Cement Composites

NATURAL FIBRE REINFORCED
CEMENT COMPOSITES
a thesis submitted
for the degree
of
DOCTOR OF PHILOSOPHY
by
YONG NI
\cc
LIBRARY
DEPARTMENT OF MECHANICAL ENGINEERING

VICTORIA UNIVERSITY OF TECHNOLOGY
AUSTRALIA
July, 1995
FTS THESIS
620.137 NI
30001004466811
N1, Yong
Natural fibre reinforced
cement composites
Acknowledgements
I wish to thank my principal supervisor, senior lecturer Dr. B. Tobias for his interest,
help and continual encouragement.
I thank my colleagues and friends in Division of Forest Products, Commonwealth

Scientific & Industrial Research Organisation (CSIRO), where I carried out this work.
A special thanks to the staff of the Pulp and Paper Group and the Fibre Composites
Group, who assisted me in my transition from student to professional researcher.
They are too many to be listed here. In particular, I wish to thank Mr. N. G. Langfors,
A. W. McKenzie (deceased), D. Menz, N. B. Clark, W. J. Chin and C. Garland, Ms.
M. McKenzie and Ms. S. G. Grover, Drs. R. Evans and V. Balodis for their assistance,
advice and criticism at various stages of this investigation.
My debt of thanks is due to Chief Research Scientist Dr. R. S. P. Goutts, who acted as
my external supervisor at CSIRO. I am further indebted to Bob in that he first
introduced me to the field of natural fibre reinforced cement composites / pulp and
paper science, and patiently "nursed" me through the techniques of these fields and
fully guided me throughout the course of this work.
Finally, I am deeply indebted to my wife, family members and friends, who have been
continuaUy understanding, encouraging and supportive during the period of this
research project.
I really wish to say "thank you all".
Yong Ni (Philip)
December 1994
NATURAL FIBRE REINFORCED CEMENT COMPOSITES

CONTENTS
Abstract
vi
List of Tables
viii
List of Figures
Chapter One: Introduction
1.1 Natural fibre reinforced cement composites (NFRC)
1.1.1. History and development of NFRC composites
1.1.2 Fabrication process
1.1.3 Property requirements for asbestos alternatives
1.1.4 Mechanical and physical properties of NFRC
1.1.5 Durability of NFRC
15
1.2 Natural plant fibre resources
16
1.3 Scopes of the present work
19
1.3.1 Bamboo fibre and bamboo-wood hybrid fibre reinforced cement

composites
1.3.2 Influence of fibre properties on composites performance
Chapter Two: Theoretical Principles of Fibre Reinforcement

2.1 Strength and toughness
19
20
22
23
2.1.1 Mixture rule for strength
23
2.1.2 TheACKtiieory
25
2.1.3 Basics of fracture mechanics
28
2.1.4 Fibre critical fracture length
31
2.1.5 Fibre aspect ratio
33
2.1.6 Fracture toughness
35
2.2 Bonding and microstructure of NFRC
38
2.2.1 Chemical bonding
39
2.2.2 Mechanical bonding
40
2.2.3 The effect of moisture on bonding
43
2.2.4 Microstructure of NFRC composites
44
Chapter Three: The Nature of Plant Fibres
52
3.1 Fibre classification
52
3.2 Structure of natural plant fibres
53
3.3 Chemical composition and durability
56
3.4 Preparation of natural plant fibres
59
3.4.1 Pulping
59
3.4.2 Refining or beating
64
3.5 Properties of natural fibres
66
3.5.1 Physical properties
67
3.5.2 Mechanical properties
69
3.5.3 The effect of moisture on fibres
75
Chapter Four: Air-cured & Autoclaved Bamboo Fibre Reinforced Cement

Composites (BFRC)
77
4.1 Experimental work
78
4.1.1 Materials
78
4.1.2 Fibre modification
78
4.1.3 Fabrication and characterisation
79
4.2 Air-cured bamboo fibre reinforced cement
80
80
83
4.3 Autoclaved bamboo fibre reinforced cement
85
85
92
11
4.4 Conclusions
93
Chapter Five: Bamboo & Wood Hybrid Fibre Reinforced Cement Composite
Materials (BWFRC)
95
96
5.1.1 Fibre preparation
96
97
5.2 Results and Discussion
98
5.2.1 Length and freeness of blended pulp
98
5.2.2 Air-cured BWFRC composites
99
5.2.3 Autoclaved BWFRC composites
103
5.3 Theoretical predication and the experimental results
107
5.4 Conclusions
108
Chapter Six: Influence of Fibre Length on Composite Properties

110
112
6.1.1 Fibre length fractionation work
112
113
113
6.2.1 Fractionation of fibre length
113
6.2.2 Influence of fibre length on composite mechanical properties
115
6.2.3 Theoretical and experimental conflict in results
121
6.2.4 Influence of fibre length on composite physical properties
122
6.3 Conclusions
123
Chapter Seven: Influence of Fibre Strength on Composite Properties
125
126
7.1.1 Fibre strength fractionation
126
7.1.2 Fibre quality evaluation
127
7.1.3 Composite fabrication and evaluation
127
128
111
7.2.1 Fibre strength variation work
128
7.2.2 Influence of fibre strength on composite properties
131
7.3 Conclusions
138
Chapter Eight: Influence of Fibre Lignin Content on Composite Properties
139
141
141
8.1.2 Fibre lignin content and evaluation of other properties
142
8.1.3 Composite fabrication and characterisation
142
8.2 Results and discussion
142
8.2.1 Fibre lignin content and other properties
142
8.2.2 Influence of lignin content on air-cured composites
144
8.2.3 Influence of lignin content on autoclaved composites
149
8.3 Conclusions
153
Chapter Nine: Conclusions and Further Work
155
9.1 Conclusions
155
9.2 Recommendations for Further Work
157
9.2.1 Pulp supply
157
9.2.2 Fibre length population distribution and cross-dimensions (coarseness)
158
9.2.3 Conformability - flexibUity and collapsibility
159
9.2.4 The impact of pulp medium consistency treatment on fibre-cement

products
160
9.2.5 Theoretical modeling
161
Appendix A: Fabrication & Characterisation Methods
162
A.l Pulp fibre preparation
162
A.1.1 Chemical pulping (Kraft pulping)
162
A.1.2 Mechanical pulping (TMP, CTMP)
165
A.l.3 Bleaching pulps (Oxygen delignification)
166
A. 1.4 Holocellulose pulp
169
IV
A.1.5 Beating
169
A.l.6 Preparation of fibres from dry lap-pulp
171
A.2 Pulp fibre characterisation
172
A.2.1 Lignin content
172
A.2.2 Drainability (Freeness)
173
A.2.3 Fibre lengtii
175
A.2.4 Handsheet preparation
177
A.2.5 Fibre strength
179
A.3 Fabrication composite materials
181
A.3.1 Materials
182
A.3.2 Slurry/vacuum dewatering and press technique
183
A.3.3 Air curing and autoclaving
184
A.4 Characterisation of composite materials
186
AppendixB: Determination of Kraft pulping parameters
188
References
189
Bibliography
202
Abstract
The health problems associated with asbestos and its related products necessitated in
finding alternative resource of fibres. Over the last two decades natural fibre (mainly
wood pulp fibre) has emerged as the most acceptable alternative reinforcement for
fibre cement products. The first three chapters of this study describe in some depth
the preparation and properties of natural fibres, the methods of incorporating such
fibres into cements and mortars, the theoretical principles of fibre reinforcement, the
properties obtained from these natural fibre (mainly wood pulp fibre) reinforced
cement composites and their applications as commercial products, especially as the
main alternatives to asbestos reinforced cement materials. Chapter four and five
discuss fabrication and performance characterisation of the resulting composites.
Whereas, chapters six, seven, eight and nine incorporate results and conclusions. The
detailed experimental procedures and methods are described in Appendix A and B.
Bamboo pulp fibre was investigated as reinforcement for incorporation into cements
and mortars. The results show that bamboo fibre is a satisfactory fibre for
incorporation into a cement matrix. The composites so formed have acceptable
flexural strength, but lack fracture toughness due to the short fibre length and the high
fines content of the bamboo pulp used in this study. Experimentation was conducted
in an attempt to improve the fracture toughness properties of bamboo fibre reinforced
cement composites. Blending bamboo fibre with varying proportions of softwood
fibre led to a range of materials with improved performance, especially with respect to
the property of fracture toughness.
VI
There is a need to be able to specify the properties of natural ceUulose fibres, in

particular wood pulp fibres, when they are to be used as reinforcement in fibre cement
products. Preliminary studies have attempted to isolate specific fibre properties such
as fibre length, fibre strength and fibre lignin content. Then these parameters were
used to relate the mechanical performance of fibre cement composites. This study has
shown that the fibre length plays a major role in the development of flexural strength
and fracture toughness of a fibre reinforced composite. As the length increases both
properties improve over a range of fibre contents. There is some concern however,
over the best manner of specifying this parameter - average fibre lengths or fibre
length distributions. Fibre strength was found to have less effect on composite
flexural strength than on fracture toughness values, for both air-cured and autoclaved
products. It was somewhat unexpected that weak fibres could provide relatively
strong, but; as might be expected, brittle materials. Fibre lignin content was shown to
have a considerable effect upon the formation of autoclaved products. There is
difficulty, however, to examine fibre lignin content as an isolated parameter, due to
interaction from fibre strength, stiffness and fibre-matrix interface bonding.
Overall, this study has provided better understanding of the complex behaviour of
natural fibre reinforced cement composites.
vu
List of Tables
Table 1.1
Comparison of properties of fibres for possible asbestos
alternatives
Table 1.2
Mechanical and physical properties of commercial WFRC
14
materials based on James Hardie's products

Table 1.3
Mechanical and physical properties of laboratory fabricated
14
NFRC
Table 1.4
Annual collectable yields of various non-wood plant fibrous
18
raw materials
Table 1.5
Availability of various non-wood plant fibrous raw materials,
18
1982
Table 1.6
Total production of various non-wood plant fibre pulps in 1982
19
Table 2.1
Effect of aspect ratio and fibre content
34
Table 3.1
Chemical compositions of natural plant fibres
57
Table 3.2
Typical softwood and hardwood fibres physical dimensions
67
Table 3.3
Some non-wood fibres physical dimensions
68
Table 3.4
Some grasses pulps physical dimensions
68
Table 3.5
Stracture and strength parameters of non-wood fibres
72
Table 4.1
Properties of air-cured bamboo-fibre-reinforced cement
81
Table 4.2
Properties of autoclaved bamboo-fibre-reinforced cement
86
Table 4.3
Fibre weighted average length (mm)
87
Table 4.4
Fibre length mass distribution percentage
87
Table 4.5
Properties of autoclaved screened long bamboo fibre reinforced
88
cement
Table 5.1
Length and freeness of blended pulp
99
Table 5.2
Properties of air-cured BWFRC
99
Table 5.3
Properties of autoclaved BWFRC
104
vui
Table 5.4
Regression analysis results
108
Table 6.1
Fibre length fractions
114
Table 6.2
Relationship between fibre length and air-cured composite
115
performance
Table 7.1
Properties of P.radiata fibre after alkaline cooking
129
Table 7.2
Composites properties reinforced with alkaline treated
131
holocellulose fibres
Table 7.3
Composites properties reinforced with acid hydrolysis fibres
136
Table 8.1
Pulping techniques
141
Table 8.2
Fibre hgnin content, freeness and fibre strength, fibre length
142
Table 8.3
Influence of fibre lignin content on air-cured WFRC
146
Table 8.4
Influence of fibre lignin content on autoclaved WFRC
150
Table A. 1
Common bleaching chemicals
168
Table A.2
Pulp test methods
172
Table A.3
Natural fibre reinforced cement composites ingredients
182
ix
List of Figures
Fig. 1.1
The Hatschek process
Fig. 1.2
Graph of flexural strength via fibre content for various WFRC
11
products
Fig. 2.1
Schematic representation of crack travelling through a fibre
23
reinforced matrix
Fig. 2.2
Tensile stress strain curves for fibre reinforced brittle matrices
26
predicted by the ACK theory (full line), and the bending

response calculated from them (broken lines)
Fig. 2.3
Theoretical model applicable to low modulus fibre-reinforced
27
cement composite at flexural failure

Fig. 2.4
Tensile load-extension curves for different failure modes of
33
sisal silvers embedded in cement

Fig. 2.5
Possible coupling mechanism between wood fibre and cement
41
matrix
Fig. 2.6
SEM showing fracture surface of WFRC preconditioned at
46
100 - 105"C for 24h

Fig. 2.7
SEM showing fracture surface of WFRC preconditioned by
47
soaking in water for 48h

Fig. 2.8a
SEM showing fracture surface of WFRC preconditioned at 50
47
5 % relative humidity and 22 2C

Fig. 2.8b
As (a) but higher magnification
48
Fig. 2.9
SEM showing cement surface at interface contains dense
49
matrix with some discontinuities

Fig. 2.10
SEM shows fractured fibres with dense material from bulk of
49
matrix up to fibre wall
Fig. 3.1
The structure of wood fibre
Fig. 3.2
The structure of bamboo
Fig. 3.3
Scanning electron micrograph of a cube of eastern white pine
54
fibre
54
55
microtomed on three surfaces

Fig. 3.4
The principle of chemical and mechanical pulping
62
Fig. 3.5
(a, top) Unbeaten fibre of P.radiata, compared to (b)
66
externally fibrillated fibre

Fig. 3.6
A pulp fibre consists of cellulose fibrUs in largely parallel
71
array, embedded in a matrix of hgnin and hemicellulose

Fig. 3.7
The maximum strength of a pulp fibre is limited by the
71
inherent properties of the cellulose fibril modified by the

spiral angle of the fibril about the fibre axis
Fig. 3.8
Laboratory beating - strength tests on chemical pulps from
73
various wood and non-wood plant fibres
Fig. 4.1
Effect of fibre content on flexural strength for air-cured
80
WFRC and BFRC

Fig. 4.2
Effect of fibre content on fracture toughness for air-cured
82
WFRC and BFRC

Fig. 4.3
Effect of fibre content on water absorption for air-cured
84
WFRC and BFRC

Fig. 4.4
Effect of fibre content on density for air-cured WFRC and
85
BFRC
Fig. 4.5
Flexural stiength as a function of percent fibre loading for
87
autoclaved BFRC and WFRC composites

Fig. 4.6
Fracture toughness as a function of percent fibre loading for
90
autoclaved BFRC and WFRC composites

Fig. 4.7
Typical Load / Deflection graph for autoclaved WFRC and
91
BFRC composites
Fig. 4.8
Density as a function of percent fibre loading for autoclaved
92
BFRC composites
Fig. 4.9
The relationship between density and water absorption for
93
autoclaved BFRC composite
Fig. 5.1
Relationship between pine fibre proportion and furnish pulp
97
XI
length weighted average

Fig. 5.2
Relationship between pine fibre proportion and furnish pulp
98
freeness value
Fig. 5.3
Influence of long fibre (pine) proportion on the air-cured
100
composites flexural strength

Fig. 5.4
101
Influence of long fibre (pine) proportion on the air-cured

composites fracture toughness
Fig. 5.5
Influence of fibre length on the air-cured composites
flexural
102
strength at total 8% fibre content

Fig. 5.6
Influence of fibre length on the air-cured composites fracture
102
toughness at total 8% fibre content

Fig. 5.7
105
Influence of long fibre (pine) proportion on the autoclaved

composites flexural strength
Fig. 5.8
Influence of fibre length on the autoclaved composites
flexural
105
strength at total 8% fibre content

Fig. 5.9
Influence of long fibre (pine) proportion on the autoclaved
106
composites fracture toughness

Fig. 5.10 Influence of fibre length on the autoclaved composites fracture
106
toughness at total 8% fibre content
Fig. 6.1
Fibre length population of Bauer-McNett technique four
114
length fraction
Fig. 6.2
Effect of fibre content on composite flexural strength for
116
different fibre lengths

Fig. 6.3
Influence of fibre length on composite flexural strength at 8%
117
fibre content
Fig. 6.4
Effect of fibre content on composite fracture toughness for
118
different fibre lengths

Fig. 6.5
Influence of fibre length to fracture toughness at 8% fibre by
118
content
Fig. 6.6
Influence of fibre WL / d on composite fracture toughness
120
Fig. 6.7
Influence of fibre length on composite density
123
Xll
Fig. 7.1
Relationship between fibre zero-span tensile strength and pulp
129
viscosity
Fig. 7.2
Relationship between fibre strength and fibre length
129
Fig. 7.3
Relationship between fibre strength and pulp freeness
130
Fig. 7.4
Relationship between fibre strength (0-span) and composite
132
strength
Fig. 7.5
Relationship between fibre strength (viscosity) and composite
132
strength
Fig. 7.6
Relationship between fibre strength (0-span) and composite
133
fracture toughness
Fig. 7.7
Relationship between fibre strength (viscosity) and composite
133
fracture toughness
Fig. 7.8
134
Fracture surface of composite reinforced with strong fibre

shows fibre pull-out
Fig. 7.9
135
Fracture surface of composite reinforced with weak fibre

shows fibre fracture
Fig. 7.10
Relationship between fibre acid treated time and composite
137
strength
Fig. 7.11 Relationship between fibre acid treated time and composite
137
fracture toughness
Fig. 8.1
Influence of fibre lignin content on air-cured WFRC
flexural
145
strength at different fibre content

Fig. 8.2
Influence of fibre lignin content on air-cured WFRC fracture
145
toughness at different fibre content

Fig. 8.3
Influence of fibre lignin content on air-cured WFRC density at
148
different fibre content

Fig. 8.4
Influence of fibre lignin content on autoclaved WFRC
flexural
149
strength at different fibre content

Fig. 8.5
Influence of fibre hgnin content on autoclaved WFRC fracture
151
toughness at different fibre content
Xlll
Fig. 8.6
Influence of fibre lignin content on autoclaved WFRC density
153
at different fibre content
Fig. A. 1
Schematic iUustration of Air-bath and 3-Iitre pulping vessels
163
Fig. A.2
Open periphery (G) and closed periphery (B) refining plates
167
Fig. A.3
The Valley Niagara beater
170
Fig. A.4
The PFI laboratory beater tackle
171
Fig. A.5
Canada Freeness tester
174
Fig. A.6
The Bauer-McNett Classifier
Fig. A.7
Kajanni FS-200 measurement principle
176
Fig. A. 8
Schematic illustration of Pulmac Zero-span tester
181
Fig. A.9
Vacuum dewatering casting box
184
Fig. AlO
Optimize autoclaving temperature and
176
time
185
Fig. Al 1 Optimize composite OPC:Silica matrix ratio for autoclaving
xiv
Chapter One:
Introduction
1.1 Natural Plant Fibre Reinforced Cement Composites (NFRC)
Since the end of the 19th century, asbestos cement has had a wide range of applications,
such as building / cladding sheets, corrugated roofing elements, pipes and tiles.
Because of the well known health risks associated with the use of asbestos fibres, together
with a possible future shortage of asbestos, there has in recent years, been considerable
research into the development of new high-performance reinforcing fibres. A variety of
fibres such as glass, steel, synthetic polymer and natural cellulose fibres have been
evaluated in both laboratory and pilot plant equipment. Among these fibres, natural
cellulose fibres (Kraft pulped soft wood fibres) demonstrate both cost effectiveness and
suitable performance to act as a replacement for asbestos fibre. James Hardie Industries in
Australia marketed non-asbestos wood fibre reinforced cement (WFRC) boards in 1981
(Anon, 1981).
1.1.1 History and development of (NFRC) composites

Although patents from the last century refer to the use of natural plant fibre as a
component of building materials made from cements and plasters (US Patent,
1884,1899,1900), interest in natural fibre (mainly wood fibre) as a reinforcement for fibre
cement has mainly taken place in the last 10-20 years.
Unfortunately the large fibre cement manufacturing companies are the real custodians of
the history of the fibre cement development and, from the obvious gap in the literature,
they have released very little information about the use of natural fibres in cement.
James Hardie and Coy Pty Ltd started manufacturing asbestos cement products in
Australia in 1917 (James Hardie Industries, 1984). James Hardie Industries took an active
interest in the use of ceUuIose, as an economic asbestos substitute, in fibre reinforced
cement in the early to mid-1940s. This work was intensified during the post-World War II
years when there was a worldwide shortage of asbestos fibre. An investigation was
conducted by Heath and Hackworthy to discover whether paper pulp could be used to
replace asbestos completely or partially in asbestos cement sheets (James Hardie and Co.
Pty Ltd., 1947). Fibres studied included bagasse, groundwood, wheat straw, cement bags
and brown paper. The experimental autoclaved sheets showed brown paper (Kraft) was
the best of the pulp sources, giving greatest strength to the composite material. However,
when asbestos supply was reinstated, this work was discontinued.
Renewed interest in wood fibres began almost inadvertently in 1960. In those days, the
asbestos fibre board, containing 15% asbestos, was made between steel interleaves. James
Hardie's was beheved to be the only group in the world which at that time was steamcuring its sheets. To make a cheap board as an alternative interleaf, boards were made up
with half the asbestos replaced by wood fibres. This board became the first generation
"Hardiflex", and full production started in 1964. From the 1960s onwards their products
have contained no more than 8% asbestos, which was about half the amount used by the
rest of the industry.
Attempts to further reduce the asbestos content by adding more wood fibre were
unsuccessful due to the ineffectiveness of these fibres, compared to asbestos, in trapping
the cement particles during formation of the sheet.
James Hardie entered into collaborative research with CSIRO in 1978 to study, among
other things the refining of cellulose fibres in an attempt to overcome the difficulties of
retaining the cement in the wood fibre reinforced cement sheet (Anon, 1981). By May
1981 the new generation of asbestos-free cement products - Hardiflex II - was being
commercially manufactured. This autoclaved product was totally reinforced by refined
Kraft wood fibres (Goutts, 1982a; Aus Patent, 1981).
In Europe during 1975 - 77, Cape Industries had made boards reinforced with 5% cellulose
and high levels of mineral fiUers called "Supalux", "Monolux" and "Vermiculux", mainly
for fire-resistant use. "Masterboard" and "Masterclad" were more dense and stronger and
used for external cladding (Harper, 1982). In 1976 Sweden banned asbestos cement
products, and all asbestos cement production was closed down. Other Scandinavian
countries were forced to use alternatives, and A /S Norcem in Norway and OY Partek AB
in Finland decided to form a joint development under the name of NOPA (Pedersen,
1980). The material produced was called "Cellcem", and contained cellulose fibres mixed
with other fibres. Manufacture of the products "Intemit" and "Pemit" started in Norway in
1977. In 1979 Finland started to produce "Minerit".
In was stated in 1985 that the UK manufacturers had replaced asbestos in about 50% of the
fibre cement sheeting products (Crabtree, 1986). James Hardie Industries by this time had
totally replaced asbestos fibre from its range of building products, which included flat
sheet, corrugated roofing and moulded products.
As well as flat sheet products, James Hardie Industries had become a world leader in
injection moulded fibre cement products and non-pressure fibre cement pipes, all based on
wood fibre as the reinforcement material. The first experimental production of WFRC
pipe was undertaken at the Brooklyn factory in September 1980. Commercial production
began in Westem Australia at the Welshpool factory in July 1984. The last asbestos pipes
made by James Hardie were manufactured in March 1987.
At the present time there is considerable activity in the patent literature concerning the use
of natural plant fibres (mainly wood fibres) or mixtures of natural plant fibres (mainly
wood fibres) with other synthetic fibres. This is taking place throughout Europe and
Japan, and companies such as Dansk Eternit-Fabrik A /S, Cape Boards and Panels Ltd,
Partek of Finland, Asano of Japan and others are involved.
1.1.2 Fabrication processes

The manufacture of asbestos fibre cement products is a mature industry. Hodgson (1985)
suggested that over 1100 sheet machines and 500 pipe machines were in production
around the globe, with total capital investment in excess of A$ 4.6 billion. The Hatschek
process (or wet process) is the most widely used method of production (see Figure 1.1).
The manufacturmg techniques are closely related to conventional heavy paper and board
making processes. An aqueous slurry of asbestos and cement matrix, about 7-10% solids
by weight, is supplied to a holding tank which contains a number of rotating screen
cylinders. The cylinders pick up the solid matter removing some of the water m the
operation. An endless felt band travels over the top surfaces of the cylinders and picks up
a thin layer of formulation from each cylinder. The " built-up " laminated ply then travels
over vacuum de-watering devices which remove most of the water. The formulation is
then wound up on a steel calender, or assimilation roll, untU a product of desire thickness
is formed. The material is further compressed by pressure rolls which are in contact with
the assimilation rolls.
Vf
SLURBYFEEO
CALENDER
PBOOUCT
SCHEEN CYLINDERS
HATSCHEK PROCESS
Fig. 1.1. The Hatschek process.
For sheet production the layer built up on the assimilation roll is automatically cut off and
drops onto a conveyor to be transferred to a stack for curing. If corrugated roofing is to be
made, the flat sheet is taken off to a corrugating station where the sheets are deposited
onto oiled steel moulds for shaping. Pipe machines are similar to the Hatschek process but
usually have only one or two vats in series. The pressure imposed on the mandrel by the
press roUs is much greater than for sheet production so as to form a dense product. The
machine may be stopped whUe the mandrel carrying the pipe is set to one side for pipe
withdrawal. The process is often referred to as the Mazza process.
The Magnani (or semi-dry) process can be used to prepare pipes and corrugated sheet.
This process has the advantage that it can provide a greater thickness of material at the
peaks and troughs of the corrugations and so increase the bending strength. The thick
slurry (about 50% soHds) of this process can flow uniformly and dkectly onto a felt
conveyor which passes over numerous vacuum boxes to dewater the formulation. In the
case of corrugated roofing the felt is compressed over a corrugated former by a shaped
roller. Pipe formation is similar to the Mazza process.
Injection moulding is now tending to replace the hand moulding of green sheets (from the
Hatschek process) for the manufacture of special fittings. A slurry of 40-50% solids is
pumped into a permeable mould and then subjected to pressures, in excess of 20
atmospheres, in a hydraulic press via a rubber diaphragm. The mix is dewatered by this
process of pressure filtration, and then has sufficient green strength for the product to be
demoulded by means of a suction lifting pad, and transferred to a pallet for curing. The
operation is very fast.
The formulation of the matrix, and hence the cure of the product, has varied from country
to country and between companies within a country. The formulations remain confidential
to the company or its licensees and only general details will be discussed here. The
autoclaved curing process has always been favoured in Australia and the USA, and in
some European countries. In the autoclave process, the matrix is usually a mixture of
ordinary Portland cement (OPC) and finely ground sand (silica), or lime and silica. The
product, after an initial pre-cure period in air, is cured in an autoclave in a steam
environment, say 8 hours at 170-180 C. The cured sheets are virtually at fuU strength
after autoclaving and can be dispatched from the factory in a short time. By contrast the
more traditional air-cured products require 14-28 days of air-curing before they can be
dispatched, this involves considerable stock inventory. The air-curing process is lower in
capital outlay, as no high pressure autoclaves and steam raising plant are required;
however, cement is more expensive than silica, and therefore material costs are higher.
1.1.3 Property requirements for asbestos alternatives

After reviewing generally the processes used to manufacture asbestos fibre cement, a
number of requirements of replacement fibres can be noted, if the existing capital
intensive equipment is to be used. For the Hatschek process a replacement fibre must be
water dispersable in a relatively dilute slurry and able to form a film on the screens. At the
same time the fibre must be able to resist chemical attack due to the high alkalinity (~ pH
13) of the matrix. If the product is to be autoclaved, resistance to temperatures above 170
C is also required. The basic essentials of cost, availability and mechanical performance
of the fibre are obvious.
As has been reported glass, steel, carbon, synthetic organic as well as natural plant fibres
(mainly wood pulp fibre) have been under examination for use in cement systems. We will
look at a comparison of the properties of these fibres as possible asbestos replacements in
existing processes in Table 1.1.
For countries committed to autoclaved products the combination of high alkalinity and
high temperature eliminates most fibres apart from natural plant fibres (mainly wood pulp
fibre), steel, carbon and aramid fibres. The cost of the latter two is almost a factor of
twenty times higher than natural plant fibres (eg. wood pulp fibre) and so look
unattractive. Steel fibres have processing problems. If one considers air-curing fibre
cement, to ehminate the temperature problems, there are still processing Hmitations. The
inorganic fibres such as steel or glass, tend to be too stiff or dense to perform well during
fUm-forming from dilute slurries; whUe the organic fibres lack a surface suitable for
bonding to the matrix and / or introduce drainage problems. Mixtures of organic fibres
(mainly PVA) and natural plant fibres (mainly wood pulp fibre) fibres were successfuUy
used to produce air-cured products in Europe (Studinka, 1989).
Table 1.1 Comparison of properties of fibres for possible asbestos alternatives (Goutts,
1988).
Fibre
Wood pulp (chem)
Wood pulp (mech)
Polypropylene
PVA
Kevlar
Steel
Glass
Mineral fibre
Carbon
Alkalia. resist.
1
2
1
1
1
1
3
3
1
Temp resist.
1
2
3
3
Process ability
Strength
1
2
3
3
2
3
3
3
3
1
2
3
1
1
3
3
3
1
Toughness
1
3
3
1
1
3
3
3
1
Price
3
3
2
2
1
2
2
3
1
1. High, 2. Medium, 3. Low.
1.1.4 Mechanical and physical properties of NFRC

As stated above, natural plant fibre (mainly wood fibre) is the most cost / performance
asbestos alternative. Natural plant fibre contains cellulose, hemicellulose and hgnin. The
cellulose fibre is the main reinforcing elements. Most natural plant fibres contain more
than 45% cellulose and some fibres even yield great than 75% cellulose (see section 3.3).
Natural fibre has been successfully employed as an asbestos fibre alternative either by
itself or as mixture with other synthetic fibre for 10 - 15 years. James Hardie Industries in
Australia manufactured a varies of autoclaved fibre cement sheets and pipes reinforced
with about 8%-10% beaten softwood (P. radiata) Kraft pulp. The Etemit Group in
Switzerland promoted various types of cellulose fibre reinforced autoclave composites in

South Africa and some air-cured cement products in Europe and Latin America reinforced
with synthetic fibres and natural cellulose fibres.
The amount of data available on natural fibre reinforced cement (NFRC) products, in the
scientific literatures, has been limited due to the fact that manufacturing interests had been
responsible for much of the preliminary work and for commercial reasons had retained the
knowledge in house or locked away in patent literature. Unfortunately, due to the
difficulty in handing theoretical treatments involving natural fibres, there was less interest
from the academic fraternity than in say cement materials containing glass, steel or
synthetic organic fibres which form the basis of a voluminous scientific literature.
This chapter will only address some selected results, relating to products containing
natural plant fibres (mainly wood fibres) as the sole reinforcement for cement matrices, in
order to give an appreciation of various effects.
1.1.4.1 Strength and fracture toughness of NFRC

During last two decades a number of papers have appeared in which wood fibres are the
sole source of fibre reinforcement. These studies have included chemical and mechanical
pulps of softwoods and hardwoods in air-cured and autoclaved matiices (Goutts,
1985,1986,1987a).
It will be seen that refined wood fibres can afford a strong, tough and durable fibre
cement, when produced commercially by traditional slurry/dewatered systems followed by
autoclaving. Such WFRC formulations can be used for the production of flat sheeting.
corrugated roofing, moulded products and low-pressure pipes which traditionally have
used asbestos fibre. Sometimes the laboratory experiments are misleading with respect to
the manufacturing processes and care must be taken in extrapolating the laboratory results
into production.
Although natural cellulose fibre (eg. wood pulp fibre) is cheap and readily processed it has
the disadvantage of being hygroscopic. The composite properties are altered by absorption
of water and, for this reason, extensive testing when both wet and dry is required. WFRC
products are generally loaded in bending and so flexural strength has more meaning than
tensile strength in the characterisation of these materials.
At CSIRO Austraha, there was an interest in using high yield pulps [(thermomechanical
pulp (TMP) and chemithermomechanical pulp (CTMP)], as an alternative to chemical
pulps, for reinforcing fibre cements. Such pulps make less demand on the forest resources
for a given quantity of pulp (yields twice that of chemical pulps), less problems with
effluent treatment, chemical requirements are much lower and processing plants are
economical at a smaller scale. Coutts (1986) reported that mechanical pulps of P.radiata
in general were unacceptable as cement reinforcement when autoclaved (with MOR less
than the matrix) but when air-cured had flexural strengths greater than 18 MPa at 8-10%
by mass of fibre. This compares poorly when one notes flexural strengths of WFRC's
containing chemical pulps of P. radiata are in excess of 20 MPa when autoclaved and 30
MPa when air-cured. When autoclaving mechanical pulps the high temperature and
alkalinity virtually "chemically pulps" the fibres releasing extractives of polysaccharides
and wood acids, which "poison" the matrix near the fibre causing poor interfacial bonds.
10
Air-curing is less drastic with respect to chemical attack, hence better properties are
evident in the final composites as shown in Figure 1.2.
30
Q KRAFT AlR-CUmai
^W
%
KRAFT AUTOCLAVED
O TMP AIR-CUREO
<
X
w
ui lO
TMP AijrOCLAVED
8
10
12
RBRE COKTENT (% BY MASS)
Fig, 1.2. Graph offlexuralstrength v. fibre content for various WFRC products (Coutts, 1988).
Although it was documented from work in the laboratory that Kraft wood fibre was
effective as a reinforcement in a cement matrix (Coutts, 1979a), the pulp performed poorly
on a pilot-plant Hatschek machine because the fibres were unable to form a web capable
of retaining cement and silica particles. The open nature of the web permitted rapid
drainage, with loss of matrix, resulting in low product strength. A collaborative project,
between CSIRO and James Hardie Industries, starting in 1978 resulted in laboratory data
which demonstrated the benefit of refining wood fibres for use in WFRC materials
(Coutts, 1982a, 1986; Aus Patent, 1981). The breakthrough that made commercial
production possible came about from the work of the Hardie's team in adapting the fibre
refining step to suit the Hatschek machme. Before launching the product in 1981 over
11
50,000 sheets had been prepared and tested on the pilot-plant, and about $10 miUion
invested in installing refining equipment in the factories.
The effect of moisture on the strength properties of WFRC composites is of importance,

and early on in the research variations in test conditions produced variations in test results.
It was evident that standard conditions must be adopted. The flexural strength of WFRC's
can be reduced to as low as 50% of dry strength values in laboratory samples, in the case
of commercial products the reduction is considerably less but is taken into consideration
for product application.
Fordos and Tram (1986) reported WFRC's containing micro silica with excellent strength
values ranging between 25-55 MPa. A very high stack compression pressure,
approximately 20 MPa, was used. Whereas most results usually report pressures of
approximately 2-3 MPa. Coutts and Warden (1990a) demonstrated the effect on
compaction on the properties of air-cured WFRC and showed that flexural strength
increased with casting pressure without resulting in a reduction in fracture toughness. Few
workers have reported the elastic modulus of WFRC materials. Andonian et.al.(1979)
have shown that both tensile and bending moduli are reduced from approximately 13 GPa
to 9 GPa as the fibre content increases from 2 to 10% fibre by mass.
Fracture toughness is perhaps the most important property for a building material.
Although strength and stiffness are important, the ability of a material to absorb impact
during handling can decide whether it will find an application in the market place. Fibre
type, pulping method, refining conditions and test conditions all have an effect on the
fracture toughness results of a given formulation. As the fibre content increases up to
12
about 8-10% by mass the fracture toughness increases rapidly then starts to taper off
Values of fracture toughness in excess of 60 times matrix values can be obtained with 10%
fibre by mass. Fracture toughness increases even further when tested wet. This effect will
be discussed when we consider bonding and microstructure. TMP pulps when compared
with chemical pulps of the same species tend to have fracture toughness values less than
half that of the chemical pulp reinforced composite. This can be partly explained in terms
of fibre number and fibre morphology (Goutts, 1986). The variation of fracture toughness
values between softwoods and hardwoods can be attributed to fibre length and fibre
morphology, which we will see is so important for fibre pull-out which takes place during
failure under load.
Beside softwood fibre, considerable research has been conducted on some hardwood
fibres, non-wood natural plant fibres and waste paper in order to search for cheaper and
more naturally available fibre resources and to better understand the natural fibre
reinforced cement based composite materials. The mechanical and physical properties of
commercial WFRC and some laboratory fabricated NFRC are listed in Table 1.2 and
Table 1.3, respectively.
It can be seen from Table 1.2 and 1.3 , when natural fibre are prepared by the chemical
pulping method, they can produce acceptable fibre cement products, when either air-cured
or autoclaved, with flexural strength properties comparable to those of softwood.
However, their fracture toughness values tend to vary and most of them less than WFRC.
Such phenomena might be explained by different fibre parameters such as fibre length {eg.
hardwood (Coutts, 1987a), bamboo (Coutts, 1994a), waste paper (Goutts, 1984a)}, fibre
13
content {eg. TMP (Coutts, 1986)}, fibre strength {eg. NZflax (Coutts, 1983c, 1994b)} and
helical angle {eg. banana (Coutts, 1990b)}, while the other phenomena remain uncertain.
Table 1.2 Mechanical and physical properties of commercial WFRC materials based on
James Hardie's products (Coutts, 1988).
Mod. of Blast. (GPa)
Hex. Strength (MPa)
Frac. Tou ghnessCkJ/m'^)
Density
W.Abs.
Product
RH
Wet
RH
Wet
RH
Wet
(g/cm^)
(%)
Villaboard II
(6 mm)
19.2(L)
13.5Cr)
16.3(av)
24(L)
12(T)
18(av)
29.4(L)
22.1(T)
25.7(av)
12.9(L)
8.6(T)
]0.7(av)
/
/
/
21.0(L)
15.9 (T)
18.4(av)
9.2(L)
8.5(T)
8.9(av)
/
/
/
16.6(L)
15.3(T)
15.9(av)
6.1(L)
5.2(T)
5.7(av)
/
/
/
12.7(L)
12.3(T)
12.5(av)
6.7(L)
3.0(1)
4.9(av)
/
/
/
4.2(L)
2.4(T)
3.3(av)
9.0(L)
5.6(1)
7.3(av)
/
/
/
19.5(L)
8.I(T)
13.8(av)
1.31
"
"
1.40
"
"
32.1
"
"
30.5
"
"
18.6
"
"
Hardiflex II
Compressed
sheet II
(12 mm)
1.62
"
"
*L: longitudinal direction of manufacture

*T: transverse direction
Table 1.3 Mechanical and physical properties of laboratory fabricated NFRC
Pulp
Kraft
"
II
tt
II
tt
"
"
"
ti
TMP
"
Fibre
Curing
softwood
air
auto
air
air
auto
air
auto
auto
air
auto
air
auto
"
hardwood
abaca
NZflax
banana
ti
sisal
bamboo
"
softwood
"
Strength(MPa)
30.3
23.1
20.3
27.3
23.2
20.0
18.5
18.3
17.0
15.5
12.3
9.5
Toughness (kJ/m^)
1.93
1.86
1.37
2.08
0.84
0.83
0.55
2.49
0.34
0.29
0.57
0.89
Density(g/cmO
1.55
1.31
1.45
1.55
1.31
15.4
W.Abs.(%)
21.1
33.9
25.8
21.9
35.0
22.9
1.37
1.59
1.41
1.47
1.29
27.6
16.7
30.7
17.8
30.5
Reference
Coutts85
Coutts 84 a
Coutts87a
Coutts87b
Coutts83
Coutts90b
Coutts90b
Coutts92a
Coutts94b
Coutts94a
Coutts86
Coutts86
*A11 composites containing 8%fibre(by mass)

*TMP denotes thermomechanical pulp
1.1.4.2 Physical properties of NFRC

The physical properties of NFRC materials can have a considerable influence on their
acceptability for use in the construction industry. If a product is strong and tough and has
low density it wiU be preferred by the workers, who handle such products on the building
14
site, compared to similar materials that are dense. At the same time due consideration
must be given to water absorption, for as the density is lowered, the void volume increases
with an associated potential increase in water absorption. Thus a load on a structure may
be considerably increased should the material become wet, with more than 30% increase
in weight occurring in some laboratory cases. High temperature mechanical pulps are very
stiff, compared to chemical or low temperature mechanical pulps, and cause poor packing
as the fibre content increases. As void volume increases with poor packing so the density
decreases and water absorption increases. Matrix material also affects the density and
water absorption. Air-cured NFRC materials are more dense than autoclaved materials.
1.1.5 Durability of NFRC

Considerable doubt has been cast on the ability of natural fibres to resist deterioration in
cement matrices, yet no evidence has been put forward to support the claims. When poor
mechanical performance of the composite has been offered as confirmation of fibre failure,
due consideration should be given to the potential of the fibre strength loss and to
"poisoning" of the matrix surrounding the fibre, resulting in weak interfacial bonding. An
extensive review by Gram (1983) reflects this uncertainty.
Lola (1986) reported the failure of natural fibre cement products after only a few years of
service. In many cases the reinforcement fibre used was aggregates of fibres and the high
alkalinity (matrix), coupled with cycles of wetting and drying, "pulped" the fibre bundles
resulting in loss of fibre strength, "poisoned" cement and weak interfacial bonds, and thus
low durability. On the other hand there are many claims which suggest that natural fibre
remforced cement products are durable after 30 years of service.
15
Sharman and Vautier (1986) have done some excellent work on the durability of
autoclaved WFRC products at the Building Research Association of New Zealand. They
discussed the possible ageing mechanisms of corrosion, carbonation, moisture stressing
and microbiological attack.
Akers and co-workers (1986) have published a series of papers which discuss the ageing
behaviour of cellulose fibres both autoclaved and air-cured, in normal environments and
accelerated conditions. Testing had taken place which showed exposure of WFRC
composites to natural weathering led to an overall increase in flexural strength and elastic
modulus after 5 years. The same workers found that air-cured WFRC products when
aged, either normally or by accelerated means, showed a marked reduction in fracture
toughness. The ageing of autoclaved materials did not result in mineralisation of the fibres
and maintained good levels of fracture toughness. The need to use synthetic fibres in aircured products was apparent, however, with aging there appears to be an increase in the
interfacial bond which leads to a greater occurence of fibre failure rather than pull-out and
thus leads to higher strengths but lower fracture toughness. A general picture is emerging
as more studies are conducted that the autoclaved WFRC products are more durable than
the air-cured hybrid composites which contain mixtures of cellulose and synthetic organic
fibres.
1.2 Natural Plant Fibre Resources

The use of natural plant fibre as reinforcement in building products has been known since
man made mud bricks using stiaw or sailed the seas in boats made from reeds and sealed
with resins or bitumen. The main use of such fibres, in developing countries, has been to
provide cheap, relatively low performance materials. The tendency in developed countries
16
was to neglect the research of natural plant fibres for use in cement composites, that is
until the "explosion of interest", as evidenced by the scientific and patent literature, which
occurred in the mid 1980's and is expanding to the present time.
Natural plant fibres exist in large quantities all over the world including wood fibre and
nonwood plant fibre. Great amounts of non-wood natural plant fibre are available and
produced in most developing countries. For example, in India alone, some 6.0 million
hectares of land is occupied with banana plantations and it was stated that 3 milhon tons of
banana fibre are available (Coutts, 1990b). Bamboo is another readily available fibre
source, there are altogether 62 genera and over 1000 species of bamboo in the world, of
which 37 genera and about 700 species grow in Asia. China has the greatest number of
bamboo species and the area of bamboo plantation in China is 3.2 million hectares, which
is one fifth of the total bamboo grove coverage in the world (Zhao, 1990).
An important factor in the availabUity of any plant for fibre is its collectable yield per unit
land area. Such estimated yields are given for a number of non-wood plants in Table 1.4,
first as collectable raw material, then as the estimated equivalent in bleached pulp.
Based on the total world-wide production of agricultural crops and the land area planted in
each crop, it is possible to make reasonably accurate estimates of the total amount of each
agricultural residue, useful as fibre, which might be collected in each country. Sunilar
estimates can be made for crops grown specifically for their fibre content. However, for
natural growing species, e.g. reeds and bamboo, such estimates are far more difficult, and
accurate data are not available.
17
Table 1.4 Annual collectable yields of various non-wood plant fibrous raw materials
(estimated) (Atchison, 1983).
Fibrous raw material
Sugar cane bagasse
Wheat straw
Rice straw
Barley straw
Oat straw
Rye straw
Bamboo, natural growth
Bamboo, cultivated
Reeds in the USSR
Kenaf-total stem weight
Kenaf bast
Crotalaria bast
Papyrus in Upper Sudan
Abaca (Manila hemp)
Seed flax straw
Cotton staple fibre
Corn stalks
Sorghum stalks
Cotton stalks
fibre
fibre
Collectable as raw material

BD metric tons per hectare year
5.0 - 12.4
2.2-3.0
1.4-2.0
1.4-1.5
2.2-3.0
1.4-2.0
1.5 - 2.0
2.5 - 5.0
5.0 - 9.9
7.4- 24.7
1.5 - 6.2
1.5 - 5.0
20.0 - 24.7
0.7 - 1.5
1.0-1.5
0.3 - 0.9
5.5 - 7.0
5.5 - 7.0
1.5 - 2.0
Equivalent in bleached pulp

BD metric tons per hectare year
1.7 - 4.2
0.7-1.0
0.4-0.6
0.4 - 0.5
0.4-0.5
0.8-1.0
0.6 - 0.8
1.0 - 2.1
2.0 - 4.0
3.0 - 9.9
0.7 - 3.2
0.7 - 2.5
5.9 - 7.4
0.4 - 0.7
0.18-0.27
0.25 - 0.86
1.55 - 1.95
1.55 - 1.95
0.60 - 0.80
Table 1.5 presents estimates giving a reasonably good indication of the tremendous
quantities of these non-wood plant fibres which can become available if economic
necessity requires their use as papermaking and reinforcing raw materials.
Table 1.5 Availability of various non-wood plant fibrous raw materials, 1982 (estimated)
(Atchison, 1983)
Raw material
Straw (wheat, rice, oat, barley, rye, seedflax,grass seed)
Sugar cane bagasse
Bastfibres(jute, kenaf, roselle, true hemp)
Core material from jute, kenaf, hemp
Leaffibres(sisal, abaca, henequen)
Reeds
Bamboo
Papyrus
Corn stalks and sorghum stalks
Cotton stalks
Potential world-wide availability BD tons

1,145,000,000
75,000,0(X)
2,900,000
8,000,000
480,000
30,000,000
30,000,000
5,000,000
900,000,000
70,000,000
An estimate of world-wide pulp production from nonwood plants, by type of fibre, is

given in Table 1.6. These figures include production for dissolving pulp. These fibres are
mainly used in paper and paperboard products. However, these fibres also provide ready
reinforcement source for composite materials, such as fibre reinforced cement products.
Table 1.6 Total production of various non-wood plant fibre pulps in 1982 (estimated)
(Atchison, 1983)
Type of no-wood plant pulp
World production (air-dry tons)
Cereal straw-mainly wheat and rye
1,390,000
Rice straw
750,000
Bamboo
960,000
Bagasse
1,600,000
Reeds
1,400,000
Cotton Unters (paper grade and dissolving pulp)
360,(X)0*
Esparto and sabai grass
120,000
Rags, abaca,flax,seed straw, hemp, sisal and other plant
fibres
1,420,000
* Includes about 60,000 metric tons for paper grade pulp, remainder for dissolving pulp & nonwovens.
1.3 Scope of the Present Work

1.3.1 Bamboo fibre and bamboo-wood hybrid fibre reinforced cement composites
The fibre cement industry has moved towards autoclaved natural plant fibre (wood)
reinforced cement mortars as the most commercially viable product to replace asbestos
cement products. In those countries without adequate forest resources but which have a
great amount of other natural plant fibre resources, then manufacturing fibre cement
products with these non-wood plant fibre would be of great advantage.
Bamboo is a rapid grown natural plant, which has good fibre qualities and is widely used
in the paper industry throughout the Asia region. Although bamboo has been used in
various forms in the construction mdustry, there is limited information in the scientific
literature conceming the use of bamboo pulp fibre. One objective of the current work is to
evaluate bamboo pulp fibre reinforced cement composites properties and study bamboo
fibre combined with wood fibre in order to improve the composites performance. The
work is covered under the following three topics:
19
1. Air-cured bamboo pulp fibre reinforced cement composites;

2. Autoclaved bamboo pulp fibre reinforced cement composites;
3. Hybrid bamboo and wood pulp fibre reinforced cement composites.
1.3.2 Influence of fibre properties on composite performance

Much has been written about the enhancement of engineering properties of cement based
products by the inclusion of fibres as reinforcement. Excellent texts and reviews on the
basic concepts of fibre reinforcement of brittle matrices are available, which deal with
synthetic fibres such as steel, polypropylene, glass and more recently kevlar and carbon.
Such fibres are homogeneous with respect to chemical composition and anatomical
structure, they have uniform cross-section area and can be obtained at a constant length,
and their tensile strength and modulus are constant or could be easily modified.
Unfortunately, natural pulp fibres are not homogeneous in chemical composition nor
uniform in shape or size even from one given species. At the same time the tensile
strength and modulus varies from fibre to fibre within a given pulp source.
The development of asbestos free fibre cement industry has made it most desirable to
obtain definite information on the relationship between the nature of the natural plant
fibres and their composites properties. It has been generally recognized that the fibre
length and strength are two of the most important factors but, because various features of
fibre morphology and chemical composition can influence composites properties, it has
been difficult to obtain a clear picture of the effect of any one property.
Further work is aimed at developing a fibre model which identifies those properties of
wood pulp fibre that are most significant for the production of fibre cement products. This
20
model may assist in identifying altemative cellulose fibre resources suitable for
reinforcement and to better understand natural fibre reinforced cement based composites.
The experiment work undertaking will be directed at better understanding of the
following:
1. Influence of fibre length on composite properties;
2. Influence of fibre strength on composite properties;
3. Influence of fibre lignin content on composite properties.
21
Chapter Two:
Theoretical Principles of Fibre Reinforcement
If the maximum benefit of composite materials as engineering materials is to be achieved,

it is necessary to understand their potential for bearing loads. Failure in a fibre composite
emanates from defects in the material. These may be broken fibres, flaws in the matrix
and debonded fibre / matrix interfaces. Figure 2.1 shows a schematic representation of a
cross-section through a fibre reinforced matrix. The diagram shows several possible local
failure events occurring before fracture of the composite. At some distance ahead of the
crack, which has started to travel through the section, the fibres are intact. In the highstress region near the crack tip, fibres may debond from the matrix (eg. fibre 1). This
rupture of chemical bonds at the interface uses up energy from the stressed system.
Sufficient stress may be transferred to a fibre (eg. fibre 2) to enable the fibre to be
ultimately fractured (as in fibre 4). When total debonding occurs, the strain energy in the
debonded length of the fibre is lost to the material and is dissipated as heat. A totally
debonded fibre can then be pulled out from the matrix and considerable energy lost from
the system in the form of frictional energy (eg. fibre 3). It is also possible for a fibre to be
left intact as the crack propagates. The process is called crack bridging.
From this simplistic approach we are immediately made aware of the importance of fibre
to matrix bond strength, frictional stress opposing pull out, tensile strength of the fibre,
fibre length and fibre content.
22
CRACK
Fig. 2.1. Schematic representation of crack travelling through a fibre reinforced matrix
2.1 Strength and Toughness

2.1.1 Mixture rule for strength
The law of mixtures for ahgned continuous fibrous composites may be modified to predict
the strength of natural fibre reinforced cements (Mai, 1978). It will be assumed that at
failure the fibres are not broken but are pulled out of the cement matrix. Thus the stress in
the fibre ((7,) is given by:
(7f =2r(l/d)
(2.1)
where Tis the fibre-matrix interfacial bond strength, I and d are the length and diameter of
the fibre. To account for the random distribution of the discontinuous short fibres,
Romualdi and Mandel (1964) suggest that the effective fibre volume fraction is 41% of the
nominal volume fraction. It is therefore possible to rewrite the ultimate tensile strength
(cTf) equation for the natural fibre reinforced cement as:
(7;= (T,,, v.,; -I- 0.410}Vf
(2.2)
23
Equation (2.2) is reduced to equation (2.3) by substituting equation (2.1) for ov:
<7; = (7r v.. + 0.82T vj(l/d)
(2.3)
In equation (2.3), <7, is the tensile strength of the un-reinforced cement mortar matrix and
Vr,i, v/ are the volume fractions of the matrix and the fibre respectively. Equation (2.3) may
be extended to predict the modulus of rupture (o),) of fibre cement since in general we
have Of, = aa, and (Jr,n> = P<ym, where a, P are constants which can be determined from
experiments and Gmh is the modulus of rupture of the cement mortar matrix in bending.
Thus,
(Jh = [a/pj (T,bV: + 0.82(ax}vf(l/d}
(2.4)
CUT may be regarded as the fibre-matrix interfacial bond stress in flexure.
Equations (2.3) and (2.4) are first given by Swamy and Mangat (1974) for steel fibre
reinforced concrete. Although they have not experimentally proven the validity of the
ultimate tensile strength as predicated by equation (2.3) they have however shown that
equation (2.4) is valid for the prediction of ultimate flexural strength of concrete
reinforced with randomly distiibuted short discontinuous steel fibres.
The values of a, /?, l/d. Cm, cj,>, and x are suggested to be 2.96, 2.81, 135, 9.71 MPa, 27.27
MPa and 0.35 ~ 0.45 MPa, respectively. Andonian and Mai (1979) attempted the first
theoretical analysis of the strength properties of WFRC composite using the mixture rule
for random fibre-cement using equations 2.3 and 2.4. The calculations gave close
predictions for the experimental results they obtained. While the theory predicts a
continuous increase in bending strength with increasing fibre mass fraction, the
experimental results show no strength improvement beyond 8% fibre mass fraction. This
24
is probably a consequence of the relatively large void fractions at larger fibre mass
fraction, which cause further reductions in the interfacial bond strength and matrix
strength due to poor compaction.
2.1.2 The ACK theory

In the case of fibres reinforced cement and gypsum plaster, however, the matrix is brittle
and fails at a strain very much lower than the failure strain of the fibres, and it is widely
accepted that the tensile strength depends on the fibre contribution alone and equation
(2.3) is given simply by:
a;^0.82Tv/l/d)
(2.5)
While there is considerable experimental support for mixture rule predicting the tensile
strength and flexural strength of fibre reinforced brittle matrices although it is difficult to
see a mechanism that would allow a matrix contribution to the tensUe strength once the
matrix has failed.
Aveston et al. (1971) have defined the salient points on the tensile stress / strain curve for
composites such as glass reinforced cement where the fibre is more extensive than the
cement, and there are sufficient fibres to support the extra load when the matrix cracks
(The "ACK theory"). Tensile stress / sU-ain curves predicted by the ACK model are shown
in Fig. 2.2 (full lines). For an " ideal" composite there is an initial elastic response after
which the matrix cracks and continues to crack at constant stress and increasing composite
strain. The multiple cracking process continues until the distance between cracks is too
small to allow transfer of sufficient load from fibre to matrix to crack it further.
Thereafter, further increase in load is taken by the fibres alone, and they extend and slip
25
relative to the matrix until they break or pull-out. Provided the volume traction is
sufficient to allow this multiple cracking process to occur, the strength of the composite
depends on the fibres alone and Equation 2.5 apphes.
Fig. 2.2. Tensile stress strain curves forfibrereinforced brittle matrices predicted by the ACK theory (full
line), andfilebending response calculated fromtiiem(broken lines) (Laws, 1983)
In bending, while the beam behaves in a linear elastic manner the neutral axis is in the
centre of the beam and the nominal stress given by simple beam theory is equal to the
26
actual stress in the beam. When the tensile stress in the surface of the beam exceeds the
elastic limit it can no longer increase hnearly with increasing strain but will follow the
tensile stress / strain response of the material; and the nominal stress calculated from
simple beam theory is no longer equal to the actual stress in the beam (Fig. 2.3).
Compression
Netural oxis
TefTsion
Stroin distribution
Stress distribution
Fig. 2.3. Theoretical model applicable to low modulusfibre-reinforcedcement composite atflexuralfailure

(Swift, 1979)
The bending curves calculated from the tensile curves in Fig. 2.2, are also shown in Fig.
2.2 (broken tines). At the critical volume fibre fraction (v^^,,;,) the fibres support the stress
after the matrix has cracked in tension, and there is a long multiple cracking region (curve
A). Over this region the bending moment continues to rise (curve A') and the ratio of
nominal bending strength or " modulus of rupture" (MOR) to ultimate tensile strength
(UTS) is high. As the volume fibre fraction increases, the composite becomes more stiff
and the MOR / UTS ratio decreases. At high volume fractions the tensile curve
27
approaches that of the fibres alone and the MOR / UTS ratio approaches unity. Thus a
relationship between bending strength and fibre volume fraction having the appearance of
a mixture rule can arise although the tensile strength / volume fibre fraction relationship
depends on the fibre contribution alone (Laws, 1983).
Swift and Smith (1979) suggest that direct tensile strength cannot be significantly
improved by low modulus fibres within the hmits of strain acceptable in a tensile member,
whereas for flexural strength it is theoretically possible to obtain a large increase resulting
from the inclusion of low modulus fibres in the composite. They have theoretically
explained and demonstrated empirically the improvement of flexural strength for sisal
slivers reinforced cement.
2.1.3 Basics of fracture mechanics

When the tensile strength of a brittle material is reached in a structure, cracking will occur.
The study of the conditions around and in front of a crack tip is called fracture mechanics.
Fracture mechanics was first studied for brittie materials such as glass. The first
applications to concrete appear to have been made by Neville (1959).
The application of fracture mechanics to concrete structures has provided new ways of
understanding and modeling phenomena which could be treated empirically before.
Fracture mechanics refers to the analysis of the fracture of materials by the rapid growth of
pre-existing flaws or cracks. Such rapid (or even catastrophic) crack growth may occur
when a system requires sufficient stored energy that, during crack extension, the system
releases more energy than it absorbs. Fracture of this type (often referred to as fast
28
fracture) can be predicted in terms of energy criterion (Romualdi, 1963; Bazant, 1985;
Bentur, 1990).
If we consider an elastic system containing a crack and subjected to external loads, the
total energy in the system, U, is
U = (-Wi + UE) + Us
(2.6)
Where, -Wi, UE and Us are the work due to the applied loads, strain energy stored in the
system and surface energy absorbed for the creation of new crack surfaces, respectively.
A crack will propagate when dU/dc < 0, where dc is the increase in the crack length.
Using this theory, one can derive the Griffith equation, which gives the theoretical fracture
strength for brittle, linearly elastic materials,
(jj,r = {2Eys/ncy'^
(2.7)
Where, Ciû c and js are the stress at first crack strain, one half of crack length and the
surface energy of the material. This is the basic equation of linear elastic fracture
mechanics (LEFM).
If we define a parameter Gc - 2'Ys = critical strain energy release rate, then the equation
(2.7) becomes,
(jancf' = (EGcf'
(2.8)
That is, fracture will occur when, in a stressed material, the crack reaches a critical size (or
when in a material containing a crack of some given size, the stress reaches a critical
value).
29
Altematively, we may define a parameter Kc = (JIJTTC)"^ = critical stress intensity factor.

K.^ = EGc
(2.9)
Kc has the units of N/m-""^, and is often referred to as the fracture toughness (not to be
confused with the term "toughness", which is used to refer to the area under the loaddeflection or stress-strain curve).
The LEFM parameters, Gc and K.c, are one-parameter descriptions of the stress and
displacement fields in the vicinity of a crack tip. In much of the early work on the
apphcations of fracture mechanics to cement and concrete, it was assumed that they
provided an adequate failure criterion. However, later research showed that even for those
relatively brittle materials. LEFM could only be apphed to extremely large sections (eg.,
mass concrete structures, such as large gams). For more ordinary cross-sectional
dimensions, non-linear fracture mechanics parameters provide a much better description of
the fracture process.
Fibres enhance the strength and, more particularly, the toughness of brittle matrices by
providing a crack arrest mechanism (see also section 2.1.6). Therefore, fracture mechanics
concepts have also been applied to model fibre reinforced cement composites. Mindess
(19 ) has reviewed the difficulties in modelling cement composites based on the fracture
mechanics approach. LEFM might be adequate to predict the effects of the fibres on first
cracking. However, to account for the post-cracking behaviour (which is responsible for
the enhanced toughness of fibre-cement composites), it is essential to resort to elasticplastic or non-linear fracture mechanics. A measure of toughness (ie., the energy absorbed
during fracture) can be obtained from the area under the stress-strain curve m tension. The
fracture mechanics concepts which could provide a more precise measure of toughness of
30
fibre reinforced cement composites include the crack mouth opening displacement
(CMOD), R-curve analysis, the fictitious crack model (PCM), and various other
treatments, all of which model (either implicitly or explicitly) a zone of discontinuous
cracking, or process zone, ahead of the advancing crack. These approaches provide
fracture parameters which are, at least, dependent on the fibre content, whereas the LEFM
parameters (G.:: or Kr) are most often insensitive to fibre content. It might be added here
that, while the J integral has often been used to describe these systems, theoretically it
cannot be applied to composite systems such as fibre reinforced concrete, where there is
substantial stress relaxation in microcracked region in the vicinity of the crack tip.
In the investigation of the fibre-crack interactions using fracture mechanics concepts, the
crack suppression, stabilisation and fibre-matrix de-bonding, three distinct issues must be
considered.
2.1.4 Fibre critical fracture length

The fibre critical fracture length is defined as twice the length of fibre embedment which
will cause fibre failure during pull-out. The fibre critical fracture length, l^. can be
calculated from equation (2.9), assuming fibre strength (Cpj), fibre diameter (^0 and the
shear stress (T.) developed at the interface are all uniform.
l,= <7fyd/2r,
(2.9)
Andonian et al. (1979) calculated the critical fracture length of P.radiata fibres in WFRC
composites to be between 18 and 23 mm, and as the measured length of the fibre is about
3.5 mm, fibre fracture was not possible. Davies (1981) and Coutts (1982b) had observed
31
fibre fracture during WFRC composite failure and concluded that the "apparent fibre
critical fracture length" must be less than 3.5 mm.
The conflicting reports on the predominance of fracture or pull-out of wood fibres from a
cement matrix prompted by Morrissey and Coutts (1985) to study a model system
consistmg of sisal slivers embedded in cement and protruding from one end of the cement
matrix. About 200 such samples were tested under tension. It was found that the slivers
did not behave in the manner predicted for uniform cylindrical fibres. After a break of the
elastic bond between the fibre and the matrix, the fibre started to puU out. The force
resisting pull-out was not proportional to the length of embedment but was dominated by
the highest local resistances present due to the fibre morphology. As pull-out proceeded,
anchor spots developed and the force required rose or fell in an apparently random
manner. When the anchorage was too strong to be dislodged by the maximum force the
fibre could carry, tensile fracture of the fibre occurred (Fig. 2.4).
There was a "critical fracture length" of embedment, which for the sisal slivers was
approximately 30 mm. When the embedment was shorter, fibre tended to be pulled out,
(Fig. 2.4) and when it was longer they tended to break. This "critical fracture length " is
not the length for which a uniformly distributed frictional stress reaches its critical values
under the maximum sustainable tensile load, as is commonly assumed (Equation 2.9), but
it corresponds to the length around which the probability of a local stiong anchorage
becomes high.
32
5
pullout
4>
10
C
<s
fracture
,
10
xtension
15
20
(mm)
Fig. 2.4. Tensile load-extension curves for different failure modes of sisal silvers embedded in cement
(Morrissey, 1985).
The "critical fracture length" of 30 mm for sisal slivers corresponds to an aspect ratio of
110 50, which is comparable with the aspect ratio of P. radiata fibres. This would
support the observation that P. radiata wood pulp fibre cement composites frequently
experience fibre fracture during failure.
2.1.5 Fibre aspect ratio

Equation 2.4 and numerous studies of steel fibre and glass fibre reinforced cements and
concretes have indicated that the major factors affecting flexural strength are the fibre
volume and aspect ratio of the fibre (l/d).
Higher values of either lead to higher values of
flexural strength (Hannant, 1978).
33
Without considering the theoretical equations leading to the above conclusions, it is

interesting to note that in cellulose fibre reinforced cement composites an increase of fibre
volume (not surprisingly) leads to increased flexural strength up to about 8% by mass of
fibre, at which stage efficient packing of the fibres becomes difficult and strength starts to
drop (Table 2.1).
Table 2.1 Effect of aspect ratio and fibre content (Coutts, 1983c, 84a, 85, 87a)
Fibre % by mass
2*
4*
6*
8*
10*
12*
2**
4**
6**
g**
10**
12**
Flexural strength (MPa)

'NZflax
P. radiata
RHt
Wet
Wet
RH
17.4
12.8
11.5
15.6
16.9
16.6
11.7
20.5
12.4
16.6
21.4
19.9
14.8
18.0
23.2
23.1
17.3
23.4
12.9
21.3
10.5
/
21.7
17.7
P. radiata
10.1
14.9
20.2
11.9
25.1
13.3
14.8
30.3
29.2
13.1
10.4
27.6
* Autoclaved mortar
T50 5 % RH, 22 2C
10.6
14.2
20.9
20.3
20.1
20.6
E. regnans
8.6
10.5
10.4
8.4
9.6
9.3
Fracture toughness (kJ/m^)

NZflax
P. radiata
RH
Wet
Wet
RH
0.24
0.37
0.31
0.19
0.71
0.88
0.57
0.34
0.94
1.15
1.81
0.48
1.74
3.15
1.86
0.84
2.15
1.92
2.97
1.19
/
2.88
2.59
2.09
0.41
0.64
1.40
1.93
2.28
2.25
P. radiata
0.64
1.52
3.72
4.51
4.60
3.60
E. i'egnans
0.33
0.25
1.00
0.51
1.61
1.06
1.49
1.37
1.83
1.46
1.79
1.68
** Air-cured cement
With respect to the aspect ratio of the fibres it has been noted by Coutts and Warden
(1985) that air-cured WFRC samples reinforced with softwood fibres (l/d = 80 - 100)
when compared to similar samples reinforced with hardwood fibres (l/d= 50 -60)
(Coutts, 1987a) displayed higher flexural strength at the same fibre content by mass (Table
2.1). Conflicting with this observation is tiie fact that autoclaved samples of mortars
reinforced with softwood fibres (Coutts, 1984a) and New Zealand flax {l/d = 200)
(Coutts, 1983c) show very similar flexural strength, but, more importantly, lower fracture
toughness (see 2.1.3). The reason that NZ flax composites do not produce better
mechanical properties is related to the fact that the fibres are weaker and the most of the
34
fibres are broken (Page, 1985), and so cr,;, is a Hmiting factor and not / /d. When tested
wet, the longer NZ flax fibres produce stronger samples than the short P. radiata fibres.
The hydrogen bonds between fibres or between fibre and matrix are destroyed (by
insertion of water molecules between the bridging hydroxyl groups) (see 2.2.3), so more
flax fibres (long fibre) can be loaded up to failure. In keeping with this, we find the very
short E. regnans fibre reinforced materials are weak when tested wet or at RH test
conditions (50 5% relative humidity, 22 2C) as short fibre is pulled out and cannot be
loaded to failure.
2,1.6 Fracture toughness

As discussed in section 2.1.3, that the post-cracking ductility imparted to the composite by
fibre addition can be considerable. Gordon (1976) states: "The worst sin in an engineering
material is not lack of strength or lack of stiffness, desirable as these properties are, but
lack of toughness, that is to say, lack of resistance to the propagation of cracks". The
origin of fracture toughness in WFRC composites was claimed to come mainly from fibre
pull-out (85% - 90%) (Andonian, 1979), although later studies by these same workers
(Mai, 1983) considered fibre fracture must be of importance in agreement with other
researchers in the field (Davies, 1981; Coutts, 1982b; Fordos, 1986).
A generalised theory has been proposed by Martson et al. (1974) where the specific
fracture resistance {R) is given by the sum of; the toughness due to fibre pull-out (Rp.o),
redistribution of stresses (RrS) and fracture of surfaces (i?,). If the fibres are pulled out
rather than fractured, then it seems appropriate to neglect stress redistribution Ry. as a
component contributing to the specific work of fracture (R). Thus
R = R,,+R,
(2.10)
35
Where, Rpo. = 0.41vffx/12d
(2.11)
Rs = VfR. + Vr, Rr + (0.41 VjflRif /d
(2.12)
Rr, Rf and Rif are the fracture energies of the cement mortar mattix, fibre and fibre-matrix
interface respectively. Normally, Rf R,,, and Rif = R, (Martson, 1974) so that equation
(2.12) is simplified to
Rs = 0.41vflRn,/d + v,R,,,
(2.13)
The specific work of fracture R is thus given by

R = 0.41vffr/ 12d + [v,, + 0.41vfl / d]Rr,
(2.14)
Toughness measurements can be conducted in several ways (Hibbert, 1982); impact testers
such as Charpy or Izod which involve stored energy in a pendulum, calculating the area
under stress - strain curves, and the use of fracture mechanics involving stress intensity or
similar parameters, or more practical tests such as dropped balls or weights, etc. Some of
these techniques are more suited to particular composites, but all have limitations which
render elusive the well-defined material properties useful to engineers and material
scientists.
Mindess and Bentur (1982) studied the fracture of WFRC products and found that
saturated samples were weaker and more comphant than air-dry specimens. It was found
that the wet samples were completely notch-insensitive, while tire air-dry specimens may
be slightiy notch-sensitive. Notch sensitivity is a requirement for the application of hnear
elastic fracture mechanics to cement composites, and so these workers concluded that
LEFM could not be used for WFRC materials.
36
Mai and Hakeem (1984a, b) have studied the slow crack growth of WFRC composites for
both dry and wet conditions, using a double-cantilever beam system. The results were
analysed using K-soIutions, and compliance measurements within the framework of
LEFM. It was concluded that LEFM concepts can be used for WFRC products.
The use of LEFM requires a linear elastic homogeneous matrix. The introduction of fibres
into matrix, to achieve ductility, must by their very nature result in a heterogeneous
material with a non-linear stress-strain curve after matrix cracking.
The fracture toughness results given in Table 2.1 for a range of NFRC composites have
all been obtained by the method of measuring the area bounded by the load / deflection
curve. As the materials are non-linear and non-elastic, this approach is useful in that the
total recorded energy will include the contributions of the work of fracture of the matrix,
the debonding and frictional slipping of the reinforcement, and any strain energy released
by fibre fracture.
For fibre cements and concretes the matrix work of fracture is virtually that of the
unreinforced cement or concrete and is less than 50 J/m^. Therefore it is assumed that any
improvement m composite toughness will depend on whether the fibres bridging the crack
are able to support the load previously cartied by the mattix, and on whether the fibres
break or pull out of the matrix.
Fibre pull-out is the most common mode of failure for steel and glass fibre cements and
concretes. This is because, to incorporate sufficient fibre into the formulations without
37
fibre tangling, the fibre aspect ratio cannot be high enough to exceed the critical fracture
fibre length.
It has been recorded that wood fibres can be loaded into the matrix material with fibre
volumes in excess of v^.>j and can be fractured during failure, hence the "apparent critical
fracture length" is exceeded. This behaviour does result in relatively high levels of
fracture toughness (Table 2.1) but does not lend itself to a conventional form of theoretical
analysis. The explanation of such properties is discussed in terms of fracture mechanics
and bonding in section 2.2.
Recentiy, Hughes and Hannant (1985) have studied the reinforcement of Griffith flaws in
WFRC in an attempt to explain their behaviour, and concluded that stresses developed are
significantly higher than would be expected using the mixture rule theory.
It is hoped that as commercial production of NFRC becomes a global activity, more

interest will be shown in the theoretical explanation of how this system works and hence
in developing ways to optimise its use in design.
2.2 Bonding and microstructure of NFRC composites

As well as the physical properties of the fibre and the matrix, a major factor which
controls the performance of a composite is the type and arrangement of bonds linking the
two materials.
38
According to composite theory, the interface (that is the region of intimate contact
between fibre and matrix) plays the dual role of tiansmitting the stress between the two
phases and of increasing the fracture energy of the composite by deflecting cracks and
delocalising stress at the crack tip.
The interfacial bond itself can be physical or chemical in nature, or a combination of both.
Too strong a bond between fibre and matrix results in a brittle material which has strength
whereas a weak bond results in a tough material lacking strength. The mechanical
performance of a NFRC composite is therefore directly related to the nature and properties
of the fibre-matrix interface.
2.2.1 Chemical bonding

The chemistry and morphology of the matrix material has been well documented (Lea,
1976) and will not be considered further, apart from stating that cement is strongly
alkaline (pH > 12) and presents metal hydroxy groups at its surface, such as -Ca-OH, -SiOH, -Al-OH and -Fe-OH (due to hydration and hydrolysis of silicates, aluminates and to a
lesser extent ferrites of calcium that are present in the cement matrix). Cellulose fibres
such as wood fibres contain covalent hydroxyl groups, -C-OH, either phenolic (from
residual lignin) or alcohohc (from the cellulose component) and carboxyhc groups, 0=QOH, due to oxidation of end groups. Hydrogen bonding and / or hydroxide bridges may
play a major role in the bonding of NFRC composites. The chemical implications wiU not
be considered quantitatively; it suffices to say that hydrogen bonds may form between
fibres or between fibres and matrix.
39
Coutts and co-workers (Coutts, 1979a, b) considered the possibility of using chemical
pretreatments of the wood fibres to enhance the bond between fibre and cement. The most
acceptable theory for the development of coupling agents is the "chemical bonding
theory", which suggests that the coupling agent acts as a link between fibre and matrix by
the formation of a chain of covalent chemical bonds (Fig. 2.5).
Although small improvements in composite mechanical performance have been recorded

from the use of pretreated fibres, the costs of such operations are currently considered
prohibitive. The initial hypothesis that some form of coupling agent or additive was
needed to achieve bonding stemmed from the general belief that the bond between wood
fibre and cement would be weak. It has subsequently been suggested that this may not be
the case (Coutts, 1984b) and that the presence of hydroxyl groups on the surface of both
wood and cement may be sufficient for adequate chemical interaction to take place via
hydrogen bonding.
2.2.2 Mechanical bonding

As well as the chemical bonding aspects of a fibre, the physical bonding potential must
also be considered. Much of the theoretical data on fibre reinforcement is based on
smooth cylindrical fibres of uniform sharp and dimensions (Section 2.1).
40
0-MX
(n-l)
{n-2)
Fig, 2.5. Possible coupling mechanism between woodfibreand cement matrix (Coutts, 1979a).
Maximum fracture energy is often achieved if frictional energy is dissipated via fibre pullout. Wood fibres are relatively long compared to their diameter and hence have aspect
ratios of say 60 - 200 (depending on whether they are hardwood or softwood, early wood
or late wood fibres), but, more importantly, the fibres are hollow and can be collapsed to
ribbons and at the same time develop a helical twist along their length (like a cork-screw).
When fibres such as these are used to reinforce a brittle matrix, an asymmetrical process
will be taking place during pull-out (after interfacial debonding has occurred). In classical
41
pull-out of straight fibres (glass, steel, etc.) the forces are symmetrically distributed around
the fibres; in the case of the contorted wood fibres, the leading edge of the hehcal fibre can
experience considerable compressive stresses resulting in a ploughing action which can
damage fibre or matrix resulting in increased fracture surfaces and hence increased
fracture energy. Similar observations were recorded by Bentur et al. (1985a,b) when they
studied the physical processes taking place during the pull-out of steel fibres, of various
geometry, from Portland cement.
Coutts and co-workers (Coutts, 1982a, 1984a) reported that mechanical refining or beating
of wood fibres resulted in improved flexural strength from autoclaved mortars reinforced
with such fibres. This phenomenon can again be discussed in terms of mechanical
bonding in that the external surface of the fibre is "unwound" and the fibrils so formed
offer extra anchoring points by which the fibres can accept stresses from the matrix and so
become more effective reinforcing elements. Although refining the fibre can assist in
mechanical bonding, its main value is in improving the drainage rate and solids retention
in the commercial Hatschek process (Anon, 1981; Coutts, 1982a).
MicheU and Freischmidt (1990) studied curly fibre in the cement and silica matrix at the
aim of improving fibre and matrix bonding. The use of curly fibres in reinforced cement
and silica sheets gave sheets with improved wet interlaminate bond strengths, relative to
sheets prepared from conventionally treated fibres but had littie effect on the values of
modulus of rupture and fracture toughness.
Mechanical bonding has been discussed in other systems, such as asbestos fibre cement, in
which Akers and Garrett (1983a) showed that asbestos can be fiberized like wood fibres,
42
steel fibres can be kinked, as we have noted in the work of Bentur et al. (1985a), or
polypropylene can be fibrillated into a nethke form, as reported by Hannant and coworkers (1978).
2.2.3 The effect of moisture on bonding

The effect of moisture on the stiength properties of NFRC composites is of importance.
Earlier in WFRC research, variation in test conditions produced variations in test results,
and it was evident that standard conditions must be adopted. Coutts and co-workers
reported that as the moisture content of a WFRC test specimen was increased, the flexural
strength of a given sample decreased while the fracture toughness increased (Coutts,
1982a). These observations have been confirmed by Mai et al. (1983). This behaviour is
common to both air-cured (Coutts, 1985) and autoclaved (Coutts, 1982a, 1984b) samples
containing a range of cellulosic fibres, be they softwood (Andonian, 1979; Coutts, 1983a),
hardwood (Coutts, 1987a), abaca (Coutts, 1987b) or NZ flax (Coutts, 1983c).
Coutts and Kightiy (Coutts, 1982b, 1984b) suggested that these observations could be
supported by the hypothesis that hydrogen bonds and / or hydroxide bridges play a major
role in the bonding and hence in the mechanical performance of WFRC composites.
Wet or dry, a wood fibre has about the same tensile strength, but its stiffness is
considerably lower when wet. Thus a dried WFRC composite has stiff, highly contorted
fibres (see section 2.2.4) locked into a rigid cement matrix which could be bonded at the
interface by a large number of hydrogen bonds or hydroxyl bridged sites. This system
when stressed can transfer the stress from the matrix to the fibres via the many interfacial
43
bonds, and hence sufficient stress may be passed on to the fibre, after the matrix has
cracked, to cause the reinforcing fibre to fracture under tensile load.
On the other hand, in a moist sample, the hydrogen bonds between fibres or between fibre
and matrix are destroyed (by insertion of water molecules between the bridging hydroxyl
groups); and, at the same time, the cellulosic fibres are swollen by water absorption and
have become less stiff Under stress this system allows the fibres to move relative to each
other or to the matrix. However, due to the pressure of swelling and the highly contorted
assemblage of fibres, considerable frictional forces are developed. If the forces are
effective over sufficient length of a fibre, they can result in the fibres being loaded to
failure; however, the number of fibres that pull out without failure is higher than when the
sample is dry, and hence the observed values of fracture toughness for wet samples is
higher than for dry samples.
2.2.4 Microstructure of NFRC composites

The examination of wood-cement composites by scanning electron microscope (SEM) is
relatively recent. Ahn and Moslemi (1980) examined the manner in which wood particles
and Portland cement bonded together in cement particle-boards and decided that
mechanical interlocking plays a significant role. This interlocking is due to crystal growth
during the hydration of cement. A similar crystal interlocking effect has been reported for
wood fibre reinforced plaster products (Coutts, 1987c).
NFRC composites have been examined by different research groups with strong interests
in the micromechanical behaviour of such composites. In a number of these SEM studies
of NFRC materials, the microstructural features of the fracture surfaces of the broken
44
composites were described, with emphasis being placed on the surface of the fibre and
how it had failed (Davies, 1981; Coutts, 1982b; Mai, 1983; Pavithran, 1987; Akers, 1989).
More recently, interest has turned to the matrix, and in particular, the interfacial region
adjacent to the fibre.
In a study of air-cured WFRC composites, containing Kraft or CTMP

(chemithermeomechanical pulp) fibres, Davies et al. (1981) proposed that the fracture
surfaces indicated that both fibre fracture and fibre pull-out were taking place during
loading to failure under ambient conditions. Andonian et al.{l919), studying autoclaved
WFRC samples, stated that fibre pull-out was the main source of fracture toughness (8090%) even though the samples had been dried in an oven at 116C for 24 h. Mindess and
Bentur (1982) also considered that a pull-out mechanism was the main process taking
place when commercially produced WFRC samples containing 20% by mass of wood
fibres were tested. These workers used only photographic evidence, stating that at 18 X
magnification (their highest), lack of focus restricted their observations due to the
unevenness of the surface. As the diameter of a wood fibre is approximately 15-40 frm,
the observation of end fractures would be difficult by optical means.
SEM studies by Coutts and Kightiy (1982b, 1984b), using autoclaved WFRC samples,
complement the earlier findings with air-cured samples, namely, that failure takes place by
mechanisms of fibre fracture and fibre pull-out. More importantiy, this study showed that
the relative importance of these mechanisms is very dependent upon the moisture content
of the test sample (Coutts, 1984a). Figure 2.6, 2.7 and 2.8 show the SEMs of fracture
surfaces of WFRC samples obtained from tiie same formulation when tested dry, wet and
at 50 5% RH. Most of the fibres protruding from the fracture surface of a sample
45
preconditioned in an oven at 105C for 24 h before testuig have fractured ends (Fig. 2.6).
By contrast. Figure 2.7 shows the fracture surface of a sample which has been
preconditioned in water for 48 h before testing and indicates considerable fibre pull-out,
although fibre fracture is still very obvious (Fig. 2.7). Figure 2.8 shows the fracture
surface of the sample conditioned at near ambient conditions (50 5% RH and 22 2C).
Both fibre fracture and fibre pull-out appear to have taken place. Higher magnification
(Fig. 2.8b) shows that considerable damage to some fibres of relatively short length has
taken place while other fibres have been pulled out of the matrix. Mai et a/.(1983), on reexamining their earlier work, are now of the opinion that fibre pull-out is not the major
source of toughness for WFRC, but indeed that fibre fracture is of considerable
importance.
The significance of moisture content, as evidenced in the above SEM work, has been
adequately described in the section 2.2.3.
Fig. 2.6. SEM showing fracture surface of WFRC preconditioned at 100 - 105C for 24h (Coutts, 1984b)
46
Fig. 2.7. SEM showing fracture surface of WFRC precondifioned by soaking in water for 48h (Coutts,
1984b)
Fig. 2.8a. SEM showing fracture surface of WFRC preconditioned at 50 5% relative humidity and 22 :
2C (Coutts, 1984b)
47
Fig. 2.8b. As (a) but higher magnification

Most of the foregoing studies have related to the condition of the fibre after failure and
rather little has been said of the most important aspect of the composite - the interfacial
region. In studies of air-cured cement matrices reinforced with aggregate particles
(Bames, 1978), steel (Pinchin, 1978; Bentur, 1985a,b) or glass fibres (Stucke, 1976;
Singh, 1981) detailed reference has been made to an interfacial region with structure and
porosity different to the bulk cement paste. This interfacial zone has, in several cases,
been considered to be 50 - 100 pm thick.
In the cases of aggregate, steel or glass fibre there is a general similarity of behaviour. At
the surface of the reinforcement a dense layer, consisting largely of hexagonal lamellar
crystals of portiandite (calcium hydroxide or CH), forms and rephcates the topography of
the reinforcement over much of its surface area. At discontinuity in this dense layer,
inclusions of hydration products such as calcium silicate hydrate (CSH) gel, ettringite and
large crystals of CH occur.
48
With increasing distance from the reinforcement surface, the proportion of CH decreases,
and a relatively porous layer (in comparison to the bulk of the matrix material) forms for
some distance before gradually becoming more dense and merging into the bulk matrix.
Fig. 2.9. SEM showing cement surface at interface contains dense matrix with some discontinuities (Coutts,
1987d).
"^^ v,,;*
^ ^ y
Fig 2.10. SEM shows fractured fibres with dense material from bulk of matrix up to fibre wall (Coutts,
1987d)
49
An SEM study of WFRC composites which focused on the region of the wood fibrematrix interface (Coutts, 1987d,e), revealed that the general characteristics associated with
the interfaces of other fibre reinforced cement composites were not observed. Coutts
noted that the interface between wood fibre and matrix is usually dense and replicates the
surface of the fibre, although discontinuities do occur, especially at higher fibre loadings
when normal packing of the fibres becomes more difficult (Fig. 2.9). Again, if one looks
at the intimate area of contact between fibre and matrix (Fig. 2.10), no obvious zone of
weak matrix material exists at the interface. It was stated that during fabrication of WFRC
composites the samples were placed under pressure to compact the structure and reduce
the water - cement ratio. It was hypothesised that, unlike the glass or steel fibres, which
are not compressible, the hollow wood fibres are compressed, and after the pressure is
removed the water level immediately adjacent to the fibre is lowered even further as the
sponge-like fibre draws excess water back into itself. This would reduce the occurrence of
voids at the interface of fibre and matrix particles, because less free water would be
present, and produce a homogenous dense matrix material from the bulk of the material
right up to the point of contact with the fibre.
Another example of a dense interface is found m the work of Akers and Garrett (1983b)
who studied the microstructure and failure mode of die fibre-matrix mterfacial region in
asbestos fibre reinforced air-cured cement. Also reported was a mutual interlocking of
asbestos fibres witii the cement hydrate, as had been observed with WFRC. But, more
importantiy, energy dispersive X-ray measurements showed no evidence of a calcium
hydroxide enriched zone adjacent to the interface, as had been reported for other fibre
cement materials. The study proposed that fiberizing asbestos fibres increased the fibre
surface available for water absorption, which helped to distribute the water-filled voids in
50
the interfacial region. Thus the nature of the fibre-cement interface need not be the same
as is described for materials such as glass, but appears to depend on the characteristics of
the reinforcing fibre.
In the case of TMP pulp reinforced composites (Coutts, 1986 and Chapter 9), the high
temperature of the autoclave (> 160C) coupled with the alkali conditions of the matrix
material (pH > 12) results in the extraction of polysaccharides and wood acids from the
reinforcing fibre (see section 3.3), which, in the case of Kraft pulp have been removed
during pulp preparation. These extractives contribute to "poisoning" of the cement (Singh,
1979; Thomas, 1983) and coating on the fibre surface thus causing poor interfacial bonds.
Hence, TMP pulp-cement composites have low strength and poor fracture toughness (see
section 1.1.4).
51
Chapter Three:
The Nature of Plant Fibres
3.1 Fibre Classification

Natural plant fibre can be classified into wood fibre and non-wood plant fibre, which
consists of:
(a) Seed hairs;
(b) Bast fibres - fibres derived from the bark of dicotyledons, which include herbaceous
plants, shmbs, and trees;
(c) Leaf fibres - fibres derived from the vascular bundles of very long leaves of some
monocotyledons. Leaf fibres are also known as " hard " fibres because they are more
lignified than bast fibres;
(d) Grass fibres - these are another group of monocotyledonous fibres where the entire
stem together with the leaves are pulped and used in papermaking. Such pulps are
composed not only of fibres, but of other cellular elements as well.
Woods are grouped into two main classes, namely softwood and hardwood. The names
hardwood and softwood are based mainly on timbers produced in the northern hemisphere.
More correctly, the softwoods or coniferous types (pines, firs and spruces) are called
gymnosperms, and the hardwoods (gums, oaks and ashes) are all classed as angiosperms.
The hardwood-softwood grouping has little meaning on a world scale, as some hardwoods
(such as balsa wood) are extremely soft.
52
3.2 Structure of Natural Plant Fibres

Natural plant fibres can be derived from wood, bast, seed leaf and grass, needless to say
there are far too many plants to describe m this thesis. There are four types of tissue in
natural plants. They are parenchyma, fibre, tracheid and vessel (or pore). However some
plants do not contain all of these four tissues, such that softwood does not contain vessels
(or pores). Each type of tissue serves one or more special functions: the parenchyma cell
conduct and store food and water; the fibre's main role is to provide mechanical support;
tracheids and vessels (or pores) conduct water and dissolved mineral salts from the roots
to the leaves as well as providing mechanical support.
In general terms it can be said that the fibre cells, which are themselves composite
structures, have a cylindrical or ribbon-like shape made up of different layers with a
hollow centre (lumen). The fibre can vary in length from less than 1 mm to greater than
250 mm. The diameter can be from less than 5 pm to greater than 80 pm (see section
3.5.1.1). When the fibres are separated from each other they can collapse flat, or the
lumen may remain open, if the walls of the fibres are thick. Fibres may develop a spiral
twist along their length, which can be of significance during fibre composite failure.
Plant fibres contain cellulose, a natural polymer, as the main material of reinforcement. In
a simplified description we can say chains of molecular cellulose are held together by
hydrogen bonds to form microfibrils, which in tum are held together by amorphous
hemicellulose and form fibrils. The fibrils are assembled in various layers of differing
thickness at different angles of orientation to build up the intemal structure of the fibre, the
main reinforcing element of interest to our research (see Table 3.5). Figure 3.1 and 3.2
show schematic structure of wood fibres and bamboo fibre, respectively.
53
Lumn
Cross section of a wood tibar
Sscondary waM
MiddU Um*l|a
lumen
middle lamlia
primary wall
secondary wall
s, layars , layar
5^ layer
Primary w i l l
---iUcinIn
rridlt lametla
^.,T
"jtlsmicaiiyissa
Fig. 3.1. The structure of wood fibre
85...90'
Fig. 3.2. The stmcture of bamboo fibre
Fibres are cemented together in the plant by lignin, a complex natural organic adhesive. In
much of the work on plant fibres, we are really discussing aggregates of fibres, which are
often incorrectiy called " fibres". As we will see, when we discuss durability, much of the
54
lack of performance of certain " fibres" can be attributed to breakdown of these aggregates,
due to the alkalinity of the matrix materials, and not to the fibre cell itself
Fig. 3.3. Scanning electron micrograph of a cube of eastern white pine microtomed on three surfaces. Note
the arrangement of longitudinalfi-acheids(tr) in radial files and the smicture of the rays. In this species resin
canals (re) are rather prominent features (C6t6, 1980).
55
The increase in size of a natural plant varies with the seasons and climatic conditions.
Such as in trees, during late autumn and winter, there is littie or no growth, but growth is
at a maximum in spring. The spring or early wood usually has wood cells with larger
diameter and thinner walls than those cells formed during periods of slow growth summer wood or late wood (Fig. 3.3).
As will be seen later in this thesis, fibres with thick cell walls have different papermaking
characteristics compared with thinner - walled fibres. Consequently, the ratio of later
wood fibres to early wood fibres can have a big influence on the abihty of a given wood
fibre source to act as a reinforcing element.
3.3 Chemical Composition and Durability

The chemical components which constitute most of the mass of natural plants are
cellulose, hemicellulose and lignin (Table 3.1). Cellulose is made up of thousands of
glucose molecular units joined in long chains, and represents 40 - 45 % of plant. It
contains 44.4 % carbon, 6.2 % hydrogen and 49.4 % oxygen, and is relatively unaffected
by alkalis and dilute acids.
Lignin occurs in the form of very complex chain polymer of phenolic building blocks
consisting of about 65 % carbon, 6 % hydrogen and 29 % oxygen, and is virtually
impossible to dissolve without first breaking it down into simpler substances. Lignin
comprises between 22 and 30 % of plant and varies in chemical composition and amount
in different species, being somewhat more abundant in softwoods than in hardwoods.
56
Table 3.1 Chemical compositions of natural plant fibres (Atchison, 1983)

Type of fibre
Stalk-straw-rice
-wheat
-barley
-oat
-rye
-cane-sugar
-bamboo
-grasses-esparto
-sabai
-reeds-phragmites communis
Bast -seed flax tow
-seed flax
-kenaf
-jute (1)
Leaf -abaca (Manila)
-sisal (agave)
Seed hull
-cotton timers
Woods
-coniferous
-deciduous
Cross & Bevan

cellulose
43-49
49-54
47-48
44-53
50-54
49-62
57-66
50-54
54.5
57.0
Alpha
cellulose
28-36
29-35
31-34
31-37
33-35
32-44
26-43
33-38
Lignin
44.75
Pentosans
Ash
Silica*
9-14
3-7
3-6
4-6.5
0.5-4
0.7-3.5
0.69
12-16
16-21
14-15
16-19
16-19
19-24
21-31
17-19
22.0
22.8
23-28
26-32
24-29
27-38
27-30
27-32
15-26
27-32
23.9
20.0
15-20
4.5-9
5-7
6-8
2-5
1.5-5
1.7-4.8
6-8
6.0
2.9
75.9-79.2
47
47-57
57-58
45.1-68.5
34
31-39
10.1-14.5
23
14.5-18.7
21-26
6.0-17.4
25
22-22.7
18-21
2.3-4.7
5
1.7-5.0
0.5-1.8
78
55-73
60.8
43-56
8.8
7.6-9.2
17.3
21.3-24
1.1
0.6-1.1
26-34
23-30
7-14
19-26
<1
<1
2.0
80-85
53-62
54-61
40-45
38-49
* The content of tramp (as distinguished from inherent) inorganics may be greatiy increased by mechanical
harvesting.
Hemicelluloses act as a matrix for the cellulose microfibrils, are of relatively low
molecular weight and are soluble in alkalis. They occur as about 15 to 30 % of the weight
of the natural plant.
All the chemical components of natural plants are formed from the sugars produced in the
leaves by photosynthesis. The water obtained from the roots is combined with carbon
dioxide from the atmosphere, by complex biochemical processes dependent on chlorophyll
which is able to absorb from sunHght the energy needed, to form simple sugars. When
these sugars are conveyed from the leaves to the plant cells, they are changed into the
much more complex cellulose, hemicellulose and lignin.
57
In addition to cellulose, hemicellulose and hgnin, natural plant also contains a wide range
(5 - 30 %) of compounds known as extractives which can be extracted by solvents. These
materials, contribute to colour, density, durability, flammability and moisture absorbency,
and include polyphenols, oils, fats, gums, waxes, resins and starches.
These extractives plus hemicellulose and lignin can inhibit or retardate the hydration and
strength development of Portiand cement. In the case of TMP pulp reinforced composites,
the high temperature of autoclave (> 160 C) coupled with the alkali conditions of the
matrix material (pH > 12) results in fibre chemical degradation and extraction of the
fibres. These extractives contribute to "poisoning" of the cement and coating on the fibre
surface. Hence, TMP pulp-cement composites have poor mechanical performance (see
section 1.1.4 and 2.2.4).
Natural fibres are susceptible, under certain conditions, to biological degradation, the
action of such decay resulting in, amongst other things, strength loss. However, in the
fibre cement system, chemical degradation would be more significant than biological
degradation due to high alkah environment (pH > 12). Sisal fibre (aggregate form) has
acceptable initial stiength. While, immersed in lime solution, the fibre suffers a 74%
reduction in stiength due to significant chemical degradation (Nilsson, 1975). Thus
composites reinforced with aggregate fibres have poor durability due to fibre degradation
(Lola, 1986).
58
3.4 Preparation of Natural Plant Fibres

3.4.1 Pulping
Natural fibre can be prepared by means of pulping. As stated, natural plant consists of
parenchyma, tracheid, vessel (or pore) and fibre tissues. The fibre cell contains cellulose,
hemicellulose and lignin. The purpose of pulping is to separate the fibres from the plant
so that they are suitable for papermaking and composite reinforcement.
Wood pulp fibre is the plant fibre of most commercial importance at the present time.
There are a number of ways of producing pulp which will be summarized briefly in the
following sections.
3.4.1.1 Chemical pulping

Chemical pulps are made by heating the wood chips at high temperatures (about 170 C)
with a solution of chemicals that dissolve most of the lignin which bonds the fibres
together in the wood. The chemicals dissolve carbohydrates as well as lignin removing
about half of the wood substance, yield is typically 45 - 55 per cent. It is not possible to
remove all the lignin during chemical pulping, therefore all unbleached chemical pulps
contain some residual lignin and are brown in colour.
The pulp is washed with water to remove pulping chemicals and dissolved wood
components. Most modern pulp mills have a chemical recovery system in which these
washings are concentrated and then burnt in a specially designed furnace to recover the
chemicals for re-use in the pulping process and to generate heat energy.
59
The Kraft process is used to produce a major portion of the world's chemical pulp.
Sodium sulphide and sodium hydroxide are the active ingredients which attack the lignin.
The process operates under alkaline conditions, and penetration of the wood is more rapid.
It is applicable to all species and so has a wide mdustrial appeal, but it creates considerable
effluent disposal problems and releases an unpleasant odour into the surrounding district.
The soda-anthraquinone (soda-AQ) process uses sodium hydroxide with a small quantity
of anthraquinone as pulping chemicals. Pulps by this process are brown in colour and are
no easier to bleach than Kraft pulps. In most respects soda-AQ pulps are similar to Kraft
pulps. The process offers no advantage with regard to bleachability of the pulps, but the
absence of sulfur offers the possibility of using new chemical recovery systems eg, the
Direct Alkah Recovery System (DARS). The successful development of DARS may
make it possible to produce soda-AQ pulp in economically viable mills of a smaller scale
than is currently necessary If the Kraft process is used.
The alkaline conditions cause less harm to the cellulose than the acidic conditions of the
sulphite process, since ceUulose is more easily hydrolysed by acids than by alkalis.
The use of chemical pulps in WFRC products removes the problem of "extractives"
which, interfere with the cure of autoclaved products and fibre - matrix bonding resulting
in materials of low perfonnance (Coutts, 1982b).
3.4.1.2 Mechanical pulping

Papermaking pulps can be manufactured by mechanically separating wood into its
constituent fibres. Chips are first steamed to soften the Hgnin and to make the separation
60
of the fibres easier. The chips are fed into a refiner where they are ground between two
plates. Altematively billets may be fed onto a grinding stone.
After this treatment the freed fibres are screened and pumped to storage. The amount of
pulp retained from mechanical processes is usually higher than 95 per cent as only the
water soluble material in the wood is removed.
Mechanical pulping in its various forms has been claimed as the pulping method of the
future (Kurdin, 1983). This method can solve the problems of limited wood supply, of
restricted capital, of environmental restrictions and of efficient utilization of mixed forests.
In its expansion from groundwood pulping, mechanical pulping has advanced into the
domain of chemical pulping blurring what was once a clear line of distinction - there is
now refiner mechanical pulp (RMP), thermomechanical pulp (TMP), chemimechanical
pulp and chemithermomechanical pulp (CTMP).
Mechanical pulping is not without its weaknesses. Means must be found to reduce the
energy consumption, increase the strength of the pulp and improve light stability. The last
factor is important in paper manufacture, but not in the use of the pulp as a reinforcement
in cement systems etc.
The principle of chemical and mechanical pulping is scheduled briefly in Figure 3.4.
61
Fig. 3.4. The principle of chemical and mechanical pulping

3.4.1.3 Non-wood pulping
Some non-wood resources provide excellent papermaking fibres. Some bast fibres such as
jute, flax and kenaf can be produced by a retting method, which allows bundles of fibres to
be freed from cellular tissue surrounding them by the combined action of bacteria and
moisture, then the fibrous material is crushed, washed and dried. Decortication, a process
used for leaf plants such as sisal and abaca, involves crushing and scraping the leaf to
remove cellular tissue, then washing and drying.
Besides these mechanical process, almost all natural plant fibres can be prepared by the
Kraft pulping process. In countiies where wood is scarce, it is quite common to find
cereal straw, bagasse (from sugar cane), bamboo and similar materials being used as a
source of papermakmg fibre. These fibres have roughly the same quality as hardwood
fibres, sometimes a bit less because the fibres have been damaged during harvesting /
62
processing, but usually adequate as a component of printing and writing papers (Li, 1992).
However, there are a few difficulties that inhibit non-wood fibres from being used more
often as a source of papermaking fibres (McKenzie, 1992).
The first difficulty is that one can harvest the crop at only one time of the year. This
means that a whole year's supply of raw material must be collected and stored within a few
weeks. Furthermore, it cannot merely be cut and stored, but must be protected from decay
immediately after harvesting. Also, the prospect of a crop failure (for whatever reason)
and the subsequent closure of a multimilhon dollar manufacturing operation for lack of
raw material is enough to deter the average investor.
Many agricultural residues have a high silica content (rice straw and bamboo are notorious
in this respect) which dissolves in alkaline pulping liquors and redeposits throughout the
pulping plant. The bulkiness of straws and similar materials compared to wood chips can
also be a problem, as the productivity of a digester depends on the amount of raw material
which can fit into it. This means that the cost of a mill designed to produce straw pulp
will probably be much higher than the cost of a wood chip mill capable of producing the
same amount of pulp of the same type.
Overall, the cost of producing pulp from non-wood material is high. It can only be
justified economically if the pulp is of superior quality, such as is the case with pulp
produced from textile or cordage grade fibres. It is significant that in many countries
where wood is scarce but agricultural residues are plentiful, it is common to import wood
pulp rather than to produce non-wood pulp from materials such as wheat straw. In other
countries, straw pulp and the like is only used because of govemment restrictions on the
63
importation of wood pulp. It is estimated that China and India produce nearly half of the
annual world total of 10 million tonnes of non-wood based paper.
3.4.2 Refining or beating

Refining or beating can be defined as the mechanical treatment of pulp carried out in the
presence of water, usually by passing the suspension of pulp fibres through a relatively
narrow gap between a revolving rotor and a stationary stator, both of which carry bars or
knives aligned more or less across the line of flow of the stock. The term "beating" is
usually apphed to a batch treatment of pulp suspension, whereas "refining" is used when
the stock is passed continuously through one or more refiners in series.
It should be pointed out that the refining of chemical pulp does not produce the same
effects as it does on mechanical pulp. Chemical pulps are relatively pure cellulose, with
the hydroxyl groups accessible. In mechanical pulps the hydroxyl groups are blocked by
the presence of lignin. The refining of mechanical pulp is really a completion of the
process of disintegration of fibre bundles down to individual fibres.
Changes observed in fibre structure as a result of the mechanical action on the fibrous
material can depend on the type of refiner, the refining conditions used, the fibre type
(hardwood or softwood) and the pulp (mechanical or chemical). However the main effects
which are observed can be classified into four areas:
(a) Intemal fibrillation or delamination
(b) External fibrillation of the fibre surface
(c) Fines formation
(d) Fibre shortening
64
Intemal fibrillation effects (a) are difficult to observe under a microscope, but they can be
considered by analogy with a piece of rope. Rope is a helical wrap of strands which
themselves are hehcal wraps of fibres. If one twists a rope in the direction of the helical
wrap the rope becomes stiffer; likewise, if the twist is in the opposite direction the rope
unwinds (or delaminates) to open up the stmcture and becomes "floppy"; such is the case
with intemal fibrillation. The main effect of intemal fibrillation is to increase fibre
flexibility and swelling. The fibres may also undergo excessive curling and twisting.
External fibrillation (b) is easily observed by scanning electron microscopy (Fig. 3.5). The
fibrils or fibrillar lamellae attached to the fibre surface can vary widely in size and sharp
(but the process is again like unravelling a piece of rope at its surface).
The last stage (c) of extemal fibrillation is the peeling off of the fibrils from the fibre
surface with the formation of fines. Depending on the forces acting on the fibre during
refining, more or less of the fibrils will be removed from the surface of the fibre.
Fibre shortening (d) is the other prunary effect attributed to refining. An indication that
fibre shortening has occurred is the change observed in particle size distribution, which is
a result of the cutting action of the blades or discs in the machinery on the single fibres.
Refining plays an important role in producing a large surface area for fibre - to - fibre or
fibre - to - matrix (in the case of composites) bonding and, more importantiy, can assist in
controlling the drainage rates of processing liquids during the fabrication of products.
This is one of natural fibre's main advantages compared to synthetic fibres such as glass,
steel, etc., in asbestos replacement.
65
Fig. 3.5. (a, top) Unbeaten fibre of P.radiata, compared to (b) externally fibrillated fibre (Coutts, 1982b).
3.5 Properties of Natural Fibres

The significance of the properties of the natural fibres in the cement composite will be
studied through out this thesis. The physical and mechanical properties of natural fibres,
such asfibresize, morphology, surface charge, strength, etc., play important roles in the
manufacturing process of fibre cements.
66

3.5.1.1. Physical dimensions
The physical dimensions of fibres are exceedingly unportant in the apphcation of fibres as
reinforcement, be it in paper, medium density fibre board or in fibre reinforced cement.
Hardwood fibres are much shorter (av. 1.0 mm) and narrower (av. 20 pm) than softwood
fibres (av. length 3.5 mm). With softwoods there is a difference in fibre diameters,
between early and late wood (av. diameter 45 |Lim and 13 |im resp.). Hardwood fibres
generally have a higher relative cell wall thickness than do early wood softwood fibres.
This implies hardwood fibres are stiffer and have a greater resistance to collapse (see
Table 3.2).
Table 3.2 Typical softwood and hardwood fibres physical dimensions (Coutts, 1988).
Fibre property
Fibre length (mm)
Fibre width (|im)
Early wood
Late wood
Cell wall thickness (|am)
Early wood
Late wood
Type of cells (%)
Fibres
Vessels
Parenchyma
P. radiata
E. regnans
2 - 6, av: 3.5
45
13
3
5
>90
/
<10
0.5- 2.5, av: 1.0

10 -23 av:20
/
/
4
/
/
65
18
17
It can be seen from Table 3.3 and 3.4 that natural fibres have great variation in length,
diameter, lumen size and fibre wall thickness. Some of the fibre properties for reinforcing
are better than those of wood fibres, for example bast fibres and leaf fibres have much
greater length than that of softwood fibres.
67
Chemical pulping processes reduce natural plant to pulp, which is then composed of
individual cells. For wood and most of non-wood pulp, the small percentage of non-fibre
cells can be washing away and the pulp contains mainly pure fibre cells. However, the
pulps of grasses (cereal straws, sugar cane bagasse, bamboo, etc.) contain fibres as well as
a great variety of other cells, such as parenchyma, epidermal and vessel segments. These
segments and pitted cell act as a filler rather than as a reinforcing material in the cement
based composites.
Table 3.3 Some non-wood fibres physical dimensions (Strells, 1967)

Fibres
Seed cotton
Hairs coir
fiax
hemp
Bast
jute
ramie
kenaf
abaca
Leaf
sisal
NZflax
Lengtii (mm)
Width (|im)
Shape
Lumen size
Wall thickness
10-40 (18)
0.3-1.0(0.7)
9-70 (33)
5-55 (25)
1.5-5(2)
60-250 (120)
2-6 (5)
2-12 (6)
0.8-8 (3.3)
2-15 (7)
12-38 (20)
12-24 (20)
5-38(19)
10-51(25)
10-25 (20)
11-80(50)
14-33 (21)
16-32 (24)
8-41 (20)
5-27(15)
ribbon-like
cylindrical
"
"
broad
"
fine
broad
thick
II
II
"
II
"
"
thin
thick
"
fine
Table 3.4 Some grasses pulps physical dimensions (Strells, 1967)

Fibre
cereal straw
sugar cane
bamboo
* thick wall
** thin wall
L. (mm)
0.68-3.12 (1.48) *
0.8-2.9**
0.8-2.8(1.7)
1.45-4.35(2.7)
2.8-3.26^'
' pitted fibres
W . (|jm)
6.8-23.8(13.3)*
27-34**
10.2-34.1 (20)
6.8-27.3 (14)
20.5-40t
Parenchyma
L. (mm) W. (|am)
0.45
130
0.85
0.25
140
65
Vessel
L. (mm) W. (|im)
1.0
60
1.35
/
150
100
3.5.1.2 Surface potential and surface area.

The Hatschek process is the major manufacturing method for fibre cement products and,
as have been shown (section 1.1.2), is very much akin to a papermaking machine. The
retention of cement particles by asbestos (or its replacement) and drainage properties of
68
the sheets during the dewatering stage are paramount unportance, hence fibre surface
potential and surface area are of considerable significance.
Zeta-potential - the electrokinetic charge on a colloidal particle - has been discussed in

many areas of science. When the particle is a wood fibre it has been postulated that the
zeta-potential plays a major role in the process of papermaking,. The zeta-potential is a
controlling parameter in filler and fines retention, dramage and pulp flow behaviour (Arno,
1974).
Work in the field of paper science (Britt, 1974; Herrington, 1986) shows quite
dramatically that zeta-potential is not a measure of surface charge, and cannot be used for
comparison of the surface charge of even very similar materials, let alone two materials as
distinct as asbestos and wood fibre. In this discussion it suffices to say that in the case of
cellulose fibres, the present evidence indicates that dissociation of ionic groups is the main
source of the charge. At very high pH values, dissociation of the hydroxyl groups can
contribute to the particle charge, but at low pH values the charge must be due to
dissociation of carboxyl groups or sulphonic acid groups, depending on the pulp yield and
on whether a sulphite or sulphate pulp is used. A cellulose fibre dispersion is negatively
charged at all pH values. When cationic hydroxylated metal species, such as occur in
cement particles, are present, charge reversal may occur, as has been established in the
case of wood fibres in the presence of papermaker's alum (Amo, 1974).

The use of natural fibres in paper, paperboard, and fibre reinforced composite materials
has created a need for better characterization of fibre mechanical properties.
69
Unfortunately, researchers often report properties of the fibre when in fact they are
studying aggregates of fibres with very different properties to those of the individual fibre.
For example, bast and leaf fibres have been widely used in the textile and cordage
industries. Tensile strength in the textile industry is defined in terms of tenacity which is
the breaking load per unit mass per unit length; for cordage applications, it can be defined
as the breaking length which is the length of fibre which can theoretically support its own
weight when suspended at one end.
Most of the pioneering work in natural fibre (wood) testing was carried out by researchers
from the paper industry, and much of the data on fibre properties interpreted from tests
carried out on paper sheets. The information on single fibres is limited, although new
techniques in recent times are providing more data as time passes (McKenzie, 1978).
3.5.2.1 Modulus of elasticity and tensile strength

Theoretical calculations suggest the modulus of elasticity of cellulose could be as high as
150 GPa. Experimental values for single fibres vary with change in fibril angle, drying
restrains and defects but range between 10 - 100 GPa (Fig.3.6 and 3.7).
Like elastic modulus, the tensile strength of single fibres is dependent on the helical angle
of the S2 layer and the presence of defects. Page and co-workers (Page 1970, 1971; Kim,
1975) have reported a value of 2000MPa for defect-free black spmce fibres with a zero
fibril angle; this figure should be considered as a maximum value. A more realistic value
would be in the range 500 - 1000 MPa for selected fibre fractions carefully prepared to
minimize fibre damage. Jayne (1959) found values for tensile strength between 350 1000 MPa.
70
Fig. 3.6. A pulp fibre consists of cellulose fibrils in largely parallel array, embedded in a matrix of lignin and
hemicellulose. The majority of the cellulose fibrils form a steep spiral around the axis of the fibre, and it is
this structure which gives to the fibre its mechanical strength
Typtal average
fibril angles for
commercial pulps
D)
C
a>
h.
*^
0)
w
c
0)
0)
il
Fiber strength\fibril angle data iskenfrom _ _ , . ,
Page et al. Pulp Paper Mag Can.,73,(8):T198(lynrJ
10
20
30
40
50
Fig. 3.7. The maximum strength of a pulp fibre is limited by the inherent properties of the cellulose fibril
modified by the spiral angle of the fibril about the fibre axis. The actual strength exhibited by the fibre in
relation to this limiting strengtii is then determined by the presence or absence of defects and dislocations
which produce "weak spots" where premature failure is induced.
71
Considerable data from zero-span tensile testing of paper sheets have been equated to
tensile values for single fibres. It would appear from all the available evidence that the
likely average tensile strength of fibres commercially prepared from reasonable quality
wood by normal methods would be not less than 700 MPa. Strengths of 800 - 900 MPa
might be anticipated if the wood supply is carefully selected to contain a large proportion
of late wood. This is not likely to be readily achieved in Australia or New Zealand where
the main softwood species is young P.radiata (McKenize, 1978).
The more common fibres, such as abaca, sisal, flax, kenaf and bamboo, range in tensile
strength from say 50 - 500 MPa, with densities of about 1.2-1.5 g/cm'. The elastic
modulus of plant fibres can range between 5 - 100 GPa (Coutts, 1992b). Unfortunately,
"fibres" are often characterized as an aggregate of fibres rather than as an individual fibre,
thus information is limited in the scientific literature. Table 3.5 lists some fibre stmcture
and strength values, it should be interpreted with care.
Table 3.5 Structure and strength parameters of non-wood fibres (Mukherjee, 1986)
Fibre
Seed coir
Bast
Leaf
Cellulose%
43
Spiral angle
45
cenL.(mm)
0.75
L/W
35
UTS (MPa)
140
Elongation%
15
fiax
hemp
jute
remie
71
78
61
83
10
6.2
8.0
7.5
20
23
2.3
154
1687
906
110
3500
780
690
550
870
2.4
1.6
1.5
1.2
banana
sisal
65
67
12
20
3.3
2.2
150
100
540
580
3.0
4.3
Pulp fibres strength properties can be reflected by comparing the pulp for tear, burst,
tensile and other sheet properties at different levels of laboratory beating. The properties
can be plotted and compared at different freeness levels; at different levels of beating time;
72
and against each other. When properties are plotted against each other, the freeness, or
beating time, or sheet bulk, is indicated. These curves can then be compared with curves
similarly obtained from other pulps, in order to assess the effects of type offibrousraw
material, and of cooking, bleaching and refining procedure, upon final strength.
Curve A: Abaca (bast fiber) urtbleached sulfate

pulp
B: U.S.A. Southern pine unbleached sulfate
pulp
C: U.S.A. Douglas-fir unbleached sulfate
pulp
D: U.S.A. southern market hardwood
bleached sulfate pulp
E: Bamboo unbleached sulfate pulp
F: U.S.A. West Coast bleached sulfite pulp
G: Esparto unbleached soda pulp
c lao
<
H: Reed unbleached sulfate pulp

I: Estimated average mixed tropical hardwoods unbleached sulfate pulp
J: Kenat whole stalk unbleached soda
pulp
K: Eucalyptus (hardwood) unbleached sulfate pulp
L: U.S.A. aspen (hardwood) bleached sulfate pulp
M: Rice straw unbleached soda pulp
N: Wheat straw unbleached soda pulp
O: Bagasse unbleached soda pulp
P: Bagasse bleached soda pulp
Q: Kenaf woody core material unbleached
soda pulp
R: Kenaf bast ribbon unbleached soda
pulp
S: Sisal pulp
40
eo
80
TAPPI IMETRICI BURST FACTOR
too
Fig. 3.8 Laboratory beating - strength tests on chemical pulps from various wood and non-wood plant fibres.
Digits superimposed on the curves represent Canadian Standard Freeness tests ^ 100 (Atchison 1983).
Figure 3.8 depicts one group of such curves, it can be seen that there are tremendous
differences in strength between the pulps prepared from these different raw materials. For
73
example, the tear strength of abaca (curve A) is outstanding; and the tear strength of most
of the softwood pulps is high compared to that of hardwood pulps and most of the nonwood plant pulps. Mixed tropical hardwood pulps compared favourable with plantation
hardwoods and with temperate zone hardwoods. Also to be noted is the strength of kenaf
bast fibre (curve R); its properties are comparable to those of softwood fibres. By contrast,
the strength of kenaf core pulp (curve Q) is very low. Its initial freeness is also very low,
meaning that it is a very slow-draining pulp.
3.5.2.2 Fibre flexibility

The flexibility of a fibre is of great importance during the preparation stage of a composite
material and also during the process of composite failure. Certainly wet fibre flexibility
has been known to be important for paper manufacture, but the property was seldom
measured quantitatively.
Tam Doo (1982) and Kerekes (1985) have developed a single-fibre flexibility test method
which ehminates the disadvantages of earlier methods. This has enabled the researchers to
measure the effect of wood species and pulping conditions on fibre flexibility and confirm
quantitatively what has long been known in a qualitative way. They showed that cedar
fibres are flexible (1.33 x 10"^ Nm"^), Douglas fir fibres are stiff (6.8 x 10"^ Nm"-) and
Southern pine pulp fibres are very stiff (11.7 x 10"2 Nm'^). It was also shown that
mechanical pulps are 20 - 30 times stiffer than chemical pulps from the same species. By
chemically treating mechanical pulps, a marked decrease in stiffness can be noted.
74
3.5.2.3 Work of fracture

The energy needed to form a fracture surface is called work of fracture, and is related to
fracture toughness by consideration of the area of the surface formed during fracture.
Gordon and Jeronimidis (1980) showed that, weight for weight, the strength and stiffness
of wood along the grain compares well with the best engineering metals, as does its work
of fracture across the grain.
Earlier Page etal (1911) showed that for single wood fibres a pseudo-plastic deformation
took place in tension. Wood fibres are hollow tubes composed of layers of fibrils wound
in a steep spiral (see Fig 3.1), and so behave as a spirally wound fibre reinforced
composite tube. Under axial tensile strain such stmctures can fail by buckling due to the
induced shear stresses. Such a failure mechanism can lead to high values of fracture
toughness. This process will be seen to be important in wood fibre reinforced cements, in
generating fracture toughness in the composite, as the fibres fail in tension.
3.5.3 The effect of moisture on fibres

It has been shown that cellulose is the primary component of the cell wall and the
crystalline microfibrils are the elements which give the fibre its tensile strength; however,
there are disordered zones which are believed to play a significant role with respect to
mechanical properties.
Hemicellulose and lignin act as matrix materials in wood fibres and are generally
considered to be amorphous. They are hygroscopic thermoplastic substances, and so
environmental conditions such as humidity and temperature have a strong influence on
them and hence tiie mechanical properties of natural fibres. With increasing moisture
75
content, the torsional stiffness of fibres decreased by about 50% (between 25% and 90%
RH at room temperature) (Nevell, 1984). This softening is related to the softening of the
hemicelluloses in the cell wall. In the longitudmal direction this effect (up to 90% RH)
results in a drop in modulus of only 11%, although the hemicellulose modulus is reduced
by a factor of a thousand. This is clearly a reflection of the fact that the stiff cellulosic
microfibrils are preferentially aligned along the fibre.
In many cases, a maximum in fibre tensile strength and tensile modulus is found at low
moisture contents. When relatively dry, fibres show low tensile properties due to a poorer
stress distribution within their stracture. Such an effect is not restricted to natural fibres
but also occurs in synthetic fibres.
When single fibres are immersed in water, the relative longitudinal stiffness drops
considerably i.e. by about 70 -80% from the value at 50% RH. This has been explained as
being due to softening of the disordered zones of the cellulose microfibrils and also to a
reduction in cell wall cohesiveness leading to a slippage between fibrils. The relative
decrease in modulus of pulp fibres, when immersed in water, is almost independent of
yield.
76
Chapter Four:
Air-cured and Autoclaved Bamboo Fibre Reinforced Cement
Composites (BFRC)
Recently, there has been a trend to use natural cellulose fibres to replace asbestos fibre in
the fibre reinforced cement industry. In Australia wood fibre (P.radiata) has replaced
asbestos fibre as a reinforcement in commercial cement product since 1981. This fibre has
a reasonably high market price. Thus considerable research effect has gone into the study
of fibre composites from fast growing cheap agricultural crops and crop residues,
especially for those countries with limited forest resources.
Bamboo is a rapid grown agricultural crop, which has good fibre qualities and is widely
used in the paper industry throughout the Asian region. Although bamboo has been used
in various forms in the constmction industry, there is limited information in the scientific
literature conceming the use of bamboo pulp fibre. Sinha et al (1975) and Pakotiprapha et
al (1983) reported that air-cured bamboo fibre reinforced cement composites had flexural
strength values close to 20 MPa, at a fibre loading of 10% by mass. However, there was
no report of the values of fracture toughness, which is as important a mechanical property
as strength or stiffness when considering building materials.
The Hatschek process followed by curing in a high pressure steam autoclave has been
commercially applied to the production of wood fibre reinforced cement (WFRC)
products. Steam curing at temperatures close to 180C enables the replacement of
between 40% to 60% of ordinary Portiand cement by less expensive silica, the latter reacts
77
with the cement to form a calcium silicate matrix of acceptable strength. The reaction is
completed within 6 to 8 hours instead of 3 to 4 weeks required with air-cured products.
This chapter discusses the preparation and mechanical and physical properties of air-cured
and autoclaved BFRC composite to establish their suitably as an alternative to wood pulp
fibre for asbestos replacement in fibre cement building products.

4.1.1 Materials
The bamboo fibre was unbleached Kraft pulp of Kappa No.26, and was prepared from
commercial packaging paper (Chang Jing Paper Mill, China). The bamboo species was
Sinocalamus affinis (Rendle) McClue. The matrix was prepared from ordinary Portiand
cement or from equal proportions of ordinary Portland cement and finely ground silica
(Steetly brand, 200 mesh, washed quartz).
4.1.2 Fibre modification

The fibres used in the composites were obtained by soaking the commercial packing paper
overnight followed by disintegration for 10 min at a speed of 2850 RPM. After
disintegration, the fibres were subjected to 3 different modifications:
1. vacuum-dewatering and crumbing [unbeaten bamboo pulp (400 Canadian Standard

Freeness {CSF})];
2. beating in a Valley Beater to 100 CSF, then vacuum-dewatering and crumbhng [beaten
bamboo pulp (100 CSF)];
78
3. disintegrating the original pulp with hot water (90-95G) for 2 min in a 3 litre NORAM
disintegrater followed by screening on 0.83 mm hole size SomerviUe screen (yield 46%)
then dewatering and crumbling [screened bamboo pulp (550 CSF)].

Air-cured and autoclaved BFRC samples were produced by a slurry / vacuum dewatering
and press technique which had proved most successful with wood-pulp fibre reinforced
cement composites (see Appendix A.3.2).
Mechanical and physical properties of BFRC such as flexural strength (MOR), fracture
toughness, void volume, water absorption and density were determined by the methods
described in Appendix A.4.
The fibre length weighted average was measured on a Kajanni FS-200 fibre length
analyser (see Appendix A.2.3). The fibre length mass distribution was converted from
reported fibre length population distribution.
Canadian Standard Freeness test method is a measurement of pulp drainage. The freeness
test used in this study was to indicate the degree of beating (see Appendix A.2.2).
79
4.2 Air-cured bamboo fibre reinforced cement

Air-cured BFRC composites were prepared from two different bamboo pulps, one beaten
(100 CSF) and the other unbeaten (400 CSF). The physical and mechanical properties of
the BFRC composites are reported in Table 4.1. These properties are also compared with
those of both softwood and hardwood pulp reinforced cements in Figures 4.1- 4.4.
36-
32-
28-
a.
24-
X
h
O
<
D
X
W
-J
U,
12-
8-
10
PERCENT FIBRE (BY MASS)
P. radiata
E. regnans
unbeaten baniboo
beaten bamboo
1
<
12
14
Fig. 4.1. Effect of fibre content on flexural stiength for air-cured WFRC and BFRC
Figure 4.1, shows the variation mflexuralstrength, of BFRC composites, as the fibre
content is increased. This same graph contains published data for air-cured WFRC
reinforced with softwood (P. radiata) (Coutts, 1985) and hardwood fibres (E. regnans)
(Coutts, 1987a). Flexural strength values for BFRC increased from about 10 MPa up to 22
MPa as the fibre content increased from 2 % - 14 % by mass. Theflexuralstrength value
80
for unreinforced cement board was suggested about 9 MPa from the graphic. Unlike
softwood and hardwood fibres which indicated maximum values of flexural strength for
the composites at about 8 % by mass, the bamboo reinforced material was still mcreasing
in flexural strength at loading of 14% by mass for both the unbeaten and beaten pulps.
Table 4.1. Properties of air-cured bamboo-fibre-reinforced cement

Fibre (w%)
Unbeaten Pulp
2
4
6
8
10
12
14
Beaten Pulp
2
4
6
8
10
12
14
MOR (MPa)
Frac. Tough(kJ/m2)
Void Vol (%)
Water Abs.(%)
Density (g/cm')
10.1 2.0
13.41.9
15.02.3
17.01.7
19.71.0
21.43.8
21.22.4
0.100.04
0.150.02
0.230.03
0.340.05
0.490.05
0.800.23
0.970.23
26.30.9
26.30.6
26.50.6
26.61.1
26.01.1
25.60.5
25.30.9
14.60.7
15.2+0.5
15.90.5
16.7+0.3
16.80.7
17.60.4
17.90.8
1.810.03
1.730.02
1.67+0.02
1.590.06
1.550.02
1.460.03
1.420.02
10.91.4
12.11.3
16.21.0
17.40.8
18.61.1
19.21.4
21.81.7
0.070.01
0.150.02
0.230.02
0.320.03
0.450.07
0.540.05
0J00.06
26.80.7
28.30.6
27.60.4
26.60.9
26.50.9
27.107
28.00.6
14.60.6
16.70.4
16.7+0.4
16.50.6
16.90.6
18.20.3
19.40.4
1.830.03
1.690.03
1.65+0.03
1.610.03
1.570.05
1.490.03
1.440.04
*A11 composite were fabricated using ordinary portiand cement, air-cured for 28 days, tested at 50+5 per cent RH and 232C.
* 3 standard deviation, sample size n=9.
The fracture toughness values of BFRC increased with increasing fibre content (Table 4.1
and Fig. 4.2). At a fibre loading of 14% by mass the fracture toughness values were ~ 1.0
kJm"2 and ~ 0.7 kJm"2, for samples containing unbeaten and beaten fibres respectively.
These values are low compared to softwood and hardwood fibre reinforced WFRC's at
fibre loading of 12% by mass which are 2.25 klm'ând 1.68 kJm'^ respectively (Fig. 4.2).
The initial indication is that bamboo fibres are unsuitable on their own as a reinforcement
for cement products as they lack the essential property of fracture toughness. The use of
hybrid fibre cement formulations containing both bamboo and long softwood fibres have
been investigated in an attempt to improve fracture toughness properties and will be
reported in Chapter 5.
81
2.50-
X P. radiata
* . regnans
2.00-
ur\beaten baniboo
beaten bamboo
B
v>
W 1.50Z
I
o
o
W
1.00
<
0.50
0.00
10
12
14
Fig. 4.2. Effect of fibre content on fracture toughness for air-cured WFRC and BFRC
An important observation with regard to the mechanical performance of the BFRC

compared to softwood and hardwood WFRC's is that the long softwood pulp fibres
produced superior properties of flexural strength (> 30 MPa) and fracture toughness (>
2.28 kJm"2) than either BFRC or hardwood WFRC, at similar loading of 8 % by mass of
fibre. This behaviour has been explained by the fact that the long softwood fibres (as
measured on a Kajanni FS-200) had a fibre length weighted average of -2.8 mm and were
more able to contribute to the reinforcement and to the process of fibre pull-out during
fracture. The shorter bamboo fibres (average fibre length weighted 1.1 mm) or hardwood
fibres (average fibre length weighted 1.0 mm) are not as effective as reinforcement.
However, this is not the complete explanation as BFRC and hardwood WFRC have
similar flexural stiength values (see Fig. 4.1), yet BFRC materials have much lower
fracture toughness values than hardwood WFRC (see Fig. 4.2).
82
In natural fibre reinforced cement composites fibre length plays a major role in the
mechanical performance of the material. At the same time other fibre properties such as
fibre coarseness, wall thickness (lumen size), fibre wall stmcture and fibre strength etc.
also effect the composite properties. For example in this study, bamboo fibre is much
weaker than softwood or hardwood fibre. Zero-span tensile index as measured on
handsheets made from pulps of bamboo, P.radiata and E.regnans were 71 Nm/g, 138
Nm/g and 138 Nm/g respectively. Bamboo fibre has a different microstructure from wood
fibre, there is a very small lumen (compared to softwood fibres) and the fibre's primary
wall is easily pealed off during refining(Wai, 1985; Wang, 1993). Furthermore bamboo
pulp contains a considerable number of segments and pitted cells (see section 3.5.1.1.),
which act as a filler rather than as a reinforcing material in the cement based composites.
The complexity of these parameters is currently being studied by investigating the
relationships between pulp fibre properties and the properties of the cement composites
derived from such pulps and some of the preliminary results will be discussed later in this
thesis.

The physical properties of void volume, water absorption and density are reported in Table
4.1. There appears to be littie difference in the physical properties of the composites
containing either beaten and unbeaten bamboo fibre. As the fibre content was increased
from 2 % - 12 % fibre by mass the water absorption only increases from 14 % to about 18
% by mass (see Fig. 4.3). It can be seen that as the fibre content increased the rate of
increase in water absorption was much lower for bamboo fibre containing materials than
for softwood (14 % - 25 %) and hardwood (18 % - 30 %) fibre reinforced products.
83
Likewise, the density of the bamboo reinforced materials decreased at a lower rate than
either softwood or hardwood WFRC (Fig. 4.4), while the void volume remained fairly
constant over the range of fibre contents studied.
32-
28-
24-
z
o
p
20-
ft.
16-
i/i
oa
<
a:
U
12-
<
X p. radiata
E. regnans
a
-i
10
unbeaten bamboo
beaten bamboo
12
14
Fig. 4.3. Effect of fibre content on water absorption for air-cured WFRC and BFRC
84
2.00-
X
^
1.80-
p. radiata
E. regnans
unbeaten bamboo
beaten bamboo
B
1.60</2
z
Q
1.40-
1.20-
10
12
14
PERCENT nBRE(BYMASS)
Fig. 4.4. Effect of fibre content on density for air-cured WFRC and BFRC
4.3 Autoclaved bamboo fibre reinforced cement

4.3.1.1 Flexural strength
1. Unbeaten fibre reinforced composites
Table 4.2 and Figure 4.5 show the variation in flexural sttength of BFRC composites, as
the fibre content is increased from 2% to 22% in steps of 2%. Figure 4.5 contains
reference data for autoclaved WFRC composites reinforced with Kraft P.radiata fibres, as
this fibre is commercially used in Austraha and is the prefcrted fibre. Flexural strength
values for autoclaved BFRC composites increased from about 12 MPa up to 18 MPa as the
fibre content was increased from 2% to 14% at which point the stiength of the composite
products start to decrease due to poor fibre distribution throughout the matrix material.
85
This observation was in general agreement with the change in flexural strength observed
with autoclaved WFRC composites.
2. Beaten fibre reinforced composites

Beaten (100 CSF) and unbeaten (400 CSF) BFRC material gave similar flexural strength
values of approximately 18 MPa at 14% fibre (Table 4.2). With the WFRC composites it
was found that beaten P radiata (550 CSF) gave a maximum flexural strength of 24.3
MPa at 10% fibre loading. This was an improvement over products reinforced with
unbeaten fibres.
Table 4.2 Properties of autoclaved bamboo-fibre-reinforced cement

Fibre (w%)
Unbeaten Pulp
2
4
6
8
10
12
14
16
18
20
22
Beaten Pulp
2
4
6
8
10
12
14
16
18
20
22
MOR (MPa)
Frac. Tough(kJ/in2)
Void Vol (%)
Water Abs.(%)
Density (g/cm.-^)
13.30.5
13.2+1.8
13.80.9
15.51.0
16.00.8
16.51.2
18.31.6
17.21.5
17.71.0
16.1 0.6
14.91.3
0.080.01
0.130.01
0.200.02
0.290.05
0.390.04
0.490.05
0.620.0I
0.77+0.07
0.960.14
1.080.11
1.200.31
35.70.5
38.00.9
39.11.0
43.1 2.2
45.1 1.7
45.70.9
45.44.9
48.6+1.4
49.21.2
50.80.8
50.90.6
22.1 0.5
24.90.9
26.71.0
30.71.1
33.91.5
35.91.3
36.66.5
41.12.3
42.82.3
45.1 2.1
49.02.3
1.620.02
1.520.03
1.46+0.02
1.410.04
1.330.02
1.27 0.02
1.26+0.11
1.180.03
1.150.03
1.130.05
1.000.05
12.1 0.5
12,21.0
14.61.3
14.9+1.2
16.1 0.9
17.11.4
18.21.3
18.21.4
16.40.9
17.11.5
16.71.2
0.100.01
0,130.02
0.220.03
0.290.04
0.400.04
0.500.02
0.500.01
0.560.06
0.820.08
0.970.08
1.03+0.11
34.2+0.6
37.7].l
39.81.5
41.1+1.1
44.21.3
44.3+1.4
42.80.8
44.71.5
46.91.1
47.00.8
47.4+1.0
21.00.6
24.81.0
27.61.4
29.1 1.1
33.11.6
34.1+1.6
32.51.2
35.92.0
40.6+1.8
40.91.1
42.21.4
1.630.02
1.520.02
1.440.02
1.41+0.02
1.340.03
1.300.03
1.320.03
1.250.03
1.160.04
1.150.01
1.120.02
*BFRC composites were fabricated using ordinary Portland cement and silica at the ratio of 1:1, autoclaved at 1.25MPa
steam pressure for 7.5h, tested at 505 per cent RH and 22+2C. BFRC composites maximum flexural strength at 14%
by mass. 3 standard deviation, sample size n=9.
86
26
,^:24J,
24
22
20
-3- 18
-u
S. 16
t 14
12
I 10
unbeaten BFRC
iS
beaten BFRC
screened BFRC
beaten WFRC
2
0
10
12
14
Percent Fibre (by mui)
16
18
20
22
24
Fig. 4.5. Flexural strength as a function of percentfibreloading for autoclaved BFRC and WFRC composites
Table 4.3 Fibre weighted average length (mm)
beaten P. radiata
2.4
unbeaten bamboo
1.0
beaten bamboo
0.8
screened bamboo
1.3
Bamboo has a fibre length less than half of that of P.radiata (Table 4.3). The bamboo
fibre used in this study was separated from commercial packaging paper and there was a
high fines content in both the beaten and unbeaten pulps. The beaten and unbeaten
bamboo pulp fines (lengths < 0.4 mm) accounted for 32.7% and 28.8% of the mass
respectively (Table 4.4). This suggests httle damage occurred during beatmg.
Table 4.4 Fibre length mass distribution percentage

Lengtli (mm) longer than
beaten P.radiata
unbeaten bamboo
beaten bamboo
screened bamboo
99.7
99.7
100
0.05
100
613
94.8
71.2
95.6
0.4
74.3
42.1
0.8
51.6
90.0
22.6
51.1
32.2
1.2
81.8
11.4
30.3
1.6
19.2
72.1
15.4
5.3
2.0
61.3
10.9
7.6
2.5
2.4
48.9
5.9
1.3
3.6
36.8
2.8
3.1
1.6
0.6
28.0
1.7
3.2
Fibre length mass distribution percentage was analysed on Kajanni FS-200 fibre length analyzer. More than
15,000 fibres were measured and analysed. Length less than 0.4mm fibre mass distribution percentage for
p.radiata, imbeaten bamboo, beaten bamboo and screened long bamboo pulp was 4.4%, 28.8%, 32.7% and
5.2% respectively.
87
3. Screened fibre reinforced composites

Bamboo pulp was washed through a SomerviUe screen (0.83 mm hole size, yield 46%) to
remove fines and to obtain the longer fibres for use as reinforcement. Compared with
unscreened unbeaten fibre the screened fibre showed improved composite flexural strength
at the same fibre loading. The maximum flexural strength of screened BFRC was 21.6
MPa at 10% fibre loading (Table 4.5).
Fibre length can make a significant contribution to the composite flexural strength.
Softwood P.radiata fibre, and screened and unscreened unbeaten bamboo fibres, had fibre
lengths (length weighted average) was 2.4 mm, 1.3 mm and 1.0 mm respectively (Table
4.3). The flexural strengths of materials reinforced with wood fibre, and screened and
unscreened unbeaten bamboo fibres were 24.3 MPa, 21.6 MPa and 16.0 MPa respectively
at 10% fibre by mass (Fig. 4.5).
Table 4.5 Properties of autoclaved screened long bamboo fibre reinforced cement
Fibre (w%)
2
4
6
8
10
12
14
MOR (MPa)
Frac. Tough(kJ/ni2)
Void Vol (%)
Water Abs.(%)
Density (g/cm')
13.31.4
39.10.8
24.70.8
1.590.02
0.100.01
14.8+1.5
41.21.0
27.71.1
1.490.02
0.23+0.02
0.370.06
]6.32.2
42.61.0
3O.01.3
1.420.03
1.420.07
18.81.2
0.54+0.12
30.1+0.9
42.60.9
32.01.4
42.71.4
1.330.02
0.710.13
21.6+2.2
37.81.4
1.230.03
46.60.6
0.790.06
19.5+1.5
39.81.7
1.180.03
1.090.14
47.00.9
18.91.9
* Unbeaten bamboo pulp was screened on 0.83mm hole size Sommerville screen (yield 46%) fibre length weighted
average (Kajanni FS200 fibre analyser) was 1.3 mm and the freeness was 550CSF. BFRC maximum flexural strength
was at 10% by mass fibre loading, which was similar fibre loading to WFRC. 3 standard deviation, sample size n=6.
The flexural strength of unscreened BFRC materials (beaten and unbeaten) increased up to
14% fibre before the maximum value was reached. This could be due to the fact that the
fine material (length < 0.4 mm) offers littie reinforcement to the composite and so a
greater mass of pulp was needed to have sufficient numbers of long fibres. When fines
(lengths < 0.4 mm) were removed, maximum stiength was achieved with fibre loading
between 8 - 10%, which is in keeping with the early study of WFRC composite. The
percentage of screened bamboo pulp fibre lengths longer than 0.4 mm, was 94.8% of its
mass. This is similar to the data for P.radiata pulp (Table 4.4).
A fibre length fractionation study of P.radiata wood pulp supports this belief that
fragments with length less than 0.3 mm, act more as a filler-diluent than as a reinforcing
fibre when used to make WFRC products (see Chapter 6). A similar behaviour regarding
the maximum flexural strength value with respect to fibre mass content (about 12%) was
noted for waste paper fibre reinforced cement products (Coutts, 1989). In that instance,
the high fibre content was required to provide sufficient mass of the longer reinforcing
fibres. This was due to the high fines content generated (lengths < 0.6 mm constituted
21.2% by mass) during processing and recycling.
There is httle difference between beaten and unbeaten BFRC composites with respect to
flexural strength, which contrasts with the observation reported in the earher research on
beaten and unbeaten P.radiata WFRC composites (Coutts, 1984a). Similar behaviour to
that of autoclaved BFRC was found in the case of autoclaved NZ flax composite (Coutts,
1983c) and air-cured bamboo reinforced cement products (Coutts, 1994b). This might be
associated with the fact that the round and small diameter NZ flax and bamboo fibres do
not have the same ability to collapse as do softwood fibres which form flat ribbons.
Altematively, it may be due to the beating, generating fines which do not contribute to the
composites stiength.
89
4.3.1.2 Fracture Toughness

The mechanism that takes place when a fibre composite is loaded to failure include fibre
fracture and fibre pull-out. The latter can have considerable influence on the value of
fracture toughness. If the fibre is short then the energy required to pull the fibre through
the matrix, after the fibre to matrix bond is broken, is low and can contribute httie to the
dissipation of energy contained in the advancing crack. Therefore the crack continues
through the sample and the material appears brittie.
2.6
-
unbeaten BFRC
2.2
beaten BFRC
screened BFRC
2.4
cT 1-8
-O
52.45
beaten WFRC
10
12
14
Percent Fibre (by mm)
16
24
Fig. 4.6. Fracture toughness as a function of percentfibreloading for autoclaved BFRC and WFRC
composites
Figure 4.6 shows tiie increment in fracture toughness values of beaten and unbeaten BFRC
composites (from 0.10 kJ/m^ to 1.03 kJ/m^ and from 0.08 kJ/m^ to 1.20 kJ/m^ respectively)
for fibre loading from 2% to 22%. The same graph has reference data for WFRC
composite fracture toughness, which are seen to have higher values (Coutts, 1984a).
Screened BFRC composites which contain more long fibres are tougher [from 0.10 kJ/m^
to 1.09 kJ/m2 for fibre loadings from 2% to 14%, (Fig. 4.6)] than unscreened composites.
Unbeaten BFRC composites are sHghtiy better than beaten ones at the high fibre loadings.
90
WFRC composites indicated better properties of fracture toughness than BFRC

composites in the three point bending load/deflection curve as shown in Figure 4.7. After
reaching the maximum load, the curve for WFRC composites went through a gradual
"tailing off" which indicated that a greater amount of fracture energy was needed to pull
the long fibres though the matiix. The curves for the BFRC composites did not show such
behaviour and were more brittle. This is indicated by a sharp "drop off of load carrying
capacity in the load/deflection curve. Fracture toughness behaviour is related to fibre
length and fibre morphology, and comparison is only valid for specimens of the same
thickness.
Fig. 4.7. Typical Load / Deflection graph for autoclaved WFRC and BFRC composites.
Beaten BFRC composites did not vary greatly in fracture toughness values from those of
the unbeaten BFRC composites. At high fibre loadings the beaten BFRC composites
showed slightiy lower values due probably to the increased amount of fines present in the
formulation (Table 4.4). As stated above, the presence of fines provides less opportunity
91
for fibre pull-out being a major component in the mechanism of failure, and hence lower
fracture toughness values are observed (see Chapter 6).
4.3.2 Physical Properties

The changing proportions of the constituent fibres and matrix affect void volume, density
and water absorption (Table 4.2 and 4.5). There is little difference in the density of the
composites when the bamboo fibre was beaten compared with the materials containing
unbeaten bamboo fibre (Fig 4.8). However, the same graph shows a slight decrease in
density, at a given fibre content, when the pulp has been screened to remove the fines.
This effect is possibly due to the increase in fines, present in the unscreened pulps, which
allows closer packing of the fibres and matrix, and hence less void volume in the
composite.
1.8
1.6
1.4
1.2
I'
f ^-^
unbeaten BFRC
0.6
HJ- beaten BFRC
0.4
screened BFRC
0.2
0
6
10
12
14
16
18
20
22
24
Percent Fibre (by mui)
Fig. 4.8. Density as a function of percentfibreloading for autoclaved BFRC composites.
The relationship between water absorption and density is depicted in Figure 4.9. The
amount of water absorbed by the cellulose fibre reinforced cement composites depends on
92
their void volume and the amount of cellulose material present; both these parameters
have an effect upon density. Thus one would expect the density to decrease and the water
absorption to increase as the fibre content is increased, due to the nature of the
hydrophilic, low density bamboo fibres. At the same time the packing of fibres and matrix
becomes less efficient, as the fibre content is increased and so void volume increases
accompanied by decreased density and increased water absorption.
I s
0.9
1.1
1.2
1.3
1.4
1.5
1.6
1.7
Demity (g'em-')
Fig. 4.9. The relationship between density and water absorption for autoclaved BFRC composite
4.4 Conclusions
Bamboofibreis a satisfactory fibre for incorporation into the cement matrix. Air-cured
BFRC composites at 14% by mass had flexural stiength values of about 22 MPa.
Altematively, autoclaved products had strength values of about 18 MPa at same fibre
loading. However, the fracture toughness was low due to short fibre length and high fines
content of the bamboo pulp.
93
By screening out "fines" contained in the original bamboo pulp, autoclaved composites
flexural strength could be improved to greater than 20 MPa while fracture toughness
exceeded 1.0 kJ/m2 at a loading of 14% fibre .
In contrast to softwood fibre reinforced cement composites, beaten bamboo fibre

composites did not vary greatly in flexural strength and fracture toughness values from
those of the unbeaten fibre composites.
94
Chapter Five:
Bamboo and Wood Hybrid Fibre Reinforced Cement
Composite Materials (BWFRC)
As we have discussed in Chapter 4 composites reinforced with bamboo fibre have lower
mechanical properties than composites reinforced with wood fibre due to relative short
fibre length of bamboo fibre. This study includes in blending fumishes of bamboo pulp
and long softwood fibre pulp to increase the average fibre length thus hopefully improving
composites mechanical properties.
Pulp blending idea is used widely in the pulp and paper industry with the aim of
modifying the final product specifications. For example, long fibre pulp blended
with short fibre pulp can improve short fibre paper properties of tear, tensile and
burst strength; short fibre pulp blended with long fibre pulp can improve long
fibre paper properties of formation and surface smoothness. Blended fumish of
some pulps at certain proportion woitid even generate synergistic effect for some
paper properties (Kuang, 1992).
Over the years, the combined use of different types of fibres to optimise the performance
of a material has also been studied and commercialised by a number of material scientists
and technologists. A considerable amount of work has been done in hybridising
polypropylene fibre, PVC fibres, glass fibre and cellulose fibre with the aim of improving
fibre-cement products fracture toughness and durabihty (Walton, 1975; Mai, 1980;
Simatupang, 1987; Gale, 1990). However, the poor temperature resistance of
95
polypropylene fibre and poor aUcah resistance of glass fibre inhibit their use in autoclaved
cement product (see section 1.1.3).
James Hardie Industiies in Austraha has a patent, which reported the combination of
cellulose fibres and a small amount of chopped cellulose fibre (preferred proportion at
about 0.1% of the total mass, and preferred length about 10 mm) (Aus Patent, 1982), but
there is no further information available regarding their products reinforced with hybrid
fibre. Fordos at Dansk Etemit in Denmark reported a study of hybrid composites
properties (1986). Kraft softwood fibre combined with eucalypt fibres at l.T ratio showed
interesting high flexural strength in air-cured cement, although detailed information was
not provided in the paper.
The purpose of this chapter is to investigate the combination of long softwood fibre with
short bamboo fibre to improve the composites properties, particularly in terms of fracture
toughness development.

5.1.1 Fibre Preparation
Bamboo and softwood pine fibres were used in this study. Bamboo fibre was prepared
from Kraft-unbleached commercial packaging paper (Jian Xi Paper Mill, China). This is a
different source of bamboo pulp to earlier work although tiie species is the same
[Sinocalamus affinis (Rendle) McClue]. Pine fibre was obtained from Austraha APM
Maryvale mill Kraft-unbleached dry lab pulp. The packaging paper and the dry lab pulp
were soaked separately in water over-night then disintegrated into individual fibres. After
disintegration the fibres were subjected to de-watering, crumbling and the moisture
96
content was measured. The wood pulp was blended with bamboo pulp in a small mixer
into various proportions as shown in Table 5.1.

Fibre cement composites reinforced with above prepared hybrid fibre were produced and
characterised as the method described in Appendix A.3 and A.4. Air-cured samples were
reinforced with 6%, 8% and 10% of blended pulp. Whereas, the autoclaved samples were
reinforced with 8%, 10%, 12% and 14% of blended pulp.
2.5
g 1 . 5
a
'o
5
...
a
% 0.5
20
40
60
80
100
Proportion of Pino in Total Pulp (%, by weight)
Fig. 5.1. Relationship between pine fibre proportion and furnish pulp length weighted average.
97

5.2.1 Length and freeness of blended pulp
The value of length weighted average and freeness of blended pulps is listed in Table 5.1
and depicted in Figure 5.1 and 5.2. These fumishes were prepared from different
proportions of bamboo and pine pulp. Both fibre length and freeness of blended pulp
increased almost linearly as the proportion of pine fibre increased. For example, bamboo
pulp alone had length about 0.91 mm and freeness around 330 CSF; while, 60% bamboo
and 40% pine blended fumish had increased length to 1.54 mm and draining abihty to 470
CSF. As discussed m Chapter 5, BFRC had fairly poor strength and toughness values due
to the short length of bamboo fibres. Blended pulp has increased average fibre length thus
one would expect composites reinforced with such hybrid fibre to have better mechanical
properties.
20
40
60
80
Proportion of Pine in Total Pulp (%, by weight)
Fig. 5.2. Relationship between pine fibre proportion and fiirnish pulp freeness value.
100
5.2.2 Air-cured BWFRC composites
Results of air-cured BWFRC composites mechanical and physical properties are listed in
Table 5.2 and shown in Figure 5.3, 5.4 and 5.5, 5.6.
Table 5.1 Length and freeness of blended pulp

Proportion of Blended Pulp
100B*+0P
90B+10P
80B+20P
60B+40P
40B+60P
20B+80P
OB+IOOP
Length Weighted Av. (mm)
Freeness (CSF)
0.91
1.22
1.31
1.54
1.82
2.05
2.37
330
390
410
470
540
630
680
* B and P represent bamboo and pine pulps respectively. The number before B and P is the weight proportions of bamboo and pine
pulp in the Wended fijrnish.
Table 5.2 Properties of air-cured BWFRC

Fibre (w%)
100B+ OP
6%
8%
10%
MOR (MPa)
Frac. Tough(kJ/m2)
Void Vol (%)
Water Abs.(%)
Density (g/cm')
17.812.3
18.1+1.1
18.5+2.0
0.8210.06
0.9710.05
1.24+0.22
33.110.9
33.6+0.5
35.3+1.4
20.610.7
21.9+0.4
23.211.2
1.6410.02
1.5610.01
1.5010.02
90B +lOP
6%
8%
10%
18.81.7
20.41.8
20.111.5
1.1110.21
1.2610.17
1.34+0.09
32.6+0.9
33.710.7
35.310.6
20.111.0
21.610.7
23.610.6
1.6310.04
1.5710.03
1.50+0.01
SOB + 20P
6%
8%
10%
17.911.5
21.0+2.2
23.312.0
0.9510.22
1.3510.19
1.6710.11
31.9+0.7
33.8+0.5
33.9+0.9
19.410.6
21.510.6
22.310.9
1.6410.02
1.57+0.02
1.5210.02
60B + 40P
6%
8%
10%
20.311.5
21.3+1.9
22.711.2
1.2110.20
1.3610.23
1.7810.25
32.510.5
34.610.5
35.611.0
20.010.4
22.210.6
23.711.1
1.6410.02
1.5610.02
1.5010.03
40B + 60P
6%
8%
10%
22.9+1.4
24.411.9
25.811.8
1.4710.27
1.6910.38
2.38+0.58
32.0+0.7
33.610.8
33.810.3
19.210.8
21.410.3
22.710.6
1.6610.03
1.5710.02
1.5210.03
20B + SOP
6%
8%
10%
23.811.8
26.612.7
25.012.4
1.4610.27
2.1510.33
2.4910.46
32.010.4
33.311.2
35.0+0.8
19.210.4
21.011.2
23.310.9
1.6610.02
1.5810.04
1.5010.03
0B+ lOOP
6%
8%
10%
24.311.4
25.511.5
26.510.9
1.9710.21
2.3810.38
2.9610.37
32.5+0.6
32.0+0.4
33.110.3
19.1+0.6
19.9+0.4
22.2+0.4
1.6610.02
1.6110.02
1.5310.02
*A11 composite were fabricated using ordinary portiand cement, air-cured for 28 days, tested at 50+5 per cent RH and 232''C.
99
Figure 5.3 shows that flexural strength of BWFRC increases with increasing of hybrid
fibre contents which is in keeping with early results in natural fibre reinforced cement.
More importantiy that strength develops almost linearly as the long pine fibre content is
increased. This behaviour was associated with increasing average fibre length which
results from adding long pine fibre.
Composites reinforced with a blend of 20 % bamboo and 80 % of wood fibre (by weight)
had similar strength values than those reinforced with wood fibre alone. The reason for
this is not clear. This could be attributed to experiment error. Coutts and Warden (1985)
demonstrated that wood fibre reinforced cement composites could achieve a flexural
strength of about 30 MPa at 8% fibre content. Their results are in good agreement with
strength developing trend shown in this work (dot lines shown m Figure 5.3).
35
30
6% of total pulp
8 % of total pulp
15 r
1 0 % of total pulp
10
20
40
60
80
100
Proportion of Pino in Total Pulp (%, by weigfit)
Fig. 5.3. Influence of longfibre(pine) proportion ontiieair-cured compositesflexuralstiengfli.
Short fibre blended with some long fibre will greatiy improve the pulps reinforcing ability.
The influence of such reinforcing ability on the composite fracture toughness property is
100
more pronounced than on the strength property. If tiie fibre is longer, more energy will
need to be consumed to pull the fibre out from the matrix, after fibre-mah-ix bond is
failure; thus, miproves composite fracture toughness. Figure 5.4 shows that fracture
toughness value of composite increases very rapidly with increasing of long softwood fibre
proportion in the reinforcing fibres.
6% of total pulp
8 % of total pulp
10% of total pulp
20
40
60
80
100
Proportion of Pina in Total Pulp (%, by waioht)
Fig. 5.4. Influence of long fibre (pine) proportion on the air-cured composites fracture toughness.
Influence of long fibre proportion on the strength and toughness properties of composite
can be understood in terms of fibre length. The relationship between fibre length and
composite flexural stiength or fracture toughness is shown in Figure 5.5 and 5.6,
respectively. The observation will be seen more clearly in Chapter 7. In that instance,
studies on the composites reinforced with various length modified pulps show the
significance of fibre length to the composites strength and toughness development.
101
0,5
1,5
2.5
Fibre Length Weighted Av. (mm)
Fig. 5.5. Influence of fibre length on the air-cured composites flexural strength at total 8% fibre content.
0.5
1.5
2.5
Fig. 5.6. Influence of fibre length on the air-cured composites fracture toughness at total 8% fibre content.
The changing of wood fibre proportion of the constituent hybrid fibre is not expected to
affect composite physical property such as void volume, density and water absorption.
102
Table 5.2 shows fair constant value of composites physical properties if the total
reinforcing fibre content remains the same.
This work is very important for utilising natural fibres as reinforcing material in fibrecement products. Most natural plant fibre (including some wood fibre) do not have good
reinforcing potential due to their short fibre length. As discussed above, short fibre
blended with certain proportions of long fibre such as Kraft pine would increase the
average fibre length thus significantly improve its reinforcing ability. One of the
weaknesses of waste paper is its unsatisfactory length distribution when utilised as a
reinforcing material (Coutts, 1989). This weakness could be overcome by adding a certain
amount of long fibre pulp. On the other hand, one would modify the products properties
by adding cheap short fibre to reduce the cost of fibre production.
In WFRC manufacture practice, natural fibre (eg. softwood) requires some degree of
refining (beating) to enhance its reinforcing ability, more importantly to control the
drainage rates of processing liquids during the fabrication of products (Coutts, 1982a).
Unrefined (unbeaten) wood fibre has very high freeness value about 700- 800 CSF.
Refining can reduce this freeness down to 500-550 CSF, which is preferred for the
manufacture process. Such an effect could also be achieved by blending long fibre with
short fibre without refining (see Table 5.1 and Figure 5.2). It might be possible to reduce
the reinforcing cost by means of fibre blending. However, this idea needs further work to
confhm in both laboratory and pilot-plant scale before it is adopted by the industry.
5,2,3 Autoclaved BWFRC composites

As in the case of the air-cured BWFRC, tiie strength property of autoclaved BWFRC
composites also increased with an increase of long wood fibre proportion as seen in Table
5.3 and Figure 5.7, 5.8. However, strength improvement was not as high as that in the
case of air-cured BWFRC. Strength of air-cured composites improved from 18.1 MPa
reinforced with bamboo fibre alone to 26.6 MPa reinforced with blended 80 proportion of
wood fibre at 8% of total fibre content, about 47 % increase. But in autoclaved
103
composites only a 37% increase of flexural strength can be observed. Over the range of
fibre loadings studied there was littie change in the strength properties of a given hybrid
formulation. However, as the pine fibre content increased there was a gradual increase in
composite strength.
Table 5.3 Properties of autoclaved BWFRC
Fibre (w%)
100B+ OP
8%
10%
12%
14%
MOR (MPa)
Frac. Tough(kJ/m2)
Void Vol (%)
Water Abs.(%)
Density (g/cm')
13.8+1.1
14511.5
14.111.2
13.712.1
0.2710.02
0.3710.04
0.4810.04
0.5410.11
43.810.8
45.611.1
46.911.1
48.611.1
32.5+1.5
35.311.6
38.2+1.8
41.4+2.3
1.3510.04
1.29+0.03
1.2310.03
1.1710.04
90B +lOP
8%
10%
12%
14%
14.611.1
15.312.6
16.412.0
15.412.3
0.4210.07
0.61+0.10
0.8010.07
0.9510.13
43.9+1.7
46.4+1.9
46.911.6
47.811.0
33.5+2.3
36.112.7
38.912.7
40.5+2.0
1.33+0.04
1.2910.05
1.2110.04
1.1810.03
80B + 20P
8%
10%
12%
14%
15.5+0.8
16.7+1.9
16.9+1.1
16.511.3
0.47+0.04
0.6310.08
0.8410.15
1.0810.15
44.911.5
45.611.7
47.1+1.0
48.811.1
33.412.1
35.312.2
38.311.7
41.7+2.6
1.3410.04
1.2910.04
1.2310.03
1.17+0.05
60B + 40P
8%
10%
12%
14%
17.211.2
17.611.5
17.7+2.3
17.8+1.8
0.6510.07
0.9710.15
1.1710.20
1.50+0.18
42.8+2.2
44.412.0
46.0+2.2
48.411.7
31.712.5
34.2+2.5
36.9+2.8
40.312.2
1.3610.04
1.3110.04
1.2510.04
1.2010.03
40B + 60P
8%
10%
12%
14%
16.511.5
18.0+1.4
19.3+1.5
18.512.5
0.7710.10
1.0910.11
1.2810.18
1.6910.10
44.311.2
45.3+1.3
46.311.4
47.6+1.2
32.5+1.6
35.711.8
37.0+2.0
39.9+2.0
1.3410.03
1.2710.03
1.2510.03
1.1910.03
20B + 80P
8%
10%
12%
14%
18.613.6
20.4+2.9
19.812.5
20.114.6
0.9510.30
1.1910.13
1.55+0.21
1.8710.50
42.812.1
45.811.2
45.9+2.1
47.811.9
31.112.4
35.011.6
36.2+2.9
39.612.7
1.3710.04
1.2910.03
1.2510.04
1.20+0.04
OB +loop
8%
10%
12%
19.912.8
20.613.1
20.9+3.3
1.5810.37
1.9010.32
2.75+0.40
42.411.4
43.111.8
44.0+1.4
31.611.7
33.6+2.0
35.211.8
1.34+0.03
1.2710.03
1.2510.03
*BWFRC composites were fabricated using ordinary Portland cement and silica at the ratio of 1:1, autoclaved at 1.25MPa steam
pressure for 7.5h, tested at 505 per cent RH and 222C. 3 standard deviation, sample size n=9.
104
B% of total pulp
1 0 % of total pulp
1 2 % of total pulp
1 4 % of total pulp
20
40
60
80
100
Proportion at Pina in Total Pulp (%. by weight)
Fig. 5.7. Influence of longfibre(pine) proportion on the autoclaved compositesflexuralstrength.
25
20
15
10
5 -
0.5
1.5
2.5
Length Weight Av. (mm)
Fig. 5.8. Influence offibrelength on the autoclaved compositesflexuralstrength at total 8%fibrecontent.
Fracture toughness value of autoclaved hybrid composites increased very rapidly with
increasing long fibre (wood) proportion as seen in Figure 5.9 and 5.10. This behaviour has
already been observed in the case of air-cured composites. Fracture toughness values of
105
the composites increased over 5-foId as the pine fibre was increased in the hybrid fibre
formulation. In contrast to strength values, the fracture toughness values were seen to
increase, for any given hybrid pulp, as the fibre content increased. This increase was of
the order of twofold as the content increased from 8% - 14%.
3.5
2.5
2
1.5
0.5
20
40
60
80
100
Proportion of Pine in Total Pulp (%, by weight)
Fig. 5.9. Influence of long fibre (pine) proportion on the autoclaved composites fracture toughness.
0.5
1.5
2.5
Length Weighted Av. (mm)
Fig. 5.10. Influence of fibre length on tiie autoclaved composites fi:acture toughness at total 8% fibre content.
106
Physical property of autoclaved products remamed fairiy constant at tiie same total
reinforcing fibre contents. But the density values was about 18% lower than that of the
air-cured counterpart. Water absorption and void volume always have reversed
relationship witii density. Hence, higher density samples would have low water absorption
and void volume.
5.3 Theoretical predication and the experimental results

We have discussed early in section 2.1 that the composite strength and toughness can be
calculated from the equations (2.4) and (2.14) respectively.
(Jt, = [a/p] cr.,,;,v, + 0.82(aT)vf(l/d)
(2.4)
R = 0.41vfft/
(2.14)
12d + [vn,. + 0.41vfl / d]Rr,
Where a, p, T, <7,h and R are constants which can be determined experimentally. The
fibre diameter d is suggested to remain constant for certain types of fibre. Thus at a
constant fibre content Vf and v,, the composite strength increases linearly, while the
toughness increases as a second order polynomial with increasing fibre length. Refer to
equations (5.1) and (5.2).
Oj, = Kn(l) + Kn
(5.1)
R = K2i(f} + K2:il) + K23
(5.2)
Where K-i, Kj-, Kzu K22 and K23 are constants.
The experimental results support the above theoretical prediction. A simple regression
analysis was performed on the data in Table 5.2 and 5.3. The results are presented in Table
5.4 which displays a relationship of fibre length, composite strength and toughness. The
relationship is valid as the statistical multiple coefficient (R"") is approximately 0.95.
107
Table 5.4 Regression analysis results

Fibre (w%)
MOR (MPa)
Relation
R"
Fract. Toughness (kJ/m')

Relafion
R/
Air-cured
6%
8%
10%
Autoclaved
8%
10%
12%
14%
5.29 L+12.35
5.74 L+13.26
6.16 L+13.48
0.917
0.900
0.830
0.24 L"-0.06 L +0.71

0.25 L^ +0.15 L +0.70
0.33 L" + 0.24 L + 0.70
0.944
0.970
0.951
4.12 L +9.98
4.47 L+10.43
4.49 L+10.68
5.44 L+-8.98
0.931
0.947
0.969
0.977
0.50L" -0.81 L +0.64

0.35 L"-0.19 L +0.30
0.89 L^- 1.56 L + 1.29
-0.45 L" +2.53 L-1.42
0.973
0.958
0.947
0.990
Table 5.1 shows that the blended pulp fibre length increased almost lineariy as the
proportion of pine fibre increased. Therefore the composite strengtii increased almost
hneariy with increasing proportions of long fibre content. On the other hand, the
composite fracture toughness improved as a second order polynomial function.
5.4 Conclusions
Bamboo pulp when blended with softwood pine pulp improves the fumish pulp's average
fibre length and drainage abihty. The improvement increased with increasmg the
proportion of wood pulp (long fibre).
The advantage of introducing the hybrid fibre was to improve and modify the composite
properties. In tiiis study increasmg the proportion of wood fibre caused improvement of
both air-cured and autoclaved composites mechanical properties, but had littie effect on
their physical properties. The improvement was minimal but more obvious in the case of
the Stiength property of air-cured products than on the autoclaved counterpart. The
108
improvement in the property of fracture toughness was more pronounced than that of
strength for the same formulations containing hybrid fibres.
109
Chapter Six:
Influence of Fibre Length on Composite Properties
The development of asbestos free fibre cement industry has made it most desirable to
obtain definite information on the relationship between the nature of the natural plant
fibres and their composite properties. It has been generally recognized that fibre length
and fibre stiength are two of the most important factors but, because various features of
fibre morphology and chemical composition can influence composite properties, it has
been difficult to obtain a clear picture of the effect of any one property.
Excellent work has been carried out in the field of paper science regarding fibre length and
papermaking properties (Watson, 1961; Page, 1969; Seth, 1990). Apart from improving
sheet formation, which indirectly affects many sheet properties, reducing fibre length has
littie direct influence on the structural and optical properties of the sheet. The major effect
of decreased fibre length is on the strength properties; most are severely reduced. Thus,
longer fibres will benefit all tensile properties, tearing resistance and folding endurance,
particularly of weakly bonded sheets or sheets wet stiength. However, fibre length can be
less important if the sheets are well bonded, because the failure in the sheet can become
controlled by the strength of the fibres.
Various methods have been used for achieving fractionation of fibre length and these
methods are designed to isolate the effect of fibre length from that of other morphological
and chemical factors. The usual method has been to separate a pulp into fractions of
different fibre length by screening. A more direct approach was used by Brown (1932) in
110
which handsheets, formed with pulps from which the fines had been removed by
screening, were cut mto narrow stiips by a sharp knife. The shortened fibres so obtained
were reformed into handsheets; these had lower strength properties than the original
handsheets.
Another method by which the effects of fibre length variation may be investigated is to
prepare pulp from individual growth rings. This has been done by Watson for P.radiata
(growth rings 2 to 12) (1952) and for P.taeda (growth rings 2 to 11, each separated into
late and early wood) (1954). This procedure gave fibres of P.radiata ranging from 1.6
mm (growth rings 2) to 3.1 mm (growth ring 11), and for P.taeda from 2.2 mm (growth
ring 2, late and early wood) to 3.2 mm (growth ring 11, late wood). An increase in fibre
length produced an hnprovement in tearing strength but it was difficult to draw any
definite conclusion regarding other strength properties. Variation in cell wall thickness as
is found between early wood and late wood of one growth ring, also had a marked
influence on strength properties.
Watson and Dads well (1961) developed another technique by cutting delignified
Araucaria klinkii chips into different lengths prior to pulping. The smaller size chip has
more cutting ends thus its pulp has more short fibres. Again their work found longer fibre
had much better tear stiength than that of short fibre.
Ill

6.1.1 Fibre length fractionation work
The fibre was prepared from unbleached high-tear P.radiata dry lab pulp (APM, Maryvale
mill, AustiaUa). The lab pulp was soaked in water over-night, disintegrated and Valley
beaten to 550 CSF freeness then subjected to fibre length fractionation.
Various fibre length fractions were prepared using the Bauer-McNett screening method,
guillotined handsheets method and Wiley grinding method.
Weighed beaten pulp was first disintegrated on the British disintegrator for 500 counter
revs and diluted to 0.2% concentration. The diluted pulp (about lOg o.d. weight) was
poured into the Bauer-McNett fibre length classifier, equipped with four series screen
compartments, for 20 minutes (10#, 30#, 50# and 80# US standard screen) (see Appendix
A.2.3). Ideally, the apparatus is designed to produce four length fractions, however in this
study only the fibres remaining on screen 10# and 80# were collected and used in the
composites due to the insufficient separation of fractions by the screen technique. This
procedure was repeated few times until the desired amount of pulp was obtained.
Handsheets made from beaten pulp were also used in tiie preparation of cut fibres. Dried
handsheets were cut into 1.0 mm strips with a sharp guillotine. Strips were soaked in
water over night, disintegrated for 1,000 rev. counts in the British disintegrator and used
for composite fabrication.
112
Air-dried beaten pulp was ground in a Wiley mill to fibre length weighted av. 0.3 mm as
the shortest fraction. Thus five fibre length fractions were ready for composite fabrication
(2 from Bauer-McNett, 1 from guillotine, 1 from Wiley mill and 1 original).

The fibre cement composite samples were produced by a slurry / vacuum dewatering and
press technique, followed by air-curing up to 28 days (see section A.3.2). Each sample
was based on a 130g dry weight of ingredients, ordinary Portland cement was used as
matrix.
Mechanical and physical properties of the composites such as flexural strength (MOR),
fracture toughness, void volume, water absorption and density were determined by the
methods described in section A.4. The fibre length fractions were measured on a Kajanni
FS-200 fibre length analyser reported as fibre length weighted average.

6.2.1 Fractionation of fibre length
A single pulp source was separated into 5 length fractions using three methods and these
length fractions are listed in Table 6.1. As mentioned before, the Bauer-McNett apparatus
is designed to produce four length fractions. In our experience, however, it was unwise to
utilise all four fractions due to the fact that diluted pulp fibre has high flexibihty and some
long flexible fibres are able to elude the small opening mesh (screen) barrier so that there
is a considerable amount of overlap in the four fractions length population distributions as
seen in Figure 6.1. Furthermore, fractionation not only separates the pulp into fractions of
different length but also effects a separation between fine and coarse fibres.
113
Table 6.1 Fibre length fractions

Length Fraction (mm)
3.13
2.66
1.59
0.74
0.30
Experiment Technique
Bauer-McNett screening
original pulp length
cutfing dry handsheets
Bauer-McNett screening
Wiley mill grinding
Fig. 6.1. Fibre length population of Bauer-McNett technique four length fraction.
Using a guillotme to cut handsheets can reduced fibre length to 1.59 mm from its original
2.66 mm. Attempts have been made to further reduce fibre length by employing a second
cut to the cut fibre formed handsheets, however, it was not successful as the second cut
only had very small amount of reduction in length compared to that from the first cut (e.g.
first cut: 1.59mm and second cut 1.50mm).
Mean fibre length may not provide the best measure of effective fibre length in both paper
sheets and composite products. The best measurement would be the fibre length
distribution. It is quite possible two different population distribution pulps would have
114
same mean arithmetic fibre length ( or even length weighted average). The observation
used in paper science suggests that of two pulps with the same mean fibre length, tiie one
with the greater proportion of long fibres will give tiie higher tearing resistance (Watson,
1961 and CoUey, 1973). The influence of length population distribution to composite
properties requires further investigation.
6.2.2 Influence of fibre length on composite mechanical properties

Table 6.2 Relationship between fibre length and air-cured composite performance.
MOR (MPa)
Frac. Tough(kJ/m2)
Void Vol (%)
Water Abs.(%)
Density (g/cm')
18.0+5.1
18.63.3
25.9 2.3
27.4+2.7
30.4+3.3
0.33+0.03
1.070.34
1.680.28
2.390.40
2.890.6I
33.30.7
33.61.0
33.00.9
34.60.4
35.00.8
18.40.6
19.70.6
20.31.0
22.30.4
23.30.8
1.810.02
1.700.03
1.630.03
1.550.01
1.500.02
L; 2.66 mm**
2
4
6
8
10
18.8+1.1
20.41.9
23.4+2.3
24.31.9
28.1+3.1
0.38+0.10
0,870.12
1.35+0.24
1.870.19
2.80+0.45
29.2+0.4
30.50.3
32.60.9
33.41.2
34.61.0
15.64+0.3
17.46+0.3
19.84+0.8
21.57+0.8
22.47+1.1
1.860.02
1.750.02
1.65+0.03
1.550.05
1.540.03
L: 1.59 mm**
2
4
6
8
10
16.2+1.1
18.I+2.I
20.91.8
22.3+0.7
22.62.3
0.22+0.02
0.590.07
0.89+0.11
1.370.17
1.680.28
29.51.0
29.40.8
30.60.6
32.4+0.6
34.3+0.8
15.90.7
16.80.6
18.10.3
20.20.6
21.9+0.6
1.860.02
I.780.02
1.700.05
1.600.02
1.570.06
L: 0.74 mm***
2
4
6
8
13.30.7
17.6+3.7
22.4+0.9
23.6+2.5
0.240.03
0.580.19
0.910.10
1.230.27
32.9+0.5
34.20.4
35.40.7
36.70.6
18.3+0.4
20.50.4
22.50.8
23.10.6
1.800.01
1.670.01
1.580.02
1.550.02
L: 0.30 mm**
2
4
6
8
13.6+1.0
1430.5
14.1+0.7
14.3+1.3
0.070.01
0.09+0.01
0.130.02
0.180.01
28.90.6
29.91.0
31.80.8
33.30.5
15.20.4
15.50.7
18.40.6
20.50.6
1.90+0.01
1.82+0.02
1.720.02
1.620.02
Fibre (w%)
L: 3.13 mm**
2
4
6
8
10
* Pulps of fibre length weighted average 3.13 mm and 0.74 mm were generated from Bauer-McNett classifier screens; pulp of fibre
length weighted average 1.59 mm was prepared from guillotined dry handsheets and pulp of fibre length weighted average 0.30 mm
was made from Wiley mill. The initial pulp for this study was Freeness 550 CSF, length weighted average 2.66 mm and kappa number
around 30 P. radiata Kraft pulp.
** 3 standard deviation, sample size n=9. *** 3 standard deviation, sample size n=6.
The results of effective of fibre length to air-cured composite mechanical and physical
properties are shown in Table 6.2 and Figure 6.2 to 6.6. Fibre lengths between 0.3 mm to
115
3.13 mm were used at weight content between 2 per cent to 10 per cent and the
comparison of properties was made at 28 days.
35
30
25
a.
'2 15
X
u.
10
10
12
Fibre Contort (% by mass)
Fig. 6.2. Effect of fibre content on composite flexural strength for different fibre lengths.
Fig. 6.2 shows the effect of fibre content on flexural sU-ength for different fibre lengths.
Except the very short length fraction (L: 0.3mm), the higher strength of the composites
with longer fibre fractions can be attributed to the increased length efficiency factor (see
section 2.1.1). The longer fibre fraction had better strength property. The length
contribution however was not that significant to the strengtii improvement. For example
at 8% fibre content, witii length fibres neariy doubled (1.59 mm and 3.13 mm) the
composites strength values increased from 22.3 MPa to 27.4 Mpa, respectively.
The porosity of the composite increases as tiie fibre content increases and fibre length
increases (see Table 6.2). This effect may be the cause of the reduction in flexural
stiength of some of tiie composites at fibre content greater than 8 to 10 per cent by weight.
116
30
Q.
25
5
JZ
D>
C
(1)
^
^^^'"^^^^
^-^"^''^
20
j
15
lex
W
"S
u.
10
^
^""^
^^,_
X
/
LL
5
0
'
0.5
- -
1.6
'
,,
2.5
,,
3.5
Fig. 6.3. Influence of fibre length on composite flexural strength at 8% fibre content
Again Figure 6.3 shows composite strength increased with fibre length increased at 8 %
fibre by mass, although tiiere were some small strength variations to composites made
from fibre length 1.59 mm and 0.74 mm. This might be explained by the fact that mean
fibre length may not be the best measure of effective fibre length. The 1.59 mm length
fraction might have a greater percentage of fines than the 0.74 mm fraction (due to
different fractionation techniques) and such fines do not contribute to composite strength
development ( see results of the 0.3 mm fraction). Density results in Table 6.2 and Figure
6.7 seen to support the above argument as composites made from the 1.59 mm fraction are
denser than those from the 0.74 mm fraction.
Fibre length contributes greatly to composite fracture toughness development. Over the
whole range of fibre contents, the longer fibres have higher values of composite toughness
(Fig. 6.4 and 6.5). The length contribution to the fracture toughness is much greater than
that to the flexural strength.
117
10
12
Fibre Content (% by mass)
Fig. 6.4. Effect of fibre content on composite fracture toughness for different fibre lengths.
0.5
1.5
2.5
3.5
Fig. 6.5. Influence of fibre length to fracture toughness at 8% fibre content.
The mechanism that takes place when a fibre reinforced composite is loaded to failure
include fibre fracture and fibre pull-out. If the fibre is strong or fibre-matrix bond is weak,
the pull-out mechanism can have considerable influence on the value of fracture
toughness. The longer the fibre the greater the energy required to pull the fibre through
118
the matrix, after the fibre to matrix bond is broken and the greater the contribution to the
dissipation of energy contained in the advancing crack.
The same proposition was developed by the histitute of Paper Chemistry (Staff of IPC,
1944) on the studies of tiie influence of fibre strength to paper tearing resistance. The
tearing resistance is the sum of the work done in breaking some fibres and the work done
in pulling the remainder out of the fibre mat. The maximum tear is obtained at a degree of
interfibre bonding such that the greatest number of fibres required a force just short of
their breaking load to pull them free, while the level of the maximum depended on fibre
length as the parameter controlling the distance over which that force is applied.
Studies of synthetic fibres such as steel, polyproylene, glass and more recentiy kevlar and
carbon reinforced cement based composites led to the same conclusion that composites
reinforced with long fibre have better strength and toughness properties. The area under a
3-point bend load-deflection curve is often described as a measure of the toughness or
energy absorbing capability of the material. Various values for toughness can be
calculated for the same curve depending upon whether the complete load-deflection curve
is used including the descending portion or whether a cut-off point is chosen. From the
point of view of serviceability of a structural unit a more meaningful value for toughness
can be obtained from the area up to the maximum load or up to a specified deflection
depending on the degree of cracking allowed m service. Hannant (1978) in his book Fibre
Cement and Fibre Concretes reported that Johnston (1975) surveyed the data for the
complete load-deflection curve and also for the area up to the maximum stress (Fig. 6.6).
It can be seen that the ability of a composite unit to absorb energy is substantial and
119
increases with fibre content and fibre length (or with length and width aspect ratio), even if
the cut off point is taken at the maximum stress.
Area under complete

load-deflection curve
Area under load-deflectioo

curve up to maximum stress
200
400
600
800
100O
WL/d
Fig. 6.6. Influence offibreWL / d on composite fracture toughness. Where W =fibrecontent by weight x
100, L =fibrelength, d =fibrediameter (Hannant, 1978).
The interesting finding in this work is that very short fibre (fragments) say, with length
less than 0.3 mm, act more as a filler-diluent then as a reinforcing fibre when used to make
composite materials. So a greater mass of pulp is needed to have sufficient number of
long fibres in the high proportion short fibre pulp reinforced composites such as bamboo
fibre reinforced cement and waste paper fibre reinforced products (Coutts, 1994a and
1989).
Due to the variation of wood fibre diameter (early wood and late wood) and the lack of
fibre diameter (coarseness) values no attempt is made to determine fibre aspect ratio
influence. Further, no definite critical fibre length value has been concluded in this study
because there is a wide length gap between 0.3 mm and 0.74 mm. Further work is needed
120
to resolve tiiese unanswered questions, however, the general picture of the effect of fibre
length is obvious.
6.2.3 Theoretical and experimental conflict in results

We have discussed early in section 2.1 that the composite strength can be calculated from
the equation (2.4) when the fibre content is less than 8% by mass.
a,^ = [a/p] cT,.iv,.,., + 0.82(aT)vf(l/d)
(2.4)
Where the a, fi, % (J,h are constants which can be determined from experiments
(Andonian, 1979). Furthermore, the fibre diameter d is suggested to be constant for a
certain type of fibre. Thus at a constant fibre content v/ and v.,,, the composite strength
increases lineally with increasing fibre length [equation (6.1)].
a, = Kj(l) + K2
(6.1)
Where Ki and K2 are constants.
However, the experimental results from this study do not support the theoretical equation
(6.1). A simple regression analysis of the data presented m Table 6.2 leads to a relation
(6.2).
Gh = 3.3(1) + 16.9 (at 8% fibre content)
(6.2)
The relationship shown in equation (6.2) is statistically unacceptable suice the statistical
multiple coefficient (R^) is only about 0.66 which is very low compared to the theoretical
estimation of approximately 0.99.
121
The non-conformability of experimental results with the underlying theory could be

attributed to the theoretical assumptions made. The fibres are pulled out of the cement
matrix instead of broken at failure. And, the fibres have uniform diameter, the fibrematrix interfacial bond strength is constant. The natural fibre, however, violates these
assumptions.
The fracture mechanics concept as discussed in Chapter 2 can not be used to correlate the
experimental results obtained in this study. If LEFM has to be valid, the stress intensity
factor, K;; in general and the fractural toughness, Ki,-in particular have to be defined in
terms of crack geometry and crack location which can not be ascertained for the kind of
composites used in this study, ie., the Ki,, is difficult to establish based on standard Kic test
procedure.
6.2.4 Influence of fibre length on composite physical properties

The changing of fibre length and proportions of the constituent fibres and matrix affect
void volume, density and water absorption (Table 6.2). Short fibre (0.3 mm) results in
close packing of ingredients, thus the composites are more dense than those made with
long fibre as showed on Figure 6.7. In general, long fibres cause formation and running
abihty problems m paper making and have "baUing up" and drainage problems in cement
products. However, the fact these problems did not appear might be due to greater control
of the sample preparation at the laboratory level. There were some density variation for
composites made from fibre length 0.74 mm and 1.59 mm fraction. This variation had
caused composite strength inconsistency, tiie reason for this could be explained by the
fibre length population distribution.
122
Water absorption and void volume change witii density changes in keeping with the eariy
results. If the composite is more dense, the amount of porosity in the composite and hence
water absorption and void volume values are lower.
1.7
1.4
0.3
0.74
1.59
2.66
3.13
Fibre Length Wsightad Av. (mm)
Fig. 6.7. Influence of fibre length on composite density.
6.3 Conclusions
A single P.radiata Kraft pulp has been divided into 5 different length fractions varying
from 0.3 mm to 3.13 mm. Fibre length makes a major contribution to tiie mechanical and
physical properties of air-cured cement composites. Both composite flexural strength and
fracture toughness values increase with increasing fibre length over the range of fibre
contents examined. The fracture toughness can be significantiy improved with long fibres.
The very short fibres (fragments) say, with length less than 0.3 mm, act more as a fillerdiluent then as a reinforcing material when used to make composite materials.
123
Composites have lower water absorption, lower void volume and higher density value
with decreasing fibre length. The conclusion concerning the effect of fibre length for
natural fibres is in general agreement with the findings of the effect of fibre length for
synthetic fibre in cement products although the composite strength doesn't increase
lineally with fibre length increases.
124
Chapter Seven:
Influence of Fibre Strength on Composites Properties
In section 2.1.1 we discussed the "rule of mixture" with respect to composite strength
which suggests that stronger fibre and higher fibre loadings lead to greater composite
strength. The influence of fibre content has been studied by a number of people and the
results suggest an optimum fibre content of about 8 - 10% by weight for softwood fibre
(Coutts, 1979a and AU Patent 1981). As well as fibre content, there is a need to study
natural fibre strength as a parameter affecting composite performance.
Lhoneux and Avella (1992) plotted a wide range of wood pulp fibre tensile strengths
against the fibre-cement composites flexural strengths. A nearly linear relationship was
observed and they suggested the stronger fibres provided better reinforcement than the
weaker fibre. The results seemed in agreement with the theory of mixture rule; however,
their work ignored the influence of other fibre factors which might also have impact on
composite strength. For example, they ignored the influence of fibre length between the
softwood and hardwood; ignored the influence of fibre chemical composition between the
chemical pulp and mechanical pulp fibres; ignored the influence of fibre fibril angle and
cross-dimension between different species.
The objective of this chapter is to investigate the influence of fibre strength on fibrecement composite properties. Therefore, the experiment is designed to vary the fibre
strength while maintaining constant as many fibre parameters as possible.
125

7.1.1. Fibre strength fractionation
As discussed before the key point of the experunent is to fractionate fibre intrinsic strength
without changing other parameters (eg. fibre length and freeness). This could be achieved
by means of cellulose degradation. Wood cellulose fibre can be reduced in strength by
alkaline, oxidative, hydrolytic and thermal methods or combinations of these conditions
which result in degradation reactions of cellulose. There are two methods used in this
study to prepare degraded fibre, namely extensive alkaline cooking and liquid phase acid
hydrolysis.
7.1.1.1 Extensive ahcahne cooking method

The fibre used in this method was holocellulose fibre, which was prepared from Kraft
pulped P. radiata (-2.6% lignin content) by chlorite dehgninification method (see
Appendix A. 1.4). After dehgnification, the fibres were subjected to further alkaline
(Kraft) cooking to produce fibres of various strengths. The details of cooking conditions
and fibre strength, length and freeness values are reported in Table 7.1.
7.1.1.2 Liquid phase acid hydrolysis method.

Liquid phase acid hydrolysis method was apphed to P.radiata Kraft pulp (-2.6% lignin).
Each lOOg of pulp was slushed m 7 litre water. 1 litre hydrochloric acid (HCl), lOM was
added with stirring. The fibre was attacked by the acid for various periods of time, such as
50h, 70h, 140h and 528h. After each desired period, the pulp was filtered and washed free
of acid.
126
7.1.2 Fibre quality evaluation

Fibre strength was evaluated as Zero-span wet (unbonded) tenstie index (ZSTI) and the
degree of polymerisation (DP). ZSTI was conducted on Pulmac Zero-span Tensile Tester.
It is a standard measurement used in pulp and paper science to evaluate pulp fibre tensile
strength (see Appendix A.2.5.2). Holocellulose fibre polymers are subjected to
degradation during further alkahne cooking process. This degradation results in fibre
strength loss and can be evaluated by means of the degree of polymerisation (DP). DP
was determined using 0.5M cupriethylenediamine as a solvent and a capillary viscometer
in this study (see Appendix A.2.5.1).
The fibre weighted average length and freeness value was measured on Kajaani FS-200
Fibre Length Analyser and Canadian Standard Freeness Tester, respectively(see Appendix
A.2.3 and A.2.2). Fibre length and freeness are two important parameters associated with
reinforced composites (see Chapter 6; Coutts, 1982). It was desirable to maintain the
constant length and freeness while the strength was altered by chemical treatment.
7.1.3 Composite fabrication and evaluation

Composite specimen were produced by a slurry / vacuum dewatering and pressing
technique with 8% (by mass) sttength modified fibres as the method described in
Appendix A.3. Both air-cured and autoclaved samples were made. In order to minimise
the matrix and curing effect, composites reinforced with different strength fibres were
cured at the same time to eliminate variations in the curing regime.
127
Mechanical and physical properties of composites such as flexural strength (MOR),

fracture toughness, void volume, water absorption and density were determined by the
methods laid down in Appendix A.4.
Specimens after fracture were straight pulled into parts and stored in the oven at 60C over
night to release moisture. The fracture surfaces were examined with a Phtiip XL 30 FEG
scanning electron microscope.

7.2.1 Fibre strength variation work
The properties of the fibres after chemical treatment are shown in Table 7.1 and Figures
7.1, 7.2 and 7.3. It can be seen from these results that the extensive cooking successfully
reduced the holocellulose fibre strength to various values. With more intensive treatment
the cellulose DP was reduced with reduction in viscosity from 16.27 mPs.s to 6.5 mPs.s;
while the strength dropped from ZSTI 100 Nm/g to 44.0 Nm/g. In addition, there was a
strong relationship between fibre zero-span tensile strength and viscosity [degree of
polymerisation (DP)] as shown in Fig. 7.1, which was in general agreement with the
findings of other workers (Gumagul, 1992). The results also showed that such a Kraft
cooking did not significantly change the other parameters of the fibres, such as length and
freeness (Figure 7.2 and 7.3). It was noticed that the length of the weak fibre (ZSTI 44
Nm/g, 2.18 mm) was only shghtly shorter than that of the strong fibre (ZSTI 100 Nm/g,
2.48 mm). However this small length variation would not be expected to have a great
influence on the composite properties (see Chapter 6)
128
Table 7.1 Properties of P.radiata fibre after alkahne cooking

Further cooking condition
ZSTI (Nm/g)
holocellulose untreated
E.A13% 30 mins to 140C
E.A26% 30 mins to K C C
E.A26% 60min to.lOmin on 180C
20
40
100
89.1
81.2
44.0
Viscosity (mPa.s)
16.27
12.96
13.66
6.50
60
W. Length (mm)
2.48
2.32
2.11
2.18
80
Freeness (CSF)
680
680
700
700
100
120
Fibre Zero-span Tensile Index (Nm g-1)
Fig. 7.1. Relationship between fibre zero-span tensile strength and pulp viscosity
3.5
20
40
60
80
100
120
Fig. 7.2. Relationship between fibre strength and fibre length
129
900
400
20
40
60
80
100
120
Fig. 7.3. Relationship betweenfibrestrength and pulp freeness
Holocellulose fibre was treated with extensive alkaline cooking to produce pulps of
various stiength values. There were two reasons to use holocellulose fibre. Firstiy these
hgnin-free fibres allowed further treatment to various fibre strengths without the
complication of products with different lignin levels, which would affect the fibre strength
determinations and composite fabrication. Secondly holocellulose fibre was found easier
to subject to further alkaline degradation than any other mechanical or chemical pulped
fibre. This may be related to the larger quantity of effective alkah (on a sample basis) in
the holocellulose cooking, or to the influence of lignin on the pulping process or to the
consumption of alkali (McKenzie, 1985).
Due to limited pulp source, the properties of pulp treated with hydrochloric acid were not
determined. However, the evidence of fibre stiength loss caused by acid attack was
observed by a number of researchers (McKenzie, 1985 and Gumagul, 1992). Gumagul
(1992) suggested that there were two types of degradation processes which control the
extent of the strength loss. Homogeneous and random degradation causes littie strength
130
loss, but locahsed degradation weakens fibre significantiy. The above two treahnents
would possibly cause localised degradation resulting in a significant loss of fibre strength.
7.2.2 Influence of fibre strength on composites properties

7.2.2.1 Alkaline treated holocellulose fibre
Table 7.2 and Figure 7.4, 7.5 shows the variation of composite flexural strength when
reinforced with fibres possessing a wide range of strength values. It is surprising that the
composites flexural strength did not decrease significantly as fibre strength reduced from
ZSTI 100 Nm/g to ZSTI 44 Nm/g in both air-curd and autoclaved WFRC specunen. The
strength of the specimens reinforced with very weak fibre (ZSTI 44 Nm/g) started to drop
off, but at the same time that point the fibre length was somewhat shorter which could
have an effect on the strength. This observation was inconsistent with the rule of mixture
of composites strength development.
Table 7.2 Composites properties reinforced with alkaline treated holocellulose fibres
ZSTI (Nm/g)
Air-cured WFRC
100
89.1
81.2
44.0
Autoclaved WFRC
100
89.1
81.2
44.0
MOR (MPa)
Tough (kJ/m2)
Void Vol (%)
Water Abs.(%)
Density (g/cm-*)
25.1+1.3
24.02.2
25.611.3
21.41.5
2.40+0.39
1.900.28
1.420.26
0.9010.14
32.9+1.0
33.310.2
32.310.9
31.7+0.5
20.711.1
21.110.1
19.911.0
19.210.6
1.5910.03
1.5810.01
1.62+0.03
1.6510.02
19.9+2.9
21.4+1.7
21.1+1.3
18.61.2
1.6210.51
1.4610.44
0.9210.19
0.5110.12
42.510.8
41.910.5
41.8+1.2
42.110.8
30.911.2
30.4+0.7
30.311.5
30.511.0
1.3710.03
1.3810.02
1.3810.03
1.38+0.02
131
3b r
(Q
0.
30
25
x:
O)
c
m
w.
1-1
20
*-
15
re
3
X
V
LL
D Autoclaved WFRC
10
Air-cured WFRC
20
40
60
100
80
120
Fig. 7.4. Relationship between fibre strength (0-span) and composite strength
35
?a. 30
25
_a_a
o> 2 0
ID
1 10
X
LL
D Autoclaved WFRC
Air-cured WFRC
10
12
14
16
18
20
Viscosity (mPs.s)
Fig. 7.5. Relationship between fibre strength (viscosity) and composite strength
The composite fracture toughness values decreased greatiy when the reinforcing pulp
contained weak fibres (Figure 1.6 and 7.7). At afibreloading of 8% by mass and fibre
strengtii decreasing from ZSTI 100 Nm/g to ZSTI 44 Nm/g, the composite fracture
toughness values were reduced from 2.40 kJ/m^ to 0.90 kJ/ for air-cured samples (about
62% reduction) and from 1.62 kJ/m^ to 0.51 kJ/m^ (about 69% reduction) for autoclaved
samples, respectively. This result suggests that the weak fibres are not suitable as a
132
reinforcement for cement products, due to the composites lack of essential fracture
toughness, even when having acceptable strength values.
2.8
CM
I
2.4
o Autoclaved WFRC
2
(/>
0)
.
O)
3
O
1.6 I-
Air-cured WFRC
1.2
0.8
U
TO
w
U.
-e-
0.4
0
20
40
60
100
80
120
Fibre Zero-span Tensile Index (Nm g-11
Fig. 7.6. Relationship between fibre strength (0-span) and composite fracture toughness
2.8
E 2.4
-)
Autoclaved WFRC
v>
Air-cured WFRC
g 1,6
t
1.2
0.4 \
o
Ifl) 0 . 8
10
12
14
16
18
20
Viscosity (mPs.s)
Fig. 7.7. Relationship between fibre strength (viscosity) and composite fracture toughness
The mechanisms that take place when a cement based fibre composite is loaded to failure
include fibre fracture and fibre pull-out. The fracture toughness value is determined by a
combination of these two mechanisms. If the fibre is weak, then the energy required to
fracture the fibre is less than that to break the fibre-matrix bonds plus pulling the fibre out
133
of the matrix. Thus the weak fibre is likely to break before being pulled out and the
composite is brittle.
The above explanation of fracture toughness was confkmed by examination of the

composite fracture surface using scanning electron microscopy. When composites were
reinforced with strong fibres, fibres were both fractured and pulled out durmg the fracture
(Figure 7.8) and the specimen had an irregular fracture surface and high fracture toughness
values resulted. Altematively, composites reinforced with weak fibres showed a fracture
surface that was flat, most fibres broke at the crack front and the crack could continue
right through the matrix. The composite showed brittle behaviour and had low values of
fracture toughness (Figure 7.9). In this study other factors such as fibre-matrix bond,
chemical composition and fibre morphology were all constant, so the difference of
composite properties were mainly caused by the difference of reinforcing fibre strength.
Fig. 7.8. Fracture surface of composite reinforced with strongfibreshowsfibrepull-out.
134
Fig. 7.9. Fracture surface of composite reinforced with weakfibreshowsfibrefracture.
By contrast it is difficult to explain the behaviour of composite strength. It could be

suggested, however, that for the high modulus cement based composites even a low
percentage of low modulus fibres has an initial crack-stopping effect, but for further
strength development a strong fibre and good fibre-matrix bond is required. In the current
study the fibre strength (even ZSTI 100 Nm/g fibre) or fibre-matrix bonding might not be
sufficient enough to produce further strength increasing. Thus the composites showed an
initial strength which was similar.
The constant density, water absorption and void volume of composites were expected
from the same matrix, fibre content and morphology. The slightly higher value of density
and lower value of void volume and water absorption for the composite reinforced with
weak fibre maybe due to the conformability which was caused by heavy degradation.
135
7.2.2.2 Liquid phase hydrochloric acid (HCl) treated fibre

The study of composites reinforced with hydrochloric acid treated fibre provided the same
conclusion that weak fibres have littie influence on the composite strength but can
significantly reduce tiie composite toughness (see Table 7.3 and Figure 7.10, 7.11). In this
instance, cellulose degradation was not as easy as that for the holocellulose. The results
showed that both composite stiength and toughness values were not greatiy affected
although the fibre had been attacked by acid for more than 140 hours. But after 528 hours
treatment, composite strength and toughness were reduced from 22.3 MPa to 19.1 MPa
and from 0.79 kJ/mHo 0.39 kJ/m', loss about 14% and 51%), respectively.
Extensive discussion in the next chapter will show that the same source of fibre pulped to
different Kappanumber (lignin level) can cause composites to have various toughness
values (see Chapter 8). This study will help us to better understand the mechanics of the
composite fracture, because different Kappa number pulp has been strongly associated
with its fibre strength (Page, 1985).
Table 7.3 Composites properties reinforced with acid hydrolysis fibres

Fibres
MOR (MPa)
Tough (kJ/m2)
Void Vol (%)
Water Abs.(%)
Density (g/cmr)
Un-treated
50 hours treated
70 hours treated
140 hours treated
528 hours treated
22.311.8
21.7+2.3
20.811.6
20.911.6
19.1+1.4
0.7910.08
0.77+0.09
0.8710.10
0.77+0.13
0.39+0.03
30.611.2
31.610.6
28.9+0.7
28.710.9
26.1+1.3
42.5+1.0
42.710.5
40.210.6
40.211.2
37.811.2
1.3910.02
1.3610.06
1.3910.01
1.40+0.02
1.4510.03
136
100
200
300
400
500
600
Acid Hydrolysis Time (hour)
Fig. 7.10. Relationship between fibre acid treated time and composite strength
1 ..
1
o
'
"^^~^~"^^-^
1 0.6 -
3
Q
1 0.4 -
0.2
100
200
300
400
500
600
Acid Hydrolysis Time (hour)
Fig. 7.11. Relationship between fibre acid treated time and composite fracture toughness
Coutts studied composites reinforced with NZ flax fibre(Coutts, 1983c). The composites
showed similar flexural strengtii values, but, lower fracture toughness values compared to
137
the composite reinforced with softwood fibre (Coutts, 1984a). This was possible due to
the fact tiie strength of NZ flax Kraft pulp fibre was weaker than that of softwood fibre.
In addition, Stevens (1992) reported that autoclaving cycle had a significant effect on
composite toughness. In his work the first autoclaving cycle produced specimens which
were quite tough. However, after ten autoclavuig cycles the samples were exhemely
brittie. The results could be explained by the fact that cellulose degraded and the fibre
strength reduced gradually after each autoclaving cycle. These weakened fibres were not
able to provide the composite with sufficient fracmre toughness.
7.3 Conclusions
Holocellulose and Kraft wood pulp fibres were chemically treated to various fibre tensile
strength without changing other parameters. The results suggested that the fibre strength
was particularly important to composite toughness but not as important to the strength. In
both air-cured and autoclaved cases, when fibre strength was reduced to about 40%, the
composites had a toughness loss more than 60%, but the strength was virtually unchanged.
Examination of SEM micrographs, of the composite fracture surfaces, showed that the
samples containing the weaker fibre had produced the expected higher population of
broken fibres than samples contaming the stronger fibres.
138
Chapter Eight:
Influence of Fibre Lignin Content on Composites Properties
Durmg the course of this stiidy, we found that pulps with different Kappa numbers (made
from the same wood chip source) could result in a significant variation of composite
properties, particularly values of fracture toughness. This finding suggested further
investigation of the influence of lignm content (Kappa number) on the composite
properties was warranted.
Along with cellulose and hemicelluloses, lignin is another major chemical component of
plants. It is present to the extent of about 24 to 32% in softwood, lower amount in nonwood plants and very little in cotton and bast fibres (see section 3.3). The influence of
lignin content on pulp and paper properties has been studied by several researchers
(Gieriz, 1961; Page, 1985), despite the extremely complex nature of hgnin chemistry. In
general, in a non-degrading pulping process, fibre strength improves with decreasing
lignin content due to the higher cellulose content. Paper formed with these strong fibres
has better mechanical properties. However, further reduction in lignin content such as
pulping to a yield corresponding to an a-cellulose content higher than 80% or bleaching
tends to reduce fibre strength, apparently due to the elitnination of this stress-equalising
matrix hgnin or due to cellulose degradation. Furthermore, because of the hydrophobic
character of lignin, it binds the fibrils firmly together (fibres become stiffer) even after
wetting. Lignin also decreases the interfiber bonding in paper.
139
Information regarding the influence of hgnin content on composites is very hmited. Mai
and co-workers (1983) studied low hgnin content bleached pulp and Kraft pulp autoclaved
composites. The composites were reinforced with commercial Wisakraft bleached pulp
and Kinleith Kraft pulp and fabricated by the Hatschek process. They found that bleached
pulp composites had higher flexural stiength but lower fracture toughness values than
those of unbleached Kraft pulp.
Coutts (1986) studied high lignin content TMP and CTMP pulps prepared from P. radiata
chips as reinforcement for autoclaved and air-cured composites (fabricated by slurry /
vacuum de-watering technique). Both pulps were unsatisfactory, for use as a fibre
reinforcement alternative to Kraft pulp in the autoclaved cement mortars, but some gave
acceptable air-cured products. It was suggested that during autoclaving the extracted
chemicals (eg. lignin) formed could cause an inhibiting effect on the curing of the cement
and the deposit of such products could effect fibre-matrix bonding.
The fibre-cement industry has already committed itself to the use of Kraft softwood pulp
as the asbestos alternative source in Australia. The level of pulping yield (lignin content)
is not identified, but Kraft pulp at kappa number around 25 (lignin content about 3.7%) is
well suited for autoclaved WFRC products.
Base on the above data, it was considered appropriate to study the relationship between
fibre lignin content and fibre cement composite properties.
140

This study was designed to investigate the effect that different hgnin containing pulps had
on reinforced fibre cement composite properties. The pulps studied contained various
levels of Hgnin and included a thermomechanical pulp (TMP), a chemithermomechanical
pulp (CTMP), three Kraft pulps and an oxygen delignification pulp. The pulps were
prepared from selected P.radiata chips (APM, Maryvale mill, Austraha) in the CSIRO,
Division of Forest Products pulping laboratory. The detailed pulping techniques have
been discussed in Appendix A. 1. While, a brief description of these techniques is listed in
Table 8.1.
Table 8.1 Pulping techniques

Pulp
Pulping Technique
presteam chips 1 -2 min at 120- 125C, followed by Asplund defibrillator 2-3 min,
then refined in Bauer refiner
CTMP
presoak chips in 10% casustic soda over 18 hours, then processed as in the case
of TMP pulp
Kraft (Kappa No.44.56) E.A. 11.51% (NaOH), 4:1 liquor to wood ratio, 2 h to 170C and 2 h at 170C in
air-bath (3L pulping vessels)
Bleached
Kraft pulp (E.A.16.63%) mixed with 1% MgC03, 2% NaOH at 10% consistency
cooked with O2 (780kpa), 75 min to 115C and 30 min at 115C
TMP
All pulps were tightly beaten in a Valley beater at condition of each 360g (o.d.) with 23
htre water without load for 20 min then with 1.5 kg bed-plate load for another 20 min.
The beating effect was evaluated by means of Canadian Standard Freeness tester (see
Appendix A.2.2).
141
8.1.2 Fibre lignin content and other properties evaluation

The pulps prepared from above techniques cover a wide range of lignin contents and there
is no single lignin evaluation method available. For high yield pulp such as TMP and
CTMP pulps, Klason Hgnin method is more suitable, whUe for Kraft and oxygen bleached
pulps. Kappa number test is normatiy applied. The percentage of lignin present is then
estimated as 0.147 X Kappa number value. Because softwood contains only a small
percentage of soluble lignin, there was no attempt to correct for soluble hgnin in Klason
Hgnin measurement. The detailed lignin content evaluation methods should be refereed to
in Appendix A.2.1.
Other fibre properties such as strength, length and freeness values are measured as wet
zero-span tensile index, Kajanni length weighted average and Canada Freeness
respectively. The testing methods are described in Appendix A.2.
8.1.3 Composite fabrication and characterisation

Air-cured and autoclaved composites were fabricated and characterised by the methods
described in Appendix A.3 and A.4.

8.2.1 Fibre lignin content and other properties
Table 8.2 Fibre Hgnin content, freeness and fibre strength, fibre length
Pulp
Lignin (%)/metiiod
Lengtii (mm)
TMP
25.52/Klason
2.17
CTMP
26.04/Klason
2.43
Kraft 44.56*
6.47/Kappa
2.53
Kraft 27.49
4.04/Kappa
2.60
Kraft 17.33
2.55/Kappa
2.36
Bleached
0.69 / Kappa
2^29
* The number after Kraft is its Kappa number
Freeness (CSF)
700
730
730
700
670
670
Strengtii as ZSTI
low
low
109
113
101
98
142
The variation of pulp Hgnin content, freeness, fibre strength and fibre length are
summarised in Table 8.2. As mentioned before due to the softwood only containing a
small percentage of soluble lignin, there was no attempt to correct for such soluble Hgnin
in Klason lignin measurement. There was an inconsistency between TMP and CTMP
Hgnin content as tiie values of 25.52 % for TMP and 26.04 % for CTMP could be
expected to be reversed. Klason lignin measures insoluble lignin on pulp base, CTMP
pulp could show a higher lignin content than that in the same amount of TMP pulp due to
pretreatment of caustic soda extracting other soluble materials from the fibres. The
principle of mechanical pulping is to mechanically separate wood into its constituent
fibres. Mild chemical treatment during CTMP pulping would not cause much lignin loss,
thus TMP and CTMP could be expected to have similar Hgnm content as that of the wood.
However, the difference in other properties of these pulps would have some influence on
the final composite performance.
Although it would be desirable to maintain a constant length, so as to limit the variables

when studying Hgnin effect, some variation was recorded. It would not be expected to
have a major influence (see Chapter 6).
In addition, the difference of fibre strength could also have some effect on composite
performance particulariy the fracture toughness as studied ui Chapter 8. The pulp with
Kappa number around 27 had the highest strength index of the six pulps. The number of
un-degraded cellulose fibres in the pulp makes the major contribution to the strength
development of the pulp. Besides cellulose content, mechanical pulps (TMP and CTMP)
and high Kappa number pulp (Kraft 44) contain a high percentage of lignin and
143
hemicellulose. There are less ceUulose fibres present in the high yield pulp compared to
the low yield pulp (e.g. Kraft 27), thus a lower strength mdex is expected. The decrease
in strength of tiie pulps with the Kappa number below 27 such as Kraft 17.33 and bleached
pulp could be due to degradation of cellulose from strong alkah cooking or bleaching.
The high freeness value suggested that 20 min, with low bed-plate load. Valley beating
would probably only improve fibre conformabihty rather than fibriUating fibre surface.
Although the TMP and CTMP pulps contained high fines proportion, they stiU showed
relative high freeness values because of the stiff nature of the fibres.
8.2.2 Influence of lignin content on air-cured composites

The influence of fibre hgnin content on air-cured composite flexural strength, fracture
toughness and density are dehneated in Figure 8.1, 8.2 and 8.3, respectively. The detailed
results is also listed in Table 8.3.
Figure 8.1 shows the flexural strength variation as fibre content was increased from 1% to
12% by mass for six different lignin content fibre-cement composites. The strength of the
composites start to decrease for fibre content over 10 % due to poor fibre distribution
throughout the matrix material. This observation was in general agreement with the
change in flexural strength observed with early studies in WFRC and other NFRC
(Coutts,1985, 1987b, 1994b).
144

^ " ^
.rSS^^"^"
X TMP
^*-
^ CTMP
O K44
K27
- K17
+ Bleached
'
10
12
14
Fibre Content |% by mass)
Fig. 8.1. Infiuence of fibre lignin content on air-cured WFRC flexural strength at different fibre content
10
12
Fig.8.2. Influence of fibre lignin content on air-cured WFRC fractiire toughness at different fibre content
145
Table 8.3 Influence of fibre Hgnin content on air-cured WFRC.

Fibre (w%)
TMP, 25.52%
2
4
6
8
10
12
MOR (MPa)
Frac, Tough(kJ/m2)
Void Vol (%)
Water Abs,(%)
Density (g/cm')
16.41.4
17.42.0
19.712.5
23.62.5
23.62.6
23.33.3
0,15+0,02
0.35+0,07
0.5210,06
0,7710,13
1,0410,17
1,2110,24
28,7+0,8
30,5+1,4
30.911,2
32.0+1.6
33.1+1,1
35,111.8
15,410,5
17,811,1
18,911,1
20,511,4
22.2+0,8
24,711,6
1.86+0,02
1,7210,03
1,64+0,03
1,56+0.04
1.49+0,02
1,4210,02
CTMP, 26.04%
2
4
6
8
10
12
15.111.7
15.912.2
20.312.7
21.2+3.4
23.314.6
20.413.5
0.2110,05
0,4310,07
0,7910,16
1,0310,29
1.3410.34
1,2910,26
29,510,6
30,611,1
33,110,8
35,1+0,5
37,511,5
38,411,1
16,010,4
17,710,9
20.610,8
23,610.6
26,711,4
28,916,6
1,84+0,02
1,73+0,03
1,6110.03
1,49+0,02
1,40+0,02
1,3410,07
K44, 6.47%
2
4
6
8
10
12
16.011.6
21.211.4
22.712.3
22.612.8
24.412,2
25.111.2
0,4810,06
0,9210,12
1.4510,20
1,7810,28
2.4310,67
2,4810,09
28,410,4
29,110,7
31,5+0,7
32,511,1
35,0+0,8
35.510.8
15,410,4
16,610,7
19,411,0
20,611,0
23.211.0
24,010,8
1,84+0,03
1,76+0,02
1,63+0,05
1.58+0,03
1,5110,03
1,48+0,02
K27, 4.04%
2
4
6
8
10
12
18.811.1
20.411.9
23.412.3
24.311,9
28,113.5
26,515,4
0,3810,10
0,8710,12
1,35+0,24
1,8710,19
2,8010.45
2,7710.67
29,210,4
30.510,3
32,6+0.9
33,411,2
34611,0
34,811,0
15,6+0,3
17,510,3
19,810,8
21.610,8
22,511,1
22,910,7
1,8610,02
1,75+0,02
!,65+0.03
1,5510,05
1,54+0,03
1,52+0.02
K17, 2.55%
2
4
6
8
10
12
15.310.9
19,810.9
22,612.9
22,612.5
24.912,3
27,211,4
0,3410,03
0,7010.10
0,8910,26
1,3810,09
1,5110,15
2,0810.10
29.1+1.0
30,210,7
32,010,7
33,2+1,1
34,111.0
34.7+0,4
15,910,7
17,510,7
19,510,7
20,811,1
22.010.7
22,910.4
1.8410,03
1.73+0,03
1,64+0,03
1,59+0,04
1.55+0.03
1,51+0,01
Bleached, 0.69%
2
4
6
8
10
12
14,211.1
20.811,5
22.9+2,5
25,014,1
26,612,4
27,114,2
0,4510,09
0,9910,20
1,5710,29
1,9710,32
2.5110,32
2,4910,46
29,7+3,3
30,410,8
30,010,8
32,9+0,8
34,410,9
34,7+0,3
16,112,7
17.1+0,7
19,110,7
20,110,8
22.111.0
22,310,4
1,86+0,09
1,78+0,02
1,68+0,08
1,63+0,03
1,5610,03
1,55+0,01
*A11 composite were fabricated using ordinary portiand cement, air-cured for 28 days, tested at 50+5 per cent RH and 232C,
Of the six different pulp reinforced composites studied, the sample pulp which contamed
pulp with a Kappa number about 27 had highest strength value. This could be due to the
fact that Kraft 27 pulp contained the strongest fibre as measured by zero-span tensile index
(see Table 8.2). If the standard deviations are considered, one would not expect a
significant strength variation amongtiiepulps at the same fibre content. The small amount
146
of variation would be caused more by different fibre length rather than by different Hgnin
content. The influence of Hgnm content will be seen more clearly when discussing
composite fracture toughness property.
The fracture toughness values of the composites mcreased with mcreasing fibre content in
all cases. However, the rate of increase and the toughness values obtained varied greatiy
between pulps as shown in Figure 8.2. Again Kraft pulp at Kappa number 27 had the
highest toughness value and a high rate of increase. This could be due to the pulp having
the strongest and longest fibres in addition to good flexibility for mixing with the cement
matrix. TMP and CTMP pulp had similar high values for lignin content, low strength
index and high fines content, one would not expect composites reinforced with these pulps
to have good fracture toughness. The slightly longer CTMP pulp composite had better
fracture toughness than that of TMP pulp. The unsatisfactory fracture toughness value of
Kraft 17 WFRC might be attributed to the smaller and weaker reinforcing fibres compared
to those in Kraft 27. Air-cured Kraft 27 WFRC had better toughness at high fibre contents
(say > 8%), although not much difference was observed between Kraft 44 and Kraft 27
WFRC at low fibre contents.
Bleached pulp composites even though they contain relatively short and weak fibres had
very competitive fracture toughness with that of Kraft 44 and 27 composites. This could
be due to bleached pulp fibre being more flexible and forming strong bonds to the cement
matrix. The high density values of WFRC reinforced witii bleached pulp supports this
statement. Fracture energy is the combination of the fibre-matrix bond breaking, fibre
fracture and fibre pull out. If the fibre is strong enough to withstand a pulling force, then a
well bonded composite should have good fracture toughness property. It was discussed in
147
Chapter 7, that fibre strength had great impact on composite toughness and that the
toughness decreases rapidly with reduction in fibre strength. In that case, the fibre was not
strong enough to withstand pulling force and the fibre failure was the controlling factor
during specimen fracture. The properties of air-cured WFRC reinforced with bleached
pulp suggested that the oxygen bleaching to Hgnin content down to about 0.7% (in
addition to wet zero-span tensile index about 100 Nm/g) was still appropriate for air-cured
products.
1.8
^^
1.6
X TMP
1.4
-----^^^^
;; CTMP
~^^^S^^
^^~~~~'~^--_
! ^
"""**
o K44
1.2
K27
K17
Bleached
10
12
0.8
14
Rbre Content (% by mass)
Fig. 8.3. Influence offibrelignin content on air-cured WFRC density at differentfibrecontent

The variation of composite density, water absorption and void volume was expected from
the flexibtiity natural of tiie pulps. The stiffness or flexibihty of the fibre mcreases with
the removal of the Hgnin, with flexibility gradually improving from TMP, CTMP, high
yield Kraft pulp, low yield Kraft pulp tiirough to bleached pulp. Fibres with a high degree
of flexibility are easy to form in the composite and provide good interfacial bonds with the
cement matiix. Such composites have high density and low water absorption and void
volume values as seen in Table 8.3 and Figure 8.3.
148
8.2.3 Influence of lignin content on autoclaved composites

Table 8.4 and Figure 8.4, 8.5, 8.6 shows the vanation of autoclaved WFRC mechanical
and physical properties influenced by fibre Hgnm content.
It can be seen from Table 8.4 and Figure 8.4 that the flexural strengtii of autoclaved
composites reinforced with low yield Kraft pulp and bleached pulp mcreased with fibre
content increasing which was in agreement with that of the air-cured products. However,
the strength of WFRC reinforced with high yield TMP pulp, at all fibre contents showed
littie improvement. CTMP pulp composite at above 8% fibre by mass content showed
some small strength development.
30
X TMP
25
-X CTMP
o K44
^z^-
^ 20
5
-1-
'
K27
_-2-
K17
i 15
'"C^
Bleached
^ ^ ^
r '
X
X
I 10
-
10
12
14
Fibre Content (% by moss)
Fig. 8.4. Influence of fibre lignin content on autoclaved WFRC flexural strength at different fibre content
149
Table 8.4 hifluence of fibre Hgnm content on autoclaved WFRC

Fibre (w%)
TMP, 25,52%
2
4
6
8
10
12
MOR (MPa)
Frac. Tough(]cJ/m2)
Void Vol (%)
Water Abs,(%)
Density (g/cm')
12,511.8
11,711.9
11.2+2,1
13,0+1,6
13,0+2,5
11,812.3
0,0710.01
0,1710,01
0,3310,05
0,4610,05
0,7310,16
0,7710,16
36,9+0,9
39,111,0
41,611,4
42,111.1
46,311,2
48,111.3
23,310,8
26,611,2
30,4+1,9
31,911,4
39,111.9
42,912,2
1,59+0,02
1,4710,03
1,3710.05
1,3210,03
1.1910,03
1,12+0,03
CTMP, 26,03%
2
4
6
8
10
12
12,0+1,4
11,210,8
13,210,7
13,912,3
16,912.8
16,212,3
0,1310,04
0.4610.07
0,6610,09
0,7910,18
0,8910,20
1,06+0,18
39,111,0
41,a+0,6
42,910,6
44,910,6
46,510,7
47,8+0,4
25,011,2
28,710,8
32,210,9
35,511,0
39,4+0,4
41,611,1
1,56+0,03
1,4310,02
1.3310,02
1.2710,02
1,1810,06
1,1410,02
K44, 6,47%
2
4
6
8
10
12
13,710.8
16,210.9
18,711,7
18,611,6
17,811,2
19,212,0
0,13+0,02
0.45+0.10
1,0910,29
2,1810,20
2,6210,21
2.6510,67
38,110,8
40.010,4
40,611,2
41,711,2
44,111,1
44,911,4
23,810,7
26,610,6
28,111,4
30,411.4
33,1+1.3
34,811.9
1,6010,02
1.5010.02
1,4410,03
1,3810,03
1,3410,02
1,2910,03
K27, 4,04%
2
4
6
8
10
12
12,110,7
17,512,0
18,011,4
19,4+1,9
20,111,4
19,010,9
0.0710,02
0.3810,05
0,6410,09
2,3610,41
2.8910,55
2.8210,12
39,311,3
41,510,7
42,710.8
43,210,5
44,411,1
44,310,6
25,211.3
28,010.8
30,511.0
32.010,9
34,311.4
34,910.9
1,5610,03
1,4810,02
1,4010,02
1,3510,02
1,3010,02
1,2710,02
K17, 2,55%
2
4
6
8
10
12
15,311,1
16,312,8
20,511,0
20,711.5
20,211,1
20,911,2
0.1510,02
0.3210,06
0.6810,03
0.8810.02
1.3810.02
1.8410,01
39,211.2
41,211,2
41,610,8
41.811.7
42,911.3
43,412.1
24.8+1.0
27,711.3
29,011.0
30,211.8
32.4+1.6
34.511,9
1,5810,02
1,4910,03
1,4410,02
1,3910,03
1,3210,03
1,2610.04
Bleached, 0,69%
2
4
6
8
10
12
14,110,5
17,311,9
19,812,2
21,712,8
21,212,4
20,411,1
0.1210,07
0,4410,06
0.5510.09
0.8110,27
1.7110.47
2.5310.27
39,1+0,6
41,511.1
41,011.0
41,311.9
44,711.2
44,4+0.6
24,710.7
27.711.3
28.211.1
29.412.0
34.111,7
34.010.9
1,5810,02
1,50+0,03
1.46+0.02
1,4110.03
1.3110,03
1,3010,01
*The composites were fabricated using ordinary Portland cement and silica at the ratio of 1:1, autoclaved at 1,25MPa
steam pressure for 7,5h, tested at 505 per cent RH and 222C, 3 standard deviation, sample size n=9.
It was discussed in section 3.3 and 3.4 that natural plant fibre (wood) consists of
considerable amount of Hgnin and this Hgnin can be dissolved by some chemical solutions
and the high temperature can accelerate the reaction rate. This is the principle of chemical
pulping. Composites subjected to autoclaving could easily dissolve Hgnin due to the high
150
curmg temperature (above 175C) and severe alkahne cement environment. This
dissolved Hgnin would eitiier poison the matrix curing or deposit on the surface of the
fibre causing poor interface bonding, or the combination of these two. The mechanism of
dissolving Hgnin is still not fully understood.
3.5
X TiMP
X CTMP
O K44
2.5
o
c
- K17
K27
/^^
+ BIssched
K 1.5
I
I
>^^^
><_^
0,5
'.f^
10
12
14
Fig. 8.5 Influence offibrelignin content on autoclaved WFRC fracture toughness at differentfibrecontent
Autoclaved composites reinforced with fibres containing low lignin appear to be more
stable than those with higher lignin content. Thus autoclved WFRC reinforced with
bleached pulp and Kraft 17 showed high flexural stiength as seen in Table 8.4 and Mai's
work (1983). Autoclaved TMP fibre reinforced composites had the lowest flexural
strength values due to the effect of Hgnin dissolving / depositing. The same observation
was reported by Coutts (1986), he suggested that high yield TMP and CTMP pulps were
unsatisfactory for use as a altemative fibre reinforcement to P.radiata Kraft pulp in
autoclaved cement mortars, but acceptable for use in air-cured products.
151
The effect of lignin content on the property of fracture toughness could be understood in
terms of fibre strength. Fibre sttength plays a dominant role to the composite fractiire
toughness as discussed in Chapter 7. Figure 8.2 and 8.5 shows composites reinforced with
strong fibre (Kraft 27) had the highest toughness value.
This remark is strengthened by tiie results of the fibre strength study in Chapter 7. The
fibres in the present study had a variation in strength properties (Table 8.2). In addition to
dissolving Hgnin, autoclaving is akin to a severe alkaline "pulping" and can further
reduces strength values due to degradation of fibre cellulose. Although the degree of such
strength loss is uncertain, the fracture toughness loss of autoclaved composites reinforced
with bleached pulp was significant compared with air-cured products. A few workers
(Lhoneux, 1991; Stevens, 1992) have studied fibre strength loss during autoclaving,
mainly using lime solution to simulate the matrix environment. This study attempts to
investigate the "real pulping" conditions present in the autoclave by means of a
reconstituted handsheet technique. Reinforcing fibre was first formed into handsheets,
then inset into cement blocks subjected to autoclave; after autoclaving, the matrix was
broken apart, the handsheets were carefully removed out and disintegrated into pulp; then
characterised for change of pulp strength and other properties. In our preHminary study it
was observed that tiie fibre strength and lignin content were reduced during autoclaving,
however extensive work is required and results wiU not be reported in this thesis.
The variation in autoclaved composites physical properties were similar to those of aircured cement products. The fibre stiffness or flexibility, which is influenced by the fibre
lignin content, became the key factor in regarding materials density, water absorption and
void volume as listed in Table 8.4 and shown in Figure 8.6.
152
1.8 I1.6
>
I 1.2
a
1
0.8
0.6
10
12
14
Fibre Content (% by mesa)
Fig. 8.6. Infiuence of fibre lignin content on autoclaved WFRC density at different fibre content
8.3 Conclusions
The results showed that both TMP and CTMP pulps did not have satisfactory mechanical
properties when used to reinforce autoclaved cement composites. Air-cured products have
better mechanical properties and could be possible used in a number of apphcations such
as renders, moulded articles or sheet products.
The Hgnin component of the pulp fibre had considerable influence on tiie composites
mechanical and physical performances. Such influences could be understood m terms of
fibre strengtii, fibre stiffness, fibre-matrix interface bonding, fibre number and lignin
dissolving mechanics during the autoclavuig process.
153
Experimental results also showed that composites reinforced with Kraft pulp with Kappa
number around 27 had the best mechanical and physical properties in both air-cured and
autoclaved products.
154
Chapter Nine:
Conclusions and Further Work
9.1 Conclusions
The following conclusions can be drawn from this work:
1. Chemically pulped bamboo fibre is a satisfactory fibre for incorporation into a cement
matrix. Mechanical and physical properties of both air-cured and autoclaved bamboo fibre
cement composites were studied. Bamboo fibre cement products have reasonable flexural
strength and competitive physical properties to their wood fibre cement counterparts.
However, the fracture toughness values are low due to short fibre length and high fines
content of the bamboo pulp. Improved properties of composites reinforced with screened
long bamboo fibre confirm this belief.
There is little difference between the properties of composites reinforced with beaten and
unbeaten bamboo pulp. This could be due to the average short fibre length or the structure
of bamboo fibre when compared with wood fibres.
2. Properties of bamboo fibre cement composites (and the other natural fibre products) can
be improved by blending long fibres (such as softwood fibre) with the bamboo
reinforcement to increase the average fibre length. Such hybrid fibres improve the
composite strength and more importantly the fracture toughness value. The extent of such
improvement is related to the increasing proportion of long fibres.
155
3. The influence of fibre length on composite performance has been extensively studied.
Fibre length plays a significant role in the behaviour of fibre cement composites. If the
fibre is long, a greater amount of fracture energy is needed to puU the fibre through the
matrix and the composite can be tougher and stronger. This work was carried out by
fractionating a single pulp mto a range of lengths. The resuHs showed that when short
fragments of fibre (weighted length 0.30 mm) were used as reinforcement and compared
with longer fibres (weighted length 1.59 mm), a drastic reduction in composite fracture
toughness and strength was observed. For example, at 8% by mass the short fibre samples
showed a seven fold decrease in fracture toughness and possessed only half the flexural
strength of samples containing the same amount of the longer fibre.
4. Research with holocellulose fibre pulp, chemically treated to vary the tensile strength of
the individual fibres, indicated that fibre strength was particularly important to composite
toughness but not to composite strength. In both cases of air-cured and autoclaved fibre
composites, when fibre zero-span tensile index (wet) and pulp viscosity values were
reduced to about 40% of fuH strength, the composites fracture toughness value also were
observed to faU to about 40% of original value. Surprisingly, the flexural strength values
were virtually unchanged. The reason behind this behaviour is not clear. Examination of
SEM micrographs, of the composite fracture surfaces, showed that the samples containing
the weaker fibre had produced the expected higher population of broken fibres than
samples containing the stronger fibres.
5. Air-cured and autoclaved composites remforced with a wide range of Hgnin content
pulps were evaluated. The Hgnin component of the pulp fibre had a great influence on the
composites mechanical and physical performances. Such influences could be understood
156
in terms of fibre strength / fibre stiffness (cellulose content), fibre-matrix mterface

bonding, fibre number and lignin dissolving mechanics during the autoclaving process.
Both TMP and CTMP pulps did not have satisfactory mechanical properties when used to
reinforce autoclaved cement composites. Bleached pulp reinforced composites had high
strength but low toughness values (compare to Kappa number 27 pulp). The composites
reinforced with Kappa number of about 27 kraft pulp had the best mechanical and physical
properties in both air-cured and autoclaved products.
9.2 Recommendations for further work

9.2.1. Pulp supply
The reinforcing potential of kraft fibres depends mainly on the properties of the raw
material used. Although fibre properties can be affected by the cooking conditions - alkali
charge affecting wet fibre flexibility and sulphidity affecting intrmsic fibre strength - these
factors do not compensate for tiie variations caused by the raw material. It is therefore
essential for both the final product quality and production costs that the interactions
between raw material and end-products properties are weU understood and the wood fibre
property variations can be measured.
The basic natural (eg. wood) fibre properties are wall thickness, fibre width, fibre length,
"weak points" in the fibre waH, S^ fibril angle, tiie index of crystalHsation and chemical
composition. These properties vary between different wood species; within one annual
ring (springwood and summerwood); between different parts of the trunk and are due to
the different growing conditions.
157
Modem forestiy has accelerated forest growth, and the trend is to harvest younger trees.
The raw material is also used more efficiently (sawmiU chips, whole tree chips and wood
residue), and the amount of imported chips from plantations seems to be increasing. This
means that quality variations in raw material are also increasing significantiy. If this is not
taken into account in pulping and stock preparation, the reinforcing potential of the raw
material is not utilised in fuU, and at the same time final product quality variations
increase.
Tasman Pulp and Paper Co. Ltd., New Zealand has made progress in matching the highly
variable wood supply to various end-products, i.e. papermaking and fibre-cement, by
means of segregation of tiieir fibre supply into density ranges. However, there is still a
long way to go to understand the true relationship between fibre quality and end-products
properties and to modify the fibre quality to match the products requirement (Williams,
1994).
9.2.2. Fibre length population distribution and cross-dimensions (coarseness)

Among the wood fibre basic properties, fibre length and fibre cross-dimensions, i.e. wall
thickness and fibre width, might be the most important factors regarding the preparation of
paper and fibre composite products. The influence of fibre length has been studied in this
thesis, however, the influence of fibre length population distribution would be more
meaningful to fibre-cement composites. This is because of the nature of non-uniformity of
fibre length, the nature of stock preparation and pulp blending.
The fibre cross-dimensions, i.e. waH thickness and fibre width, has often been
characterised using the coarseness value. Because of the close relationship between the
158
cross-dimensional properties of softwood fibres and fibre length, the variation in paper
(and fibre-cement) properties caused by the ceU wall thickness and fibre width have many
times been erroneously explained in terms of fibre length. The different cross-dimensional
properties of fibres, for example thick-walled, narrow, stiff and strong fibres and thin-wall,
wide, flexible with a large lumen and exceUent bonding abihty fibres, has great impact on
their end-products properties. The influence of fibre cross-dimensional characteristics on
paper properties has been studied by a number of researchers (Clark, 1962; Seth, 1987;
Kibblewhite, 1989), some people even suggested that in practice the cross-dimensions
(coarseness value) was the most identical characteristic to predict the papermaking
potential of different softwood fibres (Paavilainen, 1993).
Very limited work has been done regarding the effect of fibre cross-dimensions to the
fibre-cement products quality. Vinson and Daniel (1990) demonstrated that high density
summerwood fibres (coarser) were better suited for fibre cement reinforcement than low
density fibres. Coutts (1987) in studying natural fibre reinforced cement composites
suggested the importance of fibre dimensions in respect to its length and diameter aspect
ratio. However, more extensive work is required before a clear picture can emerge.
9.2.3 Conformability - flexibility and collapsibility

Fibre conformabihty, namely fibre flexibility and collapsibihty, characterises the abihty of
wet fibres to deform (plastic and elastic deformation). Both fibre flexibility and fibre
collapsibility are dependent on cross-dimensional fibre properties and on the elasticity of
the cell wall. At the same time the fine stiucture, chemical composition and water content
of the cell wall as well as the drying history and the chemical and mechanical treatments to
which the fibres are subjected aU play a role in how the fibre will respond.
159
In the case of chemical pulp fibres, it is widely agreed that collapsibility and flexibility are
important both for paper and composites properties . Furthermore the conformabihty of
the fibre also affects stock preparation and machine run-ability.
Although the importance is widely recognised, knowledge of tiie factors affecting

conformability is limited and not systematic. Also, experimental verification of the
relationship between conformability and end products properties is difficult to find in the
scientific literature (Paavilainen, 1993).
9.2.4 The impact of pulp medium consistency treatment on fibre-cement products

The processing of pulps at medium consistency (MC) has become common in the pulp and
paper industiy. The MC technology, which employs fibre mass concentrations typically in
the range 8-14%, allows considerable reduction in the volume of pulp suspension handled
during processing operations. This has led to reduced equipment size, and saving in
capital and operating costs.
The conditions required to fluidise the suspensions lead to beating, as well as the
introduction of curl and microcompression in the fibres (Seth, 1991). The MC treatment
will result in a higher sheet stietch and lower elastic modulus of the sheet. Sheet porosity
and tear wiU also be enhanced by the increased curl and microcompression (Page, 1985;
Seth, 1993). Although the influence of curl on the fibre cement composites has been
studied by Michell (1990), the implications of MC processing nevertheless needs further
investigation with both bench-scale and commercial equipment.
160
9.2.5 Theoretical modelling

There is Hmited work available on model systems in the Hterature, dealing with fibre
reinforced cement composites. A consideration should be given to the development of
models to predict composite properties based on an understanding of fibre and matrix and
their inter-relationship. Modelling allows the engineer to take into consideration scaling
and size effects. It also has the advantage of being able to indicate the direction for future
work in achieving optimum strength and toughness performance for a given system.
161
Appendix A:
Fabrication and Characterisation Methods
A.l Pulp Fibre Preparation
A.1.1 Chemical pulping (Kraft pulping)
The Kraft process is the predominant pulping process in industry today; it is tolerant to
variations in wood chip dimensions and wood quality and, because of its high strength,
Kraft pulp can be used in a wide range of end products.
The Kraft process involves heating the wood at 165-175C with a solution of sodium
hydroxide (NaOH) and sodium sulphide (Na2S) for 0.5-3 hours. Time to cooking
temperature is in the range 1-2 hours. Pulp yield 45-55%. On an industrial scale, the
spent cooking liquor (called black Hquor) is concenttated and then burnt in a fumace,
allowing the chemicals to be recovered and at the same time providing energy for the pulp
min.
In the laboratory assessment of pulpwoods, pulping conditions (temperature, time) are

chosen to resemble those of industry. Wood sample size are usually constant and
specified in terms of oven dry mass. As the volume of wood contained in each vessel
fluctuates with the density of the wood sample, sample sizes must be chosen so as to
remain within the capacity of the pulping vessels. In the air-batii used in this work six 3Htre pulping vessels are used (Fig. A. 1). Typical conditions for a cook would be two
hours at 170C, 25%o sulphidity and 4:1 liquor to wood ratio.
The alkali charge is varied to achieve the required degree of delignification, usuaUy
described m terms of Kappa number ( a measure of the residual lignin content, see section
A.2.2.1). Unbleached packaging grade pulps are usually delignified to Kappa numbers
around 40, bleachable hardwood pulps are lower, about Kappa 20. Pulps for
reinforcement in the cement composites their Kappa number are in the range of 30 to 20.
162
LOCK IXC
INC
PULPIHC
VESIELS
LOCKIMC
FLINCC
TMtHHOWrClr
fOCKCT
Fig. A.l. Schematic illustration of Air-bath and 3-litre pulping vessels.
The active chemicals in the Kraft pulping liquor are sodium hydroxide (NaOH) and
sodium sulphide (Na^S). The alkali charge is usually expressed in terms of the equivalent
quantity of sodium oxide (NaoO), although that particular chemical entity is never actually
encountered. Three parameters are used to define the chemical make up of a Kraft pulping
liquor: active alkali, effective alkali and sulphidity.
1. The active alkali (abbreviated AA) is NaOH + Na2S, expressed as Na20, and usually as
a percentage relative to the weight of oven-dry (o.d.) wood chips to be cooked.
2. The effective alkali (EA) is defined as NaOH -i- l/2Na2S, expressed as Na20.
3. The sulphidity is the percentage ratio of Na2S to active alkali, expressed as Na20.
For most new wood samples the alkaH charge required to attain the target Kappa number
will not be known. It is usually necessary to find the proper alkah charge by trial. As the
change in Kappa number with alkali charge is frequently non-linear at low Kappa
numbers, experience with similar wood samples is often the only guide. The normal alkali
requirement for softwood pulping is about 12 to 14% effective alkali on o.d. wood [8 to
10%o for hardwoods (Smook, 1982)].
163
The aUcali charge is usuaUy made up from stock solutions of sodium hydroxide and
sodium sulphide. These stock solutions have to be standardised before use because tiie
chemical concentrations change as the chemicals react with carbon dioxide in the
atmosphere.
Another essential piece of information is tiie moisture content of the wood sample. Not
only does the moisture content affect the weight of wood required hi each pulping vessel,
but also the Hquor to wood ratio. Details of the calculations required to determine the
amount of wood, chemicals and water required for a Kraft cook are given in Appendix B.
The 3-litre air bath can be preheated to a temperature well in excess of the final cooking
temperature before the pulping vessels are loaded. The air bath can also be heated after
loading the pulping vessels. It usually takes 1.5-2 hours to reach the cooking temperature
in this case. Although the air bath can be thermostatically controlled when finally at the
cooking temperature, the rise-to-temperature portion of the cycle, especially from about
160C on, is controlled manually, and some degree of judgement is required to effect a
smooth approach to the cooking temperature without significant overshoot. It is suggested
that the maximum temperature up to 180C does not significantly affect the cooking
result. From 180 to 190C, there appears to be a small reduction in yield; above 190C,
the yield and strength loss may be substantial due to attack on ceUulose (Smook, 1982).
Temperature reading are taken at intervals throughout the cook to allow the calculation of
the H-factor. H-factor, first developed by Vroom in 1957, is a means of representmg the
times and temperatures of any cooking cycle as a single numerical value. Its values lies in
aHowing the comparison of pulping procedures which incorporate differing temperature /
time profiles.
At the completion of the cooking cycle, the pulping vessels are promptly removed from
the air bath and cooled in a bath of cold water to halt furtiier deHgnification.
164
After cooling, the spent cooking liquor (black Hquor) is drained off. The softened chips
are given a brief wash in cold water to reduce foaming, then the chips are disintegrated.
The method used for disintegration often depends on the mdustrial process being
modelled. At CSIRO, the softened wood chips from the 3-Htre air bath are disintegrated in
a mixer at 2850 revolutions per minute for 10 minutes.
FoUowing disintegration, the pulp is washed with cold water in vacuum / de-water funnels
to remove residual black liquor. This is a critical step because residual black liquor will
affect the Kappa number and pH of the pulp. The washed pulp is then screened to remove
uncooked fibre bundles. At CSIRO, a Packer screen with 0.2 mm wide slots is used for
this purpose. The screened pulp is then dewatered, usually with a press, to about 15%
consistency, follow with crumbing. The crumbing, which can be done in a large Hobart
mixer, assists in distributing the moisture in the pulp evenly. This allows an accurate
estimate of the moisture content of the pulp from small samples. The moisture content of
the pulp can be measured at this stage. The crumbed pulp is bagged in a sealed plastic bag
and stored in a refrigerator to maintain the moisture content.
A.1.2 Mechanical pulping (TMP, CTMP)

A. 1.2.1 Asplund defibrator
Mechanical pulping in its various forms has been clauned as the pulping method of tiie
future. The principle of mechanical pulping is mechanically separating wood into its
constituent fibres. Chips may first be steamed or pretreated with 6-10% sodium hydroxide
(Na-^+) or calcium hydroxide (Ca+-^) to soften lignin and to make the separation of the fibre
easier.
In this experimental study, thermomechanical pulp (TMP) and chemithermomechanical

pulp (CTMP) are prepared in Asplund Type D Laboratory Defibrator, which is equipped
with facihties for pre-steaming and subsequently disintegrating wood chips under gauge
pressures up to 12 atmospheres (190C) as described by Higgins (1977). The main
variables in TMP (CTMP) are the raw material (species, basic density, chip moisture
165
content and chip geometry), steam pressure (and corresponding temperature), pre-heating
time and defibration time.
The chip charge can be as high as 400g (o.d.) wood, but in practice (e.g. hardwoods) it was
sometimes necessary to run with a charge as low as lOOg to avoid overload on the 7.5 kw
motor. P.radiata low temperature range pulps (125-135C LT-TMP) was reported to yield
well fibrillated fibres with good paper-making properties; high temperature pulps (150170C HT-TMP) to yield smooth, lignin encased, unfibritiated fibres (Higgins, 1978).
To operate the Asplund defibrator, wood chips are placed in the four compartments of the
inner vessel in approximately equal amounts, tighten the lid, check the temperature. After
presteaming the chips for 1-2 minutes at low temperature range (120-125C), start
defibrator for 2-3 minutes (1440 rpm) to break chips down into fibre bundles. These fibre
bundles are fed through Bauer refiner to form individual fibres (see section A. 1.2.2).
For chemithermomechanical pulping (CTMP), chips are soaked in 10% caustic soda
(NaOH) over 18 hours at ambient temperature and then processed as in the case of LTTMP pulps.
A. 1.2.2 Bauer refiner

Refining reduces a fibre "bundle" to individual fibres. The key part of a refuier is the
refining plates. The refining plates tiansmit the mechanical energy into the wood fibres,
so the primary function of plates is to keep the pulp between the plates.
Refining of Asplund defibrated pulp is done in the Bauer refiner fitted with 203 mm (8
inch) diameter plates (rotor and stator). Although there are two types of refining plates
(open periphery and closed periphery plates) (Fig. A.2), Asplund defibrated TMP (CTMP)
fibre bundles can be refined to individual fibres by use of closed periphery plates.
166
jS^^^K^^^
j^^^^g^^^f
lii^'>^^^^>^^^^^^^B^^'^''^^^'^îJ^^^^^^^Hl[^^'^^^^^^^^^^^H
?? r> f o p-
^g5^^Sf'^*tf^*^â
^Ki
ROTOR
^^^^^^
'^^^^^^^^^^g^^^^
H^^c c ^
^**^
Fig. A.2. Open periphery (C) and closed periphery (B) refining plates.
Refining is achieved by successive passes of the pulps at various plate clearance until the
required freeness level is reached. Refining Asplund defibrated P.radiata LT-TMP pulps
in Bauer refiner fitted with closed periphery plates (rotor: 8117-4122p, stator: 81174121p) can be done using the following passes. One pass is made at a plate clearance of 2
mm (0.08 inches), one at 1.25 mm (0.05 inches), one at 0.625 mm (0.025 inches) and two
at 0.125 mm (0.005 inches).
Pulp moisture consistency is a major factor to refining efficiency. Water should be kept to
a minimum, if too much water is used (consistency < 8%o), no work is done on tiie fibres
and therefore no refining takes place. Care should be always taken during reducing the
plates clearance not to clash the plates.
A.1.3 Bleaching pulps (Oxygen delignification)

The main objective of bleaching in papermaking is to convert a dark coloured pulp into
one which is much tighter in colour. Bleached pulp improves fibre cement composite
flexural strength due to fibres are more flexible and better fibre-matiix bonds (Mai, 1983).
Chemical pulps contain residual Hgnin. This lignin has been extensively modified by tiie
severe conditions used to make the pulps and can be quite dark in colour. It is extremely
difficult to change this Hgnin into a colouriess form with bleaching chemicals. So tiie way
to bleach a chemical pulp is to completely remove the residual Hgnin. This is done in
multistage processes using chlorine compounds or oxygen for lignin degradation, and
167
alkali for extraction of the degraded Hgnin. Often mixtures of bleaching chemicals or
sequential addition of different bleaching chemicals is used ui the same bleaching stage.
To simplify the description of these sequences, each chemical can be designated with an
appropriate letter. Table Al Hsts common bleachmg chemicals, their usual designation
and their bleaching action.
Table Al Common bleaching chemicals
Chemical
Chlorine, Ch
Hypochorite, (NaClO or CaClO)
Chlorine dioxide, ClOo
Alkali, NaOH
Oxygen, O2
Hydrogen peroxide H9O9
Designation
C
H
D
E
0
p
Bleaching action
Lignin degradation
Lignin degradation
Lignin degradation
Extraction of degraded lignin
Lignin degradation or improved delignification in E-stages
Improved delignification in E-stages
The choice and the conditions of use of the bleaching chemicals are limited because
carbohydrate degradation must be avoided. Otherwise, pulp yield would be reduced and
strength properties impaired. There are few common laboratory bleaching sequences such
as CEHD and 0D,(E0)D2.
The effect of fibre Hgnin content level on WFRC properties is studied. Low Hgnin content
in the fibres was achieved under oxygen delignification conditions. The pulp responds
weH to oxygen bleaching and about half of the Hgnin can be removed easily. The general
rule for oxygen bleaching, particulariy when applied to softwood Kraft pulps, is that about
40 per cent of the Hgnin can be removed before the strength properties of the pulps are
affected. The viscosity of the oxygen bleached pulps was measured as this property can be
indicative of fibre damage if the value is below a threshold level. A viscosity of about 22
m.Pa.s is regarded as the lower limit for an oxygen-bleached northern hemisphere
softwood Kraft pulp (Teder, 1991).
Oxygen deHgnification of the Kraft pulps was carried out with the 3L pulping vessels
which were fitted with lids incorporating valves to introduce oxygen into the vessels. Pulp
samples (lOOg o.d.) were mixed with magnesium carbonate (1% pulp basis), sodium
168
hydroxide (in the range 0.7-2.5 per cent, pulp basis) and water to give a pulp consistency
of 10 per cent. The mixtures were placed in the pulping vessels which were pressurized
with oxygen (780 kPa) and heated at 115C for 30 min (time to temperature was 75 min).
A.1.4 Holocellulose pulp

HoloceHulose is defined as lignin free, pure cellulose fibres. Holocellulose is used to
study influence of fibre strength on WFRC properties. This Hgnm free fibre allows further
Kraft cooking to various fibre strength without the complication of different lignin content
products, which might effect the fibre strength determinations.
Sodium chlorite solution at room temperature was used for the delignification. The
solution consisted of 60 g sodium chlorite, 20 g anhydrous sodium acetate and 40 ml of
glacial acetic acid, made up to a litre witii purified water. The moist, unbleached Kraft
pulp at about 20 per cent consistency is mixed with sufficient of the chlorite solution to
give about 5 per cent consistency and allowed to react at room temperature with
occasional mixing and shaking for 24 hours. The pulp was then washed and the chlorite
treatment is repeated with fresh solution for another 12 hours. After chlorite treatment the
pulp is washed thoroughly with purified water (Stone, 1960).
A.1.5 Beating
The basic purpose of beatmg is to mechanically condition tiie fibres for papermaking and
manufacturing WFRC. Addition to tiiis, beatmg plays an important role in the Hatschek
process to retain cement and silica particles (Coutts, 1982a). A more general term for
mechanical working of pulp is "refming". The term "beating" actually denotes a specific
type of refining, but is now commonly used to describe refming in the laboratory. Most
laboratory beating methods have a more selective action than mill refiners and produce
results that normaUy cannot be dupHcated in the miU. So it is necessary to optimize the
refining level with mill equipment and conditions.
169
A number of laboratory beating devices are in use around the world. The two devices
most commonly used are the VaUey beater and the PFI miU. The VaHey beater (Fig. A.3)
is essentiaUy a miniature version of the Hollander beater. Although this device has a long
tradition of use, it has some definite limitations and is gradually being displaced by the PFI
mill. The principal disadvantage is the difficulty in obtaining reproducibility with respect
to other Valley beaters and with respect to the same beater over long periods of time. The
problem relates to variable wear patterns on the metal cutting edges. However, a Valley
beater is used in our experiments to prepare pulps for composite fabrication because
sufficient amount of pulp can be treated in a single run. To operate the Valley beater, a
soaked 360g o.d. pulp with 23 litre water is beaten without load for 20 minutes then beaten
for further period of time with the bed-plate load of 5.5 kg until the desired freeness level
is obtained. After beatmg the pulp is subjected to dewatering, pressing and crumbing.
B*atr Roll
Wclghti
Fig. A.3. The Valley Niagara beater
170
Bcdpbfe
Fig. A.4. The PFI laboratory beater tackle
The PFI mill utilizes a grooved roll eccentric to a smooth trough, as illustrated in Fig. A.4.
Both the roll and "bed-plate" rotate at high speed but at different peripheral velocities; this
action induces friction, rubbing and crushing of the fibres to produce the beating effect.
Since there is no metal-to-metal contact and no edges to wear, the device does not require
calibration. It also has the advantage of requiring a relatively small amount of pulp to
carry out a complete evaluation.
A.1.6 Preparation of fibres from dry lap-pulp

In miU practise, fibres are usually prepared in tiie pulp miU and transported in a dry lappulp form. In the fibre-cement plant, the dry lap-pulp is soaked in tiie water to release
hydrogen bonds then disintegrated into pulp. Fibres experienced dry-lap distingentation
process may lose some initial strengtii. However, it is more economic and convenient for
the fibre-cement plant which has no pulping factiity.
In our experimental work, some fibres were prepared from commercial dry lap-pulps or
packaging paper. These lap-pulps or paper were torn into relatively small pieces and
171
soaked in cold water over night. Then they were disintegrated in a mixer at 2850
revolutions per minute for 10 min. Following disintegration, the pulp was dewatered,
crumbled and stored in the refrigerator until composite fabrication. The pulp moisture
content can be calculated at this stage.
A.2 Pulp Fibre Characterisation

A large number of testing methods are in common use to characterize pulps with respect to
quality, process abihty and suitability for various end uses. Many of these test procedures
are empirical in nature and provide useful behaviour information. Other more
"fundamental" tests provide the means to predict behaviour, or to explain and rationalize
the empirical test results. A summary of common test methods is given in Table A2.
Some of these methods were carried out during work for this thesis, the others are
recommended for further study (see Chapter 10).
Table A2 Pulp test methods (Smook, 1989)
Fundamental Properties
* weighted average fibre length
* intiinsic fibre strength
* fibre coarseness
* specific surface area
* wet compactability
* pulp chemical compositions
Empirical Tests
* Kappa number
* CED viscosity
* colour and brightness
* cleanliness
* drainability
* beater evaluation
A.2.1 Lignin content

A.2.1.1 Kappanumber
The non-cellulosic components (especially Hgnin) react readily with acidic permanganate
solution (KMn04). This reaction provides the bases for the Kappa number test. At a
controlled temperature, an excess of acidic permanganate is added to the pulp to be tested,
and allowed to react for a set time interval, after which the unreacted permanganate is
reduced by an excess of iodine. The Hberated iodine is then determined by reaction with
thiosulfate. The Kappa number of the pulp is equal to the number of ml of acidified 0.02
M potassium permanganate solution which would be consumed by one gram of moisture
172
free pulp in 10 min at 25C. The results are corrected to 50% consumption of tiie
permanganate added, which ensures a satisfactory relationship to the Hgnin content of the
pulp.
The test procedures to be foUowed is described ui detail in the TAPPI T236cm-85. This
method may be used for aU types and grades of chemical and semi-chemical unbleached
and semibleached wood pulps in yields under 60 per cent. However, it should be noted
that reproducibility is less for high yield pulps than for low yield pulps. For pulps such as
TMP and CTMP, Klason lignin method wiU be more suitable.
A.2.1.2 Klason lignin

Klason lignin is defined as those components of wood or pulps which are insoluble after
treatment with 72 per cent m / m sulphuric acid followed by boiling in 3 per cent sulphuric
acid. TMP and CTMP pulps are more suited to the Klason lignin method (APPITA PI Is78). In this standard, the lignin content should not be less than 1 per cent to provide a
sufficient amount of lignin, about 20 mg, for accurate weighting. It is not applicable to
bleached pulps containing small amounts of lignin.
Most woods contain some lignin which is rended soluble by the above treatment and
which is not determined by this standard. In softwoods and sulphate pulps this soluble
Hgnin content is small, about 0.2 to 0.5 per cent, but in hardwoods it can amount to 5 per
cent. Thus hardwood which has had any alkali treatment, may give a lower results than
would be obtained from the untreated wood.
A.2.2 Drainability (Freeness)

The resistance of fibres to the flow of water is an important property with respect to pulp
processing, papermaking and fibre-cement composite materials fabrication. The classical
method of determining this property in North America and Australia is by means of the
Canadian Standard Freeness (CSF) test (Fig A5). The CSF is defined as tiie number of ml
173
of water collected from the side orifice of the standard tester when pulp drains through a
perforated plate at 0.30% consistency and 20C.
Measurement of pulp drainage are know as freeness, slowness, wetness or drain time,
according to the instrument or method used. If a pulp drains rapidly, it is said to be "free"
If it drains slowly, it is said to be "slow". Freeness and slowness scales have an inverse
relationship. The Schopper-Riegler Slowness test is the principal drainage measurement
used in Europe and Asia.
lACKINCfLATC
CHAMBED
OlfCR BRACKET
SC"ECN
PLTE
LOWER BRACKET
SPREAOtB
COME
fUNNEl
1/
^ g
OLUG
SIDE
ORiricE
~"
( T J ^
/Si/
\
W
fl
BOTTOM
ORiricE
CUD
Fig. A.5. Canada Freeness tester
Freeness measurements are widely used as an indication of quality for mechanical pulps
and as a measure of the degree of refining (beating) for chemical pulps. Studies have
shown tiiat thefinesfraction (-200 mesh) is primarily responsible for changes in drainage.
The removal of the fines fraction from beaten pulps can restore the original drainabiHty,
while the pulp retains its beaten strength properties. Thesefindingsare sometimes used as
an argument against the use of drainage measurements as an index of pulp quality (Smook,
1982)
174
Although freeness measurements provide a basis for comparing similar pulps, the test does
not simulate what happens on the paper machine wire. For example, groundwood pulp
gives a lower freeness than highly beaten chemical pulp but shows faster drainage on the
paper machine. Furthermore, the same freeness value does not indicate the same degree of
fibritiation. Bamboo pulp might have same freeness value as a softwood pulp, but the
degree of fibrillation for the bamboo pulp would be much less than that for wood pulp due
to the fact that the bamboo pulp has or average short fibre length, massive pitch fines and
sensitive response to the beating force.
A.2.3 Fibre length

Fibre length is measured or indicated either by microscopic examination of a
representative number of fibres or by screen classification of a sample into different length
fractions. In the microscopic method, a known weight of fibres is projected onto a grid
pattern; all the fibres are measured and the average fibre length is calculated
mathematically.
In the classification method, a dilute dispersion of fibres is made to flow at high velocity
parallel to screen slots, while a much slower velocity passes through the slots. In this way,
the fibres are presented lengthwise to a series of screen with successively smaller mesh
openings, and only the fibres short enough not to bridge the opening pass in to the next
chamber. The Bauer-McNett Classifier is one of the traditional instruments (Fig. A.6).
In the last few years, a new optical device, the Kajanni FS-200 Fibre Length Analyzer has
become available for measuring fibre length. It is now widely used and is becoming
accepted as a standard laboratory fibre length measurement (TAPPI T271pm-91). The FS200 measures fibre length by an optical technique using polarized light and is based on the
birefringence of the wood fibres (Fig. A.7). The machine employs a measurement range
of 0 - 7.2 mm, divided into 144 classes, each of which represents a 0.05 mm interval in
length. When the pulp sample (average number of fibres 15,000 to 30,000) passes through
175
the analyzer, the number of fibres m each classification is counted. This data is fed to a
microprocessor unit which routinely records, calculates and displays the fibre length
average in tiiree modes: arithmetic, length-weighted and weight-weighted.
Constant Water Level

Supply Tank
Water Inlet
Drain Pluq
Water Overflow
Overflow Water
to Drain
Overflow Water
to Drain
'Terylene Cloth
(Muslin)
Fig. A.6. The Bauer-McNett Classifier
^^^
Capillary
Stirrer
fd
Optics
I ! I] U
\
L H
X\iw
ih
Detector
Laser
Beaker
Fig. A.7. Kajanni FS-200 measurement principle
176
There is no singulariy accepted definition of what is the average fibre length of pulp in a
sample. The generally agreed approach is to the definition best suited to the nature of the
sample under consideration.
If the sample is made up of fibres which are of fakly uniform length, the numerical or
arithmetic, average fibre length L^ is most appHcable (1). However, this definite is not
applicable when, as is most frequently the case, the sample contains a high proportion of
short fibres or fines. The use of a measurement which is weighted according to the
weights of the fibres is then preferred. The length-weighted average fibre length Li.^, is
defined for the case where the fibre coarseness, the weight per unit length,/, is assumed to
be constant (e.g. /j = C ) (2); the weight-weighted average fibre length, L.^^,-,^ is defined for
the case where the fibre coarseness is assumed to be proportional to the fibre length (e.g./;
= C-k) (3).
L, = Sl/N = I n J./Snj
(1)
h^ = Iwjl;/Iwi = I(nil:C)Vl(niljC) = Inil^/Injl

L^,, = I w J J l W i = I(nM)li/I(nili^)= lXCnd^^/l(Cn^^)=
(2)
lAd^/lA^^
(3)
Where 1 represents the average length of fibres in the i* fraction, n^ represents the number
of fibres in tiie i''^ fraction and Wj represents the weight of fibres in tiie i^" fraction.
According to Clark (1985), the coarseness of natural fibres increases with fibre length and
he therefore favour the use of the weight-weighted form. However, the use of the lengthweighted average fibre length is preferred by Kajanni researchers who consider it to give a
better prediction of the paper making potential of the fibres. Accordmgly, in the current
study the length-weighted average fibre length is employed.
A.2.4 Handsheet preparation

Many papermaking tests require more than one sheet of paper, so it is convenient to
prepare, under identical conditions, several paper sheets from each pulp sample. These
paper sheets, called handsheets, must be as uniform as possible. Two sheetmaking
177
systems are refered to in the ISO standards, conventional sheet former and the RapidKothem former. The method to make handsheets on the conventional sheet former is
introduced below.
The first step in preparing pulp for handsheets is to decide what beating points will be
used. Since beating changes the properties of pulp and paper, it is usual to make a set of
handsheets from pulp beaten to at least four different levels, one unbeaten and three at
progressively higher beating revolutions. This aUows curves to be drawn which can be
used to interpolate paper properties.
Having decided on the amount of beating, the next step is to disperse the specified quantity
of pulp thoroughly in water. This is usually achieved with a disintegrator of specified
design [ISO 5263-1979(E)].
The amount of pulp needed is determined by the grammage (mass per unit area) of the
handsheets and the number of sheets to be made. Typical handsheet grammages are
60.03.0 grams per square metre (calculated on an oven dry basis) for ISO standards
5269/1-1979 (E) using a conventional sheet former.
After disintegration, the pulp is subjected to beating and measurement of the freeness. The
pulp is diluted with water to a stock concentration suitable for the preparation of
handsheets at tiie selected grammage. This is convenientiy done using a stock divider,
which provides continuous agitation of the pulp suspension to maintain stock uniformity.
An appropriate volume of stock is taken from the stock divider to make the first
handsheet. The oven dried weight of this first handsheet can be used to adjust the amount
of water in the stock divider so tiiat later sheets have the desired grammage. The details of
the method used for sheet preparation should refer to ISO 5269/1-1979 (E).
After tiie handsheets are prepared, they are dried under conditions designed to prevent
shrinkage. In the conventional sheetmaking equipment, this is achieved by attaching the
178
wet sheet to a rigid drying plate using a press, then allowing the sheet to air-dry in contact
with the plate.
A.2.5 Fibre strength

The strength of a fibre is fundamentally attributable to the fibril angle (the angle at which
the cellulose molecules spiral about the fibre axis) and the absence of "weak spots" in tiie
fibre due to excessive breaks m the ceUulose molecular chain. Such cleavage is predominantiy a consequence of the pulping and bleaching processes employed to isolate and
purify the pulp fibre from the wood structure where it originates (see section 3.5.2). Suice
breaks in the molecular chain will reduce the average chain length, it foUows that the
viscosity test will monitor the extent of molecular chain breakage.
A standard viscosity test is conventionally used to measure pulp strength. Technically, it

measures no such thing, when carefully done, it provides a measure of the average chain
length of the cellulose molecules obtained by dissolving the pulp sample in a suitable
solvent, such as cupriethylenediamine solvent.
A zero-span tensile test on a sheet of paper measures the average strength of the fibres
which are carrying the tensile load when failure occurs. It thus depends on both the
number of fibres and their average strength and is a basic measure of fibre quality.
Comparison of fibre quality, particular in a given mill environment, are generally made at
close to the same fibre conditions. When this is the case, the zero-span tensile test is a
comparative measure of fibre strength.
It is generally understood that the loss of fibre strength at high a-cellulose contents results
from the degradation of cellulose. Homogeneous degradation is random, causes little
strength loss and can be indicated by either zero-span tensile strength or pulp viscosity
values. However, localized degradation in strength loss is unlikely to be recoverable and
measurements of pulp viscosity may be misleading (Gumagul, 1992). Thus care should
always be taken when interpreting the strength results.
179
A.2.5.1 Viscosity of pulp (captilary viscometer method)

TAPPI T230 om-89 viscosity of pulp method describes a procedure for determining the
viscosity of 0.5% cellulose solution, using 0.5M cupriethylenediamine as a solvent and a
capillary viscometer.
To measure pulp viscosity the general procedure includes:

(1) weigh 0.2500 g air-dried, moisture free pulp into dissolving bottle;
(2) add 25.00 ml distilled water and shake to disperse the pulp;
(3) add 25.00 ml cupriethylenediamine solution and purge with nitrogen for 1 min;
(4) cap the bottle and shake for few hours until the fibre is completely dissolved;
(5) conduct viscosity measurement, the viscometer size is selected to give efflux times of
over 100s, but less than 800s, and two specimens' efflux time should within 2s range.
(6) viscosity is calculate as V = Ctd
where V: viscosity of cupriethylenediamine solution at 25.0C, m-Pa-s(cp)
C: viscometer constant found by calibration
t: average efflux time, sec
d: density of the pulp solution, g/cm- (=1.052)
A.2.5.2 Zero-span tensile test

The zero-span breaking length appears to be an index of the breakmg lengtii (or tensile
stiength) of a pulp beaten to its maxhnum value under ideal conditions. Consequentiy, it
is an excellent measure of tiie "maximum strength" of a pulp, and is almost completely
independent of the laboratory beating procedure used. However, zero-span test as a dkect
measure of fibre strength suffers from the fact that changes in fibre length and / or
interfibre bonding influence the test value. A simple way to eliminate the influence of
interfibre bonding is to wet tiie sheet of paper prior to testing (wet zero-span test). Fibre
length influence can be eliminated by adjusting the initial span to zero without any finite
value.
180
In our laboratory, wet zero-span tensile sU-ength is conducted on a Pulmac Zero-span

Tester (Fig. A.8). To do this, cut the handsheet into a 20 mm by 100 mm strip, wet the
strip and clamp between two jaws (at zero span position). Then apply the tensile force to
fracture and calculate zero-span tensile strength (ZSTS, kN / m) and its index (ZSTI, Nm /
g) as foUows:
ZSTS = (P - Po) X C X 0.654
ZSTI = ZSTS X 1000 / Grammage (g/m')
Where: P is instrument reading, PQ is the conversion factor constant for zero opening and
C is the instrument factor, about 0.370.
Fig. A.8. Schematic illustration of Pulmac Zero-span tester.
A.3 Fabrication composite material

It has been reported in Chapter one that ceUulose fibre-cement sheets are commerciaUy
manufactiired on a Hatschek machine. In our smaU scale laboratory work the composite
materials are prepared by "slurry / vacuum dewatering" technique which emulates the
commercial Hatschek process. This technique includes slurrying the fibre-cement
formulation, vacuum dewatering, press and autoclaving (or air-curing).
181
A.3.1 Materials
The matrix was made from fresh commercial grade ordinary Portiand cement (OPC)
(Australian Cement, Geelong, Type A) and finely ground sihca (Steetiy brand lOOWQ).
The matrix grade and particle size might be important to material final properties,
however, such study is not included in this work. Considerable work in tiiis area would be
done by the manufacturing companies themselves.
During the manufacturing process or laboratory work, attention should be paid to storage
of the cement. At all stages up to the time of use, cement must be kept dry so as to prevent
or minimize deterioration from the effects of moisture, atmospheric humidity and
carbonation. In our experimental work, cement was purchased from local building
materials store, then batched and sealed tightiy mto several plastic jars. It is suggested that
for cement which is old than four months should be classified as "aged" and re-tested
before use (Taylor, 1969).
Table A3 Natural fibre reinforced cement composites ingredients

Fibre % by mass
2
4
6
8
10
12
water (ml)
300
350
400
500
500
500
28-day Air-curing
OPC
175C, 7.5h Autoclaving

OPC:Silica 1:1
It is usually necessary to add other raw materials or special additives to the fibre cement
fumish / products, such as flocculating agents (processing aid), PFA (cheap fiUer /
pozzolanic), ball clay (interlaminate bond agent), microsilica (void filler / pozzolanic /
interlaminate bond agent), AH, (reduction moisture movement). Hence the choice of one
or more of these special additives enables the designed properties of the end products to be
obtained. However, no additives were used in this thesis work.
182
The reinforcmg fibres were prepared by the method described in section A. 1. Wood fibre
(P.radiata) was supplied from Australia APM, Maryvale miU as specially selected high
tear wood chips and high tear Kraft pulped unbleached dry lap paper forms. Bamboo fibre
[Sinocalamus affinis (Rendle) McClue] was obtained fiom Kraft pulped unbleached
commercial packaging paper from Chang Jiang Paper MUI and Jian Xi Paper Mill, China.
A.3.2 Slurry / Vacuum dewatering and press technique

Each sample is based on a 130g dry weight of ingredients (Table A3). As the moistiire
content of the pulp is known, thus the appropriate weight of pulp fibre (oven dried base)
can be determined and then suspended in 300 to 500 ml of water (see Table A3) with
stirring for 2 minutes. In some cases, weighed moist crumbled pulp was suspended in
much more water and disintegrated for a total of 12,500 revolutions to ensure fibre
dispersion, then dewatered to required amount. The preweighed matrix ingredients were
mixed thoroughly and then slowly added to the stirred suspension of fibres. Rapid stirring
took place for 5 minutes and the mixture was poured into an evacuable casting box (125
mm by 125 mm) with filter paper and fine wire screen (Fig. A.9) so that it could be
distributed over the screen. An initial vacuum was drawn until the sheet appeared dry on
the surface and then the sample was flattened carefuUy with a tamper. A vacuum of 60
KPa (gauge) was applied for a total of 2 min. The sheet was then removed from the filter
screen. The sheet and screen were stored between two steel plates and the procedure
repeated until a stack of four sheets had been prepared. The stack of sheets was then
pressed for 10 minutes at a pressure of 3.2 MPa. During the pressing extracted water was
blown away by compressed air and the load was applied and / or released slowly to
prevent any damage to the sheets.
183
filter paper / wire

Vacuum dewafc
7777777"
Fig. A.9. Vacuum dewatering casting box
Coutts and Warden (1990) studied the influence of casting pressure and time on the
WFRC mechanical and physical properties for the slurry / vacuum dewatering technique.
They found that increasing the casting pressure and time resulted in improving
composites' mechanical properties and increasing the density of the composites. They
suggested that composites containing 8-10% fibre loadings can usually achieve
satisfactory properties when subjected to pressure for a short application time, whereas
high fibre loadings might need longer press time.
After pressing, the screens were carefully removed from the sheets which were then
stacked flat in a sealed plastic bag for 24 hours prior for curing.
A.3.3 Air-curing and autoclaving

There are few cement curing models such as water curing, air-curing and autoclaving. Aircuring and autoclaving are used for fibre reinforced cement products. Cement composites
are cured under ambient conditions and wiU take 28 days to achieve their "fuU" strength
(samples remained in the plastic bag for 7 days then cured under the ambient conditions).
The air-cured method saves on capital investment in an autoclaving facility and has
superior properties than those of autoclaved products. Air-curing model is suitable for onsite fabrication and for manufacturing polymer fibre reinforced cement products due to
mild curing conditions.
184
In both Australia and Europe, the manufacture of natural fibre reinforced cement
composites have been based on sheet products formed by techniques akin to the Hatschek
process and cured in high-pressure steam autoclaves. Steam at temperatures close to
180C enables the replacement of between 40 to 60 % of ordinary Portland cement by the
less expensive siHca, which can react with the cement to form a calcium siHcate matrix of
acceptable strength (Lea, 1976). The reaction is completed within hours [ 125 psi
(180C), 8 hours] instead of air-curing which takes weeks to achieve fuU strength. So the
storage facilities can be less as tumover is faster.
The conditions for autoclaving were initiaUy studied by Coutts and Warden (1984b).
Composites autoclaved under 180C, 125 psi steam, 8 hours at cement / sihca ratio T.l
demonstrated the best mechanical properties (see Fig. A. 10 and Al 1). At that condition,
fibre was not expected to be subjected to significant degradation while the composite
achieved adequate strength. However, there is a need to optimize the autoclaving
condition with different fibre sources, matrix ratios and fabrication process.
1:1,180*,8h(REF. 8)
1:1,170",4h
12
FIBRE CONTENT (% by mass)
Fig. A.IO. Optimize autoclaving temperature and time
185
33:67
38:62
43:57
50:50
0PC:Si02 RATIO (by mass)
Fig. A. 11. Optimize composite OPC:Silica matrix ratio for autoclaving
A.4 Characterisation of composite materials

Specimens were cut with a diamond saw to specified dimensions and stored under a
controlled atmosphere of 505% relative humidity and 222C for 7 days before testing.
Specimens measuring 125 mm by 40 mm (of varying thickness) were tested for flexural
strength and fracture toughness values. The flexural strength was measured as the
modulus of rupture (MOR) in three-point bending as: 3PL/2bd-, where P is the
maximum load recorded during the test, L is the specimen span, b is the specimen width
and d is the specimen depth. A span of 100 mm and a deflection rate of 0.5 mm/min^ was
used on an Instron testing machine (Model 1114). The results of the tests were obtained
using automatic data coUecting and processing equipment. The fracture energy was
calculated from the area under the load / deflection curve when the faUed specimen
reached 50% maximum load. The fracture toughness is given by the fracture energy
divided by the cross-section area of the specimen. The comparison of fracture energy or
fracture toughness is strictiy only valid for specimens of the similar thickness.
186
Water absorption, density and apparent void volume physical properties were obtained
using the methods laid down in ASTM C948-81, which measures specimen dry weight,
wet weight and amount of water the specimen displaced.
In all cases, at least six specunens were tested for MOR, fracture toughness, density, void
volume and water absorption. 3 standard deviations have been included for aU properties
measured.
187
Appendix B:
Determination of Kraft Pulping Parameters
Consider a wood chip sample of X% moisture content. The requirement is to pulp A gm
equivalent oven dried chips at Y% active alkali and Z%o sulphidity at a liquor to wood ratio
ofR:l.
1. Wt of air dried chips required = A/(100-X)*100 = B gms;

2. Grams of active alkaH required = A;=Y/100 = C gms;
3. At Z% sulfidity, grams sulfide required = C ;<Z/100 = D gms;
4. Vol. of Na2S soln., = D*1000/ [Na2S] = E mis;
where NajS is concentration of Na.S solution in g/1 (as NaÔ);
5. For NaOH in Na.S soln; gms NaOH in Na.S soln = E*[NaOH]71000 = F gms;
where [NaOHJj is concentration of NaOH in Na2S solution in g/1 (as NajO);
6. Grams of NaOH reqd. = C-D-F = G gms;
7. If concentration of NaOH solution is [NaOH];
then vol. NaOH = G* 1000/[NaOH]; = H mis;
8. Total liquid requked = AR = 1 mis;
9. Mis of HjO in wood chips = B-A = J mis;
10. Vol. of to HoO be added = I-(J-i-H+E) = K mis.
Hence charge per bomb is:
1. B gms of air dry chips equivalent to A gms oven dried;
2. H mis of [NaOHJi, of NaOH solution;
3. E mis of NajS solution;
4. K mis of water.
188
References:
Ahn, W.Y. and Moslemi, A.A. (1980) SEM examination of wood-Portland cement
bonds. Wood Set. 13,77-82.
Akers, S.A.S. and Garrett, G.G. (1983a) Observations and predictions of fracture in
asbestos-cement composites. J. Mater. Set. 18, 2209-2214.
Akers, S.A.S. and Garrett, G.G. (1983b) Fibre-matrix interface effects in asbestoscement composites. J. Mater. Set. 18, 2200-2208.
Akers, S.A.S., Crawford, D.,Schultes, K., and Gerneka, D.A. (1989) Micromechanical
studies of fresh and weathered fibre cement composites. Int. J. Cement Comp.
Lightweight Conor. 11, 117-123. (andrefs. therein).
Andonian, R., Mai, Y.M. and CottereU, B. (1979) Strength and fracture properties of
ceUulose fibre reinforced cement composites. Int. J. Cement Comp. Lightweight
Conor. 1, 151-158.
Anon. (1981) New - a wood fibre cement buildmg board. CSIRO Industrial Research
News 146, 1-2.
Arno, J.N., Frankle, W.E. and Sheridan, J.L. (1974) Zeta potential and its application to
filler retention. Tappi 57(12), 97-100.
Assumpcao, R.M.V. (1992) UNIDO's programme in non-wood pulpmg and
papermaking. 2nd Int No-wood Pulping Papermaking Conf Oct. 92, China, 5-19.
Atchison J.E. (1983) Data on non-wood plant fibres. Pulp and Paper Manufacture
Vol.1 Properties of fibrous raw materials and their preparation for pulping.
(Ed. Kocurek, M.J. and Stevens, C.F.B.), TAPPI, U.S.A. 1983.
Australian Patent (1981) No. 515151, issued March 19tii, 1981, James Hardie and Co.
Pty Ltd.
Australian Patent (1982) No. 554185, James Hardie and Co. Pty Ltd.
189
Aveston, J., Cooper, G.A. and KeUy, A. (1971) The properties of fibre composites.
Conf. Proc. National Physical Lab. London, Nov. 1971, P15.
Barnes, B.D., Diamond, S. and Dolch, W.L. (1978) The contact zone between Portiand
cement paste and glass "aggregate" surfaces. Cem. Conor. Res. 8, 233-243.
Bazant, Z.P. (1985) Mechanics of fracture and progressive cracking in concrete structures.
Fracture Mechanics of Concrete: Structural Application and Numerical
Calculation, 1-5.
Bentur, A., Mindess, S. and Diamond, S. (1985a) PuU-out processes m steel fibre
reinforced cement. Int. J. Cement Comp. Lightweight Conor. 7(1), 29-37.
Bentur, A., Diamond, S. and Mindess, S. (1985b) The microsh-ucture of the steel fibrecement interface. / . Mater. Set. 20, 3610-3620.
Bentur, A. and Mindess, S. (1990) Fibre Reinforced Cementitious Composites,
Elsevier Science PubHshers Ltd., 99-109.
Brown, R.B. (1932) Paper Trade J. 95, 145.
CampbeU, M.D. and Coutts, R.S.P. (1980) Wood fibre-reinforced cement composites.
/ . Mater. Set 15, 1962-1970.
Clark, J.d'A. (1962) Effect of fibre coarseness and length. 1. Bulk, burst, tear, fold and
tensile tests. Tappi 45(8), 628-634.
Clark, J.d'A. (1985) Pulp Technology and Treatment for paper. MiUer Freeman,
2nd Ed. Chap. 17.
CoUey, J. (1913)Appita 27(1), 35
Coutts, R.S.P. (1979a) Wood fibre reinforced cement composites. CSIRO Division of
Chemical Technology. Research Review 1979.
Coutts, R.S.P. and CampbeU, M.D. (1979b) CoupHng agents in wood fibre-reinforced
cement composites. Composites 10, 228-232.
190
Coutts, R.S.P. and Ridikas, V. (1982a) Refmed wood fibre cement products. Appita
35(5), 395-400.
Coutts, R.S.P. and Kightiy, P. (1982b) Microstructure of autoclaved refmed wood-fibre
cement mortars. J. Mater. Sci 17, 1801-1806.
Coutts, R.S.P. and MicheU, A.J. (1983a) Wood pulp fibre-cement composites. / . Appl
Poly. Set: Appl. Poly. Symp. 37, 829-844.
Coutts, R.S.P. (1983b) Wood fibre in inorganic matrices. Chem. in Australia
50, 143-148.
Coutts, R.S.P. (1983c) Flax fibres as a reinforcement in cement mortars. Int. J. Cement
Comp. Lightweight Conor. 5(4) 257-262.
Coutts, R.S.P. (1984a) Autoclaved beaten wood fibre reinforced cement composites.
Composite 15(2), 139-143.
Coutts, R.S.P. and Kightiy, P. (1984b) Bonding in wood fibre-cement composites.
J. Mater. Sci. 19, 3355-3359.
Coutts, R.S.P. and Warden, P.G. (1985) Air-cured wood pulp fibre cement composites.
/ . Mater. Sci. Lett. A, 117-119.
Coutts, R.S.P. (1986) High yield wood pulps as reinforcement for cement products.
Appita 39(1), 31-35.
Coutts, R.S.P. (1987a) Eucalyptus wood fibre reinforced cement. / . Mater. Sci. Lett.
6, 955-951.
Coutts, R.S.P. and Warden, P.G (1987b) Air-cured abaca reinforced cement composites.
Int. J. Cement Comp. Lightweight Conor. 9(2), 69-73.
Coutts, R.S.P. and Ward, J.V. (1987c) Microstructure of wood fibre plaster composites.
J. Mater. Sci. Lett. 6, 562-564.
Coutts, R.S.P. (1987d) Fibre-matrix mterface m air-cured, wood pulp, fibre cement
191
composites. / . Mater. Sci. Lett. 6, 140-142.

Coutts, R.S.P. (1987e) SEM study of high yield wood pulps in cement matrices.
unpublished results.
Coutts, R.S.P. (1988) Wood fibre reinforced cement composites. Natural Fibre
Reinforced Cement and Concrete Vol. 5 (Ed. R.N. Swamy), Conor. Tech. &
Design, Blackie, London, 1-62.
Coutts, R.S.P. (1989a) Wastepaper fibres m cement products. Int. J. Cement Comp.
Lightweight Conor. 11(3), 143-147.
Coutts, R.S.P. and Warden, P.G. (1989b) unpubHshed results.
Coutts, R.S.P. and Warden, P.G. (1990a) Effect of compaction on the properties of
air-cured wood fibre reinforced cement. Cement & Cone. Composites. 12, 151-156.
Coutts, R.S.P. (1990b) Banana fibres as reinforcement for building products.
/ . Mater. Sci. Lett. 9, 1235-1236.
Coutts, R.S.P. and Warden, P.G. (1992a) Sisal pulp reinforced cement mortar.
Cement & Cone. Composites. 14,17-21
Coutts, R.S.P.(1992b) From forest to factory to fabrication. Fibre Reinforced Cem &
Conor RILEM 92, p31-47 (eds R.N.Swamy), E & FN Spon, London.
Coutts, R.S.P. and Ni, Y. (1994a) Autoclaved bamboo pulp fibre reinforced cement.
Cement & Cone. Composites, in press
Coutts, R.S.P., Ni, Y., and Tobias, B.C. (1994b) Air-cured bamboo pulp reinforced
cement. / . Mater. Sci. Lett. 13, 283-285.
Cote, W.A. (1980) Papermaking Fibres, a Photomicographic Atlas. Syracuse Univ.
Press, NY.
Crabtree, J.D. (1986) Past, present and future developments in mdustrial fibre cement
technology. Keynote Speech, 2nd Int. RILEM Symp. Sheffield, 15 July.
192
Davies, G.W., CampbeU, M.D. and Coutts, R.S.P. (1981) An SEM study of wood fibre
reinforced cement composites. Holzforschung 35, 201-204.
Fordos, Z. and Tram, B. (1986) Natural fibre as remforcement in cement based
composites. RILEM FRC86 (eds R.N.Swamy, R.L.Wagstaffe and D.R.Oakley),
RILEM Technical Committee 49-TFR, Paper2.9.
Gale, D.M., Shah, A.H. and Balaguru, P. (1990) Oriented polyethylene fibrous pulp
reinforced cement composites. Thin-section Fibre Reinforced Concrete and
Ferrocement. American Concrete Institute Publication SP 124-4, 61-77.
Giertz, H.W. (1961) Effect of pulping processes on fibre properties and paper structure.
Proceedings of Cambridge / Oxford Symposia. Mech. Engineering, London.
Gordon, J.E. (1976) The New Science of Strong Materials - or Why You Don't Fall
Through the Floor. 2nd edn.. Penguin Books, 287pp.
Gordon, J.E. and Jeronimidis, G. (1980) Composites with high work of fracture.
Phil Trans. R. Soc. London Ser. A., 294, 545-550.
Gram, H-E. (1983) Durability of natural fibres in concrete. Swedish Cement and
Concrete Research Institute, Stockholm.
Gumagul, N.; Page, D.H. and Paice, M.G. (1992) The effect of ceUulose degradation
on the strength of wood pulp fibres. Nordic Pulp and Paper Research Journal
3, 152-154.
Hannant, D.J. (1978) Fibre Cements and Fibre Concretes, John Wiley, Chichester,
1-50, 61-134.
Harper, S. (1982) Developmg asbestos free calcium silicate building boards. Composites
A, 123-128.
Hibbent, A.P. and Hannant, D.J. (1982) Toughness of fibre cement composites.
Composites 13, 105-111.
193
Higgins, H.G., Gariand, C.P. and Puri, V. (1977) Thermomechanical and

chemitiiemiomechanical pulps from eucalypts and other hardwoods. Appita, 30(5),
415-423.
Higgins, H.G., Puri, V. and Gariand, C.P. (1978) The effect of chemical pretreatments
in chip refining. Appita, 32(3), 187-200.
Hodgson, A.A. (1985) Alternatives to Asbestos and Asbestos Products.
Anjalena Publications Ltd. UK, 71-85, 150-158.
Hughes, D.C. and Hannant, D.J. (1985) Reinforcement of Griffith flaws in cellulose
reinforced cement composites. / Mater. Sci. Lett. 4, 101-102.
James Hardie Industries (1984) The name behind the names 71pp.
James Hardie and Co. Pty Ltd. Report on paper pulp sheets. Internal Report Nov. 1947.
Jayne, B.A. (1959) Mechanical properties of wood fibres. Tappi 42(6), 461-467.
Johnston, C D . (1975) Steel-fibre-reinforced mortar and concrete: A review of
mechanical properties. Fibre reinforced Concrete. American Concrete Institute
Publication S.P.44, 127-142.
Kerekes, R.J. and Tam Doo, P.A. (1985) Wet fibre flexibility of some major softwood
species pulped by various processes. / . Pulp and Paper Sci. 11(2), J60-J61.
Kibblewhite, R.P. (1989) Effects of pulp drying and refining on softwood fibre wall
structural organisations. Fundamentals of Papermaking, Vol 1. Trans, of the
Cambridge Symp. (Ed. C.F. Baker & V.W. Punton), Mech. Engineering
Publication. Ltd., London, 121-152.
Kim, C.Y., Page, D.H., El-Hosseiny, F. and Lancaster, P.J. (1975) Mechanical properties
of single wood pulp fibres. III. Effect of drying sttess on stiength. Appl. Polym. Sci.
19(6), 1594-1562.
Kuang, S. J., Zhang J. and Zhang, R.L. (1992) A study on the Characteristics of
194
blended fumishes of wheat straw pulp and long fibre pulps. China Pulp & Paper
11(2), 3-10.
Kurdin, J.A. (1983) The pulp of the twenty-first century. Tappi 66(6), 9-19.
Laws, V. (1983) On tiie mixture rale for strength of fibre reinforced cement. / . Mater.
Sci. Lett. 2, 527-531.
Lea, F.M. (1976) The Chemistry of Cement and Concrete. 3rd edn., Edward Arnold,
London.
Lhoneux, B. de and Avella, T. (1992) Fibre-matrix interactions in autoclaved ceUulose
cement composites. RILEM FRC92 (ed. R. N. Swamy), published by
E & F N Spon, London, 1152-1165.
Lhoneux, B. de; Avella, T. and Garves, K. (1991) Influence of autoclaving on chemical
pulp fibre properties for fibre-cement applications. Holzforschung. 45(1), 55-60.
Li, W.L. (1992) Major R & D achievements of the paper industrial research institute
(PIRIC) in the past years (1980-1990). China Pulp & Paper 6(3), 56-60.
Lola, CR. (1986) Fibre reinforced concrete roofing technology appraisal report.
RILEM FRC86 (eds R.N.Swamy, R.L.Wagstaffe and D.R.Oakley),
Mai, Y.W. (1978) Strength and fracture properties of asbestos-cement mortar composites.
Univ. of Sydney. Technical Note S-13.
Mai, Y. W., Andonian, R. and CottereU, B. (1980) On polypropylene-cellulose fibrecement hybrid composite. In Advances in Composite Materials, Paris, 1687-1699.
Mai, Y.W., Hakeem, M.I. and CottereU, B. (1983) Effects of water and bleaching on the
mechanical properties of cellulose fibre cements. J. Mater. Sci. 18, 2156-2162.
Mai, Y.W. and Hakeem, M.I. (1984a) Slow crack growth in bleached cellulose fibre
cements. / . Mater. Sci. Lett. 3, 127-130.
195
Mai, Y.W. and Hakeem, M.I. (1984b) Slow crack growth in cellulose fibre cements.
J. Mater. Sci. 19, 501-508.
Martson, T.U., Atkins, A.G. and Felbeck, D.J. (1974) / . Mater. Sci. 9, AAl.
McKenzie, A.W. (1978) The structure and properties of paper. Part XX. The tensUe
properties of paper and papermaking fibres. Appita 32, 207-212.
McKenzie, A.W. (1985) Studies on soda-anthraquinone pulpuig. Part 1, Fibre damage
and tearing resistance. Appita. 38(6), 428-431.
McKenzie, A.W. (1992) Commonly asked questions on pulping and papermaking.
CSIRO Australia, Forestry and Forest Products Newsletter, 7, 1-3
MicheU, A.J. and Freischmidt, G. (1990) Effect of fibre curl on the properties of wood
pulp fibre-cement and silica sheets. / . Mater. Sci. 25, 5225-5230.
Mindess, S. (19 ) The application of fracture mechanics to cement and concrete:
a historical review. Fracture Mechanics of Concrete, (ed. F.H. Wittman), Elsevier,
Amsterdam, 1-30.
Mindess, S. and Bentur, A. (1982) Technical Notes: The fracture of wood fibre einforced
cement. Int. J. Cement Comp. Lightweight Conor. A, 245-249.
Morrissey, F.E., Coutts, R.S.P. and Grossman, P.U.A. (1985) Bond between ceUulose
fibres and cement. Int. J. Cement Comp. Lightweight Conor. 7(2), 73-80.
Mukherjee, P.S., Satyanarayana, K.G. (1986) An empirical evaluation of stmctureproperty relationships in natural fibres and their fracture behaviour. / . Material Sci.
21,4162-4168.
NeveU, T.P. and Zeronian, S.H. (1984) Cellulose Chemistry and its Applications.
EUis Horwood, Chichester, 506-530.
NeviUe, A.M. (1959) Some aspects of the stiength of concrete. Civil Eng., London, 54,
1153-1156
196
Nilsson, L. (1975) Reinforcement of concrete with sisal and other fibrous plaster
sheets. Swedish Council for Building Research Document D14.
PaavUainen, L. (1993) Importance of cross-dimensional fibre properties and coarseness
for the characterisation of softwood sulphate pulp. Paper and Timber.
75(5), 343-351.
PaavUainen, L. (1993) Conformability - flexibility and coUapsibiHty - of sulphate pulp
fibres. Paper and Timber. 75(9-10), 689-699.
Pavithram, C , Mukherjee, P.S., Brahmakumar, M. and Damodaran, A.D. (1987) Impact
properties of natural fibre composites. J. Mater. Sci. Lett. 6, 882-884.
Page, D.H. (1969) A tiieory for the tensUe strength of paper. Tappi 52(4), 674-681.
Page, D.H. (1970) The physics and chemistry of wood pulp fibres. Tappi 42(6), 461-467.
Page, D.H., El-Hosseiny, F. and Winkler, K. (1971) Behaviour of single wood fibres
under axial tensile strain. Nature (London), 229, 252-253.
Page, D.H.; Seth, R.S. and EI-Hosseiny, F. (1985) Strength and Chemical
Composition of wood pump fibres. Papermaking Raw Materials, Trans, of the
Oxford Symp. (Ed. V.W. Punton), Mech. Engineering PubHcation. Ltd., London,
77-91.
Page, D.H.; Seth, R.S.; Jordan, B.D. and Barbe, M.C (1985) Curi, crimps, kinks and
microcompressions in pulp fibres - their origin, measurement and significance.
Papermaking Raw Materials, Trans, of the Oxford Symp. (Ed. V.W. Punton),
Mech. Engineering Publication. Ltd., London, 183-227.
Pakotiprapha, B., Pama, R.P. and Lee, S.L. (1983) Behaviour of a bamboo
fibre-cement paste composite. J. Ferrocement 13, 235-248.
Pedersen, N. (1980) Commercial development of alternatives to asbestos sheet products
based on short fibres. Proc. Symp. on fibrous Concrete London, 16 AprU 1980,
197
189-193.
Pinchm, D.J. and Tabor, D. (1978) hiterfacial phenomena in steel fibre reinforced
cement. I: Structure and strengtii of the interfacial region. Cem. Conor. Res.
8(1), 15-24.
Romualdi, J.P. and Batson, B.V. (1963) Fibre reinforced concrete properties.
/ . Eng. Mechanics division, ASCE, 89(EM3), 147-168.
Romualdi, J.P. and Mandel, J.A. (1964) / . Am. Concrete Inst., Detroit, 61, p657.
Seth, R.A. and Page, D.H. (1987) Fibre properties and tearing resistance.
Proceedings of 1987 Int. Paper Physic Conf. Auberge Mont Gabriel, Quebec.
9-16.
Seth, R.S. (1990) Fibre quality factors in papermaking - 1. the importance of fibre
length and strength. Mat. Res. Soc. Symp. Proc. Vol.197, 125-141.
Seth, R.A., Francis, D.W. and Bennington, C.P.J. (1991) The effect of mechanical
treatment during medium consistency fluidisation on pulp properties.
TAPPI proceedings of 1991 Pulping Conf 713-722.
Seth, R.A., Francis, D.W. and Bennington, C.P.J. (1993) The effect of mechanical
treatment during medium consistency fluidisation on pulp properties.
Appita. 46(1), 54-60.
Sharman, W.R. and Vautier, P.B. (1986) Durability studies on wood fibre reinforced
cement sheet. RILEM FRC86 (eds R.N.Swamy, R.L.Wagstaffe and D.R.Oakley),
Simatupang, M.H. and Lange, H. (1987) Lingnocellulosic and plastic fibres for
manufacturing of fibre cement boards. Int. J. Cement Comp. Lightweight Conor.
9(2), 109-112.
Sinha, U.N., Dutta, S.N., Chaliha, B.P. and Iyengar, M.S. (1975) PossibiHties of
198
replacing asbestos in asbestos cement sheets by cellulose pulp. Indian Conor. J.

8, 228-232.
Singh, S.M. (1979) Investigations into the causes of poor strength of Portland cement
bonded lignocellulosic materials, / . Ind. Acad. Wood Sci. 10(1), 15-19.
Singh, B. and Majumdar, A.J. (1981) Properties of GRC containing inorganic fillers.
Smook, G.A. (1989) Handbook for Pulp and Paper Technologists. TAPPI, U.S.A.
7th printing. Chap. 7 and 22.
Staff of the Institute of Paper Chemistry (1944) Paper Tr. J. 118(5), 13.
Stevens, M.G. and CantriU, E.R. (1992) Influence of autoclaving on Kraft wood pulp
cellulose determined by differential thermal analysis. Materials Forum.l6, 359-363.
Stone, J.E. and Clayton D.W. (1960) The role of sulphide in the Kraft process.
Pulp and Paper Magazine of Canada, June, T307-313.
StreHs, I. and Kennedy, R.W. (1967) Identification of North American Commercial
Pulpwoods and Pulp Fibres. Univ of Toronto Press, Canada.
Stucke, M.S. and Majumdar, A.J. (1976) Microstructure of glass fibre-reinforced cement
composites./. Mater. Sci. 11, 1019-1030.
Studinka, J.B. (1989) Asbestos substitution in the fibre cement industry.
Swamy, R.N. and Mangat, P.S. (1974) Proc. Inst. CivilEngrs., 57, p701.
Swift, D.G. and Smitii, R.B.L. (1979) The flexural strength of cement-based composites
using low modulus (sisal) fibres. Composites 10,145-148.
Tam Doo, P.A. and Kerekes, R.J. (1982) The flexibility of wet pulp fibres.
Pulp and Paper Canada 83(2), 46-50.
Taylor, W.H. (1967) Concrete Technology and Practice. Angus and Robertson,
199
Sydney, 3rd Ed. Chap 3.

Teder, A. and 01m, L. (1991) Proceedings Inter. Confer, on Bleached Kraft Pulp
Mills, 4-7th Feb. 1991, Melboume, 59-80.
Thomas, N.L. and BirchaU, J.D. (1983) The retardmg action of sugars on cement
hydration. Cem. Conor. Res. 13, 830-842.
US Patent (1884) No. 291114, 1 Jan. 1884
US Patent (1899) No. 635221, 17, Oct. 1899
US Patent (1900) No. 658590, 25, Sept. 1900
Vinson, K.D. and Daniel, J.I. (1990) Specialty ceUulose fibres for cement
reinforcement. Thin-section Fibre Reinforced Concrete and Ferrocement.
American Concrete Institute Publication SP-124, 99-124.
Wai, N.N., Nanko, H. and Murakami, K. (1985) A morphological study on the
behaviour of bamboo pulp fibres in the beating process. Wood Sci. Technol.
19,211-222.
Wang, J.H., Guo, X.P., Xue, CY. and Wang, R. (1993) The Ultrastmcture of bamboo
fibres and its behavior during beating. China Pulp and Paper 8(4), 10-17.
Walton, P.L. and Majumdar, A.J. (1975) Cement-based composites with mixtures of
different types of fibres. Composites, 9, 209-216.
Watson, A.J., Wardrop, A.B., DadsweU, H.E. and Cohen, W.E. (1952) Aust. Pulp
Paper Ind. Tech. Assoc. Proc. 6, 243.
Watson, A.J. and Hodder, I.G. (1954) Aust. Pulp Paper Ind. Tech. Assoc. Proc.
8, 290.
Watson, A.J. and DadsweU, H.E. (1961) Influence of fibre morphology on paper
properties, part 1. fibre length. A/;/;iYa 14(5), 168-178.
WilHams, M.F. (1994) Matching wood fibre characteristics to pulp and paper processes
200
and products. Tappi Journal, 77(3), 227-233.

Zhao, J., Zheng, G.M., Wang, H.D. and Xu, J.L. (1990) The Natural History of China.
CoUuis Publication Ltd. USA, 62-65.
201
Bibliography:
Balodis, V. (1991) Notes on pulpwood quality assessment CSIRO Div. Forest
Products, Australia.
Bentiey, R.G., Scudamore, P. and Jack, J.S. (1994) A comparison between fibre length
measurement methods. Pulp and Paper Canada, 95(4) 41-45.
Coutts, R.S.P. (1980) Wood fibre cement composites prepared by "slurry/vacuum
box" techniques. CSIRO Div. Chemical Technology. Program report, CB 203.
Irvine, G.M. (1992) Monash Univ. M.Eng.Sc. pulp and paper laboratory course
Topic. 8.3 chemical pulping. CSIRO Div. Forest Products.
Mai, Y.W. (1978) Strength and fracture properties of asbestos-cement mortar composites.
Univ. of Sydney. Technical Note S-13.
McKenzie, A.W. (1989) The tear-tensile relationship in softwood pulps. Appita 42(3),
215-221.
McKenzie, A., McHenry, M. and McKenzie, M. (1991) Monash Univ. M.Eng.Sc.
pulp and paper laboratory course Topic.8.6 paper testing. CSIRO Div. Forest
Products.
McKenzie, M. and McHenry, M. (1993) Pulp evaluation lab. manual part 1:
preparation of handsheets. CSIRO Div. Forest Products. FP 267
McKenzie, A.W. (1994) A Guide to Pulp Evaluation. Division of Forest Products,
CSIRO Austraha.
Nelson, P.J., Chin, W.J. and Grover, S.G. (1993) A study of ways of increasing the
use of non-chlorine-containing bleaching chemicals, increasing the effectiveness
of chlorine dioxide, and improving the overaU efficiency of modem bleaching
sequences, part A. National pulp mills research program internal report. No.l.
Pulmac Instraments hiternational, Canada. Fiber quality and fibre strength.
202
Technology series.
TAPPI Test Methods (1994-1995). TAPPI Press, U.S.A.
Williams, W.D. (1992) Monash Univ. M.Eng.Sc. pulp and paper laboratory course
Topic.8.2 pulping, part B. high-yield pulping. CSIRO Div. Forest Products,
FP 204.
203

Cement Composites

Uploaded by

Copyright:

Available Formats

Cement Composites

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Cement Composites

Uploaded by

Copyright:

Available Formats

NATURAL FIBRE REINFORCED

DEPARTMENT OF MECHANICAL ENGINEERING

I thank my colleagues and friends in Division of Forest Products, Commonwealth

I really wish to say "thank you all".

NATURAL FIBRE REINFORCED CEMENT COMPOSITES

Chapter One: Introduction

1.1 Natural fibre reinforced cement composites (NFRC)

1.1.1. History and development of NFRC composites

1.1.2 Fabrication process

1.1.3 Property requirements for asbestos alternatives

1.1.4 Mechanical and physical properties of NFRC

1.1.5 Durability of NFRC

1.2 Natural plant fibre resources

1.3 Scopes of the present work

1.3.1 Bamboo fibre and bamboo-wood hybrid fibre reinforced cement

Chapter Two: Theoretical Principles of Fibre Reinforcement

2.1.1 Mixture rule for strength

2.1.3 Basics of fracture mechanics

2.1.4 Fibre critical fracture length

2.1.5 Fibre aspect ratio

2.1.6 Fracture toughness

2.2 Bonding and microstructure of NFRC

2.2.1 Chemical bonding

2.2.2 Mechanical bonding

2.2.3 The effect of moisture on bonding

2.2.4 Microstructure of NFRC composites

Chapter Three: The Nature of Plant Fibres

3.1 Fibre classification

3.2 Structure of natural plant fibres

3.3 Chemical composition and durability

3.4 Preparation of natural plant fibres

3.4.2 Refining or beating

3.5 Properties of natural fibres

3.5.1 Physical properties

3.5.2 Mechanical properties

3.5.3 The effect of moisture on fibres

Chapter Four: Air-cured & Autoclaved Bamboo Fibre Reinforced Cement

4.1 Experimental work

4.1.2 Fibre modification

4.1.3 Fabrication and characterisation

4.2 Air-cured bamboo fibre reinforced cement

4.2.1 Mechanical properties

4.2.2 Physical properties

4.3 Autoclaved bamboo fibre reinforced cement

4.3.1 Mechanical properties

4.3.2 Physical properties

5.1 Experimental work

5.1.1 Fibre preparation

5.1.2 Fabrication and characterisation

5.2 Results and Discussion

5.2.1 Length and freeness of blended pulp

5.2.2 Air-cured BWFRC composites

5.2.3 Autoclaved BWFRC composites

5.3 Theoretical predication and the experimental results

Chapter Six: Influence of Fibre Length on Composite Properties

6.1.1 Fibre length fractionation work

6.1.2 Fabrication and characterisation

6.2 Results and Discussion