected by
the relative density, cell wall thickness, microscopic structure, the number of
damage, etc.
Although we have the above understanding, it is still very difficult to distinguish
AE signal from different damage models in bending process. As wood is a multi-
level cell structure of biological composite materials, there is always a variety of
deformation and damage that change system energy in the same stage in the near
crack tip process zone. Therefore, double cantilever beam (DCB) experiment along
the grain of wood cracking and the compression experiment along the longitudinal
and transverse have been carried out (Fig. 6.18). And the DCB test will produce
Mode I interlaminar fracture and the compression test makes cell wall buckling and
collapse damage.
Force
(a) Force
(b) (c)
Force
Force
Force Force
AE Sensor AE Sensor
The energy release is irreversible in the process of material damage. And the same
is with AE, which is known as Kaiser effect. Kaiser effect has been described as
follows: AE only appeared obviously again when the material is loaded to the
previous unloading point. It is first considered that Kaiser effect exists in the metal
and rock. And then it is discovered that Felicity effect exists in the composite
materials and damage metal materials. Felicity effect refers to the phenomenon that
AE appears when reloading is lower than the previous loading. Felicity ratio is
defined as the ratio of repeated loading to the previous.
As a quantitative parameter, it can better reflect the extent of original suffered
material damage or the structure deficiencies. The smaller the Felicity ratio is, the
more serious of the damage or structure deficiencies is. Felicity ratio more than 1
means there exists Kaiser effect, on the contrary, Felicity effect.
Figure 6.19 is the curve of cyclic load bending test and the analysis curve of a
notched sample of P. jezoensis. There were eight times of loading and unloading in
the whole test. The first loading and unloading occurred the beginning of the
interlaminar fracture and the Felicity ratio was 0.99, which meant there existed
instability of the structure defect.
Felicity ratio was more than 1 in the second–sixth loading and unloading
because of the stress redistribution of samples as the lateral crack expansion and
gradually stable, but there was a decline trend close to maximum load. The seventh
122 6 Fractal Features and Acoustic Emission Characteristics of Wood Fracture
(a) (b)
AE cumulative (counts)
AE cumulative (counts)
Load (kN)
(c)
Felicity ratio
Loading-unloading times
Fig. 6.19 The curve of cyclic load bending test and the analysis curve of notched sample of Picea
jezoensis: a the curves of load versus time and AE cumulative counts versus time and b AE
cumulative counts versus load and c felicity rate versus loading–unloading times
loading and unloading process occurred nearby the maximum load and the eighth
loading and unloading process occurred in the third stage of toughness fracture, all
meant the material structure had been suffered serious damage. The repeated load
test behavior of the standard sample was almost similar to the notched sample. But
the Felicity ratio was more than 1 when the load was close to the maximum load in
the whole process. Therefore, the bending test of wood in low load or the most of
stage II showed Kaiser effect and in high load or toughness fracture stage presented
Felicity effect. As a result, AE technology could better monitor the wood damage in
the practical application.
It needed some clarification that the Felicity phenomenon would be different for
different species and threshold voltage. AE signal would also be produced due to
the friction of the microcracks section in the process of loading–unloading, so in
order to preclude the interference of “false damage”, the threshold voltage should
be set higher than normal. In the loading–unloading of this experiment threshold
voltage was set as 0.39 mV (about 52 dB).
References 123
References
26. Bakuckas JG, Prosser WS, Johnson WS (1994) Monitoring damage growth in titanium matrix
composites using acoustic emission. J Compos Mater 28(4):305–328
27. Katsaga T, Sherwood EG, Mp Collins et al (2007) Acoustic emission imaging of shear failure
in large reinforced concrete structures. Int J Fract 148:29–45
28. Bucur V (1995) Acoustics of wood. CRC Press, Boca Raton
29. Schniewind AP, Quarles SL, Lee H (1996) Wood fracture, acoustic emission, and the drying
process Part 1. Acoustic emission associated with fracture. Wood Sci Technol 30:273–281
30. Aicher S, Höfflin L, Dill-Langer G (2001) Damage evolution and acoustic emission of wood
at tension perpendicular to fiber. Holz als Roh Werkst 59:104–116
31. Dill-Langer G, Aicher S (2000) Monitoring of microfracture by microscopy and acoustic
emission. In: Proceedings international conference Wood and wood fiber composites,
Stuttgart, pp 93–104
32. Reiterer A, Stanzl-Tschegg SE, Tschegg EK (2000) Mode I fracture and acoustic emission of
softwood and hardwood. Wood Sci Technol 34(5):417–430
33. Reiterer A, Sinn G, Stanzl-Tschegg SE (2002) Fracture characteristics of different wood
species under mode I loading perpendicular to the grain. Mat Sci Eng A Struct 332:29–36
34. Chen Z, Gabbitas B, Hunt D (2006) Monitoring the fracture of wood in torsion using acoustic
emission. J Mater Sci 41:3645–3655
35. Choi NS, Woo SC, Rhee KY (2007) Effects of fiber orientation on the acoustic emission and
fracture characteristics of composite laminates. J Mater Sci 42:1162–1168
Chapter 7
Mechanical Characteristics of Bamboo
Structure and Its Components
Abstract In this chapter, first, the fine structure of bamboo is stated. Then the
strength and elastic modulus of bamboo fiber and ground tissue are analyzed by
mixture law and shear lag theory of meso-mechanics, and it is proved that bamboo
fibers are the main component determining the mechanical characteristic of bam-
boo. Furthermore, results of tensile tests on separated bamboo fiber bundles show
that the tensile strength of bamboo fiber obtained from the tests on bamboo blocks
is higher than that on separated fiber bundles. This might be due to the interaction
between components in bamboo in which parenchymatous ground tissue can pass
loads and uniform the stresses loaded on fibers. And the differences of structure
and strength between internodes part and node part of moso bamboo are studied.
The results indicate that either in the non-epidermis-planning samples or in
the epidermis-planning samples, the node does not take a declining effect on the
bending strength, the longitudinal shearing strength, and the compressive strength.
Instead, the node takes a reinforcing effect at different degrees. However, the node
has a significant decline effect on the longitudinal tensile strength. By analysis on
the structure of bamboo, it shows that though the vascular tissue passing bamboo
node is curved and discontinuous at different degrees, bamboo is able to increase
the bearing load area by expanding tissue at node, meanwhile, vascular bundles are
thickened and entwined circuitously. So, the ability to anti-bending and
anti-shearing of the culm under transverse load caused by snow or wind is increased
by bamboo nodes. However, the tensile load born by the culm is less than other
parts of culm during growth of bamboo. Therefore, the enlargement of node part
and the structure evolution of nodes cannot enhance the tensile strength in the
longitudinal direction.
7.1 Introduction
Through million years’ natural selection of the process of survival of the fittest, the
fine structure of bamboo has been evolved to adapt to the living environment. As
bamboo may encounter bending damage caused by wind or load during the process
of growth, various specific internal structures are evolved to intensify the ability to
resist damage. For example, the hollow circular section and stepwise changing
section of bamboo stem reflect the function selection of “equal strength bar” that
each part of bamboo stems has the same ability to resist bending deformation
(Fig. 7.1). Bamboo nodes appear at intervals of about 10 cm along the stem.
A connection between the longitudinal and the cross direction is made by the
crossing walls of nodes, which enables bamboo to achieve optimal stiffness and
stability with the least material and increases the capacity of bamboo to resist
transverse compression and shear. Although some discontinuity and different
extents in bending of the vascular tissue may affect the strength of bamboo node,
(a) (b)
Fig. 7.2 a The cross section of bamboo stem; b distribution relation between the volume fraction
of fibers (Vf) and the location parameter (x/B) on cross section of moso bamboo
the vascular bundles in node are thickened, thus the thickness of bamboo node is
always greater than that of the adjacent internodes to avoid damage at node under
external load. What’s more, there are vascular bundles procumbent circumferen-
tially or interweave together in bamboo node, which will improve the transverse
property of node.
In bamboo internodes wall most of the cells are arranged along the longitudinal
direction, and there is no transverse ray cell like that in wood [27], so bamboo
internodes wall is a typical unidirectional long-fiber-reinforced biocomposite. The
distribution of reinforcements is sparse in the inner part then denser gradually in the
outer part of bamboo wall. As seen in Fig. 7.2, exponential distribution relation is
found between the volume fraction of fibers (Vf) and the location parameter (x/B) on
cross section of moso bamboo (Phyllostachys pubescens), which is corresponding
to the bending load caused by wind or snow and according to the theory of function
adaptability.
Bamboo fiber has a complex fine structure with multilayer. Just like other plant
fibers, cellulose chain molecule of bamboo fiber is composed of carbon, hydrogen,
and oxygen. Unit lattice is constructed by the cellulose chain molecule, then cel-
lulose basic fibril with a diameter of 35–50 nm constituted of unit lattices according
to Roelofsen structure model [28]. And it is easy for cellulose basic fibrils to gather
together to form microfibrils with larger diameter which will be embedded in the
matrix composed of hemicellulose and lignin to construct lamella structures, then
they are arranged concentrically to form bamboo cell wall alternating with broad
lamella and narrow lamella. The structures of bamboo fiber cell walls can be
divided into two types. One is multilayer structure consists of narrow lamella (4–5
lamellas) alternating with broad ones. The other type has thick cell wall with narrow
cell lumina, and there are three lamellas in the secondary wall, one narrow outer
lamella and two broad inner lamellas.
In broad lamella, the microfibrils are in 10–30° to the axial direction of fiber in
right-handed helix distribution, which determine the physical and mechanical
property along the axial direction. In the narrow lamella, the microfibrils are in
7.2 Mechanical Characteristics of Bamboo 129
40–90° to the axial direction, which affect the transverse physical and mechanical
property of fiber. The ground tissue of bamboo wall consists of parenchymal cells
that distribute among vascular bundles to buffer and transmit load. For parenchymal
cell, the cell wall is easy to buckle under pressure, while, as liquid and solid coupled
cell body material, the performance of the living ground tissue can be strengthened
by the internal hydrostatic pressure.
The structure and property of natural living body may provide human many
inspirations. Bamboo has possessed the macro- and micro-structure that adapts to
the living environment best and gives full play to its wonderful and complex
function. The fine structure of bamboo plays a very important role in the bionics
design of artificial composites.
Tensile tests parallel to grain were performed at two levels. One is at macroscopic
level, i.e., on bamboo blocks; the other is at mesoscopic level, i.e., on bamboo fiber
bundles. Meanwhile, fiber volume fraction was also measured on transverse sec-
tions. The moso bamboo (Phyllostachys pubescens) used in for the tests was 4 years
old and acquired from Lujiang, Anhui Province, China. The total height was about
15 m, and the diameter at breast height was 125 mm. The average air-dry density
was 0.71 g/cm3.
7.3.1.1 Material
Specimens were taken from two adjacent internodes at the height of about 1.8 m
and the thickness of culm wall was about 10 mm. After air dried, the culm sections
were longitudinally split into 200 mm (L, longitudinal) 10 mm (T, tangential)
strips. Then, they were delaminated from the outer to the inner part of the culm
walls, after the cortex and pith periphery were removed. About six specimens were
obtained for each strip. The final shape and dimension of the specimens for tensile
tests are shown in Fig. 7.3. The dimension of the effective experiment part (middle
section) was 80 mm (L) 5 mm (T) 1.5 mm (R, radial). There are 60 speci-
mens in total with a moisture content of 12.5%.
130 7 Mechanical Characteristics of Bamboo Structure and Its Components
Displacement sensor
5mm R=1.5 mm
T=10mm
50mm
Marked length
35mm 25mm 80mm 25mm 35mm
Effective experiment part
L = 200 mm
Fig. 7.3 Shape and dimension of bamboo block used for tensile test
7.3.1.2 Theory
When the radial thickness of the specimen collected from the culm wall is not so
big, the vascular bundles can be considered as being uniformly distributed.
Normally, fibers have high strength and modulus of elasticity (MOE), while ground
tissue has low strength and MOE and high capability of deformation. It was found
that in the tensile stress (r)-strain (e) curve of bamboo block, the part from the
beginning of loading to the breaking point was almost linear. Therefore, in view of
mechanical behavior, the bamboo block can be simplified as a composite of parallel
connection model composed of two elements, i.e., fibers and ground tissue, as
shown in Fig. 7.4. It is considered that fibers and ground tissue have the same
deformation.
FC, Ff, Fm, rC, rf, and rm are noted as tensile force and stress on composite,
fibers, and matrix, and the corresponding strain and MOE are eC, ef, em EC, Ef, and
Em. The cross-sectional area of composite, fibers and matrix are AC, Af, and Am, so
the volume fractions of fibers and matrix can be calculated as Vf = Af/AC and
Vm = Am/AC, and Vf + Vm = 1. Thus, the following equations can be derived:
FC ¼ Ff þ Fm ð7:1Þ
Or
rC AC ¼ rf Af þ rm Am ð7:2Þ
rC ¼ rf Vf þ rm Vm ð7:3Þ
eC ¼ ef ¼ em ð7:4Þ
rC AC rf Af rm Am
¼ þ ð7:5Þ
eC AC eC AC eC AC
In the range of linear elasticity, the relationships among EC and MOE, volume
fraction of fibers, and matrix are
EC ¼ Ef Vf þ Em Vm ¼ Ef Vf þ Em ð1 Vf Þ ð7:6Þ
Equations (7.3) and (7.6) are the “Mixture law” of composite meso-mechanics
[29, 30].
V f = 0.204
V f = 0.115
132 7 Mechanical Characteristics of Bamboo Structure and Its Components
Fig. 7.6 Images used for fiber volume fraction (Vf) measurements. a, b, c, d, e, and f show a
series of cross sections’ images of the samples located along radial direction from inner part to
outer part of a culm wall and the corresponding values of Vf were 10.65, 14.33, 21.13, 27.30,
34.02, and 44.29%, respectively
were then measured by the image analysis software. Figure 7.6 shows an example
of a series of treated sections’ images of the samples located along the radial
direction from the inner to the outer part of a culm wall.
The ultimate stress and MOE of each bamboo block can be calculated, and then
the relationships among ultimate stress, MOE, and the volume fraction of fibers can
be obtained according to Fig. 7.7, as follows:
According to Eq. (7.7), the tensile strength and MOE of fiber and parenchy-
matous ground tissue were estimated as rf = 581.7 Mpa, Ef = 40.4 GPa,
rm = 19.0 MPa and Em = 0.22 GPa, respectively. These results suggest that fiber
plays the determinant role in tensile properties of bamboo and the ground tissue is
the low-strength-and-modulus matrix material.
To verify the feasibility of “Mixture law” used to calculate the strength and MOE of
fibers and analyze the load transmission function of ground tissue, the tensile
strength and MOE of fiber bundles separated from ground tissues are tested.
7.3.2.1 Material
Specimens were collected from the same culm walls where samples had been taken
for tensile tests on bamboo blocks. The fresh culm walls were immersed in water
7.3 The Mechanical Characteristics of the Components Bamboo 133
(a) (b)
Tensile strength (MPa)
MOE (GPa)
Fiber volume fraction Fiber volume fraction
Fig. 7.7 Relationships between fiber volume fraction (Vf) and bamboo tensile strength (a) and
MOE (b)
for one day and then split along the fiber direction into small strips. During the
splitting, some fiber bundles were detached from ground tissues and then pulled out
of the bamboo strips at an angle of 10° (the fiber bundle will be damaged or broken
if the angle is too large) between the detached fiber bundle and the bamboo strip.
The fiber bundles more than 60 mm long were chosen and air dried. Due to the
fragility of the small specimens and the difficulty of installation for the tensile test, a
special method was used with great care. The fiber bundle was fixed with strong
glue in a paper casing made of coordinate grid paper with two quadratic holes in the
middle of the two paper halves, as shown in Fig. 7.8. The paper was folded in the
middle, so that the fiber bundle was fixed with glue inside the paper casing except
for the middle part. This method protected the fiber bundle from being damaged or
broken during the installation on the tensile test machine. After being installed in
the clamps of test machine and before being loaded, the paper casing was snipped
through the middle part. Details are shown in Fig. 7.8. The tensile test conditions of
fiber bundles were the same as those for bamboo blocks. In total, 71 specimens of
fiber bundles were prepared.
Force
Paper Glue Fiber bundle Glue
Paper
20mm
Fiber Clamp
bundle
20mm
Cutting
line Scissors
20mm
Stress (MPa)
Strain
After the tensile test, the broken fiber bundle was stained with safranine, and then
air dried and embedded in resin. Cross sections were cut perpendicular to axial
direction of the fiber bundle. The cross area of fiber bundle as measured under
microscope with image analysis system. Therefore, tensile stress can be calculated.
Figure 7.9 shows an example of stress–strain curve where the part from the
beginning of loading to the breaking point was entirely linear. High rupture strength
and high MOE were observed.
However, there were two possibilities which might cause the fallibility of MOE
value calculated from the figure. First, the specimen, including the paper casing,
might have some movement inside the clamps of the test machine. Second, the
MOE of fiber bundle was much higher than that of the glue inside the paper casing
used for fixing fiber bundle. Thus, under tensile loading, the part of fiber bundle
covered by glue might have slight but not negligible strain, although no fiber bundle
was pulled out of the paper casing during the tests. Therefore, in order to estimate
the MOE of fiber bundle, according to “shear lag theory” of composites mechanics
[30], it was assumed that no movement of the specimen happened inside the clamps
of the test machine and the fiber bundle was covered uniformly by glue so that the
interface shearing stresses loaded on fiber bundle were distributed uniformly. As
shown in Fig. 7.10, Lg and L0 were noted as the lengths of the part of fiber bundle
inside glue and the middle part without glue, respectively. Af and Sf were noted as
the cross area and perimeter of fiber bundle, respectively, s as the shearing stress on
the interface of fiber bundle inside glue, r0 as the axial tensile stress on the middle
part of fiber bundle without glue. In order to keep the balance inside the glue, s and
tensile stress r(x) should meet the following equation:
Sf x s ¼ Af rðxÞ ð7:8Þ
For the middle part of fiber bundle without glue, s and r0 should meet the
following equation:
7.3 The Mechanical Characteristics of the Components Bamboo 135
x dx
σ
σo
σ(x)
x
Lg Lo Lg
Fig. 7.10 Schematic representation of stresses loaded on bamboo fiber bundle during the tensile
test
Sf Lg s ¼ Af r0 ð7:9Þ
Thus, the tensile stress r(x) loaded on position x far from the left end of the fiber
bundle should be
8 S xs
< Af ¼ Lg r0
> ð0 x Lg Þ
f x
The total length of fiber bundle under loading L = 2Lg + L0. Therefore, the total
length change of fiber bundle (DL) under the tensile stress should be
ZL ZLg ZL0
rf ðxÞdx rf xdx rf dx rf
DL ¼ ¼2 þ ¼ ðLg þ L0 Þ ð7:11Þ
E Lg E E E
0 0 0
where DL was measured as the displacement of the crosshead of the test machine,
Lg = L0 = 20 mm. Thus, according to Eq. (7.11), the MOE of fiber bundle can be
calculated.
The results of the tensile tests on fiber bundles are shown in Table 7.1. The
average tensile strength of bamboo fiber was 482.77 MPa, which was 18% less than
the value calculated by the tests on bamboo blocks. This is mainly due to the fibers
in bamboo blocks are embedded in ground tissue, which can pass loads and dis-
tribute the stresses loaded on vascular bundles. Such effectiveness is not possessed
by a single tissue. Besides, the damage on the surfaces of some fiber bundles may
be produced during separation process of fiber bundles. This could cause the low
136 7 Mechanical Characteristics of Bamboo Structure and Its Components
Table 7.1 The results of tensile tests on the separated bamboo fiber bundles
Mean Number of Standard Coefficient of
samples deviation variation (%)
Tensile strength 482.179 71 108.32 24.11
(MPa)
MOE (GPa) 33.853 71 8.16 22.44
tensile strength measured on bamboo fibers. The average MOE of bamboo fibers
was 33.85 GPa, which was also 16% less than the value measured by the test on
bamboo blocks. Such a low value may be caused by the discrepancy between the
hypothesis and the practice. During the tensile tests on fiber bundles, some slight
movements of the specimens and some slight shearing deformations of the glue
might have happened inside the clamps of the test machine. These could influence
the result of MOE of bamboo fiber measured on fiber bundles.
Figure 7.11 shows the tension fracture pictures of three specimens with different Vf.
The corresponding stress–strain curves of the three specimens are shown in
Fig. 7.5. The fracture surface of the specimen with low Vf (11.5%) was trimmed,
which indicates that the specimen was brittle. The specimen with medium Vf
(20.4%) showed an uneven fracture surface, which was caused by the interaction of
parenchymatous ground tissue and fibers. When the Vf (49.3%) was high, fracture
surface of the specimen was split along the longitudinal direction, meaning that the
specimen had high toughness.
The micrographs of the fracture surfaces obtained by scanning electron micro-
scope are shown in Fig. 7.12. The left part of Fig. 7.12a shows the ground tissue and
the right part shows the fibers. Figure 7.12b, c shows the ground tissue and fibers,
respectively, with bigger magnification. In Fig. 7.12b, the incompact multilayer
thin-walled structure of ground tissue can be seen clearly. By axial tensile stress
action, the cell walls got fractured along the cross section. Like the matrix in artificial
composites, ground tissue can pass and buffer loads among vascular bundles.
However, under loads, the cell walls of ground tissue are prone to buckle. However,
in live bamboo tissue, the buckling resistance of the cell walls would be reinforced
by the effect of the interior static hydraulic pressure caused by water. The fracture
surface of fibers was dense but rough in general (Fig. 7.12a). In the individual fibers,
different layers were also rough (Fig. 7.12c). This indicates the low strength of the
interface bond between different fiber layers and among different fibers
Fig. 7.12 Micrographs of the fracture surfaces obtained by scanning electron microscope:
a surface including parenchymatous tissue (left) and sclerenchyma fibers (right); b fracture surface
of parenchymatous tissue; c facture surface of sclerenchyma fibers
138 7 Mechanical Characteristics of Bamboo Structure and Its Components
notwithstanding the high tensile strength of bamboo fibers. Furthermore, for the
high-modulus bamboo fibers, the multilayer structure can improve the flexibility.
Bamboo internodes tissues are arranged along the longitudinal direction and are
conglutinated together by the non-cellulose component in an effective but somehow
not very effective manner, which makes the strength of the interface much weaker
than that of bamboo longitudinal direction. However, it is the weakened interface
that causes the high fracture toughness in transverse direction of bamboo [31],
making bamboo tolerant toward gnawing of wild animals and knife wound.
Meanwhile, the deficiency of the weak interface strength of culm wall can be offset
by the nodes. The unique macrostructure of a large cavity surrounded by culm wall
and the increasing distribution density of vascular bundles from the inner to the
outer part of the culm wall along the radial direction make bamboo adapt to bending
load caused by snow or wind. This reflects the optimization of nature selection, and
the structure formed is in accordance with the principle of most efficiency with least
materials.
Six moso bamboos were selected from the Anhui Province in China. They were 3
years old and about 15 m high. The diameter at breast height (dbh) varied from 110
to 125 mm and the average thickness of the culm wall was 12 mm at dbh. After
cutting, the bamboo stems were air dried in the laboratory. The moisture content of
the dried samples was about 12% and the average air dry density of internodes of
moso bamboo was 0.712 g/cm3.
The structure differences between internode and node of moso bamboo were
compared and analyzed specifically on the macroscopic forms and distribution of
fiber bundles. Mechanical properties of bamboo were tested according to the China
National Standard [32]. Tests conducted were longitudinal tensile strength, bending
strength, longitudinal shear strength, and compressive strengths across and parallel
to grain. Two types of moso bamboo were used, namely, that with culm intact (i.e.,
intact sample) and the other, with the innermost and outermost parts of culm
removed by double planing (i.e., planed sample). Samples were then divided into
four groups which were intact samples with nodes and without nodes, and planed
samples with nodes and without nodes. When preparing materials for samples with
nodes, the nodes were positioned at the center of the samples. Samples with and
without nodes were taken from neighboring sections of the same culm (Fig. 7.13),
thus, giving an almost uniform thickness of samples. All experiments were carried
out by a mechanical testing machine controlled by a computer. The speed of
crosshead was 2 mm/min. The laboratory temperature was between 15 and 18 °C,
and the humidity of the room was 60–65%.
7.4 Difference of Structure and Strength Between Internodes … 139
Fig. 7.13 Sketch of the testing samples produced from bamboo culm
For internodes of culm, it consisted of three parts, bamboo skin, middle part, and
pith ring tissue. In the culm wall, the numbers of vascular bundles increase from the
inner to the outer part of the culm, while the ground tissues decrease on the
contrary. Bamboo cell between two nodes are strictly oriented axially (Fig. 7.14),
and transverse ray cells that normally present in wood structure have not been found
here.
The node part of bamboo culm consists of a sheath scar, nodal ridge, diaphragm,
and the intra-node between nodal ridge and sheath scar [1, 2, 17]. They strengthen
to the culm erecting and conduct water. Except for the vascular tissues of the culm’s
outmost layer which are broken when bamboo primary leaf falls, most vascular
bundles are curving at different degrees when they thread a node. One part of
vascular bundles bends inward or outward, then spread along the original longi-
tudinal direction. Another part of vascular bundles changes its former direction to
transversely spread in node part; still some vascular bundles come into the dia-
phragm around periphery of the diaphragm or interweaved reticular formation. The
vascular bundles in node are thickened, thus the thickness of bamboo node is
always greater than that of the adjacent internodes (Fig. 7.14).
Bamboo blocks with node and without node were obtained from neighboring culm
and cleaved into 10 mm width bar first. And then, tensile samples were processed at
numerical control milling machine according to China national standard GB/T
15780-1995 (Fig. 7.15). The tensile samplers were divided into two types, two
groups in each type. For one type of samples, the innermost and outermost parts
140 7 Mechanical Characteristics of Bamboo Structure and Its Components
Fig. 7.14 Comparison between internodes culm and culm with a node: a radial section; b cross
section of a bamboo internodes A—A; (c) cross section of a bamboo node B—B
were milled, namely epidermis-planning sample, and the other type keeping intact
clum, namely non-epidermis-planning sample. There are 21–26 samples in each
group, and node is at the middle of sample.
Breaking tensile loads of intact samples with node and without nodes were
2435.7 and 2987 N respectively (Table 7.2); the existence of node decreased the
load by 18%. The tensile strengths for planed samples with node and without node
were 102.7 and 154.24 MPa, respectively. The strengths of samples with nodes
were lower than samples without nodes in a range of 33%. The failure surface of
samples with nodes showed that the break started from both sides of the nodes
(Fig. 7.16). This could be due to the discontinuity of vascular tissues when they
cross the node of the bamboo. Samples without nodes split longitudinally and
showed a typical ductile fracture.
Thickness
R120
2
10
80 30 60 30 80
grain
141
142 7 Mechanical Characteristics of Bamboo Structure and Its Components
Fig. 7.16 Tensile failure of a specimen without a node and a specimen with a node
groups with node and without node. The load bearing capacity is different along
different directions of sample, and generally the maximum force along radial direction
to inner Fmax,radial,inner is lower than that along radial direction to outer Fmax,radial,outer.
Prediction shows that for the intact sample with Tangential width
Thickness = 10 mm 10 mm, the strength parameters of bending sample without
node meet the following relation:
rmax;radial;inner þ rmax;radial;outer
rmax;tangential ð7:12Þ
2
The maximum loads of bending sample with node meet the following relation:
Fmax;radial;inner þ Fmax;radial;outer
Fmax;tan gential ð7:13Þ
2
Here, the direction of bending load on the sample was along tangential according
to GB/T 15780-1995 [32].
The descriptive statistics and variance analysis of the test results are shown in
Table 7.2. Maximum bending loads of intact samples with nodes and without nodes
were 1047.9 and 850.8 N, respectively, a difference of 23%. The bending strengths
for planed samples with nodes (155.7 MPa) and without nodes (150.96 MPa) were
similar. This suggests that the anti-bending ability of the node part in the intact
culm is enforced by swelling of local tissues at the bamboo node. Thus, since nodes
do not cause any negative effects on bending strength of planed samples, they
should not be considered as defects in industrial utilization.
Fig. 7.17 Bending failures: a culm without nodes and a culm with a node
7.4 Difference of Structure and Strength Between Internodes … 143
The dimension of shearing test sample was shown in Fig. 7.18. Because of augment
of local tissues at bamboo node of non-epidermis-planning samples, only the
maximum load Fmax could be compared between groups with node and without
node. The epidermis-planed sample could be compared with the shearing stress
r = Fmax/(bl) between groups with node and without node, while l and b were
length and width of shearing section, respectively.
Descriptive statistics and variance analysis of the shearing test data were listed in
Table 7.2. Maximum shearing loads of intact samples with and without nodes were
2516 and 1954 N, respectively; the shearing strength of samples with nodes was
higher by 29%. For planed samples, shearing strengths were similar, i.e., 18.9 and
18.3 MPa for nodes and without nodes, respectively. This showed that nodes could
enforce the anti-shearing ability of bamboo culm by swelling of local tissues at
nodes. For planed samples, nodes did not reduce the shearing strength. Tearing
failures of the samples are shown in Fig. 7.19. The periphery decumbent fiber
bundles were not found in the culm; only a few fiber bundles were detected at the
node. Failure plane in shear test is generally caused by slipping of the two planes
comprising ground tissues of fibers. The snipping of fiber bundles in the cross
direction did not contribute to the failure. Hence, effects of nodes on shear strength
are almost negligible for the processed samples.
20
10
11
144 7 Mechanical Characteristics of Bamboo Structure and Its Components
Fig. 7.19 Shear failure for culm a without node and b with node
Descriptive statistics and variance analysis of the experimental data were listed
in Table 7.2. Concerning the intact samples, the maximum compressive load par-
allel to grain of samples with node and without node was 14.7 and 13.7 N,
respectively. Compressive strength parallel to grain for planed epidermis samples
with node and without node was 59.8 and 56.4 MPa, respectively. Apparently, no
matter whether bamboo culm epidermis was planed or not, the compressive strength
or load of samples with node were slightly larger than that of samples without node.
The reason of this difference could be deduced from the higher proportion of
bamboo fibers in the node area (Fig. 7.14).
Because the load applied on the sample in the transverse compression test is along
radial direction of sample, the epidermis of bamboo samples must be planed. The
dimension of sample was 20 mm (Longitudinal) 20 mm (Tangential) 6 mm
(Thickness). The transverse compressive strength of samples with node and without
node was 24.9 and 18.8 MPa, respectively. The transverse compressive strength of
samples with node was 32% higher than that of samples without node. The dif-
ference between the two kinds of samples was significant. The reasons of this
difference could also be explained by a higher proportion of bamboo’s fibers in the
node area.
To sum up, the experiments of the bamboo mechanical properties have indicated
that nodes have not given any negative effects on tensile strength, bending strength,
and compressive strength parallel or transverse to grain. On the contrary, nodes
have the positive effect on the mechanical properties at different degrees. For intact
samples, nodes enhanced bending maximum load, shearing maximum load, and
compressive maximum load by 23, 19, and 7%, respectively. But the enhancement
effect initiated by nodes was not obvious for epidermis-planning samples. However,
negative effects of nodes on the tensile strength parallel to grain, for the both types
7.4 Difference of Structure and Strength Between Internodes … 145
of samples, i.e., intact and epidermis-planning samples, are significant, in which the
resistance decreased by 18 and 33%, respectively.
Bamboo is a biological compound material reinforced by long natural fibers
which give it a high resistance. Different extents in bending and some discontinuity
of the vascular tissue were observed at bamboo nodes. However, tumefaction of
bamboo nodes could increase the loading area and vascular bundles could be
thickened, circuitous, flexuose, and entwined. So, the ability to anti-bending and
anti-shearing of the culm under transverse load is increased by bamboo nodes.
However, the tensile load of the culm was relatively low during growth of bamboo.
Therefore, enlargement of nodes and the development of nodes could improve
tensile strength in the longitudinal direction.
References
1. Liese W (1998) The anatomy of bamboo culms. International network for bamboo and rattans
(INBAR), Beijing, Tech rep no. 18
2. Mohmod AL, Mustafa MT (1992) Variation in anatomical properties of three Malaysian
bamboos from natural stands. J Trop For Sci 5(1):90–96
3. Chuma S, Hirohashi M, Ohgama T et al (1990) Composite structure and tensile properties of
mousou bamboo. Zairyou 39:847–851 (in Japanese)
4. Inokuchi Y, Fushitani M, Chuma S et al (1997) Effects of volume fraction of bundle sheath on
the vibrational properties of bamboo. Mokuzai Gakkaishi 43(5):391–398 (in Japanese)
5. Inokuchi Y, Fushitani M, Kubo T et al (1999) Effects of water extractives on the
moisture-content dependence of vibrational properties of bamboo. Mokuzai Gakkaishi 45
(2):77–84 (in Japanese)
6. Inokuchi Y, Fushitani M, Kubo T et al (2002) Effects of volume fraction of bundle sheath and
water extractives on bending creep behavior of bamboo under changing moisture conditions.
Mokuzai Gakkaishi 48(6):413–424 (in Japanese)
7. Sui S, Li R (2003) A study of structure and performance of bamboo fibers. J Text Res 24
(6):535–537 (in Chinese)
8. Ahmad M, Kamke FA (2005) Analysis of calculate bamboo for structural composite
materials: physical and mechanical properties. Wood Sci Technol 39(6):448–459
9. Obataya E, Kitin P, Yamauchi H (2007) Bending characteristics of bamboo (Phyllostachys
pubescens) with respect to its fiber–foam composite structure. Wood Sci Technol 41:385–400
10. Deshpande AP, Rao MB, Rao CL (2000) Extraction of bamboo fibers and their use as
reinforcement in polymeric composites. J Appl Polym Sci 76:83–92
11. Jiang ZH, Sun ZJ, Ren HQ (2006) Application of advanced bio-composites in wind blades.
Acta Mater Compos Sin 23(3):127–129 (in Chinese)
12. Sun ZJ, Chen Q, Jiang ZH (2008) Processing and properties of engineering bamboo products.
Acta Mater Compos Sin 25(1):80–83 (in Chinese)
13. Jain S, Kumar R, Jindal UC (1992) Mechanical behaviour of bamboo and bamboo composite.
J Mater Sci 27:4598–4604
14. Zhao QS (1995) Industrial utilization of bamboo in China. China Forestry Press, Beijing, p 3
15. Nugroho N, Ando N (2000) Development of structural composite products made from
bamboo I: fundamental properties of bamboo zephyr board. J Wood Sci 46:68–74
16. Nugroho N, Ando N (2001) Development of structural composite products made from
bamboo II: fundamental properties of laminated bamboo lumber. J Wood Sci 47:237–242
17. Jiang ZH (2007) Bamboo and Rattan in the World. China Forestry Press, Bejing, pp 2–4
146 7 Mechanical Characteristics of Bamboo Structure and Its Components
18. Lakkad SC, Patel JM (1980) Mechanical properties of bamboo, a natural composite. Fiber Sci
Technol 14:319–322
19. Mohmod AL, Ariffin WTW, Ahmad F (1990) Anatomical features and mechanical properties
of three Malaysian bamboos. J Trop For Sci 2(3):227–234
20. Xian X, Xian D (1990) The relationship of microstructure and mechanical properties of
bamboo. J Bamboo Res 3:10–23
21. Mohmod AL (1993) Effects of age and height of three bamboo species on their machining
properties. J Trop For Sci 5(1):90–96
22. Yao X (1993) The microstructure of main Chinese bamboos. China Forestry Press, Beijing,
pp 6, 18
23. Suzuki K, Itoh T (2001) The changes in cell wall architecture during lignification of bamboo,
Phyllostachys aurea Carr. Trees 15:137–147
24. Yu H (2003) Study on property of bamboo culm. World Bamboo Rattan 1:5–10
25. Kamruzzaman M, Saha SK, Bose AK et al (2008) Effects of age and height on physical and
mechanical properties of bamboo. J Trop For Sci 20(3):211–217
26. Yu HQ, Jiang ZH, Hse CY et al (2008) Selected physical and mechanical properties of moso
bamboo (Phyllostachys pubescens). J Trop For Sci 20(4):258–263
27. Liese W (1998) The anatomy of bamboo culms. International network for bamboo and rattans
(INBAR), Beijing, Tech rep no. 18
28. Roelofsen PA (1959) The plant cell-wall. Gegruder Borntraeger, Berline-Nikolassee pp 126–
189
29. Jones RM (1975) Mechanics of composite materials. Scripta Book Company, USA, pp 355.
ISBN 0-07032790-4
30. Qiao SR (1997) Micromechanical properties of composites. Northwestern Polytechnical
China, University Press, p 187. ISBN 978-7-5612-0913-4
31. Gordon JE (1984) The new science of strong materials or why you don’t fall through the
floor. Princeton University Press, USA, p 287. ISBN 0-691-02380
32. China National standard (1995) Testing methods for physical and mechanical properties of
bamboo GB/T 15780. China National Institute of Standardization, Beijing
Chapter 8
Interlaminar Fracture Properties
of Bamboo
Abstract In this chapter, Mode I, II, III interlaminar fracture properties of bamboo
and the fracture mechanism are studied by double cantilever beam (DCB) method,
end-notched flexure beam (ENF) method, modified split cantilever beam (MSCB),
and split cantilever beam (SCB) method respectively. The results show that: (1) the
Mode I, Mode II, and Mode III interlaminar fracture toughness (GIC, GIIC, and
GIIIC) of bamboo are the basic attributes of bamboo that represent the capacity of
bamboo to resist the propagation of cracks; (2) On the Mode I fracture surfaces of
bamboo, smooth fibers, and plane ground tissue are found, which indicate that the
longitudinal interface strength was weak among bamboo cells. The Mode II fracture
surfaces is rougher, and ground tissue is characterized by hackle shearing defor-
mation, which indicates that a large amount of fracture energy would be absorbed
by the shear deformation of ground tissue, and GIIC 2.5 GIC; From the Mode III
fracture surface of bamboo, it can be seen that the resistance that hindered the
propagation of interlaminar crack is contributed by the transverse shear strength of
ground tissue cell wall and interface strength, and GIIIC 4.0 GIC; (3) The study on
the toughness contribution of bamboo node to the interlaminar fracture toughness of
bamboo and the mechanism show that GNode IC 2.9 GInternode
IC IIC 1.3
, GNode
GInternode
IIC , and G Node
IIIC 2.7 G Internode
IIIC , thus bamboo node can contribute a lot to
hinder the interlaminar fracture of bamboo.
8.1 Introduction
The unique tissue of bamboo culm is the material basis of perfect mechanical
characteristics. In view of macro-mechanical behavior, bamboo is a typical unidi-
rectional long-fiber-reinforced biocomposite. It exhibits significant anisotropy in
strength and stiffness corresponding to longitudinal, radial, and transverse direc-
tions. The tensile strength along the longitudinal direction can be as large as 150–
300 MPa, but the transverse tensile strength and shear strength parallel to grain are
much lower. Zeng et al. [1] reported that the tensile strength perpendicular to grain
was only about 2% of the one parallel to the grain. Therefore, once there is crack,
the propagation of delamination is controlled by the interlaminar fracture toughness
but not by strength. The interlaminar fracture toughness of bamboo represents the
ability of bamboo to resist the crack propagation along the grain. Even the crack or
lacuna perpendicular to the grain of bamboo may easily be transformed to develop
along the grain direction under loads, and thereby influencing the mechanical
behavior.
Bamboo and wood are two of the few natural composite materials that can be
used directly as structural materials. When used as structural materials, the inter-
laminar fracture toughness as well as the strength of bamboo and wood is an
important basis for the safety design of the structure. The interlaminar crack of
composites could be classified into three categories: Mode I––opening mode, Mode
II––sliding mode, and Mode III––tearing mode according to the load and fracture
characteristics.
In this chapter, the Mode I, Mode II, and Mode III interlaminar fracture
toughness of bamboo are tested, and the fracture mechanisms are studied.
8.2.1 Theory
The Mode I interlaminar fracture toughness, i.e., the critical value of energy release
rate, GIC, for delamination growth of composite material is often measured by
symmetric bending tests using double cantilever beam (DCB) specimen and cal-
culated with the compliance of the specimen [4]. Triboulot et al. [5] measured wood
fracture toughness for TL crack opening mode using DCB method, and compared it
with finite element analysis (FEA). The results tallied well with each other. Now,
the DCB method has been recommended by American Society for Testing and
Materials (ASTM) as a standard measurement for Mode I interlaminar toughness of
unidirectional fiber-reinforced composites [6].
For DCB specimen under opening loads, the strain energy U and its release rate
GIC can be calculated as follows:
8.2 Mode I Interlaminar Fracture of Bamboo 149
1 1
U ¼ Fd ¼ F 2 C ð8:1Þ
2 2
@U Fcr2 @C
GIC ¼ ¼ ð8:2Þ
@A 2b @a
The moso bamboo (Phyllostachs pubescens) used in the tests was 4 years old and
came from Lujiang, Anhui Province, China. The total height was about 15 m, and
the diameter at breast height was 125 mm. The specimens were collected from the
internodes sections located at a height between 0.6 and 6 m of the bamboo. After
air dried, the specimens were cut parallel to grain into small pieces. The moisture
content of the specimens was about 11% during the experiment.
The bamboo DCB specimens were taken from three different heights of the
culm, i.e., section 7 (height of 1.3 m), section 14 (height of 3 m), and section 19
(height of 5 m), respectively. Five or six DCB specimens per internode section
were prepared, total number 5 recommended by ASTM D5528 [6]. The
dimensions of the specimens are e = 20 mm, W = 200 mm, h = 20 mm, as shown
in Fig. 8.1. Generally, the initial crack of the wood specimen was first cut by a band
saw and then extended by a razor blade [7, 8]. In our case, to simulate a naturally
sharp crack, the crack was cleaved along the middle level parallel to the grain by a
knife as seen in Fig. 8.2. For bamboo internode specimen, the length of the initial
crack was controlled about 40 mm from the center of the loading holes; for a
bamboo specimen with node, the initial crack tip was controlled 2 mm from node to
ensure that the crack would spread through the node by on–off test. When the initial
crack was cleaved, a red dyeing reagent that had been prepared in advance was used
to dye the crack tip before the knife was pulled out, in order to obtain accurate
initial crack length.
Tests were performed on a universal metal material testing machine
(WDW-100). The test room temperature was 18 °C, and the humidity was 60%.
The DCB specimen was connected with steel U hook by steel pin (Fig. 8.3) with a
crosshead speed between 1 and 5 mm/min. At the beginning of the test, a low
crosshead speed was used, while the cantilever beams were short in order to
facilitate identifying the initial movement of the delamination. The crosshead speed
was increased when the cantilever beams were relatively long [4, 6]. A curve of the
applied load versus opening displacement (F-d) was automatically recorded by a
150 8 Interlaminar Fracture Properties of Bamboo
e 0
w
Force
Fig. 8.1 Sketch of double cantilever beam (DCB) specimen. w = 200 mm (longitudinal
direction); e = 20 mm; h = 20 mm (tangential direction); b = 8–10 mm (natural thickness, radial
direction); diameter of loading hole = 5 mm; a0 = 40–50 mm
Specimen
of bamboo
computer during the test. From start loading to maximum load Fmax, the F-d curve
remained linear until the specimen started crazing. Once crazing, the bearing
capacity of the specimen decreased sharply and the crack propagation parallel to
grain was unstable. The top point of F-d curve represented the critical point of rapid
cracking and Fmax was taken as the critical load value Fcr. After the load went
down, the test machine was stopped immediately and the recorded data was stored.
The crack tips were marked on both the sides of the specimen by means of an
optical microscope. Then, the specimen was unloaded and reloaded and the same
procedure was repeated until the specimen was fractured completely. The specimen
was taken from the test machine and the crack length after each increment of
delamination crack growth was measured. As the resistance-arresting crack prop-
agation of the outer layer is lower than that of the inner layer of bamboo, the
delamination crack lengths on each side were unequal and the crack of the outer
8.2 Mode I Interlaminar Fracture of Bamboo 151
DCB specimen
Crack
Force
layer was 5–6 mm longer than that of the inner layer. Thus, the average crack
length of two sides was taken as the actual crack length.
F-d curves corresponding to different crack lengths obtained from one of the
DCB specimens of this study are shown in Fig. 8.4, which shows the typical F-d
curve of DCB specimen. With increasing crack length, the slope of the linear part of
F-d curve decreased. The reciprocal of the slope is the corresponding compliance
(Ci) of the DCB specimen with a certain crack length (ai). The relationship between
C (mm/kN)
200
100
0
0 40 80 120 160 200
a (mm)
C and a of this specimen is shown in Fig. 8.5. A power law relationship between
C and a could be obtained as follows:
C ¼ qam ð8:3Þ
The total mean value and standard deviation of GIC in this study were 358.08 ±
61.18 J/m2 (Table 8.1). The variance analysis of GIC in different sections, i.e., at
different heights, is shown in Table 8.2 and no significant difference was found
among different heights.
In general, along the longitudinal direction, bamboo density increases from the
bottom to the top [9]. In this study, the mean air-dry densities of the samples located
at section 7 (height of 1.3 m), section 14 (height of 3 m), and section 19 (Height of
5 m) were 0.636, 0.712, and 0.729 g/cm3, respectively. The increased density could
Table 8.1 Statistical analysis of Mode I interlaminar fracture toughness (GIC) for different
sections of the bamboo
Section Location Number Number of Mean Standard Coefficient
(height, of measurements GIC (J/ deviation of variation
m) samples m2) (%)
7 1.3 6 41 371.17 52.39 14.44
14 3.4 6 52 351.55 64.20 18.02
19 5.1 5 41 367.96 64.07 18.02
Total 17 134 358.08 61.18 16.88
8.2 Mode I Interlaminar Fracture of Bamboo 153
Table 8.2 Variance analysis of Mode I interlaminar fracture toughness (GIC) for different sections
of the bamboo
Sources of variation SS df MS F-value F0.05 Significance
Between 1414.9 2 707.5 0.1914 3.0653 not significant
Within 484189.1 131 3696.1
Sum 485,604 133
where E11 is the modulus of elasticity of the specimen in the fiber direction and in
this study this value was replaced by the elastic modulus measured on each section
of the bamboo by means of a bending test according to the National Standard of
China [11]. The specimen dimension in this study met the above condition of
154 8 Interlaminar Fracture Properties of Bamboo
800
GIC (J/m2 )
600
358.08
400
200
0
20 60 100 140 180
a (mm)
Eq. (8.4). Therefore, although the value of GIC measured on DCB specimen was
affected to a certain extent by the crack length, the statistical results show that the
change of GIC due to the different crack lengths was not significant in this study.
Thereby, the interlaminar fracture toughness of Mode I can be regarded as a basic
characteristic of bamboo material representing the resistance-arresting crack
propagation.
Figure 8.7 shows the micrographs of Mode I interlaminar fracture surface obtained
by scanning electron microscope. It can be seen from Fig. 8.7a that in the Mode I
crack expanding area, there are the characteristics of smooth fiber and a matrix
(ground tissue) neatly parallel to the stem axis. Figure 8.7b shows the parenchyma
tissue of Zone A in Fig. 8.7a after 1489 magnification. Viewed from interlaminar
fracture surface, the contour of the ground tissue generally maintains integrity. The
surface is smooth or with few slice debris. Occasionally, tearing and stripping of the
whole cell wall happened. This indicates that the crack propagation among ground
tissues develops along the middle lamella or the primary cell wall. Figure 8.7c
shows the fibers of zone B in Fig. 8.7a after 14,649 magnification. It shows the
integrity of bamboo fiber and the smooth interlaminar fracture surface only with a
few traces of slight tearing in the middle lamella or primary cell wall. No tearing or
stripping was found on the secondary cell wall. It indicates that the Mode I crack
among the bamboo fibers develops through the interface between fibers, which
means that the interlaminar strength of bamboo is weak.
Bamboo is a kind of high anisotropic biomaterial with most of its tissues
arranged along the longitudinal direction and conglutinated together by the
non-cellulose component. This unique structure causes low interface strength.
However, it is just because of the weak interface strength that bamboo has high
toughness in the transverse direction, which makes the bamboo adapt to bending
load caused by snow or wind. This could be explained by the toughening mech-
anism due to the weak interface, which was first described by Cook et al. [12] and
has been widely used in designing composite material.
8.2 Mode I Interlaminar Fracture of Bamboo 155
A B C D
F/2 F/2
F/4
L L
MQ a
QM
F/2 F/4
8.3 Mode II Interlaminar Fracture of Bamboo 157
thickness of bamboo culm. Instead of being cut by a band saw then extending by a
razor, the pre-crack was cleaved along middle level parallel to grain by a knife to
form a nature-sharp (Fig. 8.2) and the crack body was TL system. The initial crack
length of each specimen was different and to avoid the influence of stress con-
centration caused by supports and loading roller on the crack tip, the initial crack
length/half span should be taken within the range of 0.25 < a/L < 0.75.
Mode II interlaminar crack was easy to propagate parallel to grain of the bamboo
outer layer. However in the inner layer, the resistance was larger and crack fragment
appeared at an angle of 45° to middle line at the surface of inner layer (Fig. 8.9). It
is because the surface of inner layer has a layer of pith ring constituted by stone
cells, which are crisp and isotropic, and it leads to the result that the pith-ring
fractures along the direction of maximum tension stress under the pure shearing
state (Fig. 8.10). Therefore, it is difficult to mark the position of crack tip in inner
layer. So different from Mode I interlaminar fracture tests that every specimen could
be loaded with repeated loading and the propagated length of crack could be
marked more than once, each specimen should be used only once in Mode II
interlaminar fracture test and a lot of specimens were needed during the test. In this
test, the number of specimen was 60 and there were 43 specimens that were
effective except the specimens diverging from the neutral layer by ±0.5 mm or
more.
Tests were performed by a computer-controlled testing machine. Temperature
was 18 °C and the humidity was 60% in laboratory. The load speed was 2 mm/min
and the curve of load–load point deflection (F-d) was recorded by computer
automatically during the test. The F-d curve of Mode II interlaminar fracture test
was shown in Fig. 8.11. At the initial stage of loading, the F-d curve kept straight,
then as the load increased, the initial crack began to crack. At this moment, because
the rigidity of specimen changed, the slope of F-d curve changed as well and the
new crack surface was rougher. Because of the friction on the crack surfaces, the
crack propagated stably and slowly in the initial stage and until the maximum load,
the unstable propagation of crack began. Since in the ENF tests of bamboo, the
crack would always propagated slowly in different extent before the specimens
fractured, there was no obvious sign of unstable propagation. Thus, the maximum
load was not the critical load when unstable propagation of crack began. Then, the
matter was how to determine the critical load. To the Mode I interlaminar fracture
test, the critical load was the corresponding load when the propagated length
Da equaled to 2% of the initial crack length.
However, in practical test the curve we obtained was F-d curve but not F-D
a curve, so the corresponding load (Da/a = 2%) should be determined on F-d
curve. As theoretical analysis and approximation [19], it was found that the slope of
secant line of the corresponding load on the F-d curve decreased 5% compared with
the slope of the line segment of the F-d curve before the propagation began, so
graphing method was used to determine the critical load on the F-d curve. The
method to determine the critical load is approved by ESIS standard [20, 21]) and
applied by the industry standard in China [22] and used by Fonselius and Riipola
[23], Demorais et al. [24], Kuboki et al. [25] and Hiroshi [26] to analyze the ENF
specimen. In this section, the method was applied.
The detail method was: at first the maximum load Fmax was read; then a straight
line (tangential line) was drawn through the points of the load–displacement (F-d)
curve having F0.1 = 0.1 Fmax and F0.5 = Fmax; third, a secant line was drawn
corresponding to a 5% reduction in the slope; at last the critical load Fcr of initial
cracking was the point of intersection of the secant and the F-d curve. If the
intersection point was beyond the maximum load Fmax, then Fmax was taken as Fcr.
8.3 Mode II Interlaminar Fracture of Bamboo 159
Fcr
F (kN)
k=0.95tan
(mm)
dD ¼ dCD
D þ dC þ hC a
BC BC
! !
a3
F F
ðL aÞ3 Fa
ðL aÞ2 F
ðL aÞ2 Fa
ðL aÞ
¼ 4
þ 2
þ 2
þ 2
þ 2
a
3EI 3EI0 2EI0 2EI0 EI0
FðL3 þ 3a3 Þ
¼
48EI
ð8:6Þ
As the rigidity of the left half beam and the right half beam is different, the
displacement of B is not the maximum displacement and there exists a corner on
section B. So we can obtain the displacement at the loading point, d, as follows:
160 8 Interlaminar Fracture Properties of Bamboo
F
L3 FL3 FL3
dA ¼ 2
¼ ¼ ð8:7Þ
3EI0 6EI0 48EI
dD ¼ dCD
D þ dC þ hC a
BC BC
! !
a3
F F
ðL aÞ3 Fa
ðL aÞ2 F
ðL aÞ2 Fa
ðL aÞ
¼ 4
þ 2
þ 2
þ 2
þ 2
a
3EI 3EI0 2EI0 2EI0 EI0
FðL3 þ 3a3 Þ
¼
48EI
ð8:8Þ
1 1
d ¼ ðdA þ hLÞ þ ðdD hLÞ ¼ ðdA þ dD Þ
2 2
ð8:9Þ
1 FL3 FðL3 þ 3a3 Þ Fð2L3 þ 3a3 Þ
¼ þ ¼
2 48EI 48EI 96EI
where I = bh3/12 is the axis moment of inertia of cross sectional in the crack region,
I0 = b(2h)3/12 = 8I is the axis moment of cross sectional in the crack free region,
E is the modulus of elasticity in the fiber direction, b is the thickness of specimen.
The loading point compliance C is defined as follows:
d 2L3 þ 3a3
C¼ ¼ ð8:10Þ
F 96EI
So strain energy U and strain energy release rate GII are represented as
1 F 2 ð2L3 þ 3a3 Þ
U ¼ Fd ¼ ð8:11Þ
2 192EI
@U 9Fcr a2 d
IIC ¼
GTest ¼ ð8:12Þ
b@a 2bð2L3 þ 3a3 Þ
where d is the displacement of the beam middle point corresponding to the critical
load Fcr when the initial propagation of interlaminar crack begins. As GTest IIC can be
calculated by the experiment parameters straightly and there is no need to test the
longitudinal modulus of elasticity E in advance, it is called experiment parameter
substitution method. What to be specified is that as unidirectional long-fiber-
reinforced composites, the crack propagation of bamboo ENF specimen will not
deviate from the initial crack direction during the test, therefore there is no need to
correct it by introducing additional crack length [27].
8.3 Mode II Interlaminar Fracture of Bamboo 161
When the ratio of span (L) and width (h) is small, the influence of transverse shear
is significant. Due to Timoshenko beam theory [19–21], the shear modulus G13
is introduced in. The strain energy release rate GIIC can be expressed as follows
[28–30]:
" #
9Fcr2 a2 E h 2
GTimo
IIC ¼ 1 þ 0:2 ð8:13Þ
16Eb2 h3 G13 a
As the propagation way of bamboo interlaminar crack is mainly type TL, the
ratio of L/h in this test ranged from 8 to 9 and the transverse shear influence on the
bending deflection. To calculate GTimo
IIC , axial elastic modulus EL should be test in
advance. In this test, according to the national standard “Testing methods for
physical and mechanical properties of bamboo” [11], the bending elastic modulus
was tested approximately as the axial elastic modulus. And the length of the
specimen was S = 160 mm, the height h = 10 mm, the span L = 120 m, and the
width was the natural thickness of bamboo culm. The variable span bending method
[31] was used to test shear modulus GTL and the length of the specimen was
S = 160 mm, the height h = 10 mm, the width was the natural thickness of bamboo
culm, and the ratio of L/h ranged from 6 to 19. Substituting the critical load Fcr and
the corresponding elastic modulus in Eq. (8.13), then the critical energy release rate
GTimo
IIC could be calculated respectively.
Davies et al. [32] and Davidson et al. [33] proposed the compliance calibration
methods. By the method, the compliance of the same material with different crack
length need be tested. The compliance of ENF specimen is generally calculated by
third-order polynomial, as follows:
CL ¼ C0 þ ma3 ð8:14Þ
C0 and m are constants related to material that gained by test date fitting. So, the
Mode II interlaminar fracture toughness can be calculated by the following formula:
Fcr2 a2
GComp
IIC ¼ 1:5 m ð8:15Þ
b
162 8 Interlaminar Fracture Properties of Bamboo
The statistic analysis result of the Mode II interlaminar fracture toughness calcu-
lated by Eqs. (8.12), (8.13), and (8.15) was shown in Table 8.3. The average value
2 2
GTest
IIC was 1303.18 J/m (SD = 116.82 J/m ); the tested value of EL/GTL was 23.00,
then the average value GIIC was 1107.54 J/m2 (SD = 125.69 J/m2); the average
Timo
value GCompIIC was 1216.06 J/m2 (SD = 166.48 J/m2). The variance analysis of the
three calculated results was shown in Table 8.4, for GTest IIC , F-value = 0.0376; for
Comp
GTimo
IIC , F-value = 0.6144 and for G IIC , F-value = 3.1987. The variance analyses
Test Timo
results of GIIC and GIIC were both less than F0.05, and the variance analysis result
of GComp
IIC was a little greater than F0.05. It can be seen that in the condition of this
test, the influence of bamboo height and specimen size on Mode II interlaminar
fracture was small. And it indicated that the Mode II interlaminar fracture toughness
could be seen as constant.
Seen from the results, the result GTest IIC calculated by Eq. (8.12) was 15% higher
Timo
than the result GIIC calculated by Eq. (8.13). It was mainly because that under the
transverse bending loads, there existed friction between the crack faces of the ENF
specimen and with the propagation of Mode II crack, the rough crack face produced
bigger friction resistance, therefore, the propagation resistance of crack increased.
In Eq. (8.12), the critical load Fcr and the deflection of loading point d were
calculated according to the actual load–deflection curve, so the critical energy
release rate calculated by Eq. (8.12) was true. However, in Eq. (8.13), although the
critical load Fcr was calculated according to the actual load–deflection curve, the
3
þ 3a3
deflection of loading point was calculated value by d ¼ FC ¼ 2L96EI but not
Timo
actually measured value. So the result GIIC , calculated by Eq. (8.13), was the
resistance of crack propagation without considering the friction between crack faces
but not the true propagation resistance of bamboo Mode II crack under transverse
loading. Although modification had been done to correct the influence of shear
stress in Eq. (8.13), the calculated result was still lower than the result calculated by
the experiment parameter straightly. That is why Eq. (8.12) is always adopted as
national test standard to test the Mode II fracture toughness GIIC of fiber-reinforced
composites in some countries. The result GTest IIC calculated straightly by experiment
parameter was 6.68% higher than the result GComp IIC calculated by Eq. (8.15) and the
difference was very small. The main reason was that when calculated GIIC by
compliance calibration method, the calculated compliance from the test curve must
be fitted to the initial crack length, and then the experimental constant m was
gained. The m should be substituted into the Eq. (8.15) to calculate the GIIC . The
calculating process was complex relatively and it led to the lower calculated result
and the higher coefficient of variation of GComp IIC relatively. So the experiment
parameter substitution method was the most suitable method to calculate the
Mode II interlaminar fracture toughness.
Table 8.3 Statistical analysis of Mode II interlaminar fracture toughness GIIC processed by different methods in different sections of bamboo
2 2
Section Location Dimensions of Number of Mean GTest
IIC (J/m ) Mean GTimo
IIC (J/m ) Mean GComp
IIC (J/m2)
(height, m) specimens (mm) samples
8 1.70 L = 60, h = 7.5 6 1310.15 (SD = 155.22) 1060.92 (SD = 90.14) 1175.78 (SD = 168.70)
15 3.70 L = 60, h = 7.5 6 1344.24 (SD = 113.73) 1146.69 (SD = 191.25) 1271.31 (SD = 196.28)
20 5.40 L = 60, h = 7.5 6 1292.41 (SD = 130.39) 1178.61 (SD = 170.61) 1202.45 (SD = 159.81)
8 1.60 L = 75, h = 9 7 1292.39 (SD = 120.44) 1093.55 (SD = 65.53) 1047.37 (SD = 52.03)
11 2.40 L = 75, h = 9 6 1328.59 (SD = 84.35) 1072.90 (SD = 31.23) 1286.93 (SD = 116.13)
14 3.40 L = 75, h = 9 6 1307.06 (SD = 152.76) 1072.65 (SD = 212.14) 1187.19 (SD = 170.66)
18 4.70 L = 75, h = 9 6 1306.28 (SD = 109.78) 1114.63 (SD = 8.46) 1369.51 (SD = 116.05)
Total 43 1303.18 (SD = 116.82) 1107.54 (SD = 125.69) 1216.06 (SD = 166.48)
8.3 Mode II Interlaminar Fracture of Bamboo
163
164 8 Interlaminar Fracture Properties of Bamboo
Table 8.4 Variance analysis of Mode I interlaminar fracture toughness GIIC of bamboo in
different sections
Sources of df F−0.05 F-value F-value F-value Significance
variation (GTest
IIC ) (GTimo
IIC ) (GComp
IIC )
Between 6 2.3637 0.0376 0.6144 3.1987 not
significant
Within 36
Sum 42
The distribution relationship of the initial crack length and the Mode II inter-
laminar fracture toughness calculated by Eqs. (8.12), (8.13), and (8.15) were shown
in Figs. 8.12, 8.13, and 8.14 respectively. It could be seen that the values of GIIC
nearly has no difference except some values dispersed for the higher a/L ratio which
were easily influenced by the stress concentration caused by the loading indenter.
So, it could be concluded that when the a/L was in the range of 0.25–0.75, the
Mode II interlaminar fracture toughness of bamboo was a constant independent to
the initial crack length basically.
GIIC(Test)(J/mm2)
a/L
a/L
GIIC(Comp)(J/mm2)
a/L
Fig. 8.15 a Image of Mode II interlaminar fracture surface of bamboo obtained by scanning
electron microscope; b 853magnification image of zone A in (a); c 1224magnification image
of zone B in (a); d 1922magnification image of zone C in (a)
166 8 Interlaminar Fracture Properties of Bamboo
Fig. 8.16 The morphologies of cracks: a crack surface of mode I; b crack surface of mode II
Figure 8.15a was the SEM photos of bamboo Mode II interlaminar fracture surface
after 55 magnification and it could be seen that the fiber bundle in the crack
propagation zone was flat, but the ground tissue was rough and uneven in surface,
and then the selected A, B, and C that located in different area in the propagating
zone were magnified. Figure 8.15b was the SEM photo of the ground tissue at A in
the concavity zone after 853 magnification and it could be seen that the ground
tissue of bamboo was thin-wall ringlike lamellar structure with obvious intercellular
space. The contour of the ground tissue generally maintained the integrity and the
surface was smooth or with few slice debris. This indicates that the crack propa-
gation among ground tissues developed along the middle lamella or the primary cell
wall. Figure 8.15c was the SEM photo of the ground tissue at B in the convexity
zone after 1227 magnification and it could be seen that multilevel parenchyma
cell was torn along crosswise and the intercellular space could be seen clearly.
Figure 8.15d was the SEM photo of the fiber at C after 1922 magnification and
the torn primary wall fragment on the surface of fiber could be seen. Thus, it was
known that the ground tissue was separated from fiber along middle lamella under
shear stress and as the increasing of the dislocation of upper and lower layer, the
thin-walled ground tissue would fracture transversely by tension, while to
thick-walled fiber cell, only middle lamella and primary wall were torn then
debonded, fragments remained.
At last, the photos of Mode I and Mode II interlaminar fracture surface taken by
common camera was shown in Fig. 8.16. By the photos, we could compare the
difference between the two kinds of interlaminar fracture of the bamboo.
Figure 8.16a showed the feature of the Mode I interlaminar fracture surface and it
could be seen that the crack propagation zone was flat. Although tearing trace could
be seen on the matrix (ground tissue surface), it was hard to identify the defor-
mation of matrix. It indicated that the interface strength of matrix was weak. The
ground tissue of bamboo is a thin-walled cell body material with low tensile
strength and there exist space among cells. High stress concentration occurred at the
8.3 Mode II Interlaminar Fracture of Bamboo 167
crack tip, when the Mode I crack propagated along grain, and it made the crack
propagate quickly. Therefore, the deformation of matrix did not develop fully and it
led to the flat fracture surface of Mode I interlaminar fracture. In Fig. 8.16b, the
photo of Mode II interlaminar fracture was shown and the ground tissue was
characteristic of hackle shearing deformation. The ground tissue of cell construction
would form large shear deformation under shear stress and then a large amount of
fracture energy would be absorbed. Generally, under static load, the speed of
Mode II crack propagation was not high, so the plastic deformation of ground tissue
would develop fully during the process of crack propagation and it led to the
laminate fracture surface.
It was found that the study on the Mode III interlaminar fracture behaviors of
artificial composites started early and the methods and theories were various, such
as crack rail shear (CRS) [34], split cantilever beam (SCB) [35], and edge crack
torsion (ECT) [36] and so on. The reason for the existence of the various methods
was that there was a considerable content of Mode II component in the measured
Mode III interlaminar fracture energy release rate (EER) and the EER varied
unstably with the propagation of crack [37, 38]. Compared with artificial com-
posites, the investigation on the Mode III interlaminar fracture behaviors of bamboo
and wood might encounter more complex problems because of the specific orga-
nization structure of natural composites. It was also one of the main reasons why
there was no research on the Mode III interlaminar fracture toughness of bamboo
and wood so far.
As known, the cells in bamboo internodes stem are all arranged strictly along the
axial direction with no crosswise ray cell, so the structure of bamboo is simple
compared with that of wood and the internodes stem of bamboo could be seen as
unidirectional long-fiber-reinforced composite with fiber bundles as reinforcement
and ground tissue as matrix. Thus, the study methods of artificial composites could
be used for reference to the investigation on the Mode III interlaminar fracture
behavior of bamboo, especially the modified methods proposed in recent years. For
example, Szekrényes [38] proposed a Modified Split Cantilever Beam (MSCB)
method, by which the deflection (d) was revised then the revised corresponding
compliance (C) could be obtained. By the finite element verification, the content of
Mode III component could be as high as 98% in the recommended range of rack
length. Mohammad et al. [39] investigated the influence of reinforcement volume
fraction on the Mode III interlaminar fracture EER of unidirectional E-glass/epoxy
composite by MSCB method. Davidson and Sediles [40] and Johnston et al. [41]
tested the Mode III interlaminar fracture toughness of unidirectional carbon/epoxy
composites by Shear-torsion-bending (STB) method and split-shear-torsion
(SST) method. Here, MSCB method and SCB method was applied to investigate
the Mode III interlaminar fracture toughness of bamboo by referencing to study the
168 8 Interlaminar Fracture Properties of Bamboo
methods of artificial composites. And the test results were compared with the
Mode III interlaminar fracture toughness of artificial composites to investigate the
mechanism of the high toughness of natural composites from the aspects of material
properties and organization structures.
8.4.1.1 Material
The material used in this test was 4 years old moso bamboo (P. pubescens) col-
lected from Huoshan, Anhui Province, China. Three adjacent internodes with two
nodes were collected at the chest height of bamboo stem, and then the nodes were
got through. After forced dried for 7 days (60 °C) in the constant temperature
humidity chamber and naturally air dried for 90 days, the sample was cut into
pieces along the grain.
The specimens were made with a size of 160 20 b (Unit: mm) for SCB
method and 160 10 b (Unit: mm) for MSCB method, where b was the
thickness of bamboo stem, about 9 mm. To simulate a naturally sharp crack, the
pre-crack was cleaved along middle level parallel to the grain by a wedge.
8.4.1.2 Method
SCB Method
As seen in Fig. 8.17, the surface of the loading end was a cambered surface, whose
center was on the center lines of the holding devices and specimen. It was a line
contact between the cambered surface and the indenter of testing machine. When
1- Grip1
3
5 2- Grip 2
3- Specimen
1 6
4- Limited Rollers
2 5- Torsion Limited Column
4
6- Base
Fig. 8.17 The test device of split cantilever beam (SCB) method
8.4 Mode III Interlaminar Fracture of Bamboo 169
loaded, load is transferred along the center lines above. The two limited rollers
played a role in limiting the free rotation of the end of specimen and keeping the
loading line at the center lines. What is more, it was also line contact between grip 1
and cylindrical rollers, which would reduce friction to make the experiment data
more accurate compared with surface contact. The torsion-limited column could
limit the torsion of specimen in order to reduce or eliminate the influence of torsion
on the Mode III interlaminar fracture behavior in the process of loading.
Tests were performed on the computer-controlled testing machine, and the load–
load point deflection (F-d) curve was recorded by computer automatically during
the test. Single-specimen with multi-points method was applied, that is, on one
specimen, repetitive loading–unloading–loading was performed. The initial loading
speed was 2 mm/min, and then it was increased properly with the propagation of
crack. At the initial stage of loading, the F-d curve was straight until the crack
began, then the slope of F-d curve increased suddenly, meanwhile, the crack
propagated slowly along the grain. After the F-d curve rose nonlinearly to the
maximum load, it began to decline. And at this moment, the testing machine is
stopped, the curve is curved, and mark the crack tip. As shown in Fig. 8.18, it was
the F-d curves of one specimen and there were 8 curves in total which meant that
the specimen was loaded for 8 times. The critical load Fcr was the load when the
crack began, namely the point the slope of the F-d curve increased. After the test
finished, length between the crack tip was marked each time and the center of the
load point was measured. Since the crack propagation amount in the outer part of
the bamboo specimen was different with that in the inner part after loaded each
time, the average value of the crack length in the outer part and that in the inner part
was taken as the final crack length. The temperature was 27 °C and the humidity
was 60% in the laboratory.
According to Irwin–Kies expression [42], the propagation resistance of Mode III
interlaminar crack along grain, namely the interlaminar fracture toughness can be
calculated by the following formula:
0.35
0.3
0.25
Load (kN)
0.2
0.15
0.1
0.05
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 5.5 6
Deflection (mm)
Fig. 8.18 The typical load–deflection curves of one bamboo by SCB method
170 8 Interlaminar Fracture Properties of Bamboo
According to the typical F-d curve of the SCB specimens, the slope of the
straight portion on the curve was inversed to crack length. And the reciprocal of the
slope was the compliance Ci corresponding to different crack length ai. By curve
fitting, the exponential relationship between the compliance and crack length was
obtained:
C ¼ qema ð8:17Þ
Fig. 8.19 Exploded view of the modified split cantilever beam. Reprinted from ref. [38],
Copyright 2018, with permission from Elsevier
8.4 Mode III Interlaminar Fracture of Bamboo 171
12P2 a2
GIIICMSCB ¼ ½fEB2 þ fTIM2 þ fFT2 þ fSV2 ð8:18Þ
b4 hE11
2 2 2
where fEB2 ¼ 1 2 s1 þa s2 þ s1 þa s2 , fTIM2 ¼ 0:1 ba GE1113 , fFT2 ¼ 0:06 11 ba GE1112 ,
1=2 1=2
fSV2 ¼ 0:32 1 s1 þa s2 ba GE1113 , and 1 ¼ 1 0:63l hb, l ¼ G 13
G12 .
b and h were thickness and half height of specimen. a was the length of crack,
and E11, G12, and G13 were the bending modulus of elasticity and shearing modulus
of elasticity.
According to the national standard “Testing methods for physical and
mechanical properties of bamboo” [11], the bending elastic modulus was tested.
The length of the specimen was S = 160 mm, the height h = 10 mm, the span
L = 120 mm, and the width was the natural thickness of bamboo culm. The vari-
able span bending method [31] was used to test shear modulus. The length of the
specimen was S = 160 mm, the height h = 8 mm, the width was the natural
thickness of bamboo culm and the ratio of L/h ranged from 6 to 17.
To test the mechanical properties of the reinforcement-bamboo fiber and the
matrix-ground tissue, the bamboo block tensile specimens and the “rule of mixture”
were applied (Chap. 7, Sect. 7.3.1). The fresh bamboo billets were cut into sections
with a length of 160 mm, and then each section was cleaved into several strips to
make tensile specimens with different fiber contents. The size in the middle part of
specimens was 40 3 1.3 (Unit: mm) and there were 28 specimens in total.
8.4.2.1 Results
Table 8.5 Mechanical properties of bamboo, bamboo fiber, and ground tissue
Item r (MPa) E11 (GPa) G12 (GPa) G13 (GPa)
Bamboo 172.43 10.30 0.35 0.33
Fiber 536.95 18.52
Ground tissue 51.67 1.62
Table 8.6 One-way variance analysis of the Mode III interlaminar fracture toughness values
obtained by the two methods
Sum of squares df Mean square F Significance
Between Groups 0.307 1 0.307 2.28 0.135
Within Groups 9.971 74 0.135
Total 10.278 75
There is no significant difference between the two groups
G IIIC-SCB(N/mm)
3
2.5
2
1.5
1
0.5
0
40 45 50 55 60 65 70 75
a (mm)
8.4.2.2 Discussion
It could be seen that the results obtained by the two methods were very close. As
seen in Table 8.6, it could be seen that there was no significant difference between
the results by one-way analysis of variance. The results not only indicated that the
applicability of MSCB method to natural fiber-reinforced composite bamboo, but
also proved the accuracy of SCB method. What’s more, the physical dimension of
specimens had no influence on the Mode III interlaminar fracture toughness of
bamboo
The distribution relationships of crack length a and the Mode III interlaminar
fracture toughness was shown in Fig. 8.20, which indicated that in the range of the
crack length, 40–72 mm, the of bamboo was kept stable with the propagation of a
crack in principle.
As seen in Fig. 8.21, the distribution relationships of crack length a and the
Mode III interlaminar fracture toughness showed that had no significant change
with the increase of crack length when the crack length was longer than 40 mm.
8.4 Mode III Interlaminar Fracture of Bamboo 173
GIIIC -MSCB(N/mm)
2.5
2
1.5
1
0.5
0
40 50 60 70 80 90 100 110 120
a (mm)
Both the distribution relationships above were in accordance with the results of
Jun [45] that when crack length was longer than 40 mm, the Mode III interlaminar
fracture toughness trended to constant.
The Mode III interlaminar fracture of artificial FRP has been studied by many
researchers by literature retrieval. The artificial FRP are mainly epoxy or polyester
reinforced by carbon fiber or glass fiber. As seen in Table 8.7, the Mode III
interlaminar fracture results of several unidirectional fiber-reinforced composites
studied by many researchers was shown. It could be seen that, compared with
bamboo, the general mechanical properties of both artificial FRP and their rein-
forcements were superior, while the Mode III interlaminar fracture EER of bamboo
was higher than that of all the artificial FRP above. Then, how to interpret that the
Mode III interlaminar fracture EER of new type artificial FRP with high perfor-
mance was lower than that of natural composite bamboo? Maybe the explanation
could be given by the comparison analysis on the organization structures of the two
kinds of materials.
The glass/epoxy composite was taken as example, which was used most widely
based on mature research. The glass fiber as reinforcement had low compatibility
and wettability with resin matrix because of its smooth columned surface, which
resulted in low interface bonding force. As seen in Fig. 8.22, it was the SEM image
of Mode III interlaminar fracture surface of E-glass/epoxy composite with
Vf = 55% [39]. Due to the difference in physical and chemical structure between
glass fiber and epoxy, abruptly changed interface properties of E-glass/epoxy
composite, and the low interface bonding force led to the phenomenon that the
exposed surfaces of glass fibers were smooth and a little deformed matrix adhered
to glass fibers, when Mode III interlaminar fracture happened. Thus, the Mode III
Table 8.7 The Mode III interlaminar fracture results of several unidirectional fiber-reinforced composites
174
Reference Method Material E11 E22 G12 m12 GIII (J/ Reinforcement Matrix
(GPa) (GPa) (GPa) (GPa) m2)
P. Robinson (1994) ISCB Carbon/ 150 9.5 1.07 0.26 1660 Carbon fiber (60%) Epoxy
epoxy
Shun-Fa Hwang (2001) SCB Carbon/ 121 9.4 6.2 0.23 1000 Carbon fiber (60%) Epoxy
epoxy
Glass/epoxy 37.9 14.3 5.6 0.29 1270 Glass fiber (60%) Epoxy
András zekrényes (2009) MSCB Glass/ 33 7.2 3 0.27 115 Glass fiber (43%) Polyester
polyester
Barry D. Davidson (2011) STB Carbon/ 147.6 – 4.31 0.32 1697 Carbon fiber Epoxy
epoxy 163.8 – 4.95 0.27 1463
Mohammad Khoshravan MSCB Glass/epoxy 41.93 5.25 3.43 0.27 120–150 Glass fiber (55%) Epoxy
(2014) 33.17 4.13 2.64 0.31 175–210 Glass fiber (46%)
23.46 2.54 2.23 – 175–475 Glass fiber (30%)
This chapter MSCB Bamboo 10.30 – 0.33 – 2040 Bamboo fiber Ground
(32%) tissue
SCB split cantilever beam; ISCB improved SCB; MSCB modified SCB; STB shear-torsion-bending
8 Interlaminar Fracture Properties of Bamboo
8.4 Mode III Interlaminar Fracture of Bamboo 175
interlaminar fracture EER was mainly contributed by the deformation of matrix and
a small amount by the debonding of the interface between matrix and glass fibers.
However, for bamboo, the connection of interfaces and the bio-assembly way of
fibers and ground tissue were special. First, the structure of both fibers and ground
tissue cells would transit gradually to decrease the abrupt change of interface
properties, namely, the cell wall of peripheral fiber possessed more lays and the size
of ground tissue cells next to fibers were smaller (Fig. 8.23). Then, for interface,
cells were connected by intercellular layer. Hydroxyl and carboxyl between cell
walls in dried bamboo were assembled by hydrogen bond or chemical bond, which
resulted in the interconnection of pectin and lignin to form complicated colloidal
complex, namely intercellular layer, and the interconnection among intercellular
layer and the primary cell walls of adjacent cells to form compound middle lamella.
Although there were significant difference in the structural features and physical
and mechanical properties between fibers and ground tissue, they could be com-
bined to form a whole by the compound middle lamella to achieve the best
performance.
As seen in Fig. 8.24, it was the SEM images of bamboo Mode III interlaminar
fracture surface. The failure characteristics of fibers were that most fibers debonded
along the interface of fibers (Fig. 8.24a), and few were peeled and fractured. As
seen in Fig. 8.24b, after fibers were debonded, ribbon-like torn traces could be
observed on the surface of the damaged intercellular layer which has obvious
feature of shear failure. For ground tissue, the failure characteristics were that in the
inner part where the content of ground tissue was high, most of the ground tissue
cells were torn and a few were debonded along the interface of ground tissue cells
(Fig. 8.24c); the ground tissue cells were nearly all torn in the outer part. In bio-
composite, cell wall is composed of multilayers and plays a role of mechanical
176 8 Interlaminar Fracture Properties of Bamboo
Fig. 8.23 The cross section of bamboo stem and vascular bundle
Fig. 8.24 The SEM images of the Mode III interlaminar fracture surface of bamboo
support. As seen in Fig. 8.24d, the multilayers of damaged ground tissue cell wall
could be observed clearly and was in shear failure mode, which would absorb large
amounts of energy during the process of Mode III interlaminar fracture.
It could be seen that the Mode III interlaminar fracture mechanism of bamboo
was that the resistance that hindered the propagation of interlaminar crack was
contributed by the transverse shear strength of ground tissue cell wall and interface
strength. The interface concluded the interface between ground tissue cells, fibers as
well as ground tissue cell, and fiber. The differences in organization structures and
interface connection between bamboo and artificial FRP were the main reason for
the high Mode III interlaminar fracture EER of bamboo compared with that of
8.4 Mode III Interlaminar Fracture of Bamboo 177
artificial FRP. Thus, the specific inner construction and organization relevance skill
of natural composite evolved by millions of years, especially the construction of
interface are valuable reference for the design of fiber-reinforced composites.
Through millions of years natural selection of the process of survival of the fittest,
the fine structure of all living beings has been evolved to go through their envi-
ronment. To overcome this disadvantage, the excellent structure—“internode +
node” is formed after millions years’ evolution. The existence of node changes the
damage pattern that when damage along the grain happens, the force to resist the
cleavage only comes from the weak interfacial resistance. Thus, the cleavage
strength along grain is greatly improved and this is also the perfect embodiment of
function adaptability of biological material. However, no study on the interlaminar
fracture of bamboo specimens with node was founded by the literature search.
Because of unique structure, there is great difference between bamboo internode
and node in structure which leads to the difference of physico-mechanical prop-
erties. Here, the Mode I, Mode II and Mode III interlaminar fracture toughness
difference between bamboo internode specimens and bamboo specimens with node
were studied aiming at exploring the influence of the existence of node on bamboo
interlaminar fracture.
8.5.1 Mode I
The material used in this test was one 4 years old moso bamboo (P. pubescens)
from Huoshan, Anhui Province, China. Three adjacent internodes with two nodes
(from the 12th internode to the 14th internode) were collected as sample chosen at
the chest height of the bamboo (Fig. 8.25). Before drying, the diaphragms in
bamboo sample were destroyed to make the sample ventilated during the process of
drying. After forced air drying for 7 days (60 °C) in the constant temperature
humidity chamber and natural air drying for 90 days, the sample was cut into pieces
along grain with a width of 22–25 cm.
As shown in Sect. 8.2, the DCB method is applied. WDW-100
computer-controlled universal mechanical testing machine was used with a test
force accuracy of ±5%, displacement resolution of 0.01 mm, and speed accuracy of
1%. 15 kN mechanical sensor, steel U hook, steel pins, etc., were also applied.
Because the bamboo culm has a hollow structure separated with nodes and the
length and diameter of each internode are limited, the number of specimens gained
was much less than that of normal composites. In this test, the DCB specimens were
collected from the 12th internode to the 14th internode, among which the internode
specimens were collected from the 12th internode, seven effective specimens (total
number 5 recommended by ASTM D5528) were obtained, and the specimens
with node were collected from the 12th–13th and 13th–14th internode, 11 effective
specimens were obtained, as seen in Fig. 8.26. Due to ASTM D5528 [6], the
dimension of the specimen was 220 20 b (mm) and b was the natural
thickness of the bamboo culm along the radial direction. The pre-crack: to simulate
a naturally sharp crack, the crack was cleaved along the middle level parallel to the
grain by a knife as seen in Fig. 8.2. The test room temperature was 25–27 °C, and
the humidity was 60%.
The DCB specimen was connected with a steel U hook by a steel pin loaded by
the computer-controlled testing machine with a crosshead speed between 1 and
4 mm/min. For bamboo internode specimens, single-specimen with multi-point
method was used, that is, for one single-internode specimen, during loading, the
crack propagated and the corresponding applied load versus load point deflection
(F-d) curve was recorded by a computer. Once the load (F) went down, the load
machine stopped, and the crack tip at both the inner part (a1inner) and the outer part
(a1outer) of the internode specimen was marked taking the average length of a1inner
and a1outer as the corresponding crack length a1. Then, it was unloaded and reloaded
and the same procedure was repeated until complete fracture appeared so the
a2
a3
Initial crack
ai
Bamboo node
220
8.5 The Toughness Contribution of Bamboo Node to the Interlaminar … 179
0.1
F (kN)
0.08
0.06
0.04
0.02
0
0 2 4 6 8 10 12
(mm)
corresponding crack length a2, a3, a4…an was obtained. The loading speed was
lifted with increasing length of crack [4]; for bamboo specimens with node,
multi-specimen with single-point method was used, that is, on one single specimen
with node, load was applied only once and the crack length a was obtained by
averaging the ainner and aouter correspondingly. For different specimens, the loading
speed was lifted properly with increasing length between node and the center of
loading holes. The curve of load–load point deflection (F-d) was automatically
recorded by a computer during the test. The typical F-d curve of one internode
specimen is shown in Fig. 8.27, and the crack length of each curve from test point
a1 to test point a29 was 43.80, 46.75, 49.10, 50.65, 53.10, 55.30, 56.50, 59.10,
61.80, 64.50, 67.55, 74.65, 76.65, 79.80, 84.95, 89.40, 94.00, 97.50, 101.35,
112.20, 115.85, 122.05, 129.55, 133.60, 138.00, and 141.70 (unit mm), respec-
tively, and for the internode specimen, 29 test points were collected; the F-d curves
of all specimens with node are shown in Fig. 8.28, and the corresponding crack
length was 45.00, 51.00, 62.50, 63.65, 70.30, 80.00, 80.75, 103.50, 117.50, 133.60,
and 136.25 (unit mm), respectively. From the two figures, it could be seen that in
the initial loading, the F-d curve kept linear until the specimen started crazing. Once
crazing, the bearing capacity of the specimen decreased sharply and then the crack
rapidly propagated along the grain. The maximum load Fmax of the F-d curve was
determined as the critical load Fcr.
After the test was finished, the specimens with node were cleaved to measure
initial crack length: at the front end of the dyeing area, five points were marked
from the outer bamboo wall to the inner bamboo wall uniformly, and the distance
between each point and the center of loading holes was measured. Finally, the
average length of the five points was gained as the initial crack length (Fig. 8.29).
180 8 Interlaminar Fracture Properties of Bamboo
F (kN)
(mm)
Fig. 8.28 Mode I interlaminar fracture F-d curves of all specimens with node
lg C ¼ lg q þ m lg a ð8:19Þ
200
2.51
180 C = 0.0007a
160 2
R = 0.99
C (mm/kN) 140
120
100
80
60
40
20
0
40 60 80 100 120 140
a (mm)
Fig. 8.30 Fitted curve of test curve compliance and crack length of one internode specimen
analysis result of lga and lgC is shown in Table 8.9. It can be seen from the
Table that lgC has a remarkable correlation to lga and the data are reliable.
Results and Discussion
According to Eq. (8.2), by DCB method, the Mode I interlaminar fracture tough-
ness of internode specimens is listed in Table 8.10 and the Mode I interlaminar
fracture toughness of specimens with node in Table 8.11. From the information
above, the average Mode I interlaminar fracture toughness of internode specimens
was 498 J/m2 (SD = 65 J/m2) with coefficient of variation of 12.9%; the average
Mode I interlaminar fracture toughness of specimens with node was 1431 J/m2
(SD = 198 J/m2) with coefficient of variation of 13.8%. The conclusion was drawn
that the Mode I interlaminar fracture toughness of bamboo specimens with node
was much higher than that of bamboo internode specimens. The relationship
between crack length (a) and Mode I interlaminar fracture toughness (GIC) of both
internode specimens and specimens with node is shown in Fig. 8.32, from which it
is obvious that the Mode I interlaminar fracture toughness of bamboo specimens
with node is much higher than that of bamboo internode specimens. The Mode I
interlaminar fracture toughness of bamboo internode specimens decreased slightly
with increasing crack length, which could be mainly due to the nonlinear phe-
nomenon caused by large deflection of DCB, i.e., so-called geometry-nonlinearity,
Table 8.8 The corresponding linear-regression analysis result of lga and lgC of one internode
specimen
Model Sum of squares df Mean square F Significance
Regression 4.31 1 4.31 15185.931 0.000
Residual 0.008 27 0.000
Total 4.317 28
Predictors: (constant), lga; dependent variable: lgC
182 8 Interlaminar Fracture Properties of Bamboo
200
2.43
180 C = 0.0012 a
2
160 R = 0.99
C (mm/kN) 140
120
100
80
60
40
20
0
40 60 80 100 120 140
a (mm)
Fig. 8.31 Fitted curve of test curve compliance and crack length of all specimens with node
Table 8.9 The corresponding linear-regression analysis result of lga and lgC of all specimens
with node
Model Sum of squares df Mean square F Significance
Regression 1.58 1 1.58 1573.66 0.000
Residual 0.009 9 0.001
Total
1.589 10
Predictors: (constant), lga; dependent variable: lgC
Table 8.10 Descriptive statistics of the Mode I interlaminar fracture toughness of all internode
specimens
Specimen Number of test GInternode
IC
Standard Coefficient of
No. points (J/m2) deviation variation (%)
1 7 523 56 10.7
2 11 538 71 13.1
3 14 509 58 11.4
4 19 547 82 14.9
5 29 475 43 9.0
6 19 473 59 12.5
7 18 470 38 8.1
as crack increases [10]; however, there was no obvious trend for the Mode I
interlaminar fracture toughness of specimens with node, which may result from the
small amount of specimens.
From the results above, it can be concluded that the Mode I interlaminar fracture
toughness of bamboo specimens with node is much higher than that of bamboo
internode specimens, which was mainly caused by the difference in structure
between bamboo internode and node (Table 8.12).
8.5 The Toughness Contribution of Bamboo Node to the Interlaminar … 183
Table 8.11 Mode I interlaminar fracture toughness of all specimens with node
Specimen No. a (mm) b (mm) Fmax (kN) C (mm/kN) dC/da GNode
IC (J/m2)
1 45.00 11.20 0.217 12.096 0.661 1396
2 51.00 11.10 0.178 17.983 0.79 1128
3 62.50 11.20 0.179 23.91 1.056 1518
4 63.65 11.05 0.181 27.377 1.084 1614
5 70.30 11.00 0.168 33.068 1.249 1599
6 80.00 10.85 0.155 42.976 1.501 1653
7 80.75 11.30 0.154 53.436 1.521 1603
8 103.50 11.80 0.126 92.945 2.167 1462
9 117.50 11.60 0.099 125.44 2.596 1105
10 133.60 12.00 0.097 164.18 3.118 1223
11 136.25 11.60 0.102 178.65 3.206 1445
Average 1431
2000
1800
1600 G ICNode =1431J/m2
1400
GIC (J/m2)
1200
1000
800 Internode
GIC =498J/m2
600
400
200
0
20 40 60 80 100 120 140 160 180
a(mm)
Fig. 8.32 Relationship between a and GIC of both internode specimens and specimens with node
Table 8.12 Statistics analysis of all test points of internode specimens and specimens with node
Specimen The number of all test GIC (mean) Standard Deviation Coefficient of
points (J/m2) (J/m2) variation (%)
Internode 117 498 65 12.9
With 11 1431 198 13.8
node
8.5.2 Mode II
The material used in this test was 4 years old moso bamboo (P. pubescens) from
Huoshan, Anhui Province, China. Three adjacent internodes were collected at the
184 8 Interlaminar Fracture Properties of Bamboo
Fig. 8.33 Schematic diagram of the selection of Mode II and Mode III specimens
chest height of the bamboo (Fig.8.33), and then got through the nodes. After forced
air dried for 7 days (60 °C) in the constant temperature humidity chamber and
naturally air dried for 90 days, the sample was cut into pieces along the grain.
ENF specimen was used, and the diagrams of specimen. Because bamboo clum
has hollow structure separated with nodes and the length and diameter of each
internode are limited, the number of specimens gained was much less than that of
normal composites. The internode specimens were collected from the 9th internode
and 17 effective specimens were gained with half span L = 60 mm, 2h = 15 mm;
the specimens with node were from 9th internode to the 10th internode and 7
effective specimens were gained with L = 55 mm, 2h = 15 mm (Fig. 8.33). Wide b
was the natural thickness of bamboo culm. And according to Yu [30], the test
results would not be influenced by the change of span.
The test method and data process methods were the same with that in Sect. 8.3.
Table 8.14 Mode II interlaminar fracture toughness GIIC of specimens with node processed by
different methods
a
No. h (mm) b (mm) a (mm) L (mm) L Fcr dcr GTest
IICðnodeÞ GTimo
IICðnodeÞ
(kN) (mm) (J/m2) (J/m2)
1 7.47 12.14 22.75 55 0.41 0.1235 2.2205 1430.11 1350.59
2 7.68 12.31 23.00 55 0.42 0.1369 1.9898 1427.07 1539.29
3 7.79 12.61 30.75 55 0.56 0.1198 1.7344 1668.82 1552.41
4 7.73 12.29 28.95 55 0.53 0.1283 1.8213 1768.97 1766.41
5 7.77 12.29 30.00 55 0.55 0.1262 1.9380 1949.10 1770.80
6 7.78 12.3 31.25 55 0.57 0.1238 1.9416 2025.11 1792.75
Table 8.15 Descriptive statistics of the Mode II interlaminar fracture toughness of all specimens
Sample Method GIIC (mean) Standard deviation Coefficient of
(J/m2) (J/m2) variation (%)
Internode Experiment 1371.54 254.13 18.53
method
Timoshenko 1350.66 117.54 8.70
method
Compliance 1276.24 212.01 16.61
method
With Experiment 1711.54 252.98 14.78
node method
Timoshenko 1628.71 177.33 10.89
method
For the Mode II interlaminar fracture of bamboo internode specimens, three data
process methods were applied and the GIIC value calculated by experiment
parameter substitution method was higher than those calculated by both
Timoshenko beam theory method and compliance calibration method and this result
agreed with the conclusion in Sect. 8.3, thus the GTest
IICðinternodeÞ value calculated by
experiment parameter substitution method was considered as the most proper GIIC
value. For the Mode II interlaminar fracture of bamboo specimens with node, the
GTimo Test Test
IICðnodeÞ value was also less than the GIICðnodeÞ value, so the GIICðnodeÞ was con-
sidered as the proper GIIC value.
The distribution relationships of a/L and the Mode II interlaminar fracture
toughness GIIC calculated by experiment parameter substitution method and
Timoshenko beam theory method were shown in Figs. 8.34 and 8.35 respectively.
It could be seen that the GIIC calculated by experiment parameter substitution
method of both internode specimens and specimens with node increased with the
increasing of a/L which resulted from the stress concentration caused by the loading
indenter when the crack tip closed to the loading indenter. However, the GIIC of
8.5 The Toughness Contribution of Bamboo Node to the Interlaminar … 187
2500
- GIIC
Test
(node)
2000 Test
- G IIC (internode)
1500
1000
500
0
0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65
a/L
Timo
- G IIC (node)
2000 Timo
- G IIC (internode)
1500
1000
500
0
0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65
a/L
The material used in this test was 4 years old moso bamboo (P. pubescens) from
Huoshan, Anhui Province, China. Three adjacent internodes were collected at the
chest height of the bamboo (Fig. 8.33), and then got through the nodes. After forced
188 8 Interlaminar Fracture Properties of Bamboo
Fig. 8.36 Mode III specimens: a internode specimens; b specimens with node
air dried for 7 days (60 °C) in the constant temperature humidity chamber and
naturally air dried for 90 days, the sample was cut into pieces along the grain.
The SCB specimens were applied and the diagram of modified SCB fixture
device was shown in Fig. 8.17. As seen in Fig. 8.33, the internode SCB specimens
were collected from the 10th internode and the size was 170 20 b (mm),
where b was the natural thickness, with 7 effective specimens; the SCB specimens
with node were from the 10th to 11th internode and the size S 20 b (mm),
where b was the natural thickness, with 9 effective specimens (Fig. 8.36). For
different SCB specimens with node, the length between node and loading points
ranged from 40 to 100 mm.
SCB method was applied. Double cantilever inversion symmetry bending load
method was used to test Mode III interlaminar fracture (Fig. 8.17). The specimen
0.5
0.45
0.4
0.35
Load (kN)
0.3
0.25
0.2
0.15
0.1
0.05
0
0 2 4 6 8
Deflection (mm)
24 y = 4.7283e
0.0183x
22 2
R = 0.9783
20
C 18
16
14
12
10
50 55 60 65 70 75 80 85 90
a (mm)
Fig. 8.38 Tipical fitted curve of test curve compliance and crack length of one internode
specimen
0.0383x
70 y = 1.6422e
2
60 R = 0.8001
50
40
C
30
20
10
0
0 20 40 60 80 100 120
a (mm)
Fig. 8.39 Fitted curve of test curve compliance and crack length of all specimens with node
Table 8.16 Descriptive statistics of the Mode III interlaminar fracture toughness of all internode
specimens
No. Number of GIIIC (internode) Standard deviation Coefficient of
measurements (J/m2) (J/m2) variation (%)
1 6 3015.92 566.72 18.79
2 4 2931.63 223.64 7.63
3 3 2475.14 149.91 6.06
4 4 2141.51 182.15 8.51
5 6 2564.89 314.87 12.28
6 4 2083.59 146.49 7.03
7 6 3275.80 525.33 16.04
As shown in Tables 8.16 and 8.17, by SCB method, the Mode III interlaminar
fracture toughness of both internode specimens and specimens with node were
calculated by Eq. (8.16) and by descriptive statistics, we got GIIIC
2 2
(internode) = 2702.79 J/m (SD = 549.76 J/m ) with variable coefficient 20.34%;
GIIIC(node) = 7387.09 J/m (SD = 2645.31 J/m2) with variable coefficient 35.81%
2
(Table 8.18). It could be seen that the Mode III interlaminar fracture toughness of
internode specimens was much less than that of specimens with node.
The distribution relationships of crack length a and the Mode III interlaminar
fracture toughness GIIIC was shown in Fig. 8.40. The effective crack length of
specimens with node varied from 38.25 to 95.90 mm; the effective crack length of
internode specimens, from 50 to 100 mm, and from Fig. 8.40, we got that the
Mode III interlaminar fracture toughness GIIIC of internode specimens had no
relationship with the increasing of crack length. According to Jun [45], when crack
length was longer than 40 mm, the Mode III interlaminar fracture toughness
8.5 The Toughness Contribution of Bamboo Node to the Interlaminar … 191
Table 8.17 Mode III interlaminar fracture toughness of all specimens with node
No. b (mm) a (mm) d1 F1 d2 F2 Fcr C (mm/ @C GIIIC(node)
@a
(mm) (kN) (mm) (kN) (kN) kN) (J/m2)
1 12.465 95.90 3.8258 0.1248 9.4913 0.2515 0.3018 44.7094 2.48 9046.74
2 12.73 71.50 4.2003 0.3662 8.6173 0.4839 0.5382 37.5206 0.97 11064.90
3 12.445 63.55 3.9051 0.4174 7.321 0.5593 0.6287 24.0829 0.72 11389.80
4 12.60 59.75 2.0493 0.178 4.7544 0.3259 0.4496 18.293 0.62 4975.14
5 12.525 42.25 1.4932 0.2988 3.4757 0.7062 0.8018 4.86575 0.32 8140.87
6 12.245 49.10 2.3717 0.2052 5.2147 0.4084 0.5603 13.9911 0.41 5286.28
7 13.025 38.25 1.5358 0.2606 2.7647 0.4054 0.6488 8.48467 0.27 4398.69
8 12.465 48.00 1.7783 0.2173 4.3631 0.4396 0.5915 11.6263 0.4 5548.68
9 13.00 45.25 0.9349 0.2042 2.7295 0.505 0.6961 5.96603 0.36 6632.62
Table 8.18 Descriptive statistics of the Mode III interlaminar fracture toughness of all specimens
Sample Number of GIIIC (mean) Standard Coefficient of
measurements (J/m2) deviation (J/m2) variation (%)
Internode 33 2702.79 549.76 20.34
With 9 7387.08 2645.31 35.81
node
- G IIIC (node)
12000
10000
- G IIIC (internode)
GIIIC (J/m2)
8000
6000
4000
2000
0
30 40 50 60 70 80 90 100
a (mm)
Fig. 8.40 Relationship between a and GIIIC of both internode specimens and specimens with node
trended to constant. However, that of specimens with node had no obvious trend
with the increasing of crack length which might be caused by the few amount of
specimens with node.
8.5.4 Discussion
For bamboo interlaminar fracture, no matter Mode I, Mode II, or Mode III, it was
obtained that the interlaminar fracture toughness of bamboo specimens with node
192 8 Interlaminar Fracture Properties of Bamboo
Table 8.19 The three modes Mode G(Node) (J/m2) G(Internode) (J/m2) GIIICðInternodeÞ
of interlaminar fracture GIIICðNodeÞ
toughness of bamboo I 1431.45 498.48 2.87
internode specimens and II 1711.53 1371.54 1.25
specimens with node
III 7387.08 2702.79 2.73
was higher than that of bamboo internode specimens (Table 8.19), which was
caused by the difference in structure between bamboo internode and node.
The bamboo internode wall consists of the outer wall, the middle of wall, and the
inner wall, and the distribution density of vascular bundles from the outer to the
b b
a-a b-b
Fig. 8.41 Structure of internode and node: a bamboo diametral plane, b cross section of bamboo
internode, c cross section of bamboo node
Fig. 8.43 The SEM image of the Mode I interlaminar fracture surface in node
Fig. 8.44 The SEM image of the Mode II interlaminar fracture surface in node
other vascular bundles change their direction in node; the vascular bundles close to
the inner wall tend to the outer wall and the vascular bundles close to the outer wall
tend to the inner wall oppositely; meanwhile, there are some vascular bundles
entering into the diaphragm circuitously, which incline procumbent circumferen-
tially or interweave together through diaphragm to the opposite wall. The bent
vascular bundles are generally thickened in node; thus, the thickness of the bamboo
wall at the node is thicker than that at internode. As seen in Fig. 8.42, the recon-
struction 3D structure graph of bamboo node reflects the structure feature of vas-
cular bundles vividly [46].
It is known that the strength of cell body biomaterial depends on the thickness of
the cell wall; however, the Mode I interlaminar fracture resistance of bamboo is
determined by the interface property between cells and wall layers. For bamboo
internode, the resistance comes from the interface resistance between ground tissue
and ground tissue, ground tissue and vascular bundle, vascular bundles and vascular
bundles; for bamboo node, the resistance not only comes from the interface
resistance between ground tissue and ground tissue, ground tissue and vascular
bundle, vascular bundles and vascular bundles but also from the contribution of the
fracture of transverse vascular bundles.
The features could be seen from the SEM images of the interlaminar fracture
surface of node. As shown in Figs. 8.43, 8.44, and 8.45, the fracture features of
8.5 The Toughness Contribution of Bamboo Node to the Interlaminar … 195
Fig. 8.45 The SEM image of the Mode III interlaminar fracture surface in node
procumbent fibers and ground tissue on Mode I, Mode II, and Mode III interlaminar
fracture surfaces of the node were shown. It could be seen that the fracture and
peeling of procumbent fibers occurred on all the three types of fracture, and it could
be found that the Mode III interlaminar fracture surface was rougher than that of
Mode II and Mode I. Especially, the damage of ground tissue cells on Mode III
interlaminar fracture surface was obviously more serious than those on Mode II and
Mode I interlaminar fracture surface. The phenomena above coincided well with the
experimental results that the Mode III interlaminar fracture toughness was higher
than Mode II and Mode I interlaminar fracture toughness.
References
1. Zeng QY, Li SH, Bao XR (1992) Effect of bamboo nodal on mechanical properties of
bamboo wood. Sci Silvae Sin 28(3):247–252 (in Chinese)
2. Ahmad M, Kamke FA (2005) Analysis of Calcutta bamboo for structural composite materials:
physical and mechanical properties. Wood Sci Technol 39(6):448–459
3. Obataya E, Kitin P, Yamauchi H (2007) Bending characteristics of bamboo (Phyllostachys
pubescens) with respect to its fiber–foam composite structure. Wood Sci Technol 41:385–400
4. Hodgkinson JM (2000) Mechanical testing of advanced fibre composites. Woodhead
Publishing and CRC Press, Cambridge
196 8 Interlaminar Fracture Properties of Bamboo
30. Zhicheng Y (1997) Study on Mode II interlaminar fracture test method of composites.
J Aeronaut Mater 17(4):54–61
31. Xu MQ, Lu ZY (2003) Determination of Loblolly’s shear modulus. Mech Eng 25:57–60
32. Davies P, Kausch HH, Williams JG (1992) Round robin interlaminar fracture testing of
carbon fiber reinforced epoxy and PEEK composites. Compos Sci Technol 43:129–136
33. Davidson BD, Teller SS (2010) Recommendations for an ASTM standardized test for
determining GIIC of unidirectional laminated polymeric matrix composites. J ASTM Int 7
(2):1–11
34. Becht G, Gillespie JW (1988) Design and analysis of the crack rail shear specimen for
Mode III interlaminar fracture. Compos Sci Technol 31:143–157
35. Donaldson SL (1988) Mode III interlaminar fracture characterization of composite material.
Compos Sci Technol 32:225–249
36. Mehrabadi FA, Khoshravan M (2013) Mode III interlaminar fracture and damage
characterization in woven fabric-reinforced glass/epoxy composite laminates. J Compos
Mater 47(13):1583–1592
37. Rizov V, Shindo Y, Horiguchi K et al (2006) Mode III interlaminar fracture behavior of glass
fiber reinforced polymer woven laminates at 293 to 4 K. Appl Compos Mater 13:287–304
38. Szekrényes A (2009) Improved analysis of the modified split-cantilever beam for mode-III
fracture. Int J Mech Sci 51:682–693
39. Khoshravan MR, Moslemi M (2014) Investigation on Mode III interlaminar fracture of glass/
epoxy laminates using a modified split cantilever beam test. Eng Fract Mech 127:267–279
40. Davidson BD, Sediles FO (2011) Mixed-mode I-II-III delamination toughness determination
via a shear-torsion-bending test. Compos Part A: Appl S 42:589–603
41. Johnston AL, Davidson BD, Simon KK (2014) Assessment of split-beam-type tests for mode
III delamination toughness determination. Int J Fract 185:31–48
42. Anderson TL (2005) Fracture mechanics—fundamentals and applications, 3rd edn. CRC
Press, Taylor & Francis Group, Boca Raton, London
43. Szekrényes A (2010) Fracture analysis in the modified split-cantilever beam using the
classical theories of strength of materials. JPCS 240:012030
44. Szekrényes A (2010) Development of an opening-tearing mode fracture system for composite
materials. EPJ Web Conf 6:42008
45. Jun X (1996) Experimental study on Mode III fracture toughness of multi-directional interface
of composite materials. J Nanjing Univ Aeronaut Astronaut 28(2):267–270
46. Yulong D, Liese W (1995) On the nodal structure of bamboo. J Bamboo Res 14(1):24–32
Chapter 9
Modeling on the Toughness Fracture
and Energy-Absorbing Mechanism
of Biomaterial—Bamboo (Phyllostachys
pubescens)
Abstract Studies showed that the whole process of the transverse bending fracture
of bamboo (Phyllostachys pubescens) involved various damage patterns: ground
tissue cracking, interface debonding, fracturing, and pulling-out of fiber bundles.
Different organization structures would contribute different toughness for different
energy wastages in the evolution of damage. In this chapter, the mesomechanics
method was applied to study the energy-absorbing mechanism of the four damage
patterns, the analysis equations of strain energy released rate were obtained, and
theoretical calculation was completed using theoretical model. The results showed
that among the four damage patterns, the work done by pulling fiber bundles out
contributed most to the fracture work, the second contribution was made by the
fracture of fiber bundles, while the energy absorbed by ground tissue cracking and
the initial interface debonding was low. The phenomena above corresponded well
with the characteristic of transverse bending fracture surface obtained by SEM
(scanning electron microscope) technique. In addition, the theoretical result was
very close to the test result of energy consumed in the whole transverse bending
fracture of bamboo. It would have great significance for the design and manufacture
of bionic material to find out the main structural factors that lead to the fine strength
and toughness properties of bamboo.
9.1 Introduction
Bamboo is studied more and more as an example of bionic material, and efforts
have been put into designing and making bionic material due to bamboo’s high
strength and toughness, as well as its regular and simpler construction [1–4].
Nowadays, the study that takes bamboo as a bionic-optimized model is mainly
derived from the following two aspects: one comes from the understanding of the
macroscopic morphology of bamboo, like hollow cylinder and the gradient distri-
bution of reinforced phase [5, 6]; the other comes from the knowledge about the
microstructure of cell wall of bamboo fiber, just like the spiral distribution of
microfibril and multistory structure [7]. In fact, the hierarchical structure of fiber cell
wall is a common feature of plant fibers, but high strength and toughness of bamboo
depend on both the excellent property of fibers and the perfect organization
structure of bamboo.
Shao et al. [8] and Wang et al. [9] studied the fracture mechanism of Mode I and
Mode II interlaminar fracture toughness of bamboo that was treated as unidirec-
tional fibers reinforced composite. Tan et al. [10] researched the propagation law
and mechanism of crack in bamboo stem by four-point bending test with
single-edge-notched bend (SENB) specimens combined with finite element model.
Habibi et al. [11] applied environmental electron microscopy (ESEM) technique in
the tensile and compress tests of bamboo to study crack path in the longitudinal
direction and radial direction. Habibi et al. [11] gave that the interface zone of
ground tissue cells and fiber cells was the preferred path of crack propagation.
References [12–16] showed that the energy absorption mechanism of the failure of
fiber-reinforced composite with the crack involved matrix cracking, interface
debonding, post-debonding friction, fiber pull-out, fiber fracture and stress redis-
tribution and so on. Kim et al. [12] pointed out that among the failure patterns
above, the toughness contributed by interface debonding and the frictional sliding
during fiber pulling-out played main contribution role for most fiber-reinforced
polymer composites.
Similarly, bamboo internodes stem could be treated as natural polymer com-
posite reinforced by unidirectional fibers that present dense distribution to sparse
distribution from outer part to the inner part of the stem. The strong anisotropy
caused by low interface strength [8] and the unequal distribution of fibers leads to
the complex process of bamboo transverse fracture. And the fracture properties
depend both on the properties of constituent materials and the meso-structure
features including the volume fraction, the distribution law and morphology of
fibers, and the properties of the interface between fiber bundles and ground tissue.
Although Tan et al. [10] have studied the toughness mechanism of crack propa-
gation in bamboo stem by both theoretical analysis and computer model, the work
was only on the energy release rate of interface cracking. Careful observation has
been done by Habibi et al. [11] on crack propagation path in the bamboo stem,
which involved no energy absorption question. As the lack of the basic data on the
mechanical properties of bamboo cells and their interfaces, quantitative analysis on
toughness contribution of bamboo organization structures has not been proposed so
far, and deep understanding on the bionic principle of cell body material
strengthened by fibers is hindered.
In this chapter, theoretical analysis model was built combined with experimental
verification to explore the toughness contribution of bamboo with different orga-
nization structures under different damage patterns. The work will be conducive not
only to the understanding of the toughness mechanism of bamboo, but also to the
research and development of advanced composite with special strength and
toughness.
9.2 Development of Theoretical Model 201
(b)
(a) F Inner part Outer part
H
e
B
S 2 mm
L
Fig. 9.1 Transverse fracture of bamboo: a diagram of bamboo transverse test; b fracture surface
of bamboo SENB specimen
In order to analyze and calculate the energy absorbed by the four damage patterns
above from the angle of mesomechanics, simplification was done to bamboo
components. As seen in Fig. 9.2, in a vascular bundle, the outer, side, and inner
202 9 Modeling on the Toughness Fracture and Energy-Absorbing …
G G
V V
Simplify
V V V V
100um 100um
Fig. 9.2 The simplification of bamboo vascular bundle. G Ground tissue; V vessel
2mm
For the deforming and cracking of ground tissue, the fracture work could be cal-
culated by a two-dimensional model, as seen in Fig. 9.4. It was assumed that the
average line densities of ground tissue and vascular bundles were km and DV
respectively. The volume fraction of ground tissue and vascular bundles were Vm
9.3 Theoretical Analysis 203
Crack
Ground tissue
Dv m Dv m Dv m
Dv
and (Vf + VT), and the average outer diameter of vascular bundle was DV = 2RV,
where Vf and VT were the volume fraction of fibers and conducting tissues. The
geometrical relationship of the parameters above was km =Vm ¼ Dv =ðVf þ VT Þ. As
ground tissue cell was a porous mass with thin wall, when loaded with tensile force,
the stress and strain of cell wall presented an approximately linear relation. Thus,
the energy absorbed by per unit volume of ground tissue was emu rmu =2 ¼
r2mu =ð2Em Þ. So, in the process of bamboo transverse fracture, the energy absorbed
by ground tissue to form unit fracture area was
where Em was the elastic modulus of ground tissue and rmu, emu were the ultimate
stress and strain.
When crack tip propagated to fiber bundles, it turned to interface zone of fiber and
ground tissue cells as shown in Fig. 9.5. According to the energy release rate
criterion of fracture mechanics, the interface debonding region was treated as
interface crack, and the propagation of crack depended on the energy balance of
interface fracture energy [16–18]. The researches of Charalambides et al. [19] and
Tan et al. [10] indicated that the energy release rate of interface crack depended on
load and had no relation with crack length. According to Williams [20, 21], Wang
and Williams [22] and Reeder [23, 24], the strain energy release rate formula of
interlaminar crack propagation of laminate composite was derived as (Chap. 4)
204 9 Modeling on the Toughness Fracture and Energy-Absorbing …
Ld
Vessel
" #
1 M12 M22 2
Gi ¼ þ ðM1 þ M2 Þ ð9:2Þ
2BEI0 ðh1 =HÞ3 ð1 h1 =HÞ3
where M1 and M2 were the moment at the tip of interlaminar crack, and I0 = BH3/
12.
For bamboo transverse fracture specimen (Fig. 9.6), M1 = 0, the bending
moment at crack tip was M = F(S − Ld)/4. So, the energy release rate of bamboo
interlaminar crack could be calculated by the following formula:
" #
3F 2 ðS Ld Þ2 1
Gi ¼ 1 ð9:3Þ
8EB2 H 3 ð1 e=HÞ3
From Eq. (9.3), it was derived that strain energy release rate decreased with the
increasing of debonding length Ld. When Gi was less than the critical strain energy
release rate of bamboo, the interlaminar crack would stop propagating. Especially,
when Ld = 0, namely the moment the transverse crack converted to side crack along
the grain, the critical strain energy released rate could be calculated by the fol-
lowing formula:
" #
3Fcr2 S2 1
Gi0 ¼ 1 ð9:4Þ
8EB2 H 3 ð1 e=HÞ3
S
L
9.3 Theoretical Analysis 205
where Fcr was the load when the crack along the grain began, E was the bending
elastic modulus of bamboo SENB specimen. In general, the duration of the prop-
agation of side crack was very short, so the load during the process of side crack
propagation was taken as F Fcr.
In the process of bamboo transverse fracture, when crack tip turned to the interface,
about half periphery of vascular bundle detached from the cluster of ground tissue,
so the homogenizing transfer of the load on fiber bundle was weakened. Under the
bending load, fiber bundle was elongated which caused the deformation of ground
tissue around the other half periphery of the vascular bundle. When the interface
crack propagated to a certain length, fiber bundles would fracture and recoil to
release elastic deformation energy. As the elastic modulus (Ef) of the fiber bundle is
much higher than that (Em) of ground tissue, when the load on fiber bundle
transferred to ground tissue after the sudden fracture of the fiber bundle, the
redistribution of strain in ground tissue was caused. Strain energy is an another
energy-absorbed mechanism, namely the fracture work of fiber bundle [25, 26].
To calculate the fracture work of fiber bundle, it was supposed that the fracture
surface of fiber bundle happened where Lp/2 far from the main crack surface (as
seen in Fig. 9.7). The elastic potential energy in a length of dx was
2
1 rfðxÞ
dUf ¼ rfðxÞ efðxÞ Af dx ¼ Af dx ð9:5Þ
2 2Ef
dx
Lp /2
u(x) x
206 9 Modeling on the Toughness Fracture and Energy-Absorbing …
After fiber bundle fractured, the work done by the diastrophism between fiber
bundle element and ground tissue that undetached from fiber bundle was
where Rf was the radius of fiber bundle, u(x) was the slipping distance between fiber
bundle element and ground tissue, and it was
Zx Zx
rfðxÞ
uðxÞ ¼ efðxÞ dx ¼ dx 0 x Lp =2 ð9:7Þ
Ef
L L
2p 2p
The total fracture work of fiber bundle element should be the sum of dUf and
dUfm, so the energy consumed by the fracture of a single fiber bundle was
ps2s ðL
p þ 2xÞ
2
dUfm ¼ pR1 ss dx uðxÞ ¼ dx ð9:10Þ
4Ef
So,
L
p
Z2
ps2s ðL
p þ 2xÞ
2
2ps2s ðL
pÞ
3
Ufu ¼ Uf þ Ufm
¼2 dx ¼ ð9:11Þ
4Ef 3Ef
L
p
2
In the same way, the energy consumed by the fracture of fiber bundles ② and ③
was
L`
p
Z2
`
ps2s ðL`
p þ 2xÞ
2
Ufu ¼ Uf` þ Ufm
`
¼2 dx ð9:12Þ
4Ef
L`
p
2
9.3 Theoretical Analysis 207
L´
p
Z2
´
ps2s ðL´
p þ 2xÞ
2
Ufu ¼ Uf´ þ Ufm
´
¼2 dx ð9:13Þ
4Ef
L´
p
2
Hence, for the fracture of a vascular bundle, the fracture work was
` ´
v
Ufu ¼ Ufu þ 2Ufu þ Ufu ð9:14Þ
After the bamboo fiber bundles were fractured, as the bending continued, the
fractured fiber bundles were pulled-out from the crack surface. As seen in Fig. 9.7,
when fiber bundle element was pulled-out from the cluster of ground tissue, the
resistance was pRf ðLp =2 xÞsu , where Rf was the radius of fiber bundle. So, the
work done to pull a single fiber bundle out was
L
p
Z2
1 pR1 ss ðL
pÞ
2
Ufp ¼ pR1 xss dx ¼ ð9:15Þ
2 16
0
Z L`
pR2 ss ðL`
p 2
`
2
pÞ
Ufp ¼ pR2 xss dx ¼ ð9:16Þ
0 8
Z L´
pR3 ss ðL´
p 2
´
2
pÞ
Ufp ¼ pR3 xss dx ¼ ð9:17Þ
0 8
Hence, for the whole vascular bundle, the work done to pull fiber bundles
out was
` ´
v
Ufp ¼ Ufp þ Ufp þ Ufp ð9:18Þ
L` pR22 R2 rf
L`
p
pR2 ss ¼ rf ; p ¼ ð9:20Þ
2 2 ss
L´ pR23 2R3 rf
L´
p
pR3 ss ¼ rf ; p ¼ ð9:21Þ
2 2 ss
It could be seen from the formula above that the fracture length was in relation to
the size of fiber bundle. Because the size and morphology of fiber bundle varied
gradually on the radial direction of bamboo stem, the fracture length of fiber
bundles in the outer part of bamboo stem was longer than that of fiber bundles in the
inner part (Fig. 9.8).
9.4 Results
Moso bamboo (P. pubescens) was taken as a sample, 4 years old, collected from Lu
Jiang, Anhui Province. After removing the cortex and pith periphery, SENB
specimens with the crack perpendicular to grain were made (Fig. 9.1a), and there
9.4 Results 209
600 S
B
L
500
Load/N
400 F0cr
300
200
100
0
0 5 10 15 20 25
Deflection/mm
Table 9.1 The size and No. H-e (mm) B (mm) UText (J)
fracture energy of SENB
specimens 1 9.44 8.59 6.4409
2 9.89 8.63 6.9065
3 9.58 8.38 6.6264
4 10.16 8.53 6.9101
5 10.17 8.87 5.2684
6 9.8 8.72 5.3467
7 9.49 8.44 7.2042
8 10.08 8.38 6.4788
9 9.62 8.84 5.6438
10 9.9 8.69 5.8968
were 10 specimens in total with a moisture content of 11%. The temperature was
15–18 °C and the humidity was 60–65% in the laboratory. Three-point bending
tests were performed on a computer-controlled testing machine with a loading
speed of 2 mm/min, and the load–deflection (F-d) curve was recorded by computer
automatically during the test (Fig. 9.9). When the interface debonding began, the
average critical load was Fcr = 400 N. The total energy consumed in the whole
bending failure process could be calculated by the following formula. The size and
test value of each specimen were shown in Table 9.1. And the average test value
was Utest = 6.272 J (SD = 0.689 J).
ZSu
Utest ¼ FðsÞds
0
The parameters in the theoretical model were actually measured according to Shao
et al. [34, 35], as seen in Table 9.2. The fracture length of fiber bundle ①, ②, and
210 9 Modeling on the Toughness Fracture and Energy-Absorbing …
③ were L ` ´
p = 3.24 mm, Lp = 3.09 mm, and Lp = 5.28 mm due to Eqs. (9.19)–
(9.21). The critical load was Fcr = 400 N, when interface debonding began in the
transverse fracture process of bamboo SENB specimen.
Due to Eq. (9.1), the energy consumed by the cracking of ground tissue was
The interface debonding could be treated as interface crack, and here it was
treated as Mode I interlaminar crack. According to Shao et al. [8], the Mode I
interlaminar fracture toughness of bamboo was GIC = 0.385 N/mm, so the fol-
lowing equation was derived:
" #
3F 2 ðS Ld Þ2 1
Gi ¼ 1 ¼ GIC ð9:24Þ
8EB2 H 3 ð1 e=HÞ3
The energy consumed by the cracking of a single layer was not high, thus no
significant change would be found on the load–deflection (F-d) curve of
transverse-bending fracture. However, multilayer cracking would happen, so much
energy would be consumed in the whole process of transverse-bending failure.
Because of the discrete distribution of fiber bundles, the laminated crack was not
straight or continuous, which brought difficulty to the calculation of the energy
consumed by multilayer cracking. As seen in Fig. 9.4, the distribution of vascular
bundles on the cross section upon the crack was shown. If it was possible that
during crack propagation, each vascular, the crack encountered, would cause a new
layering, thus it was supposed that there were about 14 layers that laminated
9.4 Results 211
cracking happened, because there were about 14 vascular bundles through the
center line of the cross section of bamboo specimen. The energy consumed by the
cracking of all layers in the process was
X
14
Uitotal ¼ Uii 1:578 J ð9:26Þ
i¼1
Due to Eqs. (9.11)–(9.14), the energy consumed by the fracture of fiber bundles
was
` ´
total
Ufu ¼ N Ufu
v
¼ N ðUfu þ 2Ufu þ Ufu Þ ¼ 1:628 J ð9:27Þ
So, the total energy consumed by four fracture patterns in the whole process was
Compared the theoretical result Umodel = 7.273 J with the result of test
Utest = 6.272 J, the theoretical result was 15.68% higher than the test result.
Among the four damage patterns in the process of bamboo transverse fracture, the
work done by pulling the fiber bundles out contributed most to the fracture work,
then it was work done by the fracture of fiber bundles and interface debonding, and
the energy absorbed by ground tissue cracking was low. The phenomena above
corresponded well with the characteristic of the transverse bending fracture surface.
The SEM images of the transverse bending fracture surface were shown in
Fig. 9.10. Figure 9.10a, b was the fracture surface of the ground tissue nearby the
inner part of bamboo, and it could be seen that ground tissue was torn transversely,
and the fracture surface was flat, so the energy absorbed was low. Figure 9.10c–e
was the SEM images of the fiber bundle pulled-out. Although single fiber cell was
presented at the fracture surface, the surface was flat relatively on the whole. It
could also be observed that there were a large amount of torn fragments on the side
face of the fiber bundle. Comparatively speaking, the fracture of fiber bundles with
high modulus did not absorb much energy and so to interface debonding for the
weak interface tensile strength, but the shearing failure caused by the pulling-out of
212 9 Modeling on the Toughness Fracture and Energy-Absorbing …
(a)
(b)
(c) (d)
(e)
Fig. 9.10 SEM images of bamboo transverse-bending fracture surface. a Ground tissue fractured
transversely; b the flat fracture surfaces of ground tissue; c fiber bundles were pulled out;
d fragments on the side of fiber bundle; e separated single fiber
the fractured fiber bundle from the ground tissue did absorb a considerable amount
of energy. The fracture surfaces demonstrated the validity of the theoretical model.
9.5 Verification
y
F
r x
Fig. 9.11 The side crack of bamboo SENB specimen was assumed to be Mode I (opening mode)
crack
Finite element was applied to analyze the crack propagation model of bamboo
SENB specimen, and it was proved that there were both normal stress and shear
stress at the tip of side crack, so the propagation of side crack was I + II mixed
mode. In general, the Mode II interlaminar fracture toughness (GIIC) of composites
was much higher than its Mode I interlaminar fracture toughness (GIC) [36], so as to
bamboo [8, 9]. Thus, from the angle of energy, for bamboo, it was thought that
when the energy release rate reached the value of GIC which was much lower than
that of GIIC, crack would preferentially propagate in Mode I pattern, as seen in
Fig. 9.11.
To measure the Mode I interlaminar fracture toughness, tests were performed
according to Shao et al. [8]. The double cantilever beam (DCB) specimen was
connected with steel U hook by steel pin with a crosshead speed between 1 and
5 mm/min. The crosshead speed was between 1 and 5 mm/min, and curves of the
applied load versus opening displacement (F-d) were recorded automatically by the
computer during the test. And once load (F) went down, the load machine is
stopped, and the crack tip at both the inner part and the outer part of DCB specimen
was marked. Then, it was unloaded and reloaded and the same procedure was
repeated until the DCB specimen fractured completely. As seen in Fig. 9.12, it
showed the F-d curves of one DCB specimen and the loading procedure was
repeated eight times, so there were eight curves and each curve corresponded to
different crack lengths. The specimen from test machine was taken off, the crack
length was measured. As resistance-arresting crack propagation of out layer
(ai − outer) is lower than that of inner layer (ai − inner) of bamboo, the average crack
length of two sides was taken as the actual crack length, namely ai ¼ aiinner þ2 aiouter .
Due to Shao et al. [8], the strain energy U and its release rate GIC can be
calculated as follows:
214 9 Modeling on the Toughness Fracture and Energy-Absorbing …
F (kN)
0.05
0.04
0.03
0.02
0.01
0
0 2 4 6 8 10 12 14 16
δ (mm)
1 1
U ¼ Fd ¼ F 2 C ð9:30Þ
2 2
@U Fcr2 @C
GIC ¼ ¼ ð9:31Þ
@A 2b @a
l 3 ð y2 y1 Þ
E¼ ð9:32Þ
4bh3 ðx2 x1 Þ
For bamboo SENB specimen, the size was 160(L) 13(H) b (Thickness)
(mm). The average value of b was 6.56 mm, and e = 3 mm, and span length
S = 120 mm. According to Eq. (9.31), the Mode I interlaminar fracture toughness
of bamboo could be obtained, and the average value was GIC = 484.22 N/m
(SD = 74.34 N/m). The average value of elastic modulus was E = 8811.20 MPa
(SD = 1051.07 MPa). The average value of load when side crack initiated was
F0 = 310.73 N. Thus, substituted the parameters above into Eq. (9.3), then the
theoretical maximum length of side crack could be obtained, Ld = 23.53 mm.
Experiments were performed to verify the rationality of theoretical result.
9.5 Verification 215
F (kN)
0.08
0.06
0.04
0.02
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Deflection (mm)
9.5.2 Experiments
0.6 B
0.5
0.4
Load (kN)
0.3
0.2
0.1
C
O
0
0 2 4 6 8 10 12 14 16 18
Deflection (mm)
Fig. 9.14 The transverse fracture F-deflection curve of one SENB specimen
Fig. 9.15 The propagation of side crack during the transverse fracture of bamboo SENB
specimen (Unit mm)
The theoretical length of side crack in the transverse fracture process of SENB
specimen was 23.53 mm, which was a little longer than the experimental measured
value 21.39 mm, about 10%, which could be explained by the comparison analysis
on the fracture surfaces.
As seen in Fig. 9.18, the scanning electron microscope (SEM) images of the
fracture surfaces of side crack and Mode I interlaminar crack were shown. It could
be seen that the surface of ground tissue presented to be hackle on the whole in the
side crack zone of bamboo SENB specimen, while the surface of most ground
tissue cells was flat and only a few ground tissue cells were damaged (Fig. 9.18a).
The failure characteristics of fibers in the side crack zone of bamboo SENB
specimen was shown in Fig. 9.18b and the ribbon-like traces on the surface of the
damaged intercellular layer of fiber cells could be observed obviously. The feature
9.5 Verification 217
15 10 15
10
5 5
Fig. 9.16 The fracture surface of failure bamboo SENB specimen (Unit mm)
35
Outer side
30 Inner side
25
C (mm)
20
15
10
5
0
1 2 3 4 5 6 7 8 9 10 11
Specimen
Fig. 9.17 The length of side crack on the outer and inner surfaces of bamboo SENB specimen
of ground tissue in the Mode I crack zone of bamboo DCB specimen was similar to
that in the side crack zone of bamboo SENB specimen (Fig. 9.18c). However, the
surface of fiber cells in the Mode I crack zone was much more smooth than that in
the side crack zone as seen in Fig. 9.18d that was in accordance with Shao et al. [8].
218 9 Modeling on the Toughness Fracture and Energy-Absorbing …
(a) (b)
Cracking direction
(c) (d)
Cracking direction
Fig. 9.18 SEM images of fracture surfaces: a, b The fracture surfaces of ground tissue and fibers
in the side crack zone of bamboo SENB specimen; c, d The fracture surfaces of ground tissue and
fibers in the Mode I zone of bamboo DCB specimen
Gi > GIC, so the theoretical result (Ld) was slightly higher than the experimentally
measured result (Le).
However, in practice, it is difficult to determine the length of side crack, the
energy absorbed by the propagation of side crack and its fracture mechanism
because of the complicated transverse fracture process and jagged fracture surface
of bamboo or FRP. Due to the observation on the propagation behavior of side
crack of SENB specimen, it can be treated as the propagation behavior of Mode I
crack to evaluate the crack length of side crack and energy absorbed by the
9.5 Verification 219
propagation of side crack of bamboo or FRP. After all, for the Mode I interlaminar
fracture of both bamboo and FRP, the test method is quite mature [38–41].
9.6 Conclusion
References
1. Ma JF, Chen WY, Zhao L et al (2008) Elastic buckling of bionic cylindrical shells based on
bamboo. J Bionic Eng 5:231–238
2. Xing DH, Chen WY, Xing DJ et al (2013) Lightweight design for thin-walled cylindrical shell
based on action mechanism of bamboo node. J Mech Design 135:014501
3. Cheng WM, Fu WG, Zhang ZQ et al (2012) The bionic lightweight design of the mid-rail box
girder based on the bamboo structure. Prz Elektrotechniczn 88:113–117
4. Xie F, Zhang ZP, Li X et al (2013) Bamboo section’s bionic design based on bamboo. AMM,
pp 385–386, 34–38. ISSN: 1660-9336
5. Ghavami K (2005) Bamboo as reinforcement in structural concrete elements. Cement
Concrete Comp 27:637–649
6. Ghavami K, Rodrigues CS, Pacionik S (2003) Bamboo: functionally graded composite
material. Asian J Civil Eng 4:1–10
7. Parameswaran N, Liese W (1976) On the fine structure of bamboo fibres. Wood Sci Technol
10:231–246
8. Shao ZP, Fang CH, Tian GL (2009) Mode I interlaminar fracture property of moso bamboo
(Phyllostachys pubescens). Wood Sci Technol 43:527–536
9. Wang FL, Shao ZP, Wu YJ (2013) Mode II interlaminar fracture properties of moso bamboo.
Compos Part B—Eng 44:242–247
10. Tan T, Rahbar N, Allameh SM et al (2011) Mechanical properties of functionally graded
hierarchical bamboo structures. Acta Biomater 7:3796–3803
11. Habibi MK, Lu Y (2014) Crack propagation in bamboo’s hierarchical cellular structure. Sci
Rep-UK 4:5598
12. Kim JK, Baillie C, Mai YW (1991) Interfacial debonding and fibre pull-out stresses: Part I
critical comparison of existing theories with experiments. J Mater Sci 27:3143–3154
220 9 Modeling on the Toughness Fracture and Energy-Absorbing …
13. Zhou LM, Kim JK, Mai YW (1992) Interfacial debonding and fibre pull-out stresses: Part II A
new model based on the fracture mechanics approach. J Mater Sci 27:3155–3166
14. Wang C (1997) Fracture mechanics of single-fibre pull-out test. J Mater Sci 32:483–490
15. Li LB (2015) Modeling the effect of interface wear on fatigue hysteresis behavior of carbon
fiber-reinforced ceramic-matrix composites. Appl Compos Mater 22:887–920
16. Meng QH, Wang ZQ (2015) Theoretical analysis of interfacial debonding and fiber pull-out in
fiber-reinforced polymer-matrix composites. Arch Appl Mech 85:745–759
17. Kim JK, Mai YW (1991) High strength, high fracture toughness fibre composites with
interface control a review. Compos Sci Technol 41:333–378
18. Chiang YC (2001) On fiber debonding and matrix cracking in fiber-reinforced ceramics.
Compos Sci Technol 61:1743–1756
19. Charalambides PG, Lund J, Evans AG et al (1989) A test specimen for determining the
fracture resistance of bimaterial interfaces. J Appl Mech 56:77–82
20. Williams JG (1988) On the calculation of energy release rates for crack laminates. Int J Fract
36:101–119
21. Williams JG (1989) End corrections for orthotropic DCB specimens. Compos Sci Technol
35:367–376
22. Wang Y, Williams JG (1992) Corrections for mode II fracture toughness specimens of
composite materials. Compos Sci Technol 43:251–256
23. Reeder JR, Crews JH (1990) Mixed-mode bending method for delamination testing. AIAA
28:1270–1276
24. Reeder JR and Crews JH (1991) Redesign of mixed-mode bending test for delamination
toughness. In: Proceedings of ICCM-8 Section, pp 13–36
25. Piggott MR (1970) Theoretical estimation of fracture toughness of fiber composites. J Mater
Sci 5:669
26. Beaumont PWR, Fitz-Randolph J, Phillips DC et al (1971) Acoustic emission studies of a
boron-epoxy composite. J Compos Mater 5:542–548
27. Thouless MD, Evans AG (1988) Effects of pull-out on the mechanical properties of
ceramic-matrix composites. Acta Metall 36:517–522
28. Hsueh CH (1994) Pull-out of a ductile fibre from a brittle matrix: PART II a simplified mode.
J Mater Sci 29:5135–5140
29. Guo SQ, Kagawa Y (2002) Tensile fracture behavior of continuous SiC fiber-reinforced SiC
matrix composites at elevated temperatures and correlation to in situ constituent properties.
J Eur Ceram Soc 22:2349–2356
30. Gao XL, Li K (2005) A shear-lag model for carbon nanotube-reinforced polymer composites.
Int J Solids Struct 42:1649–1667
31. Gao YC (1987) Debongding along the interface of composites. Mech Res Commun 14:67–72
32. Li LB, Song YD, Sun ZG (2008) Uniaxial tensile behavior of unidirectional fiber reinforced
ceramic matrix composites. Acta Materiae Compositae Sinica 4:154–160 (in Chinese)
33. Qiao RS (1997) The meso-mechanics property of composite. Northwestern Polytechnical
Press, pp 52–62 (in Chinese)
34. Shao ZP, Fang CH, Huang SX et al (2010) Tensile properties of Moso bamboo (Phyllostachys
pubescens) and its components with respect to its fiber-reinforced composite structure. Wood
Sci Technol 44:655–666
35. Shao ZP, Zhou L, Liu YM et al (2010) Difference of structure and strength between
internodes part and node part of Moso bamboo. J Trop For Sci 22:133–138
36. Jia PR, Jiao GQ (1996) On better test method of determining mixed-mode interlaminar
fracture toughness for composite materials. J Northwest Polytech Univ 14(1):105–109
37. National Technical Monitoring Bureau (1995) National standard GB/T 15780–1995: testing
methods for physical and mechanical properties of wood. China Standard Press, Beijing
38. Triboulot P, Jodin P, Pluvinage G (1984) Validity of fracture mechanics concept applied to
wood by finite element calculation. Wood Sci Technol 18(1):51–58
39. Hodgkinson JM (2000) Mechanical testing of advanced fibre composites. Woodhead
Publishing and CRC Press, Cambridge
References 221
40. American Society of Testing Materials (ASTM) (2001) Standard test method for Mode I
interlaminar fracture toughness of unidirectional fiber-reinforced polymer matrix composites.
Annual book of ASTM standards. Philadelphia, PA, D 5528-01
41. Wang FL, Shao ZP, Wu YJ, Wu D (2014) The toughness contribution of bamboo node to the
Mode I interlaminar fracture toughness of bamboo. Wood Sci Technol 48:1257–1268