Thesis Bamboo Fibres E Trujllo 24082014

POLYMER COMPOSITE MATERIALS
BASED ON BAMBOO FIBRES
Eduardo TRUJILLO DE LOS RÍOS
Supervisors:
Prof. Jan Ivens
Dr. Aart W. Van Vuure
Members of the Examination

Committee:
Prof. Adhemar Bultheel, Chairman
Prof. Jan Ivens, Promoter
Dr. Aart W. Van Vuure, Promoter
Prof. Ignace Verpoest Dissertation presented in
Prof. Peter Van Puyvelde partial fulfilment of the
Prof. Joris Van Acker requirements for the
Prof. Adriaan Beukers degree of Doctor in
Prof. Ton Peijs Engineering
Prof.
September 2014
© 2014 KU Leuven, Science, Engineering & Technology
Uitgegeven in eigen beheer, Eduardo Trujillo De Los Ríos, Belgium
Alle rechten voorbehouden. Niets uit deze uitgave mag worden vermenigvuldigd en/of openbaar gemaakt worden door
middel van druk, fotokopie, microfilm, elektronisch of op welke andere wijze ook zonder voorafgaandelijke schriftelijke
toestemming van de uitgever.
All rights reserved. No part of the publication may be reproduced in any form by print, photoprint, microfilm, electronic or
any other means without written permission from the publisher.
ISBN 978-94-6018-883-1
D/2014/7515/107
XV
Table of contents
Acknowledgements……………………………………………………………………….. I
Abstract…………………………………………………………………………………… III
Samenvatting……………………………………………………………………………… V
List of abbreviations…………………………………………………………………….... IX
List of symbols…………………………………………………………………………..... XI
Table of contents………………………………………………………………………...... XV
Chapter 1: General introduction………………………………………………………... 1

Chapter 2: Problem statement and objectives……………...…………………………... 5
Chapter 3: Natural fibres and their composites: A literature review……………….... 11
3.1 Natural fibres…………………………………………………………………….... 11
3.1.1 Fibre microstructure………………………………………………………….. 12
3.1.2 Chemical composition of natural fibres…………………………………….... 13
3.2 Mechanical properties of plant fibres……………………………………………... 14
3.3 Natural fibre composites and their advantages……………………………………. 16
3.3.1 Environmental impact……………………………………………………….. 18
3.3.1.1 Bio-composites………………………………………………………... 19
3.3.2 Coupling agents…………………………………………………………….... 20
3.3.2.1 Alkali treatment……………………………………………………….. 21
3.3.2.2 Maleic anhydride polypropylene…………………………………….... 23
3.4 Limitations in the use of natural fibre composites……………………………….... 24
3.4.1 Moisture absorption………………………………………………………….. 24
3.4.2 Thermal degradation…………………………………………………………. 25
3.4.2.1 Thermal degradation in natural fibres……………………………….... 25
3.4.2.2 Thermal degradation in natural fibre composites……………………... 30
3.5 Bamboo plants…………………………………………………………………….. 32
3.5.1 Morphology of bamboo Guadua angustifolia culm…………………………. 33
3.5.1.1 Diameter, wall thickness and internode length of the culm…………... 34
3.5.2 Anatomy of the bamboo culm……………………………………………….. 35
3.5.3 Bamboo fibres………………………………………………………………... 36
3.5.3.1 Microstructure and chemical composition of bamboo fibres……...….. 37
3.5.3.2 Mechanical properties of bamboo technical fibres………………….... 38
3.6 Extraction of bamboo technical fibre………………..……………………………. 40
3.6.1 Steam explosion…………………………………………………………….... 42
XVI
3.6.2 Mechanical extraction………………………………………………………... 44

3.6.3 Chemical extraction………………………………………………………….. 45
3.7 Bamboo fibre composites…………………………………………………………. 46
3.8 Applications of natural fibre composites………………………………………….. 48
3.9 Conclusions………..…………………………………………………………….... 50
References…………………………………………………………………………….. 51
Chapter 4: Bamboo technical fibre characterization…………………………………. 59
4.1 Introduction……………………………………………………………………….. 59
4.1.1 The modified Weibull distribution…………………………………………... 60
4.1.2 Effect of defect density distribution…………………………………………. 64
4.1.3 Effect of within-fibre diameter variation…………………………………….. 66
4.2 Dry fibre bundle test………………………………………………………………. 66
4.2.1 Theoretical background…………………………………………………….... 67
4.3 Materials…………………………………………………………………………... 69
4.4 Methods………………………………………………………………………….... 70
4.4.1 Single fibre test………………………………………………………………. 70
4.4.1.1 Measurement of the cross sectional area…………………………….... 70
4.4.1.2 Measurement of the fibre perimeter…………………………………... 70
4.4.1.3 Tensile test set up……………………………………………………... 71
4.4.1.4 Statistical calculations...………………………………………………. 73
4.4.1.5 Scanning electron microscopy (SEM) observations.……...………….. 72
4.4.2 Dry fibre bundle test (DFT)………………………………………………….. 73
4.4.2.1 Reference methodologies……………………………………….…….. 73
4.4.2.2 Preparation of DFB samples………………………………………….. 74
4.4.2.3 Testing of the bamboo DFB samples…………………………………. 76
4.5 Results and discussion………………………………………………………….…. 78
4.5.1 Fibre extraction………………………………………………………………. 78
4.5.2 Fibre cross-sectional area and perimeter…………………………………….. 80
4.5.3 Mechanical properties of bamboo technical fibres…………………………... 82
4.5.3.1 Dependency of fibre strength on fibre length………………………..... 82
4.5.3.2 Dependency of fibre strength on fibre diameter, fibre volume
and fibre surface……………………………………………..………... 84
4.5.3.3 Estimation of the modified Weibull parameters ……………………... 86
4.5.3.4 Correlation with within-fibre diameter variations …………….…….... 88
4.5.3.5 Benchmarking of single fibre mechanical properties...……………….. 89
4.5.4 Dry bamboo fibre bundle…………………………………………………….. 91
4.6 Conclusions……………………………………………………………………….. 95
References…………………………………………………………………………….. 97
XVII
Chapter 5: Unidirectional continuous and discontinuous bamboo fibre

epoxy composites...………………………………………………………….. 101
5.1 Introduction……………………………………………………………………….. 101
5.2 Overview of different models for the prediction of mechanical properties………. 103
5.2.1 Prediction of longitudinal tensile stiffness…………………………………... 103
5.2.2 Prediction of longitudinal tensile strength………………………………….... 107
5.3 Materials…………………………………………………………………………... 109
5.3.1 Bamboo fibres………………………………………………………………... 109
5.3.2 Epoxy resin…………………………………………………………………... 110
5.4 Methods………………………………………………………………………….... 110
5.4.1 Fibre patterns……………………………………………………………….... 110
5.4.2 Elaboration of the prepregs…………………………………………………... 113
5.4.3 Composite production………………………………………………………... 115
5.4.3.1 Manufacturing of the samples for tensile testing……………………... 115
5.4.3.2 Sample production for 3-point bending test…………………………... 116
5.4.4 Sample testing……………………………………………………………….. 117
5.4.4.1 Tensile test……………………………………………………………. 117
5.4.4.2 Flexural test………………………………………………………….... 119
5.4.5 Scanning electron microscopy (SEM) observations…………………………. 119
5.5 Results and discussion…………………………………………………………….. 119
5.5.1 Continuous fibre composites (UD-C)……...……………………………….... 121
5.5.2 UD-D with fixed fibre bundle overlapping ength……………………………. 123
5.5.3 UD-D with random fibre bundle overlapping length…...………...…………. 124
5.5.4 UD-D with individual technical fibres with random fibre ends…..…………. 125
5.5.5 Tensile fracture characteristics for UD-D samples…………………...…….... 125
5.5.6 Experimental stiffness compared with predicting models………………….... 127
5.5.7 Experimental strength compared with predicting models………………….... 129
5.5.8 Properties evaluated in 3-point bending test…………………………………. 134
5.6 Conclusions……………………………………………………………………….. 137
References…………………………………………………………………………….. 138
Chapter 6: Thermal degradation in bamboo fibres and bamboo fibre
polypropylene composites………....………………………………………... 141
6.1 Introduction……………………………………………………………………….. 141
6.2 Materials…………………………………………………………………………... 143
6.2.1 Bamboo fibres………………………………………………………………... 143
6.2.2 Polypropylene (PP) and maleic anhydride polypropylene (MAPP)…………. 143
6.3 Methods………………………………………………………………………….... 143
6.3.1 Thermal treatment of single technical bamboo fibres……………………….. 143
XVIII
6.3.2 Thermogravimetric analysis (TGA)....…………...………………………….. 144

6.3.3 Single fibre test………………………………………………………………. 145
6.3.4 Scanning electron microscopy (SEM) observations…………………………. 146
6.3.5 Bamboo fibre PP/MAPP composites……………………………………….... 146
6.3.5.1 UD bamboo prepregs………………………………………………..... 146
6.3.5.2 Composite production……………………………………………...…. 147
6.3.5.3 Composite production under inert atmosphere……………………….. 148
6.3.6 Sample preparation and testing………………………………………………. 149
6.4 Results and discussion…………………………………………………………….. 150
6.4.1 Thermogravimetric analysis (TGA)…………………………………………. 150
6.4.1.1 Isothermal TGA experiments…………………………………………. 152
6.4.2 Tensile properties for single fibre after thermal treatment…..………………. 154
6.4.2.1 Fibre strength…………………………………………………………. 154
6.4.2.2 Fibre stiffness and strain to failure……………………………………. 158
6.4.2.3 Influence of an inert atmosphere in mechanical properties…………… 160
6.4.2.4 Thermally treated bamboo fibres under SEM……………….………... 162
6.4.3 Bamboo fibre polypropylene composites (BFPP)………………………….... 164
6.4.3.1 Fibre degradation during composite processing…………………….... 164
6.4.3.2 Matrix degradation……………………………………………………. 166
6.4.3.3 TGA analysis for bamboo fibre polypropylene composites…………... 166
6.4.3.4 Flexural strength of bamboo fibre polypropylene composites………... 168
6.4.3.5 Bamboo fibre MAPP composites……………………………………... 174
6.5 Conclusions……………………………………………………………………….. 176
References…………………………………………………………………………….. 178
Chapter 7: Technology assessment and application potential of bamboo fibre
composites………………………………………………………………….... 181
7.1 Introduction……………………………………………………………………….. 181
7.2 Availability of the resources………………………………………………………. 182
7.3 Environmental benefits……………………………………………………………. 184
7.4 The extraction process and the effect on the ecological impact and fibre cost…… 185
7.5 Manufacturing and performance of the final composites; potential applications…. 188
7.6 Durability and recyclability……………………………………………………….. 191
7.7 Conclusions……………………………………………………………………….. 192
References…………………………………………………………………………….. 192
XIX
Chapter 8: General conclusions……….……………………...…………………………. 195

8.1 Bamboo technical fibre……………………………………………………………. 196
8.2 Unidirectional continuous and discontinuous bamboo fibre composites………..... 197
8.3 Thermal degradation in single technical fibres and bamboo fibre
polypropylene composites……………………………………………………….... 197
8.4 Potential use of bamboo fibre composites………………………………………… 198
8.5 Suggestions for future research...………………………………………………..... 199
Appendices………………………………………………………………………………... 201
Appendix 1……………………………………………………………………………. 202
Appendix 2……………………………………………………………………………. 207
Curriculum Vitae
List of publications
General introduction 1
Chapter 1
General introduction
_______________________________________________________
Most likely the reader has already noticed that the climate is changing as some
phenomenon like extreme weather, droughts, cyclones, melting of the Artic sea ice,
etc., can be experienced in daily life and others are reported in different regions of
the globe. The world's leading climate scientists agree that there is at the moment an
increase of the global temperature by 0.85 ºC since the late 19th century1 and the last
three decades have been warmer than any preceding decade. One of the main causes
of this “global warming” is the emissions of greenhouse gases, like carbon dioxide
(CO2), from different human activities. One of the main sources of CO2 in the
atmosphere is the combustion of fossil fuels - coal, oil and gas, responsible for some
63% of the total man-made global warming1.
Lately, climate experts have pointed out that if little or no action is taken to reduce
global greenhouse gas emissions by the end of the 21st century, global warming is
likely to reach a 2-5°C increase above the average temperature2 measured in pre-
industrial times. An increase beyond this range would lead to higher risk of
catastrophic changes in the global environment. To stay within this margin, the
greenhouse gas emission increase must be stopped by 2020 at the latest, to then
reduce the emissions by at least half of 1990 levels by 20503. Climate change is not
anymore a warning but a reality. Unfortunately, it is becoming more tangible, so
much so that at least 20% of the entire European Union budget for 2014-2020 will
be spent on climate-related projects and policies.
1
European Commission Climate Action website: ec.europa.eu/clima/policies/brief/causes/index_en.htm
2
European Environment Agency climate change website: eea.europa.eu/themes/climate/intro
3
UNFCCC/Kyoto Protocol website: unfccc.int
2 Chapter 1
b
This action marks a major step forward in transmuting Europe into a greener and
competitive low carbon economy. Being one of the front runners, the EU also
vouched to help developing countries to adapt to the impacts of climate change.
Nevertheless, these kinds of policies have not been adopted by all industrialized
countries.
As an example, transportation by road contributes in average about one-fifth of the

EU's total emissions of carbon dioxide (CO2) which rose by nearly 23% between
1990 and 20104. To reduce these emissions, European Union legislation requires the
greenhouse gas intensity of vehicle fuels to be reduced up to 10% by 2020, with a
reduction in CO2/km emissions of 30% in comparison with the current average5.
Another regulation that involves the transportation sector is the End of Life Vehicles
(ELV) directive 6 . It aims at making vehicle dismantling and recycling more
environmentally friendly, setting clear quantified targets for reuse and recycling of
vehicles and their components. In order to achieve these various targets, innovative
research is nowadays carried out in order to replace traditional materials by lighter
component parts with recycling potential, like fibre reinforced polymer composites
using synthetic and natural fibres.
Glass and carbon are the most common man-made fibres to reinforce polymer
composites with a great acceptance due to their good mechanical properties.
Nevertheless, they are non-renewable materials and their production requires high
energy consumption. Natural fibres such as bamboo, flax, hemp, jute and coconut
(coir) are renewable resources emerging as good alternatives for reinforcing
materials because of their bio-based character, high specific mechanical properties
that can compete with glass fibres, reasonable cost and sustainable supply. Several
studies involving life cycle assessment (LCA), agree that natural fibres are likely to
be environmentally superior to synthetic fibres in each phase of the entire life cycle
of the composite materials. Moreover at the end of life in the case of thermoplastic
matrices, they allow mechanical downcycling (recycling) of the composite (e.g. for
injection moulding).
Nowadays, several projects using natural fibre composites are being carried out with
the aim to make them commercially available in new high level applications for
different sectors. For example, flax thermoset composites are being investigated to
be used in a car trunk load floor by PSA Peugeot Citroën. Boeing is working on
sandwich panels with skins made of flax and thermoset resin matrix, while other
4
European Commission Climate Action website: ec.europa.eu/clima/policies/transport/vehicles_en.htm
5
European Environment Agency climate change website: eea.europa.eu/themes/climate/policy-context
6
European Commission Climate Action website: ec.europa.eu/environment/waste/elv_index.htm
General introduction 3
partners in the same project are investigating thermoplastic matrices such as PP and
PLA. Also, the European project “BioBuild” aims to use biocomposite materials,
including flax and hemp fibres, to bring about a step-change in the reduction of
embodied energy in building-façade and internal-partition systems, whilst being
commercially competitive. Those are a few examples of the initiatives that are now
implemented in the growing sector of green composites.
The present research “polymer composite materials based on bamboo fibres”

presents the development of an environmentally friendly material as an alternative
among other natural and synthetic fibre reinforced composites. It also seeks to
preserve and to stimulate the growth of a worldwide natural resource, bamboo, with
the advantage that it absorbs and stores large amounts of CO2. One hectare of
bamboo can sequester 62 tons of CO2/year, four times more in comparison to a
young forest and generates up to 35% more oxygen than an equivalent stand of
trees 7 . It has been found that carbon sequestration by growing forests is a cost-
effective solution to mitigating climate change.
In recent years, there has been an increasing interest to scientifically study the
potential of bamboo fibre as reinforcing material for polymer matrix composites.
Nevertheless, few efforts have been carried out to extract long bamboo fibres from
the culm, particularly because of difficulties in its extraction. In this state of mind,
this large bottleneck for the availability of long high quality technical bamboo fibres
has encouraged the development of a new extraction technique presented in this
thesis. This new process is expected to have low environmental impact because of a
low overall energy consumption and avoiding the use of chemicals. Bamboo
technical fibres are extracted from the Guadua angustifolia species, one of the three
largest bamboos in the world and the most important tropical bamboo in America.
The fibres are studied to be used as reinforcement in composite materials in
combination with thermoset and thermoplastic matrices.
In this thesis, the whole spectrum of the development of bamboo fibre reinforced
composites will be presented. The first step deals with the characterization of the
extracted technical fibre. The second one consists of the preparation techniques and
the manufacturing of high quality composites and their testing. The results are
benchmarked with other studies of natural fibres and natural fibre composites in
order to demonstrate the potential of this new material amongst well established
composite materials. As of today, particular challenges must be overcome in order to
implement the bamboo fibre composites in real high-tech applications. Some of
7
Environmental bamboo foundation: www.bamboocentral.org
4 Chapter 1
b
them, such as variability in fibre mechanical properties, the thermal degradation of

fibres during and after composite manufacturing and the optimum manufacturing
parameters for the best performance are investigated in the next chapters of the
thesis.
Problem statement and objectives 5
Chapter 2
Problem statement and objectives
_______________________________________________________
The main goal of this research is to develop a bamboo fibre composite with high
mechanical properties using technical fibres extracted from the Colombian bamboo
Guadua angustifolia. For that, two main challenges need to be tackled. The first one
consists in the extraction of long bamboo technical fibres that can meet the
requirements established in the field of reinforcements for composite materials,
while conserving the good intrinsic properties of the technical fibres at a competitive
cost. The second aspect consists in optimizing the transfer of the good mechanical
properties of the fibres into the composites performance, with both thermoplastic
and thermoset matrices. Therefore, the main objectives stated for this study are:
 To develop operating prototype machines incorporating environmentally

friendly principles for extraction, cleaning and alignment of technical fibres
that allow scale-up in industrial production;
 To perform a complete characterization of the bamboo technical fibre after
extraction and cleaning;
 To explore the processing feasibility of bamboo fibre composites by different
techniques and to study the resulting composite properties;
 To carry out a systematic study of the influence of processing parameters on
the mechanical properties of bamboo fibres and bamboo fibre composites.
In order to achieve these objectives, several activities need to be carried out through
the different chapters of this study as is schematically outlined in Figure 2-1.
6 Chapter 2
b
Figure 2-1. Schematic overview of the activities carried out in this thesis.
Before the start of this study, the principles of a new environmentally friendly
continuous process for extracting bamboo fibres were established in a previous
study in Colombia. The fibres at that stage were extracted using a manual process. In
order to obtain higher quantities of fibres needed for this research, these principles
were translated into a prototype machine for the continuous (in-line) production of
long fibres. It has to be noticed that the implemented mechanical principles can be
scaled-up for future large-scale industrial production. The technology has been
implemented for the fibre extraction of Colombian bamboo (Guadua angustifolia),
but it is expected to be equally applicable to other bamboo species due to its
common characteristics. Due to the proprietary nature, the extraction principles
cannot be disclosed in this thesis.
The cleaning-combing operation needed to remove the soft tissue present on the
fibre surface after the extraction process, was also carried out manually. To increase
the yield of clean fibres in a standardized process, a prototype machine was
developed and constructed at KU Leuven after applying a design of experiments
methodology (DOE), considering the most important process parameters for an
optimum balance between the fibre cleanness and the fibre strength. The cleaning
process is suitable to be put in line with the extraction process. These lab-scale
prototype machines allow the production of long bamboo fibres in a fast and
reproducible way, ready to be incorporated into the composites. Because of the
protection of intellectual property, both the extraction and cleaning-combing process
principles are kept secret. This also applies to a new lab-scale equipment to obtain a
unidirectional tape with randomized bamboo technical fibre ends and a business
plan for the potential commercialization of the UD bamboo fibre prepregs. For this
reason, even though some of these activities were developed in the framework of
this research, the schematics and principles of the prototypes and operations are not
presented in this thesis.
A comprehensive literature review about the key concepts of natural fibres and the
state of the art of natural fibres and natural fibre composites with emphasis on
bamboo is presented in Chapter 3. To evaluate the robustness of the new extraction
process, an extensive characterization of the mechanically extracted technical fibres
is carried out in Chapter 4. The single technical fibre (SF) test generates their most
important mechanical properties such as maximum strength, Young’s modulus and
strain to failure. Through this test, it is not only possible to obtain the fibre strength
distribution and the corresponding parameters for the modified Weibull distribution,
but also the effect on properties after exposing the bamboo fibres to relatively high
temperatures, related to polymer processing.
8 Chapter 2
b
For the thermal characterization, the effect of typical composite manufacturing

conditions in two different environments (air and inert atmosphere) is investigated,
by treating the fibres at different times and temperatures. This study is performed to
understand the mechanisms of thermal degradation of the bamboo technical fibres.
With the obtained results, it is possible to estimate the properties of the bamboo
fibre composites after manufacturing, especially with thermoplastic matrices, where
high temperatures are used during the consolidation phase. This topic is also
evaluated and complemented in Chapter 6. In parallel, SEM and XPS are applied to
give better details of the fibre surface when exposed to these high temperatures.
Complementary to the SF characterization, a dry fibre bundle test (DFB) was
developed and implemented for the first time for natural fibres, to correlate with the
results from individual fibre tests and to evaluate the advantages and disadvantages
of this technique. SF tensile test results are used later to evaluate the properties of
the bamboo fibre composites in both thermoset and thermoplastic composites.
The next step in the research includes the production and characterization of high
quality bamboo fibre composites. In order to compare with existing composite
systems, epoxy resin and polypropylene (PP) matrices were selected due to their
extensive use in industry. Special attention was paid to evaluate the feasibility of the
composite manufacturing using these matrices. Bamboo epoxy composites are
presented in Chapter 5. First, the properties of unidirectional bamboo fibre
composites have been determined, without the presence of fibre ends inside the
specimens. Then, the effect of fibre ends was introduced, and the effect of the
distribution of these fibre ends on the UD properties was investigated and
benchmarked with a fully unidirectional fibre composite. This is important for the
development of continuous tape (prepreg) that inevitably will contain fibre ends.
Tensile tests and 3-point bending tests were performed for the evaluation of the
properties.
In Chapter 6, bamboo fibre thermoplastic (TP) matrix composites using PP and

maleic anhydride grafted polypropylene (MAPP) were produced by compression
moulding at different temperatures. The aim of this chapter is to maximise the
composite properties, taking into account the effect of the processing conditions.
Due to the high temperatures during the TP manufacturing process, a reduction in
composite properties is expected due to thermal degradation of the fibres. In order to
minimise fibre damage, two different actions are applied. The first is to find the
most effective process parameters (temperature and time of exposure) to obtain good
mechanical properties. The second action consists of a more ambitious procedure,
investigating for the first time in natural composites, the use of inert atmosphere (i.e.
argon) for the delay of fibre degradation. The goal is to explore the possibility of
using higher processing temperatures, thus expanding the processing window for the
use of other TP matrices, currently restricted to low melting temperature matrix
systems. The mechanical properties of the unidirectional bamboo fibre composites
are obtained from 3-point bending tests. The transverse three point bending test
(T3PB) is used as an indirect method to determine the fibre-matrix interface quality.
The technical assessments, the potential application of bamboo fibre and bamboo
fibre composites as well as their competitive advantages over other natural fibres
and glass fibres, are discussed in Chapter 7. Also, important aspects such as the
availability of the raw material, ecological and environmental impacts of the
resource and the composite material, manufacturing and performance of the final
material and its recyclability are presented. Finally, the potential use of bamboo
technical fibres for the reinforcement of polymeric matrices in high performance
applications is summarized, together with the future work, in the general
conclusions in Chapter 8.
10 Chapter 2
b
Literature review 11
Chapter 3
Natural fibres and their composites:
A literature review
_______________________________________________________
3.1 Natural fibres
Commonly, natural fibres are classified in three categories according to their origin:
plant fibres, animal fibres and mineral fibres. Plant fibres are further divided in
several main groups depending on the place they are extracted (stem, leaves, fruit,
seeds, etc.). Figure 3-1 shows a schematic representation of these categories with a
number of examples for each group.
Figure 3-1. Classification of natural fibres according to their origin and the location within the plant [1].
12 Chapter 3
b
The term “fibre” covers materials whose length (L) is several times their diameter
(d), and therefore, they have a high L/d or aspect ratio. For example, hemp and jute
fibres have this ratio between 100 to 1000, cotton and wool have it between 1000
and 300 [2-4]. Synthetic fibres can reach any value of L/d ratio due to their
continuous manufacturing process.
3.1.1 Fibre microstructure
With natural fibres, a distinction between elementary fibre and technical fibre has to
be made. The first one refers to the fibres of a few millimetres long that can be
extracted for example with chemical treatments. The technical fibres correspond to
the fibres obtained after some standard extraction process (e.g. mechanically), and
they are composed itself of elementary fibres held together by means of an interface
of pectin and lignin. Technical fibres are the ones used in this study for single fibre
characterization and composite manufacturing. On a microscopic scale, elementary
fibres have two types of cell walls consisting in a thin outer primary wall and an
inner layered secondary wall arranged as concentric ellipsoids with a small channel
in the middle called lumen [5] (see Figure 3-2). Each elementary fibre has a complex
layered structure where the primary wall is encircling the secondary wall with
random orientation of the cellulose microfibrils.
b c
a
Figure 3-2. Structural architecture of a natural fibre. a) sisal fibre rope, b) sisal technical fibre, c) mesoscopic
scale: technical fibre, compose of elementary fibres and d) microscopic scale: elementary fibre consisting in a
thin outer primary wall and an inner layered secondary wall arranged as concentric cylinders with a small
channel in the middle called lumen. Adapted from [6].
The secondary wall is divided in multiple layers (S1, S2 and S3) based on the
orientations of the microfibrils. S1 is made up of a cross-fibrillar network, whereas
S3 has a more transverse orientation. Both sections have fibrils oriented at large
angles to the fibre axis. The middle part of the secondary wall (S2) is characterized
by helicoidal wound cellular microfibrils formed from long chain cellulose
molecules that are approximately parallel and longitudinally directed. The
mechanical properties of the fibre are determined by this section, because it makes
up ~80% of the total thickness and thus acts as the main load bearing component [7,
8]. The angle between the fibre axis and the micro fibrils is called microfibrillar
angle [9]. Vegetable fibre is in itself as a composite material, consisting mainly of
cellulose microfibrils embedded in a matrix that has lignin and hemicellulose as
primary constituents [3].
3.1.2 Chemical composition of natural fibres
According to Bledzki et al. [10] and McKendry [11], hemicelluose, cellulose and
lignin are the three main components of biomass. In general, they cover respectively
20-40, 40-60 and 10-25 wt % for all natural fibres. The amount of these three
constituents varies considerably among the plant fibres (Table 3-1). Other
components, usually regarded as surface impurities, are both the pectin, which
provides flexibility to the plant, and waxes which consist in different types of
alcohols [12]. Climatic conditions, age and the maturation process influence not
only the structure of fibres, but also, their chemical composition.
Cellulose Hemicellulose Lignin

Fibre
(%) (%) (%)
Banana 65 - 5
Coir 43 0.15–0.25 45
Cotton 82.7 5,7 -
Flax 71 16,7 2
Hemp 59-78 18–22 6
Henequen 60 28 8
Jute 45-63 18-21 21-26
Kenaf 50.5 21 17
Ramie 68.6 13,1 7
Sisal 70 12 12
Bamboo 60.8 24-28 32.2
Table 3-1. Chemical composition of some natural fibres [10, 13-15].
Cellulose is a natural polymer whose elementary macromoleule unit is the anhydro-

d-glucose. It contains three hydroxyl groups (-OH) (see Figure 3-3). Two of these
hydroxyls groups form intermolecular bonds, while the third one forms
14 Chapter 3
b
intramolecular hydrogen bonds [10, 12, 16]. Due to the presence of these hydroxyl
groups, the natural fibres, in general, have hydrophilic properties. The fibre moisture
level can reach 3-13% and can lead poor interface properties when used to reinforce
hydrophobic matrices [17].
Figure 3-3. Cellulose structure [7].
Several studies on various plant species predict that the tensile strength and Young’s
modulus increase with increasing cellulose content and decreasing microfibrillar
angle [12, 17]. This is because cellulose is an oriented long-chained, linear polymer
which is packed into a crystalline lattice over much of its length. Thus, it possesses
two of the basic strengthening elements: high degree of crystallinity and high
molecular weight with a degree of polymerization of around 10,000. Cellulose has
been estimated to have a theoretical modulus of 140 GPa [10, 12, 18].
The second main component, hemicellulose, comprises several sugar units and
exhibit a high degree of chain branching. It is an amorphous polymer with
significantly lower molecular weight than cellulose [19]. It is strongly bound to
cellulose fibrils presumably by hydrogen bonds and it forms the supportive matrix
for cellulose microfibrils. Hemicellulose is very hydrophilic and mainly responsible
for the moisture sorption behaviour of the fibres, soluble in alkali, and easily
hydrolyzed in acids [16].
The third main constituent is lignin, with a complex three-dimensional copolymer of

aliphatic and aromatic constituents. It is totally amorphous and hydrophobic in
nature. It is considered to be a thermoplastic polymer with very high molecular
weight that gives rigidity to the plants and trees [12].
3.2 Mechanical properties of plant fibres
Table 3-2 presents some characteristics and results in mechanical properties of

natural fibres compared to glass and carbon fibres. The strength and stiffness of
plant fibres mainly depend on several factors such as chemical constitution of the
fibre, the microfibril angle, the plant age, the environmental growing conditions and
the defects introduced during extraction. When the density of the material is taken
into account, the specific properties of some natural fibres are comparable with those
of glass fibres. [20]. More information about mechanical properties of natural fibres
can be found in Summerscales et al. [21, 22] and Bledzki et al. [10].
Microfibril Elongation Tensile Young’s

Density Diameter
Fibre angle at failure strength modulus
(gr/cm3) (µm)
(deg.) (%) (MPa) (GPa)
Bagasse - 490 - - 70 -
Coir 1.2 - 30 - 49 30 175 4-6
Cotton 1.6 20 - 7.0 – 8.0 287 - 597 5 - 12
Curaua 1.3 66 - 3.9 913 30
Flax 1.5 50 - 100 10 2.7 – 3.2 345 - 1035 50 - 70
Hemp 1.1 120 6 1.6 389 - 900 35
Henequen - 180 - 3.7 – 5.9 430 - 570 10 - 16
Jute 1.3 260 8 1.5 –1.8 393 - 773 26
Kenaf 1.3 106 - 1.8 427 - 519 23 - 27
Pineapple 1.3 - - 2.4 608 - 700 24 - 29
Ramie 1.5 34 7 3.6 – 3.8 400 - 938 24 - 32
Sisal 1.5 50 - 80 20 - 25 2.0 – 2.5 337 - 413 8 - 10
Bamboo 0.8– 1.1 100 - 200 10 - 391-713 18 - 55
E-Glass 2.5 9 -15 - 2.5 1200 – 1500* 70
Carbon (PAN) 1.4 5-9 - 1.4 – 1.8 4000 230 – 240
Table 3-2. Mechanical properties of some plant technical fibres compared to glass and carbon fibres [10, 23-
26]. *The strength of glass fibres highly depends on its previous handling history: a recent extruded fibre can
exhibit a strength of around 3500 MPa. However, prior to composite production, the strength in some cases
can be reduced approximately to 1250 MPa (due to all manipulations) and furthermore, the strength may
become as low as 900 MPa during the composite fabrication [27].
The industrial adoption of natural resources for reinforcing composites is an active

subject of research. The acceptance of natural fibre reinforced plastics in technical
applications depends on the availability of material data and specific design
information. Establishing the reliability needed for final product applications
requires extensive testing and a substantial amount of research [28]. In fact, one of
the main concerns for massive industrial application of natural fibres, is their
variability in mechanical properties [29]. In comparison with synthetic fibres,
natural fibres have a significantly higher variation in diameter between technical
fibres and along of individual fibres [29-32]. In addition, the fibre strength was
found to be negatively correlated to fibre diameter and gauge length [32]. This
variation in fibre properties can be characterized and predicted when quality
management is used, but also controlled when this information serves as a feedback
for optimizing the extraction and further fibre preparation.
16 Chapter 3
b
3.3 Natural fibre composites and their advantages
Cellulosic materials are the most abundant form of biomass and most likely to be
used as reinforcement fibres [33]. Moreover environmental friendliness, natural
abundancy and renewability make natural fibres a possible alternative to synthetic
reinforcing fibre materials, while offering good specific properties such as strength
and stiffness [34]. Additionally, unlike the traditional engineering fibres (e.g. glass
and carbon fibres), lignocellulosic fibres cause no health hazards due to fibre cutting
and low machine wear during post-processing and composite finishing. In composite
use, the thermal conductivity of the natural fibre composite is low; therefore they
make a good thermal barrier [35]. Also, the hollow nature of vegetable fibres leads
to a better acoustic insulation and damping properties once combined to a certain
types of matrices [36].
Nowadays, a substantial amount of research is carried out to analyse the possibility

of using natural fibres as reinforcement in polymer matrices, often shortened to
“composites”. Most composites have strong, stiff fibres embedded in a thermoset
(e.g. epoxy or unsaturated polyester resin) or thermoplastic (e.g. polypropylene)
matrix. These polymers are weaker in comparison to the fibres, but when the two
phases are combined, it is possible to generate a light and strong composite material.
In this structure, the fibres will bear the majority of the load and the polymers are
mainly used as fibre binder, as load transfer medium between the fibres, to protect
the fibres against e.g. abrasion and to prevent the fibres from buckling.
The composite material performance in the final application, depends on several

aspects such as: fibre volume fraction (Vf), mechanical properties of fibre and matrix,
impregnation, adhesion between fibre and matrix, fibre length (L) and fibre
orientation [2]. Regarding this last aspect, the unidirectional (UD) fibre
configuration achieves the maximum performance in the final composite, since the
fibres are aligned and placed in specific directions according to the required working
loads in the use phase. Structural and semi-structural composite parts are generally
manufactured with continuous aligned long fibres. Tables 3-3 to 3-6 show relevant
studies of (UD) natural fibres composites with thermoset and thermoplastic matrices
evaluated in flexural and tensile tests.
Before treatment After treatment

Vf Increase*
System Alkali treatment applied Flex. strength Flex. modulus Flex. strength Flex. modulus Ref.
(%) (%)
(MPa) (GPa) (MPa) (GPa)
UD Flax + epoxy 3% NaOH, 20 min at 20 °C 40 218 18 283 22 30/22 [37]
260 17
UD Jute + epoxy 26% NaOH, 20 min at 20 °C 178 9 46/89 [38]
40
239 15
UD Jute + vinylester 5% NaOH, 4h at 30 °C 199 12 20/25 [39]
35
225 12
UD Hemp + epoxy 22% NaOH, 1h at RT 33 148 6 52/100 [17]
100
UD Hemp + polyester 6% NaOH, 48h at 19 °C 60 77 7 9 30/27 [40]
UD Kenaf + polyester 120 13
6% NaOH, 48h at 19 °C 60 30 4 300/225 [40]
UD sisal + epoxy
2% NaOH, 4h at 60°C 37 168 12 225 16 34/33 [41]
UD sisal + uns. polyester
- 40 65 1.9 - - - [42]
Table 3-3. Flexural properties of natural fibres composites with thermoset matrices. *Flexural strength / Flexural modulus.

Vf Tensile Tensile Tensile Tensile Increase*
System Alkali treatment applied Ref.
(%) strength modulus strength modulus (%)
-
UD Flax + epoxy 20 127 17 - - - [43]
UD sisal + epoxy 2% NaOH, 4h at 60°C 37 200 8 19/-17 [41]
238 7
UD sisal + uns. polyester - 40 99 3 - - - [42]
UD Jute + epoxy 26% NaOH, 20 min at 20°C 40 159 15 35/53 [38]
215 23
Table 3-4. Tensile properties of natural fibres composites with thermoset matrices. *Tensile strength / Tensile modulus.
18 Chapter 3
b
System Process parameters Vf (%) % MA Flexural properties Ref.

σ (MPa) E (GPa) ε (%)
UD Jute – PP 140, 150 and 160°C 49 - 101 11 - [44]

(coated* yarns) 20 Bar 15 min
UD Jute yarns-PP 180°C 10 min 21 - 122 9 - [45]

(MB** + coated* 20 Bar
yarns)
UD Flax-PP 12 min pre-heating, 17 - 90 ±25 11±5 1.8 [46]
(retted fibres) 220 °C (2.5 min)
UD Flax-PP 12 min pre-heating, 27 - 129±19 16± 2 1.2 [46]

(boiled fibres) 220 °C (2.5 min)
UD Flax-MAPP 12 min pre-heating, 28 0.6 164±56 14± 6 2.0 [46]

(retted fibres) 205°C 2.5 min
UD Flax-MAPP 12 min pre-heating, 32 0.6 212±14 23± 2 1.4 [46]

(boiled fibres) 205 °C (2.5 min)
UD Glass-MAPP 200°C 45min 58 10 1037 32 3.3 [47]

25 Bar
UD Glass-PP 200°C 45min 58 - 774 38 2.1 [47]

25 Bar
Table 3-5. Flexural properties for UD natural and glass fibre composites with thermoplastic matrices.
*PVA/PP (polyvinyl alcohol/polypropylene), **MB (micro braiding): continuous jute yarns were used as a
core and PP fibres were braided around the reinforcing yarn. MA= maleic anhydride (functional group grafted
onto one end of the PP chain that acts as a compatibilizer).
System Process parameters Vf (%) % MA Tensile properties Ref.

σ (MPa) E (GPa) ε (%)
UD Hemp-MAPP 180 °C 3min 30 10 37 12 0.5 [48]

3 Bar
UD Hemp-PP Film stacking 35 - 125 - - [17]
UD Hemp-MAPP Film stacking 35 0.5 122 - - [17]
Table 3-6. Tensile properties for UD natural fibres composites with thermoplastic matrices. MA= maleic
anhydride (functional group grafted onto one end of the PP chain that acts as a compatibilizer).
3.3.1 Environmental impact
Traditionally glass and carbon fibres are the most common fibres to reinforce
polymer composites. They have a great acceptance due to their good mechanical and
thermal properties and have been used for several years in many applications, which
can vary from aerospace components to sporting goods. Nevertheless, the end of life
disposal of these composites is not that evident.
Natural fibres are environmentally friendly, not only during their growth stage but
also after their lifetime because they are biodegradable. A comparative study
including life cycle assessment (LCA) studies of natural fibres and glass composites
was carried out by Joshi et al. [49]. His study reveals that natural fibre composites
are likely to be environmentally superior to glass fibre in most cases because:
 Carbon dioxide fixation of natural fibres is an important issue in the reduction of

the greenhouse effect;
 In many bulk composites applications, natural fibre composites have higher fibre
content for equivalent performance, reducing more polluting base polymer
content;
 The light-weight natural fibre composite in comparison with classical materials,
e.g. steel, improves fuel efficiency and reduces emissions in the use phase of the
component, especially in automotive applications;
 The end of life incineration of natural fibres results in recovered energy and
carbon credits; being basically carbon neutral;
 Finally, their production has a low environmental impact as compared to glass
fibre production, particularly because of low energy utilisation., during the
production of natural fibres, the energy consumption required to produce a
natural fibre mat, including cultivation, harvesting and fibre extraction, amounts
to just less than one-fifth of the energy required to produce glass fibre mat [49,
50].
The environmental and ecological advantages of bamboo plantations and their use as
reinforcement in composite materials for their beneficial properties in comparison to
glass fibres, e.g. CO2 and energy consumption, will be further discussed in Chapter
7.
3.3.1.1 Bio-composites
The new worldwide environmental legislations and the general tendency towards an
appropriate use of the renewable resources have produced a great interest in the use
of biomaterials that can be environmentally sustainable and compatible with the
environment. In response to this necessity, many countries have been stimulating
“green chemistry” and the production and use of “green products” derived from
nature. Composites are not the exception to this new paradigm and there is, therefore,
considerable interest from manufacturers in developing new “green” composites.
20 Chapter 3
b
The development of bio-composites based on biopolymers and natural fibres as

reinforcements, offers totally or partially biodegradable composites with mechanical
properties that can compete in some applications with composites based on
traditional polymers such as PP. It has been reported by Oksman et al. [51], that the
composite strength of polylactic acid (PLA) matrix and flax fibres are about 50%
better compared to similar PP/flax fibre composites. Nevertheless, the developments
of these materials are still in their infancy with most of the related process or
product research initiated since the late 1980s. Nowadays, only a few fully
commercial applications have been released in the market [52].
As stated by Nishino [33] and Oksman et al. [51], most sustainable plastics cannot
compete economically with conventional petroleum-derived plastics in their present
state. Economically favourable composites, therefore, are expected to be made from
costly sustainable plastics in combination with relatively inexpensive natural
reinforcement fibres. Until now, many studies have been made on the potential of
natural fibres, but just a moderate number of investigations have been made on the
possibilities to use bio-based polymers as matrix for such fibres. One of the
challenges in replacing conventional composite materials with bio-composites is to
design materials that exhibit structural and functional stability during storage and
usage. Then, these materials might be able to degrade once disposed of after their
intended lifetime [53]. A more detail overview regarding this topic is given by
Plackett [52] and Averous et al. [54].
3.3.2 Coupling agents
Studies on composite materials have shown that the properties of composite

materials depend on their individual components and their interfacial compatibility.
Bonding between the reinforcing fibre and the matrix has a significant effect on the
properties of the composite; it means that the stress transfer and load distribution
efficiency at the interface is determined by the degree of adhesion between the
components. Generally, coupling agents facilitate the optimum stress transfer at the
interface between fibres and matrix [2, 35].
According to Gassan and Bledzki [7, 15], to improve the properties of natural
composites, the structure and surface of the natural reinforcing fibres can be
modified in two different ways. In one hand, the physical treatments change
structural and surface properties of the fibres, without changing their chemical
composition, and thereby influence the mechanical bonding with the matrix. In this
case, surface cross-links can be created, the surface energy can be changed and
reactive groups can be generated. Electric discharge methods, such as corona and
cold plasma [55], thermo-treatment [56] and mercerization [57] are examples of
these treatments.
On the other hand, when two materials are incompatible, it is often possible to
increase compatibility by introducing a third material that has intermediate
properties between those of the other two by means of chemical methods. They can
be divided in two subgroups: “compatibilization” and “coupling” methods. The first
one can be polymers with functional groups grafted onto the chain of the polymer.
Graft copolymerization (e.g. maleic anhydride (MA)), it is common treatment used
in thermoplastic matrices [12]. The second method, mostly used in thermoset
polymers, typically utilises a difunctional chemical reagent to enable a chemical
reaction between the fibre surface and the matrix polymer. Isocyanates, silane and
organosilanes coupling agents, are some examples of this treatment [7, 58].
3.3.2.1 Alkali treatment
Alkali treatment in single fibres
Alkalization cannot be considered as an authentic physical modification since the

chemical composition of the plant fibre is altered after treatment, nor can it be
classified under the pure chemical methods, because no additional coupling agent is
introduced into the composite [57]. The alkalisation process (mercerization) is one
of the most commonly used chemical for bleaching and/or cleaning the surface of
plant fibres [16], where the fibres are immersed in a NaOH solution at specific
concentration for a specific period of time. This chemical treatment produces
modifications on the fibres related with the surface morphology and fibre internal
structure that will be reflected in the mechanical properties of the composite material
[59, 60]. The following reaction takes place as result of alkali treatment [12, 19]:
Fibre – OH + NaOH  Fibre – O- – Na+ + H2O
The important modification is the removal of hydroxyl groups in the network

structure. As a result, the penetration of sodium hydroxide into crystalline regions
(cellulose I), cellulose II is formed. This is an irreversible exothermic process
resulting in the modification (rearrangement) of micro-fibrils that specially happens
at high concentrations of the solution [61]. In addition, it has been reported that the
mercerization also removes part of the hemicellulose, certain amount of lignin, wax
and oils covering the external surface of the technical fiber [7, 12]. This creates a
rough fibre surface topography promoting a good mechanical interlocking between
fibre and matrix [16, 40, 62]. It is important to select the right concentration of alkali
22 Chapter 3
b
solution, the temperature and the time of exposure according to each fibre, in order
to avoid weakening or damage of the fibre because of the excessive delignification
[12, 19]. The hemicellulose and lignin in the technical fibre can be considered as the
“glue” of the elementary fibres. This surface cleaning makes the interfibrillar region
less dense and less rigid allowing the fibrils to re-arrange along the fibre major axis,
resulting in a better load sharing between them with further improvement in the
technical fibre strength [58, 63, 64]. Generally, this benefit was observed together
with an improvement of the aspect ratio (L/d) of the technical fibres [65, 66], and
can be seen in Table 3-7 for some cases when comparing the fibre strength before
and after alkali treatment. For technical bamboo fibres, it was also reported that the
increase of the fibre strength and stiffness due to the alkali treatment could be
caused specifically by the removal of waxes and some amount of lignin from the
fibre surface, resulting in a slight reduction of the fibre diameter but more
importantly, a diminution of weak non-cellulosic materials [67].

Alkali treatment
Fibre Fibre Tensile Tensile Tensile Tensile
(% NaOH, time Ref.
Diam. Strength Modulus Strength Modulus
and temperature)
(µm) (MPa) (GPa) (MPa) (GPa)
Sisal 150 391 15 496 13 2%, 4h at 60°C [41]
Sisal - 500 16 810 27 16%, 48h at RT [18]

Sisal - 365 - 481 - 4%, 24h at RT [68]
Hemp 26 514 25 347 20 10%, at RT [69]
Curauá 94 913 30.3 736 27 5%, 1h at RT [70]
Palm 180 - 208 233 7 366 5 6%, 3h at 95 °C [71]
Table 3-7. Strength and stiffness values after single fibre test before and after alkali treatment.
Effect of the alkali treated fibres on natural fibre composite properties
Several studies show the benefits of alkali treatment on the composite performance
in flax [37, 43], jute [7, 39], hemp [17, 40], kenaf [40] and sisal [41] fibres. Tables
3-3 and 3-4 illustrate the positive effect of this treatment on the tensile and flexural
properties in a number of natural fibres and thermoset matrix composites. Besides of
the mechanical interlock, this improvement can be explained not only by the
eventual improvement of the intrinsic properties of the technical fibres, but also, by
their increased aspect ratio after treatment [38, 64, 72]. Moreover, it is a result of an
improved adhesion due to the activation of the hydroxyl groups on the fibre surface,
thus increasing the number of possible reaction points for better chemical bonding
[64].
According to some studies [10, 57], low concentrations of NaOH (<3%) are not
strong enough to cause significant swelling of the cell wall but it is sufficient to
remove non-cellulose (impurities) on the surface of the fibre. Moreover, it was
found that the mechanical properties reach an optimum for a certain NaOH
concentration followed by a drastic decrease. Van de Weyenberg [57] treated flax
fibres with NaOH at several concentrations. For the longitudinal strength, the
highest improvement was achieved with a 3% NaOH concentration. For the
longitudinal modulus, the transverse strength, and the transverse modulus, the 1%
NaOH concentration presented the best results. Jacob et al. [73] examined the effect
of NaOH concentration (0.5, 1, 2, 4 and 10%) for treating sisal fiber-reinforced
composites and found that the maximum tensile strength was obtained with the 4%
NaOH treatment at RT. A similar result was found by Mishra et al. [74], who
reported that 5% NaOH treated sisal fiber-reinforced polyester composites gave
better tensile strength than the 10% treated composites.
3.3.2.2 Maleic anhydride
Maleic anhydride (MA) grafted polypropylene (MAPP) is a coupling agent that also
acts as a compatibilizer. It consists of long polymer chains with a MA functional
group grafted onto one end of the PP chain. MAPP acts as a bridge between the
nonpolar polypropylene matrix and the polar fibres by chemically bonding with the
cellulose fibres through the MA groups, and bonding to the matrix by means of
polymer chain entanglement [69].
The mechanism of reaction of maleic anhydride with PP and fiber can be explained
as the activation of the copolymer by heating (170 °C) before fiber treatment (Figure
3-4a) and then the esterification of cellulose fiber (Figure 3-4b) [58]. The coupling
occurs at one side through covalent bonds and hydrogen bonds between the
hydroxyl groups of the cellulosic fibres and the acidic anhydride groups of MA,
while on the other side a co-crystallisation and entanglement occurs with the
polymer [75].
Compatibilization with MA in sisal [75], flax [46] and jute [76, 77] fibre composites,
results in composites with enhanced mechanical properties and reduced water
absorption due to an increased interaction between the anhydride groups and the
hydroxyl groups on cellulose. In this case, a better wettability and higher fibre-
matrix interfacial adhesion is promoted, as a result of the chemical bonding.
24 Chapter 3
b
a. b.
Figure 3-4. Schematic representation of the interphase between MAPP and the hydroxyl groups of a cellulose
fibre in two steps: a) activation of the copolymer by heating (T= 170 °C) (before fibre treatment) and b)
esterification of the cellulose [10, 19, 58].
According to Chattopadhyay et al. [78], short bamboo fibre reinforced

polypropylene containing 5 wt% MAPP has a 37% increase in impact strength, 81%
increase in the flexural strength, 150% increase in the flexural modulus, 105%
increase in the tensile strength, and 191% increase in the tensile modulus. Chen et
al. [79], found that the tensile modulus, the tensile strength, and the impact strength
of short bamboo fibres and PP increased significantly incorporating 0.5 wt % maleic
anhydride in the PP matrix. In Tables 3-5 and 3-6 show flexural and tensile UD
natural fibre thermoplastic composites with different concentrations of MA.
3.4 Limitations in the use of natural fibre composites
3.4.1 Moisture absorption
The effect of moisture on the mechanical properties of natural fibre composites

during long term service in outdoor or exposed to a moist environment, is nowadays
an active field of research [65, 80]. According to Jindal [81] and Stamboulis et al
[5], the intrinsic hydrophilicity character of technical natural fibres is a major
obstacle which prevents the extensive application of natural fibres in composites.
This originates moisture absorption and may cause dimensional instabilities
(swelling) in the produced parts, creating some stresses at the interface and
preventing the fibre properties to be transferred into the composite [57].
By reducing the moisture sensitivity with a treatment, the technical fibre becomes
more hydrophobic [82] by the reduction of hemicellulose, largely responsible for the
water absorption behaviour exhibited by the fibres [83]. Hence, the compatibility
with hydrophobic polymeric matrices such as PP increases beneficing the interfacial
adhesion, and therefore the void content in the composite can be reduced, giving less
chance to moisture to penetrate into the material.
3.4.2 Thermal degradation
Natural fibres are lignocellulosic polymers subjected to thermal degradation at

elevated temperatures, e.g. during composite processing [7, 84]. The majority of
natural fibres have low degradation temperatures (~185 °C), which make them
inadequate for processing temperatures above 200 °C [85]. For this reason natural
fibres limit the choice of the polymers as a matrix [76]. Both thermoset resins (e.g.
epoxy and unsaturated polyester resins with curing temperatures below 180 °C), and
some thermoplastic polymers (e.g. PP and MAPP) can be used with all range of
natural fibres [57]. It is thus of practical significance to understand and control the
thermal decomposition process of natural fibres to keep the thermal stability of the
components during the manufacturing process [86]. The most important aspects to
consider regarding this topic will be explained in the next paragraphs.
3.4.2.1 Thermal degradation in natural fibres
It has been established that no fibre damage occurs at temperatures below to 160°
[87] independent of the exposure time. Li et al. [19] stated that thermal treatments
between 170°C and 180°C will not affect the tensile properties if such temperatures
are maintained for less than one hour. According to Wielage et al. [88], Rachini et
al. [89] and Placet [30], the first degradation typically occurs at temperatures above
180°C and more severe damage occurred within a temperature range between 215
and 310°C [84]. For natural fibres, the thermal degradation is strongly influenced by
its chemical composition [90]. A typical thermogravimetric analysis (TGA) and its
first derivative (DTG) curves for the pure main components of natural fibres, such as
cellulose, hemicellulose and lignin, are shown in Figure 3-5.
Cellulose
Hemi-
cellulose Lignin
Figure 3-5. Thermal analysis for pure hemicellulose, cellulose and lignin determined by TGA and DTG [91].
26 Chapter 3
b
According to Yang et al. [91], these three components behave significantly different
behaviour during pyrolysis in air environment. Hemicellulose starts with an early
decomposition between 220 and 315 °C, with a maximum mass loss rate (0.95
wt.%/°C) at 268 °C. Cellulose begins at a higher temperature range (315 - 400 °C)
with the maximum weight loss rate (2.84 wt.%/ °C) attained at 355 °C. Lignin was
found to be the most difficult one to decompose, doing it slowly under the whole
temperature range from ambient to 900 °C. Similar results for those three
constituents were found by Martin et al. [92] and Rachini et al. [89]. Generally, the
contribution of hemicellulose and cellulose on the thermal degradation is more
important than the one of lignin [89]. A typical TGA (and its corresponding DTG)
thermal decomposition process for a natural fibre is shown in Figure 3-6.
Figure 3-6. Typical thermogravimetric decomposition process of a natural fibre [1].
For a typical natural technical fibre it has been found that the thermal degradation
registered during TGA, with a heating rate of 10 °C/min, can be divided in three
individual stages according to several studies, see Table 3-8. In this table typically
the first peak occurs at temperatures below of 100 °C and corresponds to the initial
moisture evaporation. The second peak is attributed to the hemicellulose
decomposition that occurs between 220 and 290 °C. Cellulose degradation shows
the third peak between 310 and 400 °C. For flax fibres it was found that 2nd and 3rd
peaks occurred at around 344 and 459 °C respectively. Significant lignin
decomposition can be expected at around 425 °C, nevertheless, this 4th peak is not
always visible [93].
Temperature (°C)
Fibre 1st 2nd 3th Techni- Flow rate Heating Ref.
peak peak peak que (mL/min) rate
(°C/min)
Sisal 75 222 310 DTG 20 10 [92]
Curauá 86 225 363 DTG 50 10 [85]
Flax 90 344 459 DTG 20 10 [94]
Hemp 87 253 395 DTG - 10 [89]
Jute <100 287 353 DTG - 10 [77]
Table 3-8. Initial weight loss and initial temperature degradation for some natural fibres and PP determined by
DTG in air environment.
Fibre thermal decomposition mechanism
At relatively high temperatures, the thermal degradation is a complex process that

mainly depends on the structure and the chemical composition of the natural fibre
and the intensity of the treatment, e.g. temperature and time of exposure. In air, the
thermal degradation mechanism of cellulosic fibres can be divided in two stages (see
Figure 3-7):
 At low temperatures, between 100 °C and 250 °C, some of the changes in
physical properties of the fibres can be explained in terms of alterations in the
structure such as depolymerisation, hydrolysis, dehydration and
decarboxylation and recristallysation [87]. Besides that, the formation of free
radicals is also noticed. They can contribute to the formation of
hydroperoxide groups, responsible to a large extent to the depolymerisation
of cellulose (by bond scission). Besides depolymerization, a variety of
oxidation and decomposition reactions take place, leading to the formation of
carbonyl, carboxyl, lactone and aldehyde functional groups [93]. All these
reactions, specially the oxidative degradation of cellulose, takes place
primarily in the amorphous region of the cellulose [88] giving a drastic
decrease in molecular weight and consequently in mechanical performance.
 At higher temperatures (~300 °C), the cellulose thermal degradation is caused

by the destruction of hydrogen bridges, changes in crystallinity, and some
processes mentioned before, i.e. the formation of free radicals, carbonyl
groups, and carboxyl groups but in a higher scale, especially in air [93]. The
decomposition under inert environment is mainly due to the presence of free
radicals [89, 95].
28 Chapter 3
b
Figure 3-7. Schematic stages for the thermal degradation of a cellulosic fibre [61, 87, 93, 96].
Single fibre properties after thermal treatment
Kohler et al. [97] evaluated the tensile strength of flax treated at 200°C during
different exposure times. A drop in tensile strength was observed after 30 minutes.
Mieck et al. [98] reported major damage to flax fibres after four minutes of exposure
at 240°C. Additionally, in a study carried out by Van de Velde et al. [94], flax
technical fibres were treated at different temperature-time of exposure couples (120
and 180°C during 15, 30, 60 and 120 minutes). As a result, no significant decrease
was observed in tensile strength after 120°C treatment, independent of the time of
exposure. However, at higher temperatures and longer exposure times the general
tendency is a remarkable decrease of the mechanical properties. Moreover, it has
been reported that relatively low temperatures (150 °C) and longer exposure times
(up to 4h), significantly reduced the tenacity of natural fibres [87]. It was also
noticed that alkali treated natural fibres are more thermally stable compared to the
untreated fibres due to the partial removal of amorphous cellulose, known to have
low thermal resistance [1, 99].
For bamboo technical fibres, a study of Ochi et al. [100] revealed that for thermal
treatments below 140°C there is no significant decrease in mechanical properties.
Around temperatures of 160°C, a gradual decrease of the tensile strength is noticed
after 30 min of exposure. At temperatures between 180°C and 200°C, the tensile
strength drops by around 25% during the first 10 min and tends to stabilize after 30
min. The Young’s modulus for the bamboo fibres was found to be independent of
heat treatment in the mentioned ranges of temperature and time. Figure 3-8 shows
the decrease of the technical fibre tensile strength when exposed at different
temperature-time of exposure couples for bamboo and flax fibres. Another
consequence of the heat treatment is that lignin migrates to the fibre surface due to
its softening, that begins at around 214°C [89, 101]. This phenomenon was
examined by AFM observations in the bamboo technical fibres surface, before and
after they were treated in an autoclave (3 bar at 150 °C) for around 1 hour [102].
When this occurs, the hydrophobicity of the fibre surface increases [55] leading for a
better compatibility with thermoplastic matrices and hence reduced moisture
sensitivity in the final composite material.
Figure 3-8. Effect on the fibre strength of thermally treated flax and bamboo technical fibres, when exposed at
different temperature-time of exposure couples [87, 88, 97, 98, 103].
Effect of air and inert atmosphere during thermal treatment
In general, the degradation reactions are much less significant in an inert atmosphere
than in air. Thermal resistance of flax [94], jute [86] and hemp [56, 89] fibres was
found to be higher under inert atmosphere since the process of decomposition of
cellulose occurs much quicker in air environment. According to Prasad et al. [56],
after giving a heat treatment of 30 minutes at 180 °C in air and nitrogen environment,
an undesirable general degradation of the mechanical properties, in some cases of
around 15%, was noticed especially in air. Nevertheless, technical fibre
30 Chapter 3
b
defibrillation occurs in both cases. By comparing untreated hemp fibres with the
inert thermally treated, no remarkable degradation in terms of mechanical properties
was noticed, with a maximum strength difference of around 7%. In another study,
the same author [101] noticed under optical and SEM observations, a more severe
damage on the primary wall when the technical fibre was treated in air, compared
with the effects under nitrogen environment when the thermal treatment was carried
out at 220 °C for 30 minutes.
3.4.2.2 Thermal degradation in natural fibre composites
The main decomposition range of various natural fibres overlaps with the processing
temperatures of some thermoplastics. As mentioned before, the thermal
decomposition for lignocellulosic fibres is estimated to start around 200 °C and
strongly depends of the exposure time. Classically, during the manufacturing of
natural fibre composites, the process temperature is chosen in accordance with the
rheological properties of the polymer matrix, e.g. viscosity [30, 92]. This
temperature must not lead to a loss of the fibres’ integrity. Table 3-9 gives an
overview of commonly used thermoplastics with their melting temperature. Until
now, thermoplastics with a low melting and shaping temperature have been used as
matrix in natural fibre composites (e.g. PP, PE, PVDF).
Melting Shaping
Thermoplastic
Temperature [Tm] (°C) Temperature* (°C)
Polyethylene (PE) 120-145 150-200
Polypropylene (PP) 165 185-200
Polyvinylidene Fluoride (PVDF) 170 190-230
Polyamide 6 (PA6) 220 240
Polybutylene Terephthalate (PBT) 225 250-280
Polyethylene Terephthalate (PET) 255 280-300
Polyetheretheketone (PEEK) 343 365-380
Table 3-9. Common used thermoplastics and their melting and shaping temperature [104]. *Common used
temperature to produce parts in manufacturing, usually between 25 to 30 °C above Tm.
During production of natural fibre reinforced plastics, shorts periods of exposure to

high temperatures are allowed without a significant drop in the properties of the
fibre [94, 98]. This is a positive aspect when considering industrial processes such as
pre-impregnating fibres with thermoplastics (“prepregs”). For that, the fibres are
stabilized by a thermoplastic in molten state exposing the fibres at high temperature
for a relatively short time. Thermoplastic prepregs are matrix-fibre combination with
the exact amount of polymer necessary to achieve the desired fibre volume fraction.
They are ready to be used for the manufacturing of laminated composite structures
with well defined fibre orientations.
The use of inert atmosphere can be adapted for the manufacturing of natural fibre
composites in a well controlled process at relatively low cost [56]. According to
Wielage et al. [88], the consolidation process under a non-oxidative environment
reduces the decomposition not only of the natural fibres, but also, of the
thermoplastic matrix itself, e.g. PP and MAPP. Nevertheless, the latter statements
had been based on single technical fibre analysis. To the best author’s knowledge,
no studies evaluating the mechanical behaviour of natural fibre thermoplastic
composites manufactured under inert atmosphere have been made.
Thermal oxidative degradation of polypropylene
At high temperatures the components of the chain back bone of the polymer start to
separate (chain scission), resulting in a reduction of the molecular weight and
mechanical properties over time. In addition to thermal degradation, the chains can
be broken mechanically by shear stresses induced during the process [69]. The main
mechanism of degradation of PP is random-chain scission via free-radical transfer
process (auto-oxidation) [105]. This process involves several steps such as initiation,
propagation, chain branching and termination [106]. The thermal degradation of PP
is highly dependent of the environment [88]. In nitrogen, the process starts around
300 °C, with a maximum weight loss at 431 °C. However in air, the degradation
process shifts towards lower temperatures (235 °C) [86].
Antioxidants and stabilisers are usually added to the formulation of the PP to

prevent thermal oxidative degradation during processing and service (special care
must be taken for food or medical applications) [105]. In the case of the stabilizers,
they are used to keep the polymer chains and the original molecular structure intact,
thus enabling properties such as strength, stiffness and toughness to be retained over
a longer period of time. Most stabilizers slow down polymer degradation by reacting
rapidly with available peroxide radicals to produce another less active radical. Other
stabilizers interrupt the thermal oxidative degradation cycle at the hydroperoxide
propagation step, to slow down or prevent the cycle from completing [69]. Primary
antioxidants inhibit the oxidation reaction by combining with the formed free
radicals. Hindered phenolics (e.g. butylated hydroxytoluene (BHT)) are commonly
used as primary antioxidants. Secondary antioxidants, also called peroxide
decomposers, inhibit oxidation of PP by decomposing hydroperoxides. Phosphites
and thioesters are commonly used as secondary antioxidants, and are usually
combined with primary antioxidants to produce a synergistic protection against
oxidation [104].
32 Chapter 3
b
3.5 Bamboo plants
Bamboo plants belong to the Gramineae family which essentially means that they
are grasses (not trees), with more than 90 genera and over 1200 different species
around the world. Bamboo plants are naturally distributed in tropical and subtropical
areas, between 46° North and 47° South latitude and are thus commonly found in
Asia, Africa and Latin America. However, it has been introduced to the north,
including Central America [107, 108], see Figure 3-9.
Figure 3-9. Distribution of woody bamboo in the world. Source: www.eeob.iastate.edu. According to
Lobovikov et al. [107] the percentage world contribution from the different continents is: 65% Asia, 28%
America and 7% Africa.
The bamboo species Guadua angustifolia belongs to the woody bamboos and is the
largest and economically most important bamboo in the Western Hemisphere [109],
especially in South America, where Colombia has more than 51.500 Ha of
plantations [110]. According to Londoño [108], bamboo guadua is not only one of
the three largest species, but also one of the most important bamboo in the world due
its excellent mechanical properties with several advantages:
- High growth rate (11 to 21 cm per day). Culms reach their final height in the
first 6 months of growth, and come to maturity when they are 4 to 6 years old
[111]. In comparison with a wood tree, bamboos have a very short growth
cycle and major productivity [112]. Moreover, bamboo forests control
erosion and regulate the level of rivers;
- This plant plays and important economic role in the places where it grows. In
South America, including Colombia, it has been used specifically for building
and handicraft purposes, but unfortunately its exploitation has not had an
additional value and only a few technological developments have been
associated with this natural resource;
- Guadua angustifolia is one of the tropical species that have been identified as
having great potential to fix atmospheric carbon dioxide. The carbon fixation
estimated for 400 clumps/ha of Colombian bamboo, for a growth period of 6
years is 54.3 tonnes [110]. Therefore, bamboo is an effective plant in terms
of global warming management in comparison with young forests. According
to Lybeer et al. [113], bamboo is one of the most important non-timber forest
products and one of the more important agricultural non-annual plants in the
world.
3.5.1 Morphology of bamboo Guadua angustifolia culm
In general terms, the bamboo culm is a thin and hollow cylinder divided in to several
internodes each one with its own internal diaphragm that separates the internodes.
The fibres in the nodes are not unidirectional but randomly oriented in an entangled
manner giving isotropic properties to the nodes and providing additional
reinforcement to the culm [114] (see Figure 3-10). The function of these nodes is the
prevention of buckling due to bending and they may also play the role of axial crack
arresters due to the fact that the fibres in the internodes are oriented along the
bamboo's culm. Among plants, bamboo has a unique structurally smart structure that
helps it to withstand extreme natural environmental conditions and it is a unique
example of a natural unidirectional fibre reinforced composite [34].
Figure 3-10. a) Different segments of bamboo G. angustifolia plant, b) parts in which is divided the bamboo
culm and c) discontinuity of the bamboo fibre within the node (discontinuity drawing according to [115]).
34 Chapter 3
b
3.5.1.1 Diameter, wall thickness and internode length of the bamboo culm
As shown in Figure 3-11a and b, the bottom segment has the largest diameter and
wall thickness, but the average internode length is lower in comparison with the
other segments. These characteristics make the base of the culm suitable to carry
compressive loads. The diameter and thickness of the wall decrease from the bottom
to the top [115-117]. It was found for Moso bamboo (Phyllostachys pubescens) that
the length of the internodes reaches the maximum length value of 32 cm in the
middle section of the bamboo plant [116], see Figure 3-11c.
Figure 3-11. a) Internodal length for bamboo G. angustifolia, adapted from [118], b) Wall thickness variation
(G. angustifolia) across the culm, adapted from [117] and c) intermodal length vs intermodal number for
Moso bamboo (Phyllostachys pubescens) [114].
For bamboo G. angustifolia the maximum average culm diameter is found on the
bottom part of the culm. The average biggest wall thickness and longer internode
distance, are found in the bottom and middle part of the culm respectively [117], see
Table 3-10. For this species, the total high of the bamboo plant range between 20
and 23 m, whereas the commercial height is in average 15 m [117, 118], see Figure
3-10a.
Segment of the Average Average wall Average internode

culm* diameter thickness length
(cm) (cm) (cm)
Bottom 11.5 1.8 20.2
Middle 11.1 1.3 34.9
Top 5.9 0.9 33.4
Table 3-10. Average values for the main anatomical variables in bamboo G. angustifolia in three different
segments [117]. *See Figure 3-10a.
3.5.2 Anatomy of the bamboo culm
Figure 3-12. Bamboo G. angustifolia microstructure [67]: (a) bamboo culm, (b) cross-section of the culm
showing the fibre distribution through the wall thickness, (c) vascular bundle, the main anatomical constituent
of the plant, composed of vesselsI, floem II, protoxilemIII, parenchyma tissueIV, and fibre bundlesV, (d) bamboo
technical fibre composed by several elementary fibres, (e) elementary fibres with pentagonal or hexagonal
shape, and (f) model of polylamellae structure of a thick-walled elementary fibre proposed by Liese [115],
where, in the thick lamellae (L1–L4), the cellulose fibrils are oriented at a small angle to the fibre axis,
whereas the thin ones (N1–N3) show mostly a more transverse orientation, P, primary wall, O, external sheet
of secondary wall.
The gross anatomical structure of a transverse section of any culm internode is

determined by the shape, size, arrangement and number of the vascular bundles.
They are the basis to understand the mechanical behaviour in the different sections
of the culm because they are closely related to the fibres [115]. As can be seen in
Figure 3-12b and c, the vascular bundles can be detected easily because of their dark
colour. These bundles are distributed densely in the outer region of the wall and
sparsely in the inner region, which contains mainly parenchyma tissue with fewer,
36 Chapter 3
b
large vascular bundles [116]. Through the height, they are concentrated in the upper
part of the culm compared with the base. For bamboo Guadua, in the periphery zone
of the culm (2.5 mm of thickness) the number of vascular bundles per cm2 is in
average between 346 and 530, whereas in the interior layer, for the same thickness,
there are 52 to 96 vascular bundles [115, 117].
3
Around the vascular bundles there is a spongy tissue called parenchyma, see Figure
3-12c, composed of large cells that are mixed with some short cubic cells. Inside
these cells, some nutrients in the form of starch grains are stored filling 50 to 70% of
the tissue [119]. The amount of parenchyma tissue decreases along the culm length
and across the cross section (decreasing from the inside to the outside of the culm
wall) [114, 115, 120].
3.5.3 Bamboo fibres
A typical cross section of a technical bamboo G. angustifolia fibre shows an

irregular “beam” shape (Figure 3-12c and d), having a size variation depending on
their position across the surface culm section. Each fibre, irrespective of its position,
contains several elementary fibres [67]. The cross-section of these primary fibres is
either pentagonal or hexagonal and they are arranged in a honeycomb pattern,
separated by a thin wall of matrix (Figure 3-12e) [34, 67, 121].
According to Amada et al. [116], the average density of bamboo is 0.8 g/cm3 and the
average fibre volume percentage of fibres in a piece of bamboo is approximately
32% taking into account the reduction of the diameter and wall thickness over the
height. Considering the gradient of the fibres on the wall thickness, the fibre volume
fraction (Vf) is about 15 – 20% and 60 – 65% at the inner and outer surface
respectively. Fibres constitute about 40-50% of the weight of the total tissue and
between 60 - 70% of the weight of the total culm tissue [115]. As mentioned before,
the bamboo culm resembles a unidirectional fibre reinforced composite as is shown
in Figure 3-13.
Figure 3-13. Cut bamboo (G. angustifolia) culm in the field, showing some bamboo technical fibres vertically
oriented. The bamboo culm resembles a unidirectional fibre reinforced composite.
3.5.3.1 Microstructure and chemical composition of bamboo fibres
According to Liese [122], the elementary fibre size decreases from the bottom to the
top along the culm length. Across the culm wall, the fibre length often increases
from the periphery, reaching its maximum at about the middle and decreases toward
the inner part. The fibres in the inner part of the culms are always much shorter (20 -
40%). The elementary fibres are long and tapered at both ends with a small hole in
the centre called lumen, see Figure 3-12e [123]. Longitudinally within the internode,
the shortest elementary fibres are always near the nodes and the longest in the
middle part, this a variation of more than 100%. The elementary fibre length is
positively and strongly correlated with the average fibre diameter, cell wall
thickness (layers) and internode diameter. For bamboo G. angustifolia the mean
fibre elementary length is 2.1 mm and its diameter varies from 10 to 25 µm [124].
According to several studies [115, 120, 125, 126], bamboo fibres have a specific
layered structure that is unique amongst the vegetable fibres. This polylamellate
structure is depicted in Figure 3-12f. It consists of an alternation of thick and thin
layers. In the thick layers, the fibrils are oriented at a small angle towards the length
axis of the fibre, corresponding with a microfibril angle between 2 and 10°. The thin
layers show a more transverse orientation of the fibres. The amount of layers
depends on the position and maturity of the fibre. The lamellar structure leads to the
high mechanical characteristics of bamboo fibres.
According to Londoño et al. [117], the composition of the culm of the Guadua
angustifolia, on average, is 51% parenchyma tissue, 40% fibre and 9% conducting
cells. Compared to other tropical and subtropical species, this species exhibits a
typical percentage of fibres. Bamboo consists of about 50 - 70% of cellulose, 30%
38 Chapter 3
b
pentose and 20 - 25% lignin, and typically has a silica content ranging from 0.5 to
4 % (mostly is deposited in the skin zone). There are some differences in these main
constituents according to the species’ conditions of growth [120]. After cellulose,
lignin represents the second most abundant constituent in the bamboo. Bamboo
lignin is a typical grass lignin, which is built up from three phenyl-propane units, p-
coumaryl, coniferyl and sinapyl alcohol, interconnected through biosynthetic
pathways [122, 127]. The lignification and the thickening of the cell walls are the
main changes during maturation, both increase downward from top to bottom,
whereas lignification increases from inside to outside. Full lignification of bamboo
culm is completed within one growing season [128].
3.5.3.2 Mechanical properties of bamboo technical fibres
Bamboo fibres are often called “natural glass fibres” because of their properties [25,
67]. Among the well-known natural fibres, bamboo has a favourable combination of
low density and high mechanical properties giving good specific properties (i.e.
stiffness and strength). Several studies have been carried out to characterise single
technical bamboo fibres. A summary of these results, with the corresponding
extraction method, is presented in Table 3-11. Nevertheless, these results should be
taken as indicative values because in many studies the authors do not specify
important details such as: the bamboo species, the age and growing conditions of the
plant, part of the culm where the fibres were extracted from, etc. Also, some missing
information about the methodology for single fibre testing was omitted, including:
the testing gauge length, the loading rate, the applied technique for the determination
of the fibre’s cross section, etc. All of this makes impossible a direct comparison
between the different studies in single bamboo fibre properties.
Numerous investigations on strength properties in relation to the culm age have not
shown consistent results [129]. Whereas some reports show higher strength values
for one year old bamboo culm compared to older ones, other studies revealed a
general increase with age up to 6 to 8 years old, followed by a decrease in all
strength properties in culms of ten years old [130]. It is commonly accepted that a
three-year-old bamboo is ready to be used for industrial purposes, but a study carried
out by Lybeer et al. [113] concludes that younger bamboos would be also
appropriate for industrial purposes.
Fibre physical properties Mechanical properties

Extraction Diameter Length Density Strength Young’s Strain to
method (µm) (mm) (gr/cm3) (MPa) Modulus failure
(GPa) (%)
Mechanical -
100 - 200 - 0.8 - 1.1 391 - 713 18 - 55
(beating)
Mechanical 220 –
100-600 - 270 - -
(rolling mill) 270
Mechanical
- - 1.4 450-800 18 - 30 -
(grinding)
Mechanical 262 ± 160 - - 420 ± 170 38 ± 16 9.8

(pin roller)
Steam
- - - 516 17 -
explosion
Steam
366 ± 74 ~250 1.3 726 33 2 ± 0.6
explosion
Steam
15 - 210 - - 441  220 36  13 1.3
explosion
Steam
195 ± 150 - - 308 ± 185 26 ± 14 2.5
explosion
Manually
300-900 “Long” - 106-203 - -
stripped
90- 400 HC:
Chemical - 0.9 341 19 1.7
200-350
Chemical 270 10 1.3 450 18 -
Chemical 230 - - 395 ± 155 26 ± 14 2.8
Chemical + 50-400
>10 0.8 - 0.9 645 - -
compression HC: 150-250
Chemical +
HC: 50-100 120-170 - 370 - -
roller mill
No mentioned 150-350 - - 200 - 800 - -
Table 3-11. Physical and mechanical properties of bamboo fibres according to the extraction method [13, 32,
81, 100, 131-138]. HC: highest fibre diameter concentration.
Moreno et al. [139], performed a mechanical characterization of bamboo G.

angustifolia fibres, taking into account three different variables: age, height and wall
thickness of the culms. With respect to the positioning of the fibre on the wall
thickness, fibres extracted from the outer third of the wall present the lowest values
of tensile strength (Figure 3-14). Considering the age of the culms, young and
mature culms present similar strength values (655 and 663 MPa respectively). The
best condition for obtaining the highest strength values can be reached by extracting
the fibres at 2/3 of the wall thickness from the inside of the wall thickness, from a
40 Chapter 3
b
culm of 3 years old (769 MPa). The total average strength was found to be 643 ± 12
MPa (100 mm gauge length). The tensile apparent Young’s modulus did not present
significant differences across the height of the culm with an average value of 27 GPa.
Nevertheless, this result contrasts with the study of Bangarshetti et al. [140], who
found a significant variation in strength properties of bamboo along the length and
also across the cross section.
Figure 3-14. Both a) tensile strength and b) tensile modulus properties of bamboo (G. angustifolia) fibres
depending on the age of the culm and their position in the wall thickness [139]. * From the outside towards
the inner side.
In general, the fibre extraction process and further fibre manipulation significantly
determine the fibre quality and mechanical properties [8, 30, 141]. An adequate
procedure results in a much better and easier extraction of the fibres from the plant,
and therefore, minimizes the mechanical loading and subsequently damage to the
fibres. Mechanical overstraining of the fibres during the extraction process can result
in the formation of defects [8]. Different methods for extracting bamboo fibres are
explained in the next paragraphs.
3.6 Extraction of bamboo technical fibre
It is technically difficult to obtain large bamboo technical fibre quantities with a

consistently uniform shape (i.e. long fibres). Tung et al. [142] reported that it is very
difficult to extract very thin, but long fibre bundles without causing serious damage
to them by the processing. Little effort has been devoted to the extraction of long
bamboo fibres because of the difficulty of extracting good quality (technical)
bamboo fibres from the culm. As a consequence, there is a limited availability of the
fibre and only a few research groups have investigated this material as a potential
UD reinforcement in polymer matrix composites [13, 34, 81, 134, 136, 143, 144].
To practically apply the benefits of bamboo fibres, it would be necessary to develop

a process to extract long high quality fibres from bamboo plants. This extraction
process might be carried out in a standardized way, and should be able to compete
with existing methods for other natural fibres in terms of production volume, cost,
environmental impact (e.g. waste water and chemicals) and uniform consistency in
fibre properties. Nowadays, there are several extraction methods used for extracting
bamboo fibres such as: steam explosion, chemical extraction, retting, mechanical
methods using hammer mills, toothed rollers, compression (crushing), shearing, etc.
Table 3-11 shows various extraction processes and the obtained fibre characteristics.
From this comparison, it is clear that the fibre separation process determines the
morphology and mechanical properties of the extracted fibres. Figure 3-15 shows
bamboo fibres used in different studies as reinforcement in polymeric matrices.
a. b. c.
d. e. f.
g. h. i.
Figure 3-15. Bamboo fibres used in other studies extracted by several methods. Chemical extraction (alkali
treatment): a) [145]*, b) [146]*, c) [147]*, d) 1% concentration and 70° for 10 hours [137]. Mechanical
extraction: e) milled bamboo [144], f) [148]*, g) pin rollers [137]. Steam explosion plus mechanical process:
h) 100 min 170°C, 0.8 MPa, 8 times [13], i) 175 °C and 0.8 MPa for 60 minutes. The process war repeated 9
times [137]. *Details about the process are not specified.
42 Chapter 3
b
3.6.1 Steam explosion
This technique is a very common method to extract natural fibres in general (e.g.
flax [149] and hemp [150] fibres). It has been known as an effective method to
separate the lignin from the woody materials and is also applied in the pulp industry.
A schematic view of the process is shown in Figure 3-16. A typical procedure to
obtain the fibres consists first in removing both the nodes and the skin from the culm
followed by the cutting of the internodes in strips. Then, the bamboo strips are
placed into an autoclave with over-heated steam at 175°C and 0.8 MPa for around
60 minutes and suddenly depressurizing the autoclave to atmospheric pressure. This
operation causes micro steam explosions inside the parenchyma, facilitating the
extraction of the fibres of a few hundreds of microns diameter. The process needs to
be repeated several times (9-12 cycles) in order to obtain fine technical fibres [142].
Figure 3-16. Schematic process for the extraction of bamboo fibres by steam explosion.
It has been stated that after using this extraction process, a further mechanical
process is necessary to improve the quality of the bamboo fibres [133]. The need of
this additional step, was also reported by Okubo et al. [13], because the technical
fibres were not effectively separated and a large amount of impurities and
parenchyma remained on the surface, see Figure 3-17a. Therefore an additional
mechanical process consisting in rubbing the technical fibres was applied for further
separation, see Figure 3-17b. Ashimori et al. [20] used steam explosion repeating the
cycle 10 times. They also complemented the extraction with a mechanical process,
where the bamboo strips were scratched to refine the fibres, in order to reach
diameters between 125 and 425 µm. A second attempt from the same author with
the aim to obtain finer bamboo fibres, consisted in freezing the steamed-exploded
fibres and then crush them mechanically. They found that freezing fibres did not
cause splitting in fibre bundles and it is not very effective method to reduce the fibre
diameter, even at cryogenic temperatures.
Soft tissue Elementary fibre

.
a. b.
Figure 3-17. SEM micrographs of a) bamboo fibres extracted by steam explosion and b), same fibres after
additional mechanical process (rubbing) in order to clean the fibre. After this process the fibres appeared as a
“cotton” fibres obtaining diameters between 10 – 30 µm [13].
A fibre quality comparison between steam exploded and two types of mechanical
extraction (i.e. crushing and shearing) methods for bamboo fibres was made by
Ogawa et al. [144]. They founded a great difficulty in controlling the diameter,
length and aspect ratio of the bamboo fibres by those three methods. The crushing
method (Figure 3-18a) had a lower efficiency due to the necessity of a screening
process (by sifting) to obtain uniform diameter. In steamed fibres (Figure 3-18b), the
heat may damage the fibres, decreasing the mechanical properties. A reduction in
strength of about 20% due to the heat exposure during the steam explosion
technique was reported by Okubo et al [25]. Finally, fibre dimensions obtained by
the shearing (rubbing) method varied widely (Figure 3-18c). According to Phong et
al. [137], the steam-explosion technique offered lower moisture contents than
untreated fibres at all relative humidity levels (i.e 50, 60, 70, 80 and 90% RH). This
can be attributed to the chemical changes on the fibre during the thermal treatment.
Liu et al. [151] compared the fibre diameter distribution of bamboo (D.
membranaceus) of two batches of fibres obtained by steam explosion and
mechanical extraction (rolling mill). In both cases, before the extraction, the bamboo
strips were treated with alkali to facilitate the extraction process of the fibres. In
general, under visual inspection, the steamed fibres had a dark color and rough
surfaces. In contrast, the ones extracted mechanically presented brighter color and
smoother surfaces. The fibre diameter distribution of these fibres is shown in Figure
3-19. In this study the mechanical process gives a normal distribution with a peak
value between 120 and 140 μm. On the other side, no regularity in fibre diameter
distribution can be found for the steam exploded fibres, giving higher diameters.
44 Chapter 3
b
Figure 3-18. Bamboo fibre characteristics after three different extraction processes [144].
Figure 3-19. Fibre diameter distribution of bamboo (D. membranaceus) fibres extracted by mechanical
extraction and steam explosion techniques. A picture of mechanically extracted fibres (above) and steam
exploded fibres (down side) can be seen. The fibre diameter determination was based on the average apparent
density of the fibres, and individual fibre mass and length [151].
3.6.2 Mechanical extraction
Different mechanical extraction processes for bamboo fibres were explored by

Deshpande et al. [136]. They applied a compression technique (CT) and a roller mill
technique (RM) in combination with alkali treatment (0.1 N NaOH for 72 h) in order
to extract bamboo fibres. The CT method, consisting in compressing the strips with
a constant load of 10 tons during 10 s, yielded fibres with larger diameters and larger
deviations compared with fibres from RM. The average tensile strength and its
correspondent standard deviation were higher with fibres obtained from CT.
Additionally, they found that bamboo fibres with larger diameters had a higher
strength compared to fibres with lower diameters. This behaviour is different from
what is generally observed in natural fibres and must be attributed to increased fibre
damage upon extracting finer fibres.
Ogawa et al. [144] obtained short bamboo fibres by end-milling and found that the
fibre shape can be controlled by adjusting the end-mill parameters and using a spiral
path (see Figure 3-20). As mentioned before in section 3.6, other methods include
beating, shearing, rolling, gridding and the use of tooth rollers and pin rollers to
extract the fibres [81, 133, 134, 137]. More details about the fibre characteristics and
mechanical properties obtained by the mechanical extraction technique are shown in
Table 3-11.
Figure 3-20. End-mill technique to extract short bamboo fibres [144].
3.6.3 Chemical extraction
Typically in chemical extraction, bamboo culm pieces are soaked in a chemical

solution (e.g. NaOH) with the purpose of dissolving the soft tissue that binds the
technical fibres. Rao et al. [132] examined a chemical process for extracting bamboo
fibres combined with retting. The results of the investigation show that the fibres
after chemical extraction show a large number of major defects along the length.
In literature, the further industrial adoption of this method has been found
contradictory. Some drawbacks have been pointed out by Prasad et al. [56]. On the
one hand, it is an expensive process that will not be able to compete with the
46 Chapter 3
b
existing ones (e.g. mechanical and steam explosion methods); and on the other hand,
it was found that after treatment, the chemical nature of the fibres can change and
this might not be desired.
From another point of view, Phong et al. [137] found that in comparison with steam
explosion and mechanical extraction techniques, the chemical technique has some
advantages such as cheaper equipment, higher aspect ratio of the fibre, relative low
energy consumption and better control of fibre properties. They treated bamboo
strips of around 2-3 mm thickness, in 1, 2 and 3% NaOH solution at 70°C during 10
hours and obtaining the higher technical fibre mechanical properties at 1%
concentration. Comparisons between chemical extraction, mechanical extraction and
steam explosion are shown in Table 3-12.
Method Cost of Energy Installing Environ- Fibre Fibre Control of

equipment consumption area mental l/d diameter fibre
impact (µm) property
Steam
Expensive High Large No High 195 ± 150 Good
explosion
Mechanical Reasonably Relatively Small No Low - Acceptable

expensive low
Chemical Cheap
(alkali (excluding Medium Very large Yes High 230 ± 180 Good
treatment) waste water
treatment)
Table 3-12. Comparison of different extraction methods for bamboo fibres [137].
3.7 Bamboo fibre composites
Bamboo fibres are suitable to be used as reinforcement in polymeric composites in

order to manufacture light and strong composites due to their excellent mechanical
properties and they could be appropriate to be used for structural and semi-structural
composites applications [67]. Several studies on bamboo fibre reinforced composites,
mainly with random technical fibre configuration, have been published using
thermoset [14, 42, 81, 136], thermoplastic [13, 78, 133-135, 152, 153] and
biodegradable [78, 80, 100, 135, 154, 155] matrices. A summary of these studies is
shown in Tables 3-13 and 3-14, discriminated by tensile and flexural tests properties.
Nevertheless, the results of these studies are in general rather poor and suggest that
the intrinsic good properties of the bamboo technical fibres were not being fully
transferred to the composite. Apparently improvements in fibre quality and better
interface fibre-matrix material are still required.
Fibre characteristics Before treatment After treatment

Fibre
F. diam F. length Treatment Vf Tensile Tensile Tensile Tensile
extraction Ref.
System (µm) (mm) applied (%) strength modulus strength modulus
method
UD
Unsat. polyester + BF* 90 - 400 - Retting + mechanical - 40 126 2.5 - - [42]
Hand-lay-up
43 425
UD 100- Milling machine +
220-270 - 37 381 20.3 - - [81]
Araldite matrix + BF* 600 razor blade
25 320
Hand-lay-up
Random
Steam Alkali
Starch-based resin + BF 200 25 50 45 - 60 -
explosion treatment [154]
Hot press
Random
- 0.2-0.5 Grinding bamboo chips MAPP 20 20 2.1 22 2.1 [134]
PP + BF*, Injection
Random
150 6-12 MAPP 30 15 2.5 17 2.6 [153]
PP + BF*, Hot press Grinding bamboo chips
Random 125- Steam

- MAPP 41 - - 30 3.6 [13]
PP + BF*, Hot press 210 explosion
Random 10- Steam explosion +

- MAPP 41 - - 35 4.7 [133]
PP + BF*, Hot press 30 mechanical
Random
70 0,5 Not mentioned LDI** 30 29 2.6 46 3.9 [80]
PLA + BF*, Comp.mould
Random 70 0,5 Not mentioned LDI** 30 21 23 38 25 [80]

PBS+ BF*, Comp.mould
Random
90- 125 1-6 NaOH (20%) + MAPP 50 29 1.7 50 1.5 [78]
MAPP + BF*, Comp. mould
pressure, 1 h.
Table 3-13. Overview of the tensile mechanical properties of bamboo technical fibre composites. BF*:bamboo technical fibres. ** Lysine-based diisocyanate.
48 Chapter 3
b
Fibre
Bending properties
characteristics Fibre
Vf
System F. F. extraction Flexural Flexural Ref.
(%)
diam. length method Strength Modulus
(µm) (mm) (MPa) (GPa)
15 NaOH (0,1 N) 15 144 8.6
UD
250 - during 72 h + 15 162 8.0
Polyester + BF* [136]
400 Compression 20 100 1.2
Hand lay-up
UD
150 NaOH (0,1 N) 15 158 3,5
Polyester +BF* - [136]
250 during 72 h + 15 143 3,2
Hand lay-up
Roller mill
UD
900 - Retting +
Unsat. polyester + BF* - 40 128 3.7 [42]
400 mechanical
Hand-lay-up
Chemical
Orthogonal plies treatment
- 10-
PP + BF* (Alkali) 50 - 2.3 [135]
70
Compression moulding no specified
NaOH (20%)
Random
90 - under pressure
PP + BF* 1-6 50 49 2.8 [78]
125 during 1 h
Compression moulding
Random NaOH (20%)
90 -
MAPP + BF* 1-6 under pressure 50 68 4.1 [78]
125
Compression moulding during 1 h
Random
Steam explosion
PLA + BF* - 1.1 50 107 - [155]
+ mech. Process
Compression moulding
Random Chemical
Biodegrad. Resin (CP- - 10- treatment 50 4.3
- [135]
300) + BF* 70 (Alkali) 60 5.1
Compression moulding no specified
Table 3-14. Overview of the flexural mechanical properties of bamboo technical fibre composites.
BF*:bamboo technical fibres.
Glass/bamboo fibre hybrid composite has been also developed by Thwe et al. [134,
153] and Dieu et al. [156]. They conclude that hybridisation could enhance the
durability of natural fibre composite under environmental aging and also, that it is
possible to include 25% of bamboo fibres in weight without sacrificing the
mechanical properties, while the density is reduced about 13%.
3.8 Application of natural fibre composites
Natural fibres have been used nowadays as substitution of synthetic fibres in several
applications and different sectors. Moreover, they exhibit a favourable nonbrittle
fracture on impact, which is an important requirement in some applications (e.g.

passenger compartments) [53], good material damping [157], good specific
properties and an advantegeous “green image”. For an overview of different
applications of natural fibres, see Figure 3-21. According to the fibre length, their
use can range from high-end products such as surf boards, fishing rods, bicycle
frames to short fibre reinforcement in different applications [158]. Current
applications for natural-fibre-reinforced composites mainly focus on the
construction and automotive industries, although there are a number of niche
applications where low volumes of material are used. Figure 3-22 shows some
commercial applications of natural fibre components in automotive components,
part of bicycle frames and sports goods.
Figure 3-21. Different applications for natural fibres according to the fibre length [158].
The automotive industry is now looking for “green” composites. The two most
important factors now driving the use of natural fibres by the European automotive
industry, are cost and weight; with the consequent improvements in fuel efficiency.
Some of the produced parts are interior panels, bumpers, spoilers and body parts
[158]. DaimerCrysler already commercialised the first flax fibre reinforced polyester
composites in engine and drive train covers in buses and car passengers satisfying
the cost, weight and durability (after 50.000 km of road tests) requirements [50].
50 Chapter 3
b
Figure 3-22. Applications in natural fibre components in automotive parts (source: http://www.ircomas.org),
and recent application of natural fibres in bicycle frames and tennis rackets for increased material damping
(source: http://www.lineo.eu/).
3.9 Conclusions
There is in general an increasing interest in the use of natural fibres in several

sectors for new high-end applications driven by cost, good mechanical properties,
“green” image and governmental environmental regulations.
Bamboo technical fibres can be an attractive alternative to other natural fibres and
glass fibres in composite applications. Bamboo fibre can be used for a variety of
structural and semi-structural applications due to its good specific properties,
renewability, fast growing and other environmental benefits, which include a large
CO2 capture and low energy consumption per kg of fibres. Moreover they offer a
continuous supply along the year. These advantages are becoming very important
selection criteria for the materials of the future.
Despite showing a good potential, long high quality bamboo technical fibres have
not been used as a unidirectional reinforcement in polymeric composites materials
due to the non availability of the fibres. Unidirectional distribution of the fibres is
the most suitable configuration for high end composite applications in order to
obtain the maximum efficiency of the technical fibres into the composite, according
to the load conditions in each particular case.
Mechanical extraction methods (e.g. rubbing, crushing and compression) are used as
complementary operations for steam explosion or chemical extraction and vice versa.
Each of these extraction methods has their own advantages and disadvantages in
terms of cost, energy consumption environmental pollution and fibre quality. “Pure”
mechanical extraction is only found when obtaining short bamboo fibres (e.g.
milling, grinding). To practically apply the benefit of bamboo fibres, it is necessary
to develop a process to extract long high quality fibres from the bamboo culm in
industrial scale, with low environmental impact (e.g. low water consumption, no use
of chemicals or high temperature) and low cost (e.g. in line process).
Several limitations have been found for the use of the natural fibres in composite
materials, such as the variability in mechanical properties, thermal degradation
during composite manufacturing and their hydrophilic behaviour. Furthermore,
moisture absorption causes swelling of the fibre and thus reduces the interface
between fibres and composites. Deep understanding of each of these phenomena is
needed in order to minimize their impact in the mechanical properties. By doing so,
it will be possible to establish the correspondent limits, and the conditions that need
to be taken into account for an appropriate characterization manufacturing and use
of the natural fibre composites. When these limitations are understood, the use of
natural fibres will widen and will be better accepted by various high-performance
composites industry.
References
[1] Azwa ZN, Yousif BF, Manalo AC, Karunasena W. A review on the degradability of polymeric
composites based on natural fibres. Materials & Design. 2013;47(0):424-42.
[2] Jayamol K, Thomas S. Short fibre-reinforced plastics composites. In: Cheremisinoff N, editor.
Handbook of Engineering Polymeric Materials1997. p. 811-37.
[3] Jayaraman K. Manufacturing sisal - polypropylene composites with minimum fibre degradation.
Compos Sci Technol. 2003;63(3-4):367-74.
[4] Saiza E, Riande E. Materiales de altas prestaciones. Revista de Plásticos Modernos.
1998;75(501):250-6 [Spanish].
[5] Stamboulis A, Baillie CA, Peijs T. Effects of environmental conditions on mechanical and
physical properties of flax fibers. Compos Part A: Appl Sci Manuf. 2001;32(8):1105-15.
[6] Silva FvdA, Mobasher B, Filho RDT. Cracking mechanisms in durable sisal fiber reinforced
cement composites. Cement and Concrete Composites. 2009;31(10):721-30.
[7] Bledzki AK, Gassan J. Composites reinforced with cellulose based fibres. Prog Polym Sci.
1999;24(2):221-74.
[8] Bismark A, Mishra S, Lampke T. Plant fibers as reinforcement for green composites. In:
Mohanty A, Misra M, Drzal L, editors. Natural Fibers, Biopolymers and Biocomposites: CRC Press;
2005.
[9] Kavelin KG. Investigation of natural fiber composites heterogeneity with respect to automotive
structures 2005.
[10] Bledzki A, Gassan J. Composites reinforced with cellulose based fibres. Prog Polym Sci.
1999;24(2):221-74.
[11] McKendry P. Energy production from biomass (part 1): overview of biomass. Bioresource
Technology. 2002;83(1):37-46.
[12] John MJ, Anandjiwala RD. Recent developments in chemical modification and
characterization of natural fiber-reinforced composites. Polym Compos. 2008;29(2):187-207.
52 Chapter 3
b
[13] Okubo K, Fujii T, Yamamoto Y. Development of bamboo-based polymer composites and their
mechanical properties. Composite Part A: Appl Sci Manuf. 2004;35(3):377-83.
[14] Jain S, Kumar R, Jindal UC. Mechanical behaviour of bamboo and bamboo composite. J Mater
Sci. 1992;27(17):4598-604.
[15] Gassan J, Bledzki AK. The influence of fiber-surface treatment on the mechanical properties of
jute-polypropylene composites. Compos Part A: Appl Sci Manuf. 1997;28(12):1001-5.
[16] Mwaikambo LY, Ansell MP. Chemical modification of hemp, sisal, jute, and kapok fibers by
alkalization. J Appl Polym Sci. 2002;84(12):2222-34.
[17] Bledzki AK, Fink HP, Specht K. Unidirectional hemp and flax EP- and PP-composites:
Influence of defined fiber treatments. J Appl Polym Sci. 2004;93(5):2150-6.
[18] Mwaikambo LY, Ansell MP. Mechanical properties of alkali treated plant fibres and their
potential as reinforcement materials. I. hemp fibres. J Mater Sci. 2006;41(8):2483-96.
[19] Li X, Tabil L, Panigrahi S. Chemical Treatments of Natural Fiber for Use in Natural Fiber-
Reinforced Composites: A Review. J Polym Environ. 2007;15(1):25-33.
[20] Ashimori M, Kamkaya K, Aoyama E, Nagai S. Study on splitting of bamboo fibres due to
freezing and tensile of FRTP using bamboo fibres. JSME International Journal Series A Solid
Mechanics and Material Engineering. 2004;47(4):566-9.
[21] Summerscales J, Dissanayake NPJ, Virk AS, Hall W. A review of bast fibres and their
composites. Part 1 - Fibres as reinforcements. Compos Part A: Appl Sci Manuf.41(10):1329-35.
[22] Summerscales J, Dissanayake N, Virk A, Hall W. A review of bast fibres and their composites.
Part 2 - Composites. Compos Part A: Appl Sci Manuf.41(10):1336-44.
[23] Rao K, Rao K. Extraction and tensile properties of natural fibers: Vakka, date and bamboo.
Compos Struct. 2007;77(3):288-95.
[24] Cao Y, Shibata S, Fukumoto I. Mechanical properties of biodegradable composites reinforced
with bagasse fibre before and after alkali treatments. Compos Part A: Appl Sci Manuf.
2006;37(3):423-9.
[25] Okubo K, Fujii T. Eco-composites using natural fibers and their mechanical properties. In:
Brebbia C, De Wilde W, editors. High performance Structures and Composites. Southampton: WIT
Press; 2002. p. 77-85.
[26] Brower W. Natural fibre composites in structural components: alternative application for sisal?
In: FAO, editor.2004.
[27] Dwight D. Glass Fibre Reinforcements. In: Zweben AKC, editor. Comprehensive Composite
Materials: Elsevier Ltd.; 2000. p. 231-86.
[28] Baur E, Otremba F. Design, materials properties and databases. In: Müssig J, editor. Industrial
applications of natural fibres, structure, properties and technical applications: Wiley Publishers;
2010. p. 399-406.
[29] Joffe R, Andersons J, Wallstrom L. Strength and adhesion characteristics of elementary flax
fibres with different surface treatments. Composite Part A: Appl Sci Manuf. 2003;34(7):603-12.
[30] Placet V. Characterization of the thermo-mechanical behaviour of hemp fibres intended for the
manufacturing of high performance composites. Composite Part A: Appl Sci Manuf.
2009;40(8):1111-8.
[31] Zhang Y, Wang X, Pan N, Postle R. Weibull analysis of the tensile behavior of fibers with
geometrical irregularities. J Mater Sci. 2002;37(7):1401-6.
[32] da Costa LL, Loiola RL, Monteiro SN. Diameter dependence of tensile strength by Weibull
analysis: Part I bamboo fiber. Revista Matéria. 2010;15(2):110-6.
[33] Nishino T. Natural fibre sources. In: Baillie C, editor. Green composites: Polumer Composites
and the environment: Woodhead Publishers; 2004.
[34] Ray AK, Das SK, Mondal S, Ramachandrarao P. Microstructural characterization of bamboo. J
Mater Sci. 2004;39(3):1055-60.
[35] Kozlowski R, Wladyka-Przybylak M. Use of natural fiber reinforced plastics. In: Wallenberger
FT WN, editor. Natural Fibers, Plastics and Composites: Kluwer Academic Publishers Group; 2004.
[36] Pietak A, Korte S, Tan E, Downard A, Staiger MP. Atomic force microscopy characterization
of the surface wettability of natural fibres. Appl Surf Sci. 2007;253(7):3627-35.
[37] Van de Weyenberg I, Chi Truong T, Vangrimde B, Verpoest I. Improving the properties of UD
flax fibre reinforced composites by applying an alkaline fibre treatment. Compos Part A: Appl Sci
Manuf. 2006;37(9):1368-76.
[38] Gassan J, Bledzki AK. Possibilities for improving the mechanical properties of jute/epoxy
composites by alkali treatment of fibres. Compos Sci Technol. 1999;59(9):1303-9.
[39] Ray D, Sarkar BK, Rana AK, Bose NR. The mechanical properties of vinylester resin matrix
composites reinforced with alkali-treated jute fibres. Compos Part A: Appl Sci Manuf.
2001;32(1):119-27.
[40] Aziz SH, Ansell MP. The effect of alkalization and fibre alignment on the mechanical and
thermal properties of kenaf and hemp bast fibre composites: Part 1: polyester resin matrix. Compos
Sci Technol. 2004;64(9):1219-30.
[41] Rong MZ, Zhang MQ, Liu Y, Yang GC, Zeng HM. The effect of fiber treatment on the
mechanical properties of unidirectional sisal-reinforced epoxy composites. Compos Sci Technol.
2001;61(10):1437-47.
[42] Ratna Prasad AV, Mohana Rao K. Mechanical properties of natural fibre reinforced polyester
composites: Jowar, sisal and bamboo. Materials & Design. 2011;32:4658-63.
[43] Charlet K, Baley C, Morvan C, Jernot JP, Gomina M, Bréard J. Characteristics of Hermès flax
fibres as a function of their location in the stem and properties of the derived unidirectional
composites. Compos Part A: Appl Sci Manuf. 2007;38(8):1912-21.
[44] Khondker O, Ishiaku U, Nakai A, Hamada H. Fabrication mechanical properties of
unidirectional jute/PP composites using jute yarns by film stacking method. J Polym Environ.
2005;13(2):115-26.
[45] Khondker OA, Ishiaku US, Nakai A, Hamada H. A novel processing technique for
thermoplastic manufacturing of unidirectional composites reinforced with jute yarns. Compos Part
A: Appl Sci Manuf. 2006;37(12):2274-84.
[46] Van de Velde K, Kiekens P. Effect of material and process parameters on the mechanical
properties of unidirectional and multidirectional flax/polypropylene composites. Compos Struct.
2003;62:443-8.
[47] Rijsdijk H, Contant M, Peijs A. Continuous-glass-fibre-reinforced polypropylene composites: I.
Influence of maleic-anhydride-modified polypropylene on mechanical properties. Compos Sci
Technol. 1993;48(1-4):161-72.
manufacturing of high performance composites. Compos Part A: Appl Sci Manuf. 2009;40(8):1111-
8.
[49] Joshi SV, Drzal LT, Mohanty AK, Arora S. Are natural fiber composites environmentally
superior to glass fiber reinforced composites? Composite Part A: Appl Sci Manuf. 2004;35(3):371-6.
[50] Schloesser T. Natural fiber reinforced automotive parts. In: Wallenberger F, Weston N, editors.
Natural fibers, plastics and composites. Norwell: Kluwer Academic Publishers; 2004. p. 275 -85.
[51] Oksman K, Skrifvars M, Selin JF. Natural fibres as reinforcement in polylactic acid (PLA)
composites. Compos Sci Technol. 2003;63(9):1317-24.
[52] Plackett D. Biodegradable polymer composites from natural fibres. In: Smith R, editor.
Biodegradable polymers for industrial applications: Woodhead Publishing 2005.
[53] Mohanty A, Misra M, Drzal L. Sustainable bio-composites from renewable resources:
Opportunities and challenges in the green materials world. J Polym Environ. 2002;10(1-2):19-26.
[54] Averous L, Boquillon N. Biocomposites based on plasticized starch: thermal and mechanical
behaviours. Carbohydrate Polymers. 2004;56(2):111-22.
[55] Strom G, Carlsson G. Wettability of kraft pulps-effect of surface composition and oxygen
plasma treatment. J Adhes Sci Technol. 1992;6(6):745-61.
[56] Prasad B, Sain M. Mechanical properties of thermally treated hemp fibers in inert atmosphere
for potential composite reinforcement. Mater Res Innov. 2003;7(4):231-8.
[57] Van de Weyenberg I. Flax fibres as a reinforcement for epoxy composites [Doctoral thesis].
Leuven: University of Leuven; 2005.
54 Chapter 3
b
[58] Bledzki AK, Reihmane S, Gassan J. Properties and modification methods for vegetable fibers
for natural fiber composites. J Appl Polym Sci. 1996;59(8):1329-36.
[59] Sreekumar P, Kuruvilla J, Unnikrishnan G, Sabu T. Surface-modified sisal fiber-reinforced
eco-friendly composites: Mechanical, thermal, and diffusion studies. Polym Compos.
2011;32(1):131-8.
[60] Jahn A, Schroder MW, Futing M, Schenzel K, Diepenbrock W. Characterization of alkali
treated flax fibres by means of FT Raman spectroscopy and environmental scanning electron
microscopy. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy.
2002;58(10):2271-9.
[61] Ouajai S, Shanks RA. Composition, structure and thermal degradation of hemp cellulose after
chemical treatments. Polym Degradation Stab. 2005;89(2):327-35.
[62] Bisanda ETN, Ansell MP. The effect of silane treatment on the mechanical and physical
properties of sisal-epoxy composites. Compos Sci Technol. 1991;41(2):165-78.
[63] Mukherjee PS, Satyanarayana KG. Structure and properties of some vegetable fibres. J Mater
Sci. 1984;19(12):3925-34.
[64] Thamae T, Baillie C. Influence of fibre extraction method, alkali and silane treatment on the
interface of Agave americana waste HDPE composites as possible roof ceilings in Lesotho. Compos
Interfaces. 2007;14(7-9):821-36.
[65] Das M, Chakraborty D. Influence of alkali treatment on the fine structure and morphology of
bamboo fibers. J Appl Polym Sci. 2006;102(5):5050-6.
[66] Das M, Chakraborty D. Effects of alkalization and fiber loading on the mechanical properties
and morphology of bamboo fiber composites. II. Resol matrix. J Appl Polym Sci. 2009;112(1):447-
53.
[67] Osorio L, Trujillo E, Van Vuure AW, Verpoest I. Morphological aspects and mechanical
properties of single bamboo fibers and flexural characterization of bamboo/ epoxy composites. J
Reinf Plast Compos. 2010;30(5):396-408.
[68] Thimothy T, Baillie C. Influence of fibre extraction method, alkali and silane treatment on the
interface of Agave americana waste HDPE composites as possible roof ceilings in Lesotho.
Compos Interfaces. 2007;14:821-36.
[69] Beckermann G. Performance of hemp-fibre reinforced polypropylene composite materials
[Doctoral thesis]: University of Waikato; 2007.
[70] Gomes A, Goda K, Oghi J. Effects of alkali treatment to reinforcement on tensile properties of
curaua fiber green composites. International Journal Series A Solid Mechanics and Material
Engineering. 2004;47:541-6.
[71] Abdal-hay A, Suardana N, Jung D, Choi K, Lim J. Effect of diameters and alkali treatment on
the tensile properties of date palm fiber reinforced epoxy composites. International Journal of
Precision Engineering and Manufacturing. 2012;13(7):1199-206.
[72] Li Y, Mai Y-W, Ye L. Sisal fibre and its composites: a review of recent developments.
Compos Sci Technol. 2000;60(11):2037-55.
[73] Jacob M, Thomas S, Varughese KT. Mechanical properties of sisal/oil palm hybrid fiber
reinforced natural rubber composites. Compos Sci Technol. 2004;64(7-8):955-65.
[74] Mishra S, Misra M, Tripathy S, Nayak S, Mohanty A. Graft copolymerization of acrylonitrile
on chemically modified sisal fibers. Macromolecular Mater Eng. 2001;286(2):107-13.
[75] Joseph PV, Rabello MS, Mattoso LHC, Joseph K, Thomas S. Environmental effects on the
degradation behaviour of sisal fibre reinforced polypropylene composites. Compos Sci Technol.
2002;62(10-11):1357-72.
[76] Mohanty S, Nayak S, Verma S, Tripathy S. Effect of MAPP as a coupling agent on the
performance of jute–PP composites. Reinforced Plastics and Composites. 2004;23(6):625-37.
[77] Doan T, Gao S, Mader E. Jute/polypropylene composites I. Effect of matrix modification.
Compos Sci Technol. 2006 66(7-8):952-63.
[78] Chattopadhyay SK, Khandal RK, Uppaluri R, Ghoshal AK. Bamboo fiber reinforced
polypropylene composites and their mechanical, thermal, and morphological properties. J Appl
Polym Sci.119(3):1619-26.
[79] Chen X, Guo Q, Mi Y. Bamboo fiber-reinforced polypropylene composites: A study of the

mechanical properties. J Appl Polym Sci. 1998;69(10):1891-9.
[80] Lee S, Wang S. Biodegradable polymer/bamboo fiber biocomposite with bio-based coupling
agent. Composites Part A. 2006;37:80-91.
[81] Jindal UC. Development and testing of bamboo-fibres reinforced plastic composites. J Compos
Mater. 1986;20(1):19-29.
[82] Pott G. Natural fibers with low moisture sensitivity. In: Wallenberger F, Weston N, editors.
Natural Fibers, Plastics and Composites: Kluwer Academic Publishers Group; 2004. p. 104-22.
[83] Davies G, Bruce D. Effect on environmental relative humidity and damage on tensile
properties of flax and nettle fibres. Textile Research Journal. 1998;68(9):623-9.
[84] Yao F, Wu Q, Lei Y, Guo W, Xu Y. Thermal decomposition kinetics of natural fibers:
Activation energy with dynamic thermogravimetric analysis. Polym Degradation Stab.
2008;93(1):90-8.
[85] Araújo J, Waldman W, De Paoli M. Thermal properties of high density polyethylene
composites with natural fibres: Coupling agent effect. Polym Degradation Stab. 2008;93(10):1770-5.
[86] Doan T, Brodowsky H, Mader E. Jute fibre/polypropylene composites II. Thermal,
hydrothermal and dynamic mechanical behaviour. Compos Sci Technol. 2007;67(13):2707-14.
[87] Gassan J, Bledzki AK. Thermal degradation of flax and jute fibers. J Appl Polym Sci.
2001;82(6):1417-22.
[88] Wielage B, Lampke T, Marx G, Nestler K, Starke D. Thermogravimetric and differential
scanning calorimetric analysis of natural fibres and polypropylene. Thermochim Acta.
1999;337:169-77.
[89] Rachini A, Le Troedec M, Peyratout C, Smith A. Comparison of the thermal degradation of
natural, alkali-treated and silane-treated hemp fibers under air and an inert atmosphere. J Appl
Polym Sci. 2009;112(1):226-34.
[90] Bilbao R, Millera A, Arauzo Js. Thermal decomposition of lignocellulosic materials: influence
of the chemical composition. Thermochim Acta. 1989;143(0):149-59.
[91] Yang H, Yan R, Chen H, Lee DH, Zheng C. Characteristics of hemicellulose, cellulose and
lignin pyrolysis. Fuel. 2007;86:1781-8.
[92] Martin A, Martins M, da Silva O, Mattoso L. Studies on the thermal properties of sisal fiber
and its constituents. Thermochim Acta. 2010;506:14-9.
[93] Van De Velde K, Kiekens P. Thermal degradation of flax: The determination of kinetic
parameters with thermogravimetric analysis. J Appl Polym Sci. 2002;83(12):2634-43.
[94] Van de Velde K, Baetens E. Thermal and Mechanical Properties of Flax Fibres as Potential
Composite Reinforcement. Macromolecular Mater Eng. 2001;286(6):342-9.
thermal properties of kenaf and hemp bast fibre composites: part 2 - cashew nut shell liquid matrix.
[96] Gao M, Dai Q. Studies on thermal degradation of cellulosic fibers treated with flame retardants.
The Chinese Journal of Processing Engineering. 2006;6(2).
[97] Kohler R, Wedler M. Nichttextile Anwendungen von Flachs. Techtextil Symposium 6.
Frankfurt (Germany)1994. p. 1-8.
[98] Mieck K, Nechwatal A, Knobelsdorf C. Anwendungsmöglichkeiten von naturfarbstoffen bei
verbundmaterial Melliand Textileberichte. 1994;11:892-8.
[99] Ray D, Sarkar BK, Basak RK, Rana AK. Study of the thermal behavior of alkali-treated jute
fibers. J Appl Polym Sci. 2002;85(12):2594-9.
[100] Ochi S, Takagi H, Niki R. Mechanical properties of heated-treated natural fibers. In: Brebbia
C, De Wilde W, editors. High performance Structures and Composites. Southampton: WIT press;
2002. p. 117-23.
[101] Prasad M, Sain M, Roy N. Structure Property Correlation of Thermally Treated Hemp Fiber.
Macromolecular Mater Eng. 2004;289(6):581-92.
56 Chapter 3
b
[102] Fuentes CA, Tran LQN, Dupont-Gillain C, Van Vuure AW, Verpoest I. Interfaces in Natural
Fibre Composites: Effect of Surface Energy and Physical Adhesion. Journal of Biobased Materials
and Bioenergy. 2012;6(4):456-62.
[103] Andersons J, Joffe R, Hojo M, Ochiai S. Glass fibre strength distribution determined by
common experimental methods. Compos Sci Technol. 2002;62(1):131-45.
[104] Tripathy D. Practical guide to polypropylene. UK2002.
[105] Ray S, Cooney P. Thermal degradation of polymer and polymer composites. In: Kutz M,
editor. Handbook of Environmental Degradation of Materials: William Andrew; 2005. p. 215-42.
[106] Achimsky L, Audouin L, Verdu J. Kinetic study of the thermal oxidation of polypropylene.
Polym Degradation Stab. 1997;57:231-40.
[107] Lobovikov M, Shyam P, Piazza M, Ren H, Wu J. Non-wood forest products 18 world
bamboo resources. A thematic study prepared in the framework of the Global Forest Resources
Assessment. In: FAO, editor. Rome 2007.
[108] Londoño X. La Guadua un bambú importante de América. First Bamboo Seminar. Guayaquil,
Ecuador 2001. p. 12-8 [Spanish].
[109] Young S, Wool L. Systematics of the Guadua angustifolia (Poaceae: Bambusoideae). Annals
of the Missouri Botanical Garden. 1992;79(4):737-69.
[110] Riaño N, Londoño X, López Y, J G. Plant growth and biomass distribution on Guadua
angustifolia Kunth in relation to ageing in the Valle del Cauca - Colombia. J Am Bamboo Soc.
2002;16(1):43-51.
[111] Londoño X. A decade of observations of a "Guadua angustifolia" plantation in Colombia.
The Journal of the American Bamboo Society. 1998;12:37-42.
[112] Jiafu L. The development strategy of bamboo resource and bamboo industry in China.
International training workshop on sustainable bamboo management and processing techniques for
small size bamboo enterprises. Hangzhou, China 2000. p. 1-21.
[113] Lybeer B, Acker J, Goetghebeur P. Variability in fibre and parenchyma cell walls of
temperate and tropical bamboo culms of different ages. Wood Science and Technology.
2006;40(6):477-92.
[114] Amada S, Ichikawa Y, Munekata T, Nagase Y, Shimizu H. Fiber texture and mechanical
graded structure of bamboo. Compos B: Eng. 1997;28(1-2):13-20.
[115] Liese W. The Anatomy of Bamboo Culms (Inbar Technical Reports)1998.
[116] Amada S. Fiber texture and mechanical graded structure of bamboo. Compos Part B-Eng.
1997;28(1-2):13-20.
[117] Londoño X, Camayo G, Riaño N, López Y. Characterization of the anatomy of Guadua
angustifolia (Poaceae: Bambusoideae) culms. J Am Bamboo Soc. 2002;16:18-31.
[118] Arbelaez A. La estructura morfológica del culmo de la Guadua angustifolia Kunt. In: FAS,
editor. Medellin, Colombia: Universidad Nacional de Colombia; 1998. p. 3-18 [Spanish].
[119] Janssen J. Bamboo in building structures [Doctoral thesis]: Eindhoven University of
Technology; 1981.
[120] Liese W. The structure of bamboo in relation to its properties and utilization. Bamboo and its
use. International Symposium on Industrial Use of Bamboo. Beijin, China1992.
[121] Abdul Khalil HPS, Bhat IUH, Jawaid M, Zaidon A, Hermawan D, Hadi YS. Bamboo fibre
reinforced biocomposites: A review. Materials & Design. 2011;42(0):353-68.
[122] Liese W. Anatomy and properties of bamboo. Recent Research on Bamboos - Proceedings of
the 2nd International Workshop. Hangzhou, China 1985.
[123] Osorio L, Trujillo E, Van Vuure AW, Lens F, Ivens J, Verpoest I. Composite micromechanics
explain the behaviour of bamboo fibres (pending of publication). 2013-2014.
[124] Osorio L, Trujillo E, Van Vuure A, Lens F, Ivens J, I V. The relation between bamboo fibre
microstruture and mechanical properties. In: Proceedings of ECCM-14 Conference. Budapest, June,
2010. p. 1-10.
[125] Murphy R, Alvin K. Variation in fibre wall structure of bamboo. IAWA Bulletin.
1992(13):204.
[126] Parameswaran N, Liese W. On the fine structure of bamboo fibres. Wood Science and
Technology. 1976;10:231-46.
[127] Higuchi T. Chemistry and biochemistry; bamboo for pulp and paper of bamboo. In: G.
Lessard ACI, editor. Bamboo Research in Asia. Ottawa 1980. p. 51-6.
[128] Itoh T, Shimaji K. Lignification of bamboo culm during its growth and maturation. In: Group
XIC, editor. Bamboo Production and Utilization. Kyoto1981. p. 104-10.
[129] Mohmoud A. Effects of Age and Height on Selected Properties of Three Malaysian Bamboo
Species: Universiti Putra Malaysia; 1991.
[130] Zhou F. Studies on physical and mechanical properties of bamboo woods. Journal of Nanjing
Technological College of Forest Products. 1981;02.
[131] Sano O, Matsuoka T, Sakaguchi K, Kamkaya K. Study on the interfacial shear strength of
bamboo fibre reinforced plastics. In: Brebbia C, De Wilde W, editors. High Performance Structures
and Composites. Southampton: WIT Press; 2002. p. 147-56.
[132] Rao K. Extraction and tensile properties of natural fibers: Vakka, date and bamboo. Compos
Struct. 2007;77(3):288.
[133] Shito T, Okubo K, Fujii T. Development of eco-composites using natural bamboo fibres and
their mechanical properties. In: Brebbia C, De Wilde W, editors. High performance Structures and
Composites. Southampton: WIT Press; 2002. p. 175-82.
[134] Thwe MM, Liao K. Environmental effects on bamboo-glass/polypropylene hybrid composites.
J Mater Sci. 2003;38(2):363-76.
[135] Shibata S, Fukumoto I. Effects of bamboo and kenaf fibres on the flexural modulus of bio-
composites. Technical Article, University of the Ryukyus, Okinawa, Japan. 2004:14-7.
[136] Deshpande AP, Bhaskar RM, Lakshmana RC. Extraction of bamboo fibers and their use as
reinforcement in polymeric composites. J Appl Polym Sci. 2000;76(1):83-92.
[137] Phong N, Fujii T, Choung B, Okubo K. Study on how to effectively extract bamboo fibres
from raw bamboo and wastewater treatment. Journal of Materials Science Research. 2012;1(1):144-
55.
[138] Defoirdt N, Biswas S, De Vriese L, Tran N, Van Acker J, Ahsan Q, et al. Assessment of the
tensile properties of coir, bamboo and jute fibre. Composite Part A: Appl Sci Manuf.
2010;41(5):588-95.
[139] Moreno L, Osorio L, Trujillo E. Estudio de las propiedades mecánicas de haces de fibra de
Guadua angustifolia. Ingeniería & Desarrollo. 2006;20:125-33 [Spanish].
[140] Bangarshetti P, Lakshmana Rao C. Constitutive modeling of Bambo fiber bundle. National
Symposium on fracture mechanics. Bangalore2004.
[141] Bos H, Van Den Oever M, Peters O. Tensile and compressive properties of flax fibres for
natural fibre reinforced composites. J Mater Sci. 2002;37(8):1683-92.
[142] Tung N, Yamamoto H, Matsuoka T, Fuji T. Effect of surface treatment on interfacial strength
bamboo fibre and PP resin. JSME International Journal Series A Solid Mechanics and Material
Engineering. 2004;47(4):561-5.
[143] Tanaka R, Okubo K, Fuji T, Ono M, Sakurai A. Development of a pleasure boat using
bamboo fibre reinforced plastics. 16th International Conference on Composite Materials. Kyoto,
Japan2007. p. 1-6.
[144] Ogawa K, Hirogaki T, Aoyama E, Imamura H. Bamboo fibre extraction method using a
machining center. Journal of Advanced Mechanical Design, Systems, and Manufacturing.
2008;2(4):550-9.
[145] Kinoshita H, Kaizu K, Fukuda M, Tokunaga H, Koga K, Ikeda K. Development of green
composite consists of woodchips, bamboo fibers and biodegradable adhesive. Compos B: Eng.
2009;40(7):607-12.
[146] Shibata S, Cao Y, Fukumoto I. Flexural modulus of the unidirectional and random composites
made from biodegradable resin and bamboo and kenaf fibres. Compos Part A: Appl Sci Manuf.
2008;39(4):640-6.
[147] Yamashita O, Imanishi H, Kanayama K. Transfer molding of bamboo. J Mater Process
Technol. 2007;192:259-64.
58 Chapter 3
b
[148] Rowell R, Norimoto M. Acetylation of bamboo fiber. Journal of the Japan Wood Research
Society (Japan). 1987.
[149] Kessler R, Becker U, Kohler R, Goth B. Steam explosion of flax - a superior technique for
upgrading fibre value. Biomass and Bioenergy. 1998;14(3):237-49.
[150] Vignon M, Dupeyre D, Garcia-Jaldon C. Morphological characterization of steam-exploded
hemp fibers and their utilization in polypropylene-based composites. Bioresource technology.
1996;58(2):203-15.
[151] Liu C. Morphological and mechanical characterization of bamboo fibres [Master thesis].
Leuven, Belgium: University of Leuven; 2010.
[152] Katayama T, Ashimori M. Improvement of strength of bamboo fiber reinforced composites
by freezing filamentization. In: Brebbia CA, Wilde WPd, editors. High Performance Structures and
Materials II, 2004. p. 433-40.
[153] Thwe MM, Liao K. Durability of bamboo-glass fiber reinforced polymer matrix hybrid
composites. Compos Sci Technol. 2003;63(3-4):375-87.
[154] Takagi H, Ichihara Y. Effect of fibre length on mechaical properties of "green"composites
using a starch - baed resin and short bamboo fibres. JSME International Journal Series A Solid
Mechanics and Material Engineering. 2004;47(4):551-5.
[155] Okubo K, Fuji T, Yamashita N. Improvement of interfacial adhesion in bamboo polymer
composite enhance with micro-fibrillated cellulose. JSME International Journal of Applied
Mechanics. 2005;48(4):199-204.
[156] Dieu T, Liem N, Mai T, Tung N. Study on fabrication of BMC laminates based on
unsaturated polyester reisn reinforced hybrid bamboo/glass fibres. JSME Int J A-Solid M.
2004;47(4):570-3.
[157] Bledzki A, Mamun A, Faruk O. Abaca fibre reinforced PP composites and comparison with
jute and flax fibre PP composites. Express Polym Lett. 2007;1(11):755-62.
[158] Suddell B, Evans W. Natural fiber composites in automotive applications. In: Mohanty AK,
Mishra S, Drzal L, editors. Natural Fibers, Biopolymers, and Biocomposites Boca Raton FL: CRC
Press, Taylor & Francis; 2005.
Bamboo technical fibre characterization 59
Chapter 4
Bamboo technical fibre characterization
_______________________________________________________
4.1 Introduction
The cylindrical shape of the bamboo culm (see Figure 4-1) limits its direct use in
several engineering applications [1]. A more flexible alternative is the extraction of
the bamboo fibres from the culm, to be used as reinforcement of polymeric matrices.
Figure 4-1. Bamboo (Guadua angustifolia) culm in the plantation.
The industrial adoption of natural resources for reinforcing composites is an active

subject of research. The acceptance of natural fibre reinforced plastics in technical
applications depends on the availability of material data and specific design
information. Establishing reliable properties of the fibres to be safely used in final
product applications, requires extensive testing and a substantial amount of research
[2]. In fact, the variability in mechanical properties of natural fibres is one of the
60 Chapter 4b
main concerns for their industrial applicability [3-5]. In comparison with synthetic
fibres, natural fibers have a significantly higher variation in diameter between the
fibres and within a fibre [4-6]. Nevertheless, this variation in fibre properties can not
only be characterized and predicted when quality management is used [7], but also
controlled when this information serves as a feedback for optimizing the extraction
and further fibre preparation.
In spite of their benefits, studies on bamboo fibre reinforced plastics are relatively
scarce because fibres are not readily available [8-10]. A step forward in the
extraction of high quality, long bamboo fibres has been achieved in earlier work at
KU Leuven. The complete characterization of the fibre strength is the subject of this
chapter, with the purpose to explore the possibility to use bamboo fibres as
reinforcement in polymeric matrices. Two methods were used to estimate the fibre
properties; the single technical fibre (SF) test and the dry fibre bundle (DFB) test.
For the first method, geometrical and strength data were obtained for a large
population of single bamboo technical fibres. The fibres were individually tested in
tension using various gauge lengths for further statistical analysis. To determine the
strength distribution of the fibres, the fibre data were analyzed using the modified
Weibull model [11]. To use the model it must be determined if the defect density is
to be presented as function of the fibre length, fibre volume or fibre surface area to
obtain the best fitting. The within-fibre diameter variation parameter λ was also
determined. These results provide a practical, quantitative predictive model for the
strength of the currently studied bamboo technical fibres. On the other hand,
mechanical properties and Weibull parameters were also obtained from the DFB test
in order to estimate the fibre properties of a larger population of bamboo technical
fibres and to be compared with the SF measurements.
The tests results should provide a better understanding of the intrinsic properties of
the material and a more accurate appreciation of its potential. The generated data can
further be used for the design and modelling of bamboo fibre composites, in order to
fulfil the correspondent safety demands for end products. Also, the presented
methodology is sufficiently generic to be applied to other natural fibres.
4.1.1 The modified Weibull distribution
The Weibull model [12] is a widely used statistical approach for describing the
tensile strength of brittle materials such as carbon and glass fibre [13-15]. As shown
in Table 4-1, it was also applied to a wide range of natural fibres such as jute, hemp,
sisal, flax, coconut and bamboo. Two assumptions that underlie the theory are that
the material is brittle and that the strength is governed by the most serious flaw [16,
17]. The brittleness assumption is satisfied if the material has elastic behaviour up to
failure. The weakest link theory can be applied to a technical bamboo fibre by
assuming the fibre as a chain consisting of several segments with certain strength.
These segments can be regarded not only as a concatenation of elementary fibres,
where the fibre ends are considered as “weak points”, but also, as the minimum
length or volume where any type of defect can be found. Strength distribution of
technical natural fibres is usually described by means of a two parameter Weibull
distribution [18]. To improve the accuracy of the prediction, a modified Weibull
model had been introduced by the implementation of a third parameter [11, 19]:

 V   
m

P  1  exp       (4-1)
  V0  0  
where P is the probability of failure of a fibre of volume V at a stress less than or

equal to σ. The parameters of the distribution are the scale parameter σ0, the shape
parameter m and the volume sensitivity β. The scale parameter σ0 is a measure of the
characteristic strength of the fibre. It corresponds to the fracture stress with a failure
probability of 63.2% of a fibre with reference volume V0. The shape parameter m
mainly defines the variability of the distribution. The higher m, the more narrow and
right-skewed the distribution, and thus the more consistent the quality of the fibres.
For natural fibres this value ranges between 1 and 6 and synthetic fibres usually
have shape factors between 5 and 15 [20, 21]. Andersons et al. [14] found a shape
factor between 5.1 and 5.4 for glass fibres. The strength variability was caused by
the inherent flaw distribution along the fibre and by the fibre-to-fibre strength
variability within a batch. This was found mainly due to variations in the production
process and the damage introduced during the handling of the fibres.
The parameter β is a measure of the sensitivity of the strength with respect to the
tested volume V. It is an empirical parameter that was introduced to improve the
predictive power of the Weibull model with respect to experimental data [11, 19].
The implementation of this parameter was found in several studies [6, 11, 20, 22]
with positive results. If β=1, the conventional Weibull distribution is obtained, in
which each part of volume V contributes equally to the failure probability. Choosing
β between 1 and 0 weakens the volume-dependency: the lower β, the lower the
decrease of strength for increasing volume V. The volume dependency entirely
disappears if β=0 [23]. A final parameter is the reference volume V0. V0 is
independent of the other parameters and therefore it has to be selected arbitrarily.
Some studies based the selection on the dimensions of an elementary fibre [21, 24].
More commonly, V0 is chosen to correspond to a standard unit. The latter convention
is also adopted in this work, which uses a reference volume V0 of 1 mm3.
62 Chapter 4b
Loading Average Gauge Shape Scale Average

Extraction Fibre cross Scale Weibull
Fibre rate diameter length parameter parameter strength
method sectional area variable estimation
(mm/min) (µm) (mm) (m) σo (MPa) (MPa) Ref.
determination
Enzyme
5 3.0 3041 1597 L
Flax retted - Digital images 0.5 19 MLE* [4]
5 3.5 253 1489 V
manual
5 4.2 1459 -
Optical -
Flax Green 1 8 4.0 1359 - L LR** [17]
microscope -
10 3.3 1356 -
-
- -
5 3.6 1168
Optical - -
Flax Dew retted - 8 3.3 1093 L LR**
microscope - - [17]
10 2.2 1543
Enzyme - 5 960
Optical SW: 5.2 SW: 1430 MLE*
Flax retted, 0.5 - 10 870 L [25]
microscope MW: 2.8 MW: 1400
manual - 20 740
selection
Alkali Optical 0.8 140 10 2.7 412 367 L LR** [26]

Bamboo
microscope
Manually 350 80 2.9 228 203 Weib.
Bamboo stripped Profile projector 1.3 550 80 4.0 159 144 - analys. [1]
off 850 80 2.7 119 106 program
Steam By weight and - 15 9.3 855 813 LR** [21]

Bamboo 0.1, 1.5 V
explosion density - 35 3.5 683 639
Brown - 343 [21]

-
By weight and
0.1, 1.5 -
15 9.3 360 V LR**
coir density 35 3.7 206 186
1 6.3 - 428
Optical 8 3.0 - 384 [27]
Coir Retting 20 250 L LR**
microscope 35 4.5 - 211
50 4.9 - 162
Retting 44 5 2.2 436 384

and further Digital 54 10 1.4 415 365 [28]
Jute - L LR**
alkali microscope 47 15 1.3 410 363
treatment 48 20 1.2 377 340
Hemp Manual Optical [29]

0.01 40 10 2.7 - 285 L LR**
Extraction measurement
Weib.
Water Optical - 1.5 3.4 876 786
Hemp 0.5 L analys. [20]
retted microscope - 10 4.2 745 677
program
Mech. 10 4.6 - 391
process 20 3.7 - 392 MLE*
Sisal SEM 0.1 200 - [30]
(decorticati 30 3.6 - 385
ons) 40 3.0 - 400
Agave Traditional Image analysis [31]

20 3.1 7 5.0 205 126 L LR**
Americ. method software
-
20 5.5 3810
Provided by the - MLE*
- 1.5 23 40 5.5 3880 L [14]
E-glass manufacturer -
80 5.1 4440
-
Provided by the 20 4.0 6300 5700
Carbon - 1 20 L LR [32]
manufacturer 40 2.9 5600 5000
(T700S)
Carbon Provided by the 20 5.1 4000 3600

- 1 20 L LR
(G34- manufacturer 40 6.1 3600 3400
700)
Table 4-1. Weibull distribution parameters for some natural and synthetic fibres. *MLE: maximum likelihood method. **LR: linear regression method. L: length.
V: volume. SW: standard Weibull distribution. MW: modified Weibull distribution.
64 Chapter 4b
From Equation 4-1 the average strength (σv), the standard deviation (σsd) and the
modus (Ms), the value with higher repetition into the data, can be calculated
respectively with Equations 4-2 to 4-4 [21]. Γ corresponds to the gamma function.

V m  1
 v   0  V0   1  m  (4-2)
 2
   1     v
2 2
(4-3)
 
sd 0
m
1
 m 1 m 
M  0  V m
(4-4)
 m 
s
According to Equation 4-1, for a constant fibre diameter the average strength (σ2) at
certain volume (V2), can be estimated from a known strength (σ1) at its
corresponding volume (V1) and shape parameter (m) as follows [11]:

V 2  m
 2
  
1
 (4-5)
V 1 
This equation has been useful for predicting fibre strength [22], especially in
analyses of the micromechanical fibre fragmentation test, commonly used to
evaluate the interfacial shear strength at extremely short fibre lengths [33].
4.1.2 Effect of defect density distribution
Equation 4-1 assumes that the defect density is homogeneously distributed over the
volume of the material. Weibull [12] noted, however, that other defect density
distributions are also plausible. In the fibre literature (see Table 4-1) it is commonly
assumed that the defect distribution is only function of the length of the fibre. The
resulting modified Weibull distribution is identical to Equation 4-1 except that
volumes are substituted by lengths (L):

  L   
m

P  1  exp       (4-6)
  L0  0  
The derived statistics (Equations 4-2 to 4-5) can similarly be adapted by substituting
volume by length. While some studies use as reference length L0 the length of an
elementary fibre [21, 24], this study uses 1 mm, as it is more practical.
For fibres with different diameters, the volume-based model and the length-based
model lead to different strength predictions, as is illustrated in Figure 4-2. The figure
divides fibres into segments with equal failure probability for length-based defect
densities (left) and volume-based defect densities (right). If fibres have the same
length, a length-based defect density predicts a failure probability that is
independent of the diameter, while a volume-based defect density gives rise to more
defects for thicker fibres and thus to a higher probability of failure. If the fibres have
the same volume, wider fibres are also shorter, leading for the length-based model to
less defects and thus to a higher strength.
a) Considering: L b) Considering:
V
L 0
V 0
Same length and
different volume
Same volume and
different length
Figure 4-2. Probability of failure for a single (technical) bamboo fibre under tension when it is
considered as a concatenation of small pieces taking into account the a) length (L) and b) volume
(V) as a chain link.
The defect distribution within bamboo fibres can be affected by their morphology
and by the extraction process. The length and diameter of elementary fibres are
homogeneously distributed within the fibre, and the elementary fibre ends are also
distributed homogeneously (see section 4.5.3.2). Since the fibre ends form “weak
links” between the elementary fibres, this fibre architecture suggests a volume-based
defect distribution [21, 24, 34]. The extraction process, however, may add defects
that are distributed differently. For example, certain compression, rubbing or
combing-based extraction steps may systematically introduce damage every certain
distance along the fibre length, thus favouring a length-based defect distribution. In
combing processes involving needles, thicker fibres may have a higher probability
66 Chapter 4b
of needle-attack and undergo higher tensile stresses during the combing process.
Some chemical processes may introduce damage on the fibre surface. The latter two
processes thus favour yet another defect distribution: a surface-based one. Given that
the bamboo fibre diameter can range between 150 and 250 µm for the same batch
[35], it is thus meaningful to consider how defects are distributed within the fibres
and to relate this to the fibre extraction method.
4.1.3 Effect of within-fibre diameter variation
Some studies [6, 36] argue that the reduction in mechanical properties in wool fibres
with increasing fibre length is not only caused by the accumulation of defects along
the fibre but also by diameter variations along the length of a fibre. Therefore,
Zhang et al. [6] and Xia et al. [28] incorporate in their studies on respectively wool
and jute fibres a within-fibre diameter variation parameter λ that replaces β in
Equation 4-6 and obtaining:
  L     m 
P  1  exp        (4-7)
  L0    0  
Where λ is the slope of the line obtained when plotting the logarithm of the
coefficient of variation of the fibre diameter (CVFD) versus the logarithm of the
measured fibre length. This parameter represents the exponential parameter of the
change of the within-fibre diameter variation over the fibre length. Zhang et al. [6]
obtained the CVFD after measuring the fibre diameter variation of a number of fibres
every 40 μm intervals along the fiber length (10, 20, 50 and 100 mm). The
parameter λ is positive, which reflects that there is an increase of the within-fibre
diameter variation (CVFD) with increased fibre length. This parameter must not be
confused with β, mentioned in section 4.1.1 and which is determined by curve fitting
using equation 4-6. λ is an empirical parameter that was introduced to improve the
predictions of the Weibull model with respect to the fibre strength at different gauge
lengths.
4.2 Dry fibre bundle test
An alternative for the estimation of the Weibull parameters and the evaluation of the
mechanical properties of the technical fibres, is the dry fibre bundle (DFB) test [37,
38]. This method is based on the study of the rupture of a bundle (cluster) of fibres,
assuming spatial random distribution of failure of the fibres. The methodology for
the dry fibre bundle test was initially developed by Daniels [39] and Coleman [40]
and later on applied in several studies, particularly on synthetic fibres as for example
carbon [38], E-glass [41-43], Kevlar® [44] and ceramic filaments [45].
According to the mentioned studies, the DFB test helps to overcome some of the
drawbacks found when testing individual (synthetic) fibres. Some of these
disadvantages are a tedious selection of very small diameter fibres from a tow
inducing damage to the samples, preparation and gluing samples onto a paper frame
as well as the large amount of required samples to obtain reliable results
(normally >30 is recommended) [38, 44]. All this preparation becomes very time-
consuming for the SF testing. For these reasons, the DFB test has gained increased
support in recent years, since the statistics concerning the fibre strength are more
conveniently obtained using fibre bundles and are more relevant to the situation
which might prevail in the finished fibre reinforced composite material [44].
To the best knowledge of the author, Weibull parameters and mechanical properties
for natural fibres have not been reported in literature using DFB testing methods,
even though the SF test is known to be a time consuming preparation technique.
4.2.1 Theoretical background
For the implementation of the DFB test several conditions have to be taken into
account [38, 41, 44]. The strength distribution of a technical single bamboo fibre
under tension is assumed to follow the two-parameter Weibull distribution; also that
the relationship between the applied stress (σf) and the strain (ε) for single fibres
follows Hooke’s law up to fracture as shown in Equation 4-8.
 f  Ef  (4-8)
From equations 4-6 and 4-8 it is possible to obtain Equation 4-9.
P( )  1  exp  L ( /  0 )m 

(4-9)
Where P(ε) is the failure probability of a single fibre under strain no greater than ε.
ε0 is the scale parameter of the Weibull distribution for the strain. ε0 can be obtained
from Equation 4-10 [44]:
 o   max  Lm 
1/ m
(4-10)
68 Chapter 4b
εmax corresponds to the maximum strain at maximum load (Fmax) reached by the
bundle. At an applied strain, the number of surviving fibres (N) in a bundle which
initially consist of No fibres is shown in Equation 4-11 [38]. N can be related to the
applied load (F) on the bundle by Equation 4-12.
N  No 1  P( )  No exp  L ( /  0 )m  (4-11)
F ( )   AN  ANo E f  exp  L ( /  0 )m  (4-12)
This latest expression (Equation 4-12) is the load-strain (F-ε) relationship for a
bundle of fibres under tension, where A is the average cross-sectional area of a
single fibre and Ef is the Young's modulus of the fibre. The Weibull shape parameter
(m) can be obtained via two methods; the first one [38, 44] is through the maximum
load using Equation 4-13, where So corresponds to the initial slope of the load-strain
(F-ε) curve.
m  ln( max So / Fmax )

1
(4-13)
The second method to determine m consists in applying a graphical method with the
F-ε curve as is shown schematically in Figure 4-3 using So, and S* [37]. This last
value (S*) corresponds to the slope of the straight line defined by the origin of the F-
ε curve connected to the maximum load. Then, the Weibull modulus is obtained
with Equation 4-14. This method gave similar values for the m parameter in
comparison with those obtained by other studies using the maximum load method
described above [44]. The maximum load method (Equation 4-13) will be preferred
in case the maximum strain is not clear to determine from the graphic method.
F(N)
1/ m  ln(So / S  )
εmax
ε(%)
Figure 4-3. Graphic method to determine the fibre scale parameter (m) from the load-strain (F-ε) curve [37].
1/ m  ln(So / S  ) (4-14)
The second Weibull scale parameter (σo), can be obtained from equation 4-15 as
follows [38, 41]:
 o   max / m1/ m (4-15)
The bundle strength (σb) corresponds to the Fmax divided by the area of the individual
fibres multiplied by the total number of fibres present in the bundle (Equation 4-16).
In this study, the bundle strength does not represent the individual technical fibre
strength (σf). Coleman [40] established his theory on the basis of Weibull theory and
Daniels’ theory [39]. The relationship between the strength of the fibre bundle and
strength of the individual fibers is indicated in Equation 4-17 [46], where e is the
base of the natural logarithm. Finally, the fibre Young’s modulus can be obtained
from Equation 4-18 [44].
 b  Fmax / No A
(4-16)
1
 f   b (me) (1  1/ m)
m
(4-17)
E f  So / ANo (4-18)
4.3 Materials
Bamboo culms (Guadua angustifolia) were collected from a typical bamboo

plantation in Colombia, specifically from the Coffee Region, at 1.300 meters above
sea level, annual average temperature of 23 °C, annual average precipitation of
2.200 mm and relative humidity of 80% according to the environmental authorities
of the region. Technical fibres were extracted by the author from the bamboo culms
using a proprietary purely mechanical extraction process that neither uses chemicals
nor high temperature. The maximum length of the extracted fibres was the internode
length, which for 48 month culms is reported to be between 20 and 35 cm [47].
70 Chapter 4b
4.4. Methods
4.4.1 Single fibre test
4.4.1.1 Measurement of the cross sectional area
A first important source of uncertainties when obtaining the strength properties of

the fibres is the methodology for determining the cross sectional area, due to the non
circular cross-section, that can lead to a large difference in the reported tensile
strength measurements [48]. For this reason, special care was taken for the
measurement of the cross sectional area of the individual technical bamboo fibres.
This included not only strict standard conditioning of the fibres before
measurements, but also the comparison of two different techniques implemented for
15 randomly selected technical fibres, to verify the accuracy of the results. The first
method (Aw) comprises of weighing the fibres and then dividing the weight by the
fibre length and apparent density (1.4 gr/cm3 [35]), obtained by a gas pycnometer
Beckman 930 at ~3 mm fibre length and following the methodology described by
Tran L. [49]. The diameter can be estimated assuming that the fibres have a constant
cross-section and circular shape.
The second method (AS), is considered more accurate, consisting in the direct
measurement of the cross sectional area with image analysis of the tested fibres. For
this, each fibre is collected at the end of the tensile test and cut carefully near the
breaking point. The fibre is then vertically encapsulated in soft resin, polished and
imaged by Scanning Electron Microscopy (SEM XL30 FEG). The images are
processed using the image analysis software Digimizer 3.0 to measure the cross-
sectional area. The area of all other fibers, used in single fibre test (420 fibres), was
only measured using the first method (Aw).
4.4.1.2 Measurement of the fibre perimeter
The variation in perimeter was determined along the length of 12 randomly selected
individual technical bamboo fibres. For this, wetting measurements in n-Hexane
were performed with a Krüss K100SF tensiometer using the Wilhelmy technique
according to the methodology described by [50]. The perimeter measurement was
done every 100 µm along the technical bamboo fibre length of 20 mm at a speed of
1 mm/min with a detection speed of 3 mm/min (approaching speed) and a detection
sensitivity of 1µg. With the obtained data, the cumulative coefficient of variation of
the fibre perimeter (CVFp) was calculated for 5 different lengths along the fibre (3.8,
7.8, 11.6, 15.5 and 20 mm), see a schematic representation in Figure 4-4. In
principle this technique allows measuring the fibre perimeter continuously along the
fibre with a reasonable accuracy of less than 8% relative error [51]. From the
perimeter, the fibre diameter can be estimated, assuming circular shape.
As mentioned before in section 4.1.3, according to Zhang et al. [6], the slope of the
line obtained when plotting the logarithm of the coefficient of variation of the fibre
diameter (CVFD) versus the logarithm of the measured fibre length, corresponds to
the within-fibre diameter variation parameter (λ) along the technical fibre length.
Total measured distance = 20 mm
100 µm
Technical fibre #1
Perimeter Perimeter Perimeter Perimeter

1 2 n n+1
Technical fibre #2
Technical fibre # n
Length 1
Average CVFP1
Length 2
Average CVFP2
Length n
Average CVFPn
Figure 4-4. Schematic representation for the calculation of the CVFP at 5 different fibre lengths (for a total
length of 20 mm). The measurements were performed on 12 randomly selected bamboo technical fibres using
the Wilhelmy technique. The arrows represent the technical fibre perimeter measurement at this point.
4.4.1.3 Tensile test set up
Before tensile testing, randomly selected fibres were visually inspected. This
selection consisted in making a manual loop with the technical fibre to verify the
absence of major damage along the length of the technical fibre, see Figure 4-5. This
fibre selection was carried out with the purpose of testing only technical fibres in
good condition without pre-damage due to external reasons such as packaging,
72 Chapter 4b
transportation or manipulation of the fibres after the extraction. The good condition
of the fibres will assure more reliable results. The fibres considered suitable to be
tested were pre-conditioned at 21°C ± 2°C and 50 ± 2 %RH for at least 24 hours
before the experiment in order to assure they reached a moisture equilibrium.
The single fibre tensile test was carried out using a mini-Instron 5943 under standard
environmental conditions (21°C ± 2°C and 50 ± 2 %RH). The fibres were
pneumatically clamped at 5 bar of pressure using rubber-faced clamps of 10 x 30
mm. These clamps allowed reducing fibre manipulations and avoiding the use of a
paper frame. A vertical visual reference was used during placement of the samples
on the grips in order to avoid fibre misalignment. The latter can lead to bending
stresses at the grips, and thus causing premature failure. The crosshead speed was
0.85 mm/min with a 1kN load cell. The load was registered during the entire test.
Fibres of gauge length 1, 2, 5, 10, 20, 30 and 40 mm were tested. The number of
samples and average fibre diameter per batch are given in Table 4-3 in the results
section.
a. b.
Figure 4-5. Visual inspection (applying manual loop) of the bamboo technical fibres before tensile testing, a)
technical fibre suitable to be tested and b), a rejected technical fibre due to major damage (indicated by the
arrows).
The strain rates on the fibres are different for each span length considering that the
displacement rate used in this study was always constant (0.85 mm/min). The strain
rate varied between 0.00035 s-1 and 0.014 s-1 for 1 mm and 40 mm fibre gauge
lengths respectively. These values can be considered to be sufficiently low in order
to expect no significant variations in the fibre strength due to the differences of the
strain rate, especially for a brittle bamboo technical fibre.
4.4.1.4 Statistical calculations
All statistical calculations were performed using MATLAB 2009b (The MathWorks
Inc., Natick, Massachusetts, US). The statistical inferences assume a confidence
level of 95% (α= 0.05).
4.4.1.5 Scanning electron microscopy (SEM) observations
Micrographs of fibres and composites were made by scanning electron microscopy

(SEM30 XL FEG). The samples were sputter coated with gold for further
observations using secondary electrons using a voltage between 10 and 15 kV.
4.4.2 Dry fibre bundle test
4.4.2.1 Reference methodologies
Through scientific literature regarding the dry fibre bundle (DFB) testing, no
international standard or norm was found to perform the DFB test applied to
synthetic fibres. This is even though the methodology described in section 4.2.1, has
been developed and implemented by several authors exclusively for this type of test.
The bundle characteristics and testing parameters for glass and carbon fibres ranged
from one study to another in terms of the number of filaments in the roving
(between 194 [41] and 4000 [42]), the gauge length (between 30 and 100 mm [38,
41-44]) and the displacement rate (between 0.02 [37] and 0.05 [41] mm/min). It
might be considered that these DFB characteristics differ significantly in comparison
with the DFB samples made of bamboo fibres. These differences are related to the
amount of fibres in the samples and the regular “perfect” cross sectional area found
in individual synthetic filaments, instead of an irregular geometry and the larger
diameter of technical bamboo fibres.
For natural fibres, standard test methods for wool (ASTM D1294 [52] and D2524
[53]) and cotton fibres (ASTM D1445 [54]) were found. These ASTM standards
specify the methodology to calculate the bundle tensile strength (σb), the Young’s
modulus (Ef) and the strain at maximum load (εmax). The basic differences between
the procedures employed in method D1294 and D2524 are in the gauge lengths
(25.4 and 3.2 mm respectively) and the methods for clamping the bundle sample. In
test method D2524, specific clamps are required whereas in D1294, conventional
clamps may be used. The ASTM D1445 standard also uses 3.2 mm of gauge length
(flat bundle) using a standard clamping system. Nevertheless, the preparation
methodology for the DFB samples in these standards was found to be poor.
74 Chapter 4b
The technique to uniformly load the fibres consists in aligning them by hand and to
stretch the fibres by holding them with the fingers and keeping them together with
masking tape. In the next section, the importance of an appropriate DFB preparation
in order to obtain reliable data will be explained.
4.4.2.2 Preparation of the bamboo DFB samples
The first aspect to be taken into account during the DFB preparation is the alignment
of the fibres in order to avoid a non-uniform loading. The theoretical background
shown before in section 4.2.1 is valid only if assuming that the fibres in the dry
bundle, undergo the same strain when loading (see Equation 4-11 and 4-12) among
the N surviving fibres before breakage. An inadequate alignment causes an uneven
distribution of the load throughout the fibres in the bundle during tensile testing
infringing the theoretical assumption and leading to a premature failure of the
bundle and thus to a sub-estimation of the fibre parameters. The influence of a non-
uniform loading on the bundle is shown schematically in Figure 4-6a. It might be
also noted that an accurate measurement of the Young’s modulus (Ef) requires that
the fibres in the bundle are perfectly aligned and uniformly strained [37].
a. b.
Figure 4-6. a) Schematic sequential representation of a non-uniform loading during the DFB test due to the
misalignment of the fibres in the sample. b) effect of misalignment in the grips during the tensile testing [55].
For the preparation of the samples, the bamboo fibres where carefully stretched,
aligned and evenly spread by hand reaching an average areal density of 232 ± 24
g/m2 and thickness of 0.47 ± 0.03 mm as shown in Figure 4-7a. The bundles were
glued with cyanoacrylate glue at the ends between aluminium foils (griping zone), to
load as uniformly as possible all the fibres during the tensile test. The distance
between these aluminium foils corresponded to the desired gauge length. The tested
flat bundle configuration (see Figure 4-7b), allows an appropriate and uniform
holding of the bundle to be used with conventional tensile test grips, without a
special set-up to perform the tensile test In this study, the influence of the gauge
length and the amount of fibres in the bundle was evaluated. For that, four types of
bamboo DFB samples were prepared varying the gauge length (40 and 100 mm) and
the width (10 and 250 mm), as presented in Table 4-2.
a. b.
Figure 4-7. a) Front and b) lateral view of the bamboo dry fibre bundle.
Type of bamboo dry fibre bundle
Properties TkL TkS ThL ThS
Gauge length (mm) 100 40 100 40

Width (mm) 25 ± 0.5 25 ± 0.7 10 ± 0.6 10 ± 0.6
Average thickness (mm) 0.44 ± 0.10 0.51 ± 0.10 0.49 ± 0.10 0.43 ± 0.10
Average areal density (g/m2) 215 ± 25 256 ± 23 223 ± 22 236 ± 26
Average number of fibres* 195 ± 15 211 ± 17 64 ± 15 76 ± 13
Number of successfully tested samples 6 5 8 8
Table 4-2. Geometry and general characteristics of the bamboo DFB test specimens. *Calculated from the
correspondent fibre length, the total mass of the fibres present in the bundle (after testing, the fibres were
carefully cut corresponding to the gauge length and weighed), and the average fibre cross sectional area
(calculated independently from the 420 samples in SF test).
76 Chapter 4b
4.4.2.3 Testing of the bamboo DFB samples
The accurate measurement of the load-strain (F-ε) curve is one of the most
important aspects for a reliable evaluation of the fibre Weibull parameters and the
mechanical properties for individual technical bamboo fibres through DFB testing.
This condition states the necessity of measuring the true strain during the DFB
tensile test, which is generally calculated using the displacement of the cross-head
measurement of the testing machine [37]. In this research, an optical extensometer
was used to obtain the true strain values of the bamboo DFB at different places on
the sample. This was achieved using correlation of digital images taken of the
sample during loading (4 images per second) by the Vic2D software (LIMESS
Messtechnik und Software GmbH). To achieve better image recognition by the
software, the samples were first painted with black random speckles to give a unique
pattern to the surface, following the methodology described by [56]. For optimal
results, it is necessary to obtain an adequate speckle pattern where it must have a
considerable quantity of black speckles with different shapes and sizes. Such
patterns can then be identified by the strain mapping software program, see Figure
4-8.
a. Speckles on the bamboo

Dry fibre dry fibre bundle surface
bundle
b.
Camera
5 mm
Figure 4-8. a) Set up for the DFB tensile test including the camera for the optical extensometer in order to
register the deformation of the sample and b), speckles pattern on the sample.
As the tensile test proceeds, the system takes subsequent images and analyses them
in comparison to either the initial image or the previous image. The correlation
between the images allows the determination of local displacement and strain values.
The difference of using the displacement of the grips and the optical extensometer to
determine the bundle strain (ε), for the same bamboo DFB sample, is clearly shown
in Figure 4-9. The measured gauge length was 50 mm for TkL and ThL specimens,
and 20 mm for TkS and ThS samples. An Instron 4467 testing machine with a load
cell of 30 kN was used to perform the tensile test with a crosshead displacement rate
of 0.3 mm/s with a correspondent strain rate of 0.0075 s-1 and 0.003 s-1 for the short
(40 mm) and long (100 mm) DFB samples, respectively.
Using the
displacement from
the grips
Using optical
extensometer
Figure 4-9. Comparison of the load-strain (F-ε) curves for the same DFB demo sample (TkS), using optical
extensometer and the displacement of the grips for the determination of the strain (ε).
Strain 1 Grip
Strain 2
Strain 3
DFB sample
Strain 1
Strain 2
Strain 3
Grip
Figure 4-10. Comparison of the strain measurements with an optical extensometer in different zones (strains 1,
2 and 3) of the DFB sample during the tensile test to verify the homogeneous deformation of the bundle to
obtain reliable results.
78 Chapter 4b
To avoid misalignment of the bundle sample in the grips which could cause a
premature failure of the fibres in the bundle, as is shown in Figure 4-6b, a visual
reference served as a guide to assure proper positioning of the DFB samples in the
grips. To verify the uniform loading of the DFB during the tensile test, the sample
strain was measured in at least three different zones as is shown in Figure 4-10.
These measurements should be close to each other to assure reliable results.
4.5 Results and discussion
4.5.1 Fibre extraction
An innovative method for extraction of long technical bamboo fibres was co-invented
by the author and a prototype is currently operating at pilot scale. This machine was
designed with mechanical principles that can be scaled-up for future large volume
industrial production. The technology has been implemented for Colombian bamboo
(Guadua angustifolia), but it is equally applicable to other bamboo species. This
novel fibre extraction technique presents several advantages compared to the other
methods described in section 3.7.1 (e.g. steam explosion, chemical and mechanical
extraction), such as:
 Production of long bamboo technical fibres. The length of the extracted fibres is
the internode length, which for 48 month old culms is reported to be between
20 and 35 cm [47].
 During the extraction process the use of high temperature, high pressure or
chemicals is not needed, reducing both the damage introduced to the fibres and
the amount of energy required for the extraction of bamboo fibres. These
aspects make this new process a more environmentally friendly method
 It is a standardized in-line process that avoids storage or batches during the

process and assures a continuous production. See also Chapter 7 for more
details about the and technical and environmental assessments of the extraction
method.
A second operation required to remove the soft tissue (parenchyma) present on the
fibre surface of the extracted technical fibres (see Figure 4-11), is the cleaning
process. For this procedure, a proven lab-scale prototype machine, which can also be
scaled up for industrial purposes, was developed and constructed at KU Leuven.
Figure 4-12 shows cleaned bamboo technical fibres after the extraction and cleaning-
combing process. The extracted technical fibres described in this section, were used in
this study not only for SF and DFB characterization, but also for the production of
composites with thermoset and thermoplastic matrices (Chapters 5 and 6).
Elementary fibres for Guadua angustifolia have an average length of 2.1 mm and an
average diameter of 17 µm [57]. The single technical fibre can be regarded as a
concatenation of elementary fibres whose pentagonal or hexagonal cross-sections
are arranged in a honeycomb pattern, see Figure 4-13. The mechanical properties of
the mechanically extracted fibres as well as some fibre morphology aspects will be
discussed in the next sections.
a.
a.
b.
b.
Figure 4-11. a) Mechanically extracted technical bamboo fibre with soft tissue (parenchyma) and b), bamboo
technical fibre surface after cleaning.
Figure 4-12. Bamboo technical fibres after mechanical extraction (maximum length between 20 and 35 cm).
80 Chapter 4b
20 µm
Figure 4-13. Technical bamboo fibre composed of elementary fibres.
4.5.2 Fibre cross-sectional area and perimeter
The cross sectional area of 15 fibres was determined using two methods. The first
one (Aw), was based on the weight of individual fibres and the apparent average
density of bamboo fibres. The second one (AS), was based on analysis of SEM
images (Figure 4-14a). The average relative error between both methods was 3.8 ±
1.5%. For the area determination for all tested fibres (420 specimens), only method
Aw was used because of its practical simplicity. Fibre diameters were found to range
from 100 to 220 µm, as shown in Figure 4-14b. The fibre diameter distribution of all
tested fibres is shown in Figure 4-15 (diameter obtained from area by assuming
circular shape), also distinguishing between a thick and a thin fraction. The
coefficient of variation was 20%. Figure 4-16 presents a typical perimeter variation
profile along the length of a single technical bamboo fibre using the Wilhelmy
method. The measured fibres with this technique (12 samples) show a low perimeter
variation with an average of 0.68 mm ± 0.1 and a average within perimeter variation
along the measured length (20 mm) of 0.078 mm.
a. b.
Figure 4-14. a) SEM image of a bamboo fibre cross-section with contour perimeter line and b) different fibre
diameters found in this batch of fibres.
Nt= 210
Number of fibres
Fibre diameter range (µm)
Figure 4-15. Fibre diameter distribution for all single bamboo technical samples indicating two halves of the
fibre population (Nt=420) (diameter obtained from cross-sectional area by assuming circular shape).
Figure 4-16. Typical variation of fibre perimeter along the length of a technical bamboo fibre, measured
every 100 µm using the Wilhelmy technique.
82 Chapter 4b
4.5.3 Mechanical properties of bamboo technical fibres
4.5.3.1 Dependency of fibre strength on fibre length
For each gauge length, individual fibre strength values are plotted in Figure 4-17 and
average results are shown in Table 4-3. These data show a decrease in fibre strength
with increasing gauge length. ANOVA revealed statistically significant strength
differences between the groups with gauge length 1 and 2 mm, 5 and 10 mm and 20,
30 and 40 mm. The fibres with gauge length 1 mm have not only a higher strength,
but also a lower variance. Because the gauge length is shorter than the length of an
elementary fibre (2.1 mm), it is likely that entire elementary fibres are clamped
between top and bottom. This is illustrated in Figure 4-18. In this case the crack has
to run through the cell wall breaking the elementary fibres (Figure 4-18a), instead of
causing debonding between them (Figure 4-18a). As a consequence, the strength
increases and a different fracture type can occur, especially at very short gauge
lengths (i.e. 1 mm). The same may also occur for gauge length 2 mm, but to a much
lesser extent.
Tested Average N° Average *Average *Average **p value (t- ***R2

gauge fibre samples strength strength strength test) for “thin” (linear
length diameter all fibres “thin” fibres “thick” and “thick” regression)
(mm) (µm) (MPa) (MPa) fibres* fibres (α=0.05)
(MPa)
1 132 ± 33 37 943 ± 94 942 ± 100 945 ± 90 0.935 0.006
2 137 ± 36 34 898 ± 124 960 ± 118 837 ± 100 0.002 0.180
5 161 ± 27 50 833 ± 113 802 ± 120 863 ± 98 0.055 0.026
10 156 ± 27 43 821 ± 125 839 ± 143 802 ± 101 0.338 0.023
20 164 ± 32 54 754 ± 72 738 ± 74 770 ± 66 0.099 0.001
30 112 ± 8 159 733 ± 121 718 ± 131 749 ± 109 0.117 0.038
40 156 ± 23 43 748 ± 115 787 ± 111 707 ± 107 0.021 0.055
All 146 ± 26 420 790 ± 132 790 ± 144 791 ± 118 0.949 0.007
Table 4-3. Fibre diameter and strength for bamboo fibres tested at different gauge lengths. *Each group is
divided in ‘thick’ fibres, having diameter above the median, and ‘thin’ fibres for strength comparisons. **Two
sided t-test (for p≥0.05, there is no significant difference). ***Linear regression between fibre diameter and
strength.
For larger gauge lengths, the number of elementary fibre ends increases. The
lamellar interphase that bonds the elementary fibres is typically weaker than the
elementary fibre. The interphase, which is rich in lignin [51] may become the critical
site through which fibre failure takes place in tension, as has been observed for flax
fibres [58]. The larger the gauge length, the larger the number of critical sites, and
thus the higher is the probability of failure. Recall that critical flaws can also have
other origins such as imperfections developed during growth and defects introduced
during extraction [27, 59].
Figure 4-17. Bamboo fibre strength at different gauge lengths and their correspondent PDF’s (Probability
distribution functions).
a. b.
Figure 4-18. Schematic view of a single technical bamboo fibre tested at two different gauge lengths with a
typical type of fracture depending of the gauge length. For a) the crack has to run through the cell walls
breaking the elementary fibres and b), the crack runs through the lignin layer surrounding the elementary
fibres due to a weaker interphase.
84 Chapter 4b
It was possible to carry out this Weibull analysis without fractioning the analysis in
“short” (i.e. 1 and 2 mm) and “long” (>2 mm) gauge lengths, with no concern about
the critical discontinuities at short lengths. The technical bamboo fibres are
composed by hundreds of elementary fibres, for this reason, even at short gauge
lengths, the typical discontinuities of the elementary fibres at certain points will be
bridged by many other elementary fibres placed around, as seen in Figures 3-11 and
4-13. This situation contrasts, for example, with the flax fibre, where the technical
fibres are assembled only by 10–40 elementary fibres of 33 mm average length [60].
In this case, the use of Weibull statistics at short gauge lengths and the effect of
these discontinuities need to be verified first to ensure reliable results.
To obtain more information about the origin and mechanisms of fibre failure, SEM
micrographs were made of the post-mortem fibre fracture surface. As shown in
Figure 4-19, mainly three types of fracture were distinguished: (a) a straight crack
right through the elementary fibres, (b) a crack proceeding along the primary wall
layer that surrounds the elementary fibres and (c), a combination of the two. For
fracture type (a) and (b), observed also by [58], the stress-strain curve is linear until
failure. For fracture type (c) the stress-strain curve showed, for a number of samples,
small drops in the load during the test before reaching ultimate failure. It was
analyzed whether there was a correlation between type of failure, fibre diameter and
gauge length. The latter two parameters appeared not to influence the occurrence of
the failure types except for very short fibre gauge lengths.
Figure 4-19. Fracture micrographs showing a) clear transversal fracture (b) crack propagation through the
primary layer of the elementary fibres and (c) a combination of the two cases for bamboo fibres after single
fibre tensile test.
4.5.3.2 Dependency of fibre strength on fibre diameter, fibre volume and fibre
surface
Figure 4-20 shows the strength of individual technical fibres versus their
corresponding fibre diameter for all samples included in this study. Table 4-3 reports
the average fibre diameter for all fibres and for each set of different gauge length.
ANOVA (α=0.05) revealed no statistically significant differences in strength as

function of fibre diameter, even when evaluated at different gauge lengths. In
addition, the fibres were divided into a group of ‘thin’ fibres and a group of ‘thick’
fibres, based on whether the fibre diameter was respectively lower or higher than the
median fibre diameter. If the fibre diameter would influence the strength, then a
systematic difference in strength should be found for each gauge length. Table 4-3
shows only a statistically significant difference, according to the Student’s t-test, for
the group with gauge lengths 2 and 40 mm. When all ‘thin’ and ‘thick’ fibres
(Figure 4-20) are compared together, almost no difference in strength is found. The
‘thick’ fibres do, however, have a lower variance than the thin fibres. When the
largest group (30 mm) is further divided into four groups with increasing diameter,
ANOVA does not show a significant difference between these groups. These results
indicate that, notwithstanding a factor 3 difference between the lowest and the
highest tested fibre diameter, there is hardly any effect of the average fibre diameter
on strength.
Figure 4-20. Single technical fibre strength versus fibre diameter where the median for fibre strength and fibre
diameter are shown.
For the length-based model, predicted average strength and modus, Figure 4-17
presents the probability density functions and 95% confidence intervals for the
experimentally obtained data. For the surface-based and volume-based models,
Figure 4-21 plots the predicted average strength, modus and 95% confidence
intervals. It seems in both cases a similar trend as for fibre length is followed. As no
effect of fibre diameter was found, the effects of fibre surface area and fibre volume
on fibre strength must be attributed to the effect of fibre length, as both area and
volume are linearly related to length.
86 Chapter 4b
Figure 4-21. Predictions of fibre strength for all fibres as function of a) fibre surface area and b) fibre volume.
4.5.3.3 Estimation of the modified Weibull parameters
The modified Weibull distribution was used to quantify the strength of bamboo
fibres. It must first be established whether the defect density distribution should be
described as a function of fibre volume (Equation 4-1), length (Equation 4-6) or
fibre surface area. Subsequently three parameters must be estimated. These
parameters were estimated for each of the three defect density distributions based on
the entire dataset using maximum likelihood estimation (MLE). The obtained
parameter values are shown in Table 4-4.
Scale Scale Shape Sensitivity

variable parameter parameter parameter β
σ0 m
(MPa)
Length 982 ± 29 7.6 ± 0.5 0.48 ± 0.09
Surface area 761 ± 39 6.7 ± 0.5 0.93 ± 0.44
Volume 637 ± 85 6.7 ± 0.5 0.46 ± 0.22
Table 4-4. Maximum likelihood estimates of the modified Weibull parameters for all fibres, assuming that the
defect density distribution is function of fibre length, fibre surface area or fibre volume. The reported
variations are 95% confidence intervals.
It can be seen that the parameter values are physically acceptable for all three defect
density distributions. The 95% confidence intervals on the parameters are overall the
smallest for the length-based model. This indicates that the length-based model
yields the most precise prediction. The wider confidence intervals for volume and
surface area also support the finding in the previous section that the fibre diameter is
poorly correlated with strength. For the length-based model, the predicted average
strength, modus, probability density functions and 95% confidence intervals (for
experimental points) were shown in Figure 4-17.
It was noticed earlier that for gauge length L=1 mm, the failure behaviour may be
positively biased, because it is shorter than the elementary fibre length. To
investigate this influence, the MLE analysis was repeated for gauge lengths L ≥ 2
mm. For the three models, the results were similar, with the lowest differences
occurring for the length-based model (σ0= 987 ± 42 MPa, m=7.5 ± 0.6, β=0.49 ±
0.11). From this insensitivity it can be concluded that the length-based model
reproduces the strength increase towards lower gauge lengths. It is thus a practical
predictor of bamboo fibre strength over the entire range of measured gauge lengths.
Mod. Weibull (length-based model, Eq. 4-6)

Weibull (original length-based model)
Experimental data
Figure 4-22. Comparison between experimental strength and predictions of the modified and non-modified
Weibull model. Predictions are based on the results of all fibres. Error bars 95% confidence intervals on the
average strength.
The modified Weibull parameters obtained for the length-based model are
physically meaningful. The scale parameter σ0 of 982 MPa is somewhat higher than
the average strength for gauge length L= 1 mm, as it corresponds to a failure
probability of 63.2% at that gauge length. The shape parameter m of 7.6 is relatively
high in comparison with other natural fibres and some glass fibres. This indicates
that the bamboo fibres obtained by the current extraction process have relatively low
strength variability. The sensitivity exponent β of 0.48 ± 0.09 indicates that bamboo
fibre strength decreases significantly less with length than predicted by Weibull’s
original model, which has β=1. As shown in Figure 4-22, when using Weibull’s
original model (σ0=1148 and m=8.6, β=1) it had a stronger length-dependency than
88 Chapter 4b
the experimental data and the modified Weibull length-based model. It also
mispredicted the average strength for short (≤5 mm) and long (≥30 mm) gauge
lengths. The modified Weibull distribution predicts more accurately the
experimental values in comparison to the original model. Repeating the Weibull
fibre strength predictions taking into account fibre volume yielded no improvement.
A potential explanation in terms of within-fibre diameter variations is further
explored in the following section. The Weibull parameters calculated by linear
regression (LR) are shown in Appendix 1.
4.5.3.4 Correlation with within-fibre diameter variations
Variations in cross-sectional area along the length of a fibre induce variations in

stress that may lead to a different overall fibre strength compared to a constant fibre
cross-section with the same average area. As a measure of the within-fibre diameter
variation with gauge length, the parameter λ was shown in section 4.1.3. This
parameter is graphically determined as the slope of the plot of the logarithm of the
coefficient of variation of the fibre diameter CVFD versus the logarithm of the
measured length along the fibre. Because the fibre perimeter was more readily
measured than the fibre diameter, this study uses fibre perimeter. The obtained
coefficient of variation of the fibre perimeter CVFP is closely related to the CVFD,
since perimeter and diameter are proportional if the cross-sectional shape does not
vary.
Figure 4-23. Average CVFP along the fibre length.
Figure 4-23 shows that CVFP increased with fibre length and the exponent λ=0.47
(R2= 90%). This value was very close to parameter β (0.48) obtained earlier,
although this may be coincidental, as there is to our knowledge no direct rigorous
relation between both parameters. In literature it was found that for wool [6] and jute
[28] λ was found to be 0.18 and 0.33 respectively. According to these studies, an
improved strength prediction of the original Weibull’s model is found if the fibre
diameter variation λ is taken into account. In this study however, the difference
between β and λ was very small (0.48 vs 0.47), and hence the suggestion from
literature to use λ could not be confirmed.
4.5.3.5 Benchmarking of SF mechanical properties
A correlation between the mechanical properties (i.e. fibre strength and Young’s
modulus) and various corresponding extraction processes can be seen in Figure 4-
24. In this figure, starting at the lowest values for the strength, one finds the
manually stripped fibres and the mechanical extraction carried out with a rolling mill
machine. For the first method, even though the fibres do not experience any
extraction, the low strength can be explained because the fibres are not completely
cleaned. Very often the cross sectional area of the fibres is overestimated due to
impurities and soft tissue attached to the surface that do not contribute in carrying
the load and give an underestimation of the mechanical properties. For the second
case, milling clearly shows to be an aggressive extraction process that severely
damages the technical fibre with an additional disadvantage of producing relatively
short fibres (<5 mm) [61].
The next group concerns chemical extraction processes giving strength values
ranging from 340 to 450 MPa and a Young’s modulus close to 19 GPa. With these
methods, in some cases, it is necessary to add a mechanical operation for removing
the soft tissue and impurities still present on the fibres after the chemical treatment
for the final refinement of the fibres. Steam explosion gives a higher modulus than
the previous methods, with a rather constant value of 35 GPa with a few exceptions
and with a strength ranging from a moderate value of around 500 MPa to a good
value of 720 MPa reported by Defoirdt et al. [21]. Some mechanical processes,
using different principles (e.g. grinding) in combination with an additional mild
chemical treatment, reach strength values of 625 MPa and a Young’s modulus of 36
GPa at best. The mechanical extraction process carried out in this study introduced
little damage to the fibres and preserved the intrinsic good properties of the fibres as
is shown in the upper right corner in Figure 4-22. Figure 4-25 shows the variability
of technical bamboo fibre strength and Young’s modulus reported in literature
represented by a “box area” whose sides represent the values for the standard
deviation of the data. For the case when the authors did not report the standard
deviation, a single point in the graph indicated the average values of their results.
90 Chapter 4b
Young’s modulus (GPa)

Fibre strength (MPa)
Figure 4-24. Overview of extraction processes and corresponding fibre mechanical properties; the strength
values for the currently studied bamboo technical fibres were determined at 5 mm gauge length (see Table 4-3)
[21, 62-69].
This study*
Figure 4-25. Strength vs Young’s modulus from some studies on technical bamboo fibres [8, 21, 62-65, 67, 70]
*Mechanical properties measured at 5 mm gauge length (see Table 4-3). As a reference, the E-glass fibre has
an effective tensile strength of around 1500 MPa and a Young’s modulus of 73 GPa.
The results show that the bamboo fibres extracted in this research with the author’s
proprietary mechanical method, have high values of strength and modulus, being at
the top right of the graph. The specific properties of the bamboo fibres (normalized
to the material’s density) characterized in this study are 595 kN∙m/kg and 31 x103
kN∙m/kg for the strength and Young’s modulus respectively. These values are
comparable to the specific properties of E-glass, one of the most used synthetic
fibres used nowadays to reinforce polymer composites. This means that bamboo
fibres can potentially replace this synthetic fibre in several applications.
4.5.4 Dry bamboo fibre bundle
A typical load-strain (F-ε) curve for a TkL (Thick-Long) sample is shown in Figure
4-26 in comparison with a theoretical curve obtained with Equation 4-12. In general,
the experimental curves exhibited a regular increase of the force until a well defined
maximum force value indicated by point A. This characteristic is the main indication
of an appropriate failure of the bundles.
A C
Figure 4-26. Comparison of the (typical) experimental and theoretical load – strain (F-ε) curves for a TkL
sample. The parameters to obtain the theoretical curve using Equation 4-12 were taken from Table 4-5.
Moreover, the experimental curve showed a deviation from the initial linear
behaviour at a certain load value, as seen at point B in Figure 4-26. This behaviour
was also observed in similar studies concerning DFB tests on carbon [37], E-glass
and Kevlar [44]. A gradual change in the slope but with a continuous curve (without
slacks) is an indication that the fibres effectively fail in an individual and random
way [37, 44]. Figure 4-26 shows that the Weibull parameters (i.e. m and εo) and the
92 Chapter 4b
test conditions, especially the alignment of the fibres and the determination of the
true deformation of the bundle, were correctly chosen or determined. This is proved
by the good correspondence between the experimental and theoretical curves for
most cases. It must be noted that failures of technical fibres near the grips were
rarely observed, which confirms the absence of stress concentrations on the bundle
due to the clamping system. Figure 4-27 shows a bamboo DFB specimen after
failure.
a. b.
Figure 4-27. Bamboo DFB specimen after failure a) TkS and b) TkL sample.
During the tensile test, after the maximum load point (Fmax), see Figure 4-26 point A,
the fibres abruptly moved due to multiple fibre fractures occurring at the same time
as can be seen in Figure 4-28. These alterations made it difficult to continue tracking
the real displacement (ε) of the bundle by the camera, and generated some erroneous
data (noise) as shown in Figure 4-26 point C. This aspect did not affect the
calculation of the data because this deviation occurred beyond Fmax. The results of
the DFB test including bundle and technical fibre strength and Weibull parameters
for single technical fibres are shown in Table 4-5.
Figure 4-28. Tensile test sequence on bamboo DFB test at different times (TkS sample). Some of the visible
damages, which normally include movement of the broken fibres, are shown with the arrows.
According to the equation, (4-17), it can be seen that the fibre strength only depends
on the Weibull shape parameter m. For this reason it is possible to determine this
property (σf) derived from the DFB test, without the necessity to perform single fibre
test to find the Weibull parameters (e.g. the shape parameter m). For comparisons,
the fibre strength was calculated for each type of bundle using two m values: a) the
corresponding m value calculated from the DFB test according to the sample type
and b), using the average m value from the single fibre test calculated at different
gauge lengths (m=7.6). Table 4-5 shows that the fibre strengths calculated with those
two criterions are different. The first case gives around 20% and 15% higher average
values for the long (100 mm) and short gauge lengths (40 mm) respectively. As the
m parameter calculated form selected fibres is normally higher, indicating low
strength variability, it can be understood that it does not represent the non selected
batch of fibres present in the DFB sample. For this reason, the m value calculated
from the DFB test will be preferred for the technical fibre strength calculation, as is
more representative for the conditions of the experiment.
Type of bamboo dry fibre bundle
Properties TkL TkS ThL ThS

a
Bundle strength [MPa] (σb) 223 ± 38 335 ± 38 262 ± 34 336 ± 49
b
Fibre strength [MPa] (σf) 379 ± 41 536 ± 43 448 ± 56 539 ± 78
c
Fibre strength [MPa] (σf) 312 ± 22 468 ± 32 366 ± 38 470 ± 32
d
Fibre Young’s modulus [GPa] (Ef) 38 ± 4 45 ± 6 37 ± 5 37 ± 5
e 3.6 ± 0.9 4.4 ± 1.3 3.5 ± 1 4.4 ± 1.1

Shape parameter (m)
f 3.7 ± 1.2 4.6 ± 1.1 3.6 ± 1 4.7 ± 0.9
Shape parameter (m)
g
Scale parameter [MPa] (σo) 319 ± 38 480 ± 37 375 ± 49 482 ± 65
Shape parameter for the strain 8.5 11.1 8.7 11.2

h
[mm/mm *10-3] (εo)
Table 4-5. Dry fibre bundle strength and properties of individual technical bamboo fibres derived from the dry
fibre bundle test. aEquation 4-16, bEquation 4-17 using the corresponding m value obtained from the DFB test
according to the case, cEquation 4-17 using the average m value obtained from the SF test at different gauge
lengths (7.6), dEquation 4-18, eEquation 4-14 (graphic method), fEquation 4-13 (maximum load), gEquation 4-
15, hEquation 4-10.
94 Chapter 4b
The fibre strength (σf) calculated from the DFB test was consistently lower (~30 %)
than the values obtained in single fibre tests at the same gauge length (see Figure 4-
29). It was the same case when comparing the shape parameter m. This was an
expected result taking into account that the strength variation of the technical fibre is
averaged out in the DFB test, obtaining smaller standard deviations. In addition, no
pre-selection of the fibres was done, in contrast to what was done for the SF test.
This means that a bigger amount of fibres was tested from a specific batch and in
their original conditions, giving more “realistic” strength values. The Young’s
modulus of the technical fibre was easily determined from the DFB test with a few
samples for each case. The average value for all bundle geometries was 39 ± 4 GPa,
showing a good correspondence with the value of 43 ± 2 GPa calculated from single
fibre tests after the machine compliance correction applied by Osorio et al [35] for
the same type of bamboo technical fibres.
Values from the SF test

Values from the DFB test
40 mm gauge length 100 mm gauge length
Figure 4-29. Fibre strength comparison between values obtained by single technical bamboo fibre (SF) test
and dry fibre bundle (DFB) test at different gauge lengths. The SF strength reported at 100 mm gauge length
was extrapolated from the experimental results at 40 mm gauge length using Equation 4-5 with β=0.48 and
using the fibre length (L) instead of the fibre volume (V). The fibre strength from the DFB test corresponds to
the average value of the ThL and ThS shown in Table 4-5.
A drop in strength properties was also reported in literature when comparing results
from DFB and single fibre properties [44]. This can be explained mainly by two
reasons. A cooperative fibre failure denominated as “doublets” or “multiplets”,
where one fibre failure can trigger the next one, has been reported in literature when
testing synthetic DFB specimens [44]. This phenomenon can reduce the bundle
strength (σb), and hence affect the calculation of the Weibull parameters and the
strength properties of the individual fibres. Nevertheless, in the present study, this
effect was expected to be much less significant due to the significantly lower fibre
density present in the dry bundle samples in comparison with a roving of synthetic
fibres. Also, because of the reduced neighbouring fibres associated with the flat
bundle geometry instead of the traditional round DFB shape.
On the other hand, as was reported by Coleman [40], the ratio of the tensile strength
of a bundle to the mean tensile strength of the constituent filaments decreases in case
of an increasing strength dispersion of the constituent filaments. According to the
same author, the tensile strength of a dry bundle has the same order of magnitude,
but is less than the mean strength of the fibres. For this reason, it cannot be expected
that the value of the bundle strength (σb) gives reliable values for the individual
technical fibre strength (σf). This last value needs to be determined using the fibre
Weibull parameters combined to Equation 4-17.
4.6 Conclusions
In this chapter, mechanically extracted bamboo single technical fibres were

characterized with two different approaches for a wider overview of the mechanical
properties of this fibre. The first method consisted of SF (single fibre) testing carried
out to evaluate the effect of defects introduced by the extraction process as function
of different scale variables: fibre length, fibre surface area and fibre volume
applying the Weibull defect distribution analysis. It was found that Weibull’s
original model (β = 1) does not predict accurately the fibre strength distribution of
the technical fibres leading to an overestimation of the fibre strength at short gauge
lengths. The necessity to incorporate a length sensitivity exponent (β = 0.48) giving
as a result the modified Weibull distribution, indicates that bamboo fibre strength
decreases significantly less with length than predicted by Weibull’s original model.
The same distribution (length-based model) gives a more accurate prediction of the
strength distribution, and it can also be used for predicting fibre strength at different
gauge lengths. For this reason the length (L) is the recommended scale parameter to
be used for the Weibull strength distribution. The average technical fibre strength
was found to decrease with increasing gauge length, from 943 MPa at L=1 mm to
733 MPa at L=40 mm. Furthermore, the fibre strength was found to be nearly
independent of the mean fibre diameter. Diameter variations within the fibre may be
better correlated with strength, but further investigation is required to establish this
relation. Finally, the Weibull shape parameter (m) was found to be 7.6 for all tested
fibres (after visual selection), which means that the strength distribution is narrower
(having lower scatter) than that of most natural fibres and some synthetic fibres.
96 Chapter 4b
These results confirm that the quality of the bamboo fibres produced by the current
mechanical extraction method is high.
The second method for the technical bamboo fibre characterization was the dry fibre
bundle (DFB) test. The developed methodology is generic enough to be used for
other natural fibres but certain conditions need to be fulfilled such as:
- The straightness of the technical fibres in the DFB;

- The clamping of all technical fibres present in the sample;
- The accurate mounting of the specimen relative to the vertical axis of the
universal machine, in order to avoid misalignment in the grips and consequently
inhomogeneous loading during the tensile test;
- The accurate measurement of the strain during the test.
This procedure allowed to characterize a bundle of fibres under tension, avoiding

intensive preparation of individual samples for the SF test. The results from this
experiment gave systematically lower strength in comparison with SF tests due to
the fact that no particular selection of the fibres was carried out. Also, less scatter is
obtained and the results are calculated from the average behaviour of a large number
of fibres, which is statistically more representative. In general, this procedure
maximized the reliability of the results with minimal preparation time and amount of
material. Nevertheless, the preparation of the samples (i.e alignment of the fibres)
and the methodology for the testing (i.e accurate determination of the bundle strain)
play a critical role in the results. Moreover, it is necessary to previously have
determined the Weibull parameters for the individual technical fibres. The
sensitivity of the methodology to the bundle geometry was assessed by testing four
combinations of cross sectional area and length. The results indicated that the
amount of fibres is not critical but the length of the fibres was the most dominant
variable for the fibre strength calculations.
The strength properties and the morphology of bamboo technical fibres are highly
dependent on the extraction process, since defects introduced by the extraction
method reduce the tensile strength of the fibres. Also, the fibre’s aspect ratio
influences the final properties of bamboo fibre reinforced polymers. For this reason,
fibre properties cannot be generalized and should be linked with the correspondent
extraction method and the tested gauge length in order to avoid unfair statements
and comparisons. After an extensive literature review, no major progress was found
in the extraction of long bamboo fibres, even though this research topic had been
studied since the 80’s. This could be evidence that it has been a relatively short
period of time to develop a fully standardized extraction process, which could be
improved through time, as it is the case for other traditional natural fibres such as
flax, hemp or jute.
The novel mechanical extraction applied proved to be an effective method to

produce long high-quality bamboo fibres suitable to reinforce polymer composites.
The extracted bamboo fibres exhibit good mechanical properties among other
natural fibres and comparable specific properties (normalized to the materials
density) to glass fibres under tensile loading. Parallel research to this work suggests
that bamboo fibres are relatively insensitive to humidity [57]. Apart from this, there
are some distinctive aspects that present bamboo fibres as a good alternative for the
composite industry: the aesthetic properties of the final composite (the “bamboo
look”), environmental benefits due to the environmentally friendly extraction
process and the possibility of downcycling (recycling) of the composites (e.g. for
injection moulding), a marketable “green” image, and potentially low cost.
References
[1] da Costa LL, Loiola RL, Monteiro SN. Diameter dependence of tensile strength by Weibull
analysis: Part I bamboo fiber. Revista Matéria. 2010;15(2):110-6.
[2] Baur E, Otremba F. Design, materials properties and databases. In: Müssig J, editor. Industrial
applications of natural fibres, structure, properties and technical applications: Wiley Publishers;
2010. p. 399-406.
[3] Andersons J, Sparnins E, Porike E, Joffe R. Strength distribution of elementary flax fibres due to
mechanical defects. In: 11th Inorganic-Bonded Fibre Composites Conference. Madrid, 2008. p.
247-53.
[4] Joffe R, Andersons J, Wallstrom L. Strength and adhesion characteristics of elementary flax
fibres with different surface treatments. Composite Part A: Appl Sci Manuf. 2003;34(7):603-12.
[5] Barkakaty B. Some structural aspects of sisal fibers. J Appl Polym Sci. 1976;20(11):2921-40.
[6] Zhang Y, Wang X, Pan N, Postle R. Weibull analysis of the tensile behavior of fibers with
geometrical irregularities. J Mater Sci. 2002;37(7):1401-6.
[7] Cescutti G, Müssig J. Natural Fibres: Industrial Quality Management. Kunststoffe Plast Europe.
2005;95(1):97-100.
[8] Thwe MM, Liao K. Environmental effects on bamboo-glass/polypropylene hybrid composites. J
Mater Sci. 2003;38(2):363-76.
[9] Jain S, Kumar R, Jindal UC. Mechanical behaviour of bamboo and bamboo composite. J Mater
Sci. 1992;27(17):4598-604.
Mater. 1986;20(1):19-29.
[11] Watson A, Smith R. An examination of statistical theories for fibrous materials in the light of
experimental data. J Mater Sci. 1985;20(9):3260-70.
[12] Weibull W. A statistical theory of the strength of materials. Proceedings of the Royal Swedish
Institute of Engineering Research. 1939;151:1-45.
[13] Pickering KL, Murray TL. Weak link scaling analysis of high-strength carbon fibre. Composite
Part A: Appl Sci Manuf. 1999;30(8):1017-21.
[14] Andersons J, Joffe R, Hojo M, Ochiai S. Glass fibre strength distribution determined by
common experimental methods. Compos Sci Technol. 2002;62(1):131-45.
98 Chapter 4b
[15] Paramonov Y, Andersons J. A family of weakest link models for fiber strength distribution.
Composite Part A: Appl Sci Manuf. 2007;38:1227-33.
[16] Wu D, Zhou J, Li Y. Methods for estimating Weibull parameters for brittle materials. J Mater
Sci. 2006;41(17):5630-8.
[17] Zafeiropoulos NE, Baillie CA. A study of the effect of surface treatments on the tensile
strength of flax fibres: Part II. Application of Weibull statistics. Composite Part A: Appl Sci Manuf.
2007;38(2):629-38.
[18] Weibull W. A statistical distribution functions of wide applicability. J Appl Mech.
1951;18:293-7.
[19] Gutans J, Tamuzs V. Scale effect of the Weibull distribution of fibre strength. Mech Compos
Mater. 1984;20:1107-9.
[20] Pickering KL, Beckermann GW, Alam SN, Foreman NJ. Optimising industrial hemp fibre for
composites. Composite Part A: Appl Sci Manuf. 2007;38(2):461-8.
2010;41(5):588-95.
[22] Gurvich MR, DiBenedetto AT, Ranade SV. A new statistical distribution for characterizing the
random strength of brittle materials. J Mater Sci. 1997;32(10):2559-64.
[23] Creasy TS. A method of extracting Weibull survival model parameters from filament bundle
load/strain data. Compos Sci Technol. 2000;60(6):825-32.
[24] Van de Weyenberg I. The relation between flax fibre characteristics and the mechanical
properties of flax fibre reinforced epoxy composites [Master Thesis]. Belgium: KU Leuven; 1999.
[25] Andersons J, Sparnins E, Joffe R, Wallstrom L. Strength distribution of elementary flax fibres.
Compos Sci Technol. 2005;65:693-702.
[26] Cao Y, Wu Y. Evaluation of statistical strength of bamboo fiber and mechanical properties of
fiber reinforced green composites. J Cent S Univ Technol. 2008;15(0):564-7.
[27] Kulkani A, Satyanarayana K, Rohatgi P. Weibull analysis of coir fibres. Fibre Sci Technol.
1983;19:59-76.
[28] Xia ZP, Yu JY, Cheng LD, Liu LF, Wang WM. Study on the breaking strength of jute fibres
using modified Weibull distribution. Composite Part A: Appl Sci Manuf. 2009;40(1):54-9.
2009;40(8):1111-8.
[30] Silva FA, Chawla N, Filho RD. Tensile behavior of high performance natural (sisal) fibers.
Compos Sci Technol. 2008;68(15-16):3438-43.
[32] Deng S, Ye L, Mai Y, Liu H. Weibull fibre strength parameters derermined by single fibre
fragmentation tests. In: Murray S, editor. Proceedings 11th International conference on composite
materials. Queensland, Australia: Woodhead, Publishing Limited; 1997. p. 578-86.
[33] Dibenedetto A, Gurvich M, Ranade S. On in-situ evaluation of fibre strength distribution. J
Mater Sci Lett. 1997;16(22):1791-2.
[34] Van de Weyenberg I, Ivens J, Verpoest I. Parametric study of the relationship between the fibre
and the composite properties of flax fibre reinforced epoxy. In: Proceedings ECCM-9 conference.
Brighton, June, 2000.
[36] Zhang Y, Wang X. The effect of along-fibre diameter variation on fibre tensile behaviour.
Wool Tech Sheep Bree 2000;48(4):303-12.
[37] R'Mili M, Bouchaour T, Merle P. Estimation of Weibull parameters from loose-bundle tests.
[38] Chi Z, Chou T-W, Shen G. Determination of single fibre strength distribution from fibre
bundle testings. J Mater Sci. 1984;19(10):3319-24.
[39] Daniels H. The statistical theory of the strength of bundles of threads. I. Proceedings of the
Royal Society of London Series A Mathematical and Physical Sciences. 1945;183(995):405-35.
[40] Coleman B. On the strength of classical fibres and fibre bundles. J Mech Phys Solids.
1958;7(1):60-70.
[41] Cowking A, Attou A, Siddiqui A, Sweet M, Hill R. Testing E-glass fibre bundles using
acoustic emission. J Mater Sci. 1991;26(5):1301-10.
[42] Evans K, Caddock B, Ainsworth K. Statistical changes during the corrosion of glass fibre
bundles. J Mater Sci. 1988;23(8):2926-30.
[43] R'Mili M, Moevus M, Godin N. Statistical fracture of E-glass fibres using a bundle tensile test
and acoustic emission monitoring. Compos Sci Technol. 2008;68(7):1800-8.
[44] Hill R, Okoroafor E. Weibull statistics of fibre bundle failure using mechanical and acoustic
emission testing: the influence of interfibre friction. Composites. 1995;26(10):699-705.
[45] Calard V, Lamon J. Failure of fiber bundles. Compos Sci Technol. 2004;64(5):701-10.
[46] Lu Y, Hao L, Hirohashi M. Study of strength and its reliability of SiC fiber bundle by
experimental and Monte-Carlo simulation approach. Materials Science and Engineering: A.
2011;528(12):4080-5.
[48] Thomason J, Genltes F, Brennan A. Natural fibre cross sectional area effects on the
determination for fibre mechanical properties. In: Proceedings of ECCM-15. Venice, June, 2012.
[49] Tran LQN. Polymer composite materials based in coconut fibres. Leuven University of Leuven
[Doctoral thesis]; 2012.
[50] Tran LQN, Fuentes CA, Dupont-Gillain C, Van Vuure AW, Verpoest I. Wetting analysis and
surface characterisation of coir fibres used as reinforcement for composites. Colloids Surf A:
Physicochem Eng Aspects. 2011;377:251-60.
[51] Fuentes CA, Tran LQN, Dupont-Gillain C, Vanderlinden W, De Feyter S, Van Vuure AW, et
al. Wetting behaviour and surface properties of technical bamboo fibres. Colloids Surf A:
[52] ASTM. D 1294. "Tensile Strength and Breaking Tenacity of Wool Fiber Bundles", ASTM
International, West Conshohocken, PA, 2001, DOI: 101520/D1294-05R132001.
[53] ASTM. D 2524. "Standard Test Method for Breaking Tenacity of Wool Fibers", ASTM
International, West Conshohocken, PA, 2003, DOI: 101520/D2524
2003.
[54] ASTM. D1445 "Test Method for Breaking Strength and Elongation of Cotton Fibers", ASTM
International, West Conshohocken, PA, 1999, DOI: 101520/D1445M-121999.
[55] Parkins D. Development and validation of a dry-fiber bundle test method. Master thesis in:
Montana State University; 2000.
[56] Cintron R, Saouma V. Strain measurements with the digital image correlation system vic-2D.
System. 2008;106:2D.
[57] Osorio L, Trujillo E, Lens F, Ivens J, Van Vuure A, Verpoest I. A complete microstructural
and mechanical characterization of bamboo technical and elementary fibres. Proceedings of ICCM-
19 Montreal-Canada, 2013. 28:1-7.
[58] Bos H, Van Den Oever M, Peters O. Tensile and compressive properties of flax fibres for
natural fibre reinforced composites. J Mater Sci. 2002;37(8):1683-92.
[59] Bismark A, Mishra S, Lampke T. Plant fibers as reinforcement for green composites. In:
Mohanty A, Misra M, Drzal L, editors. Natural Fibers, Biopolymers and Biocomposites: CRC Press;
2005.
[60] Baley C. Analysis of the flax fibres tensile behaviour and analysis of the tensile stiffness
increase. Compos Part A: Appl Sci Manuf. 2002;33(7):939-48.
100 Chapter 4b
[61] Ogawa K, Hirogaki T, Aoyama E, Imamura H. Bamboo fibre extraction method using a
machining center. Journal of Advanced Mechanical Design, Systems, and Manufacturing.
2008;2(4):550-9.
[62] Rao K, Rao K. Extraction and tensile properties of natural fibers: Vakka, date and bamboo.
Compos Struct. 2007;77(3):288-95.
[63] Okubo K, Fujii T. Eco-composites using natural fibers and their mechanical properties. In:
Brebbia C, De Wilde W, editors. High performance Structures and Composites. Southampton: WIT
Press; 2002. p. 77-85.
[64] Ochi S, Takagi H, Niki R. Mechanical properties of heated-treated natural fibers. In: Brebbia C,
De Wilde W, editors. High performance Structures and Composites. Southampton: WIT press; 2002.
p. 117-23.
[66] Okubo K, Fuji T, Yamashita N. Improvement of interfacial adhesion in bamboo polymer
composite enhance with micro-fibrillated cellulose. JSME International Journal of Applied
Mechanics. 2005;48(4):199-204.
[67] Shibata S, Fukumoto I. Effects of bamboo and kenaf fibres on the flexural modulus of bio-
composites. Technical Article, University of the Ryukyus, Okinawa, Japan. 2004:14-7.
[68] Shito T, Okubo K, Fujii T. Development of eco-composites using natural bamboo fibres and
their mechanical properties. In: Brebbia C, De Wilde W, editors. High performance Structures and
Composites. Southampton: WIT Press; 2002. p. 175-82.
[70] Amada S, Ichikawa Y, Munekata T, Nagase Y, Shimizu H. Fiber texture and mechanical
graded structure of bamboo. Compos B: Eng. 1997;28(1-2):13-20.
Unidirectional continuous and discontinuous bamboo fibre - epoxy composites 101
Chapter 5
Unidirectional continuous and discontinuous bamboo
fibre – epoxy composites
_______________________________________________________
5.1 Introduction
Discontinuity of the fibre in the

nodes [1]
Internode length
(± 25 – 35 cm)
Figure 5-1. Fibre bundle discontinuity into the bamboo culm. Schematic of the node (left) given by [1].
A large bottleneck that impeded the introduction of bamboo fibres as composite

reinforcing materials for many years has been the extraction of undamaged long
fibres [2-6]. Recently, a new environmentally friendly mechanical process was
developed producing high quality long bamboo technical fibres suitable to be used
as reinforcement in polymeric matrices. Now, with the availability of the fibres, a
second obstacle to be overcome towards the large-scale industrialization of bamboo
fibre composites, apart from the challenges mentioned in section 3.4, is to produce
continuously a unidirectional (UD) prepreg (preform) to be used as reinforcement in
the composite industry.
102 Chapter 5b
This step is necessary due to the inherent nodal constitution of the bamboo culm,
where the reinforcing fibres entangle in the nodes and contribute as such to the
buckling strength of the culm [1, 7]. This structural morphology, limits the extracted
technical fibres to the internode length (between 25 and 35 cm for the species
Guadua angustifolia [8]), see Figure 5-1.
To overcome this situation, the production of a continuous bamboo yarn cannot be

considered because the fibres have a high bending stiffness that impedes twisting
without fibre damage. Moreover, the twisting itself not only damages the fibre, but
also reduces drastically the stiffness of the yarn in the composite due to the fact that
the fibre will be off-axis with respect to the load [9], see Figure 5-2. As an
alternative to produce an endless yarn, certain mixed yarn technologies could be
considered, where the bamboo fibres would constitute the core and a thermoplastic
matrix yarn would be wrapped around them. But, because of the considerable
thickness of bamboo fibres (between 90 and 250 µm), the resulting yarn will be
quite coarse and prepregs or textiles based on these yarns (e.g. woven or braided)
would result in thick composite plies with low fibre volume fraction.
a. b.
Figure 5-2. a) Schematic representation of a twisted and non-twisted yarn [10], b) experimental ( ) and
simulated stiffness results from different models for flax fibre bundles when loaded at different twist angles
[9].
The aim of this chapter is to translate the intrinsic good bamboo technical fibre
properties, shown in Chapter 4, into good composites. Therefore, a novel approach
is intended in this study to use the discontinuous bamboo technical fibres for large-
scale composite applications by the development of continuous unidirectional
preform, in which the bamboo fibre ends are distributed in a defined overlap pattern.
For this, a systematic study was conducted on the effect on mechanical properties of
different UD preforms with discontinuous fibre patterns, ranging from technical
fibre bundles with different overlapping lengths, to individual technical fibres with
randomly placed fibre ends. The use of a continuous preform or prepreg tape will
allow the use of existing technologies to produce high performance composite parts
and to overcome the actual restriction of having discontinuous fibres. The results
will be benchmarked with the mechanical behaviour of a fully continuous
unidirectional bamboo fibre-epoxy composite (UD-C), to show the feasibility of the
use of technical bamboo fibres in continuous preforms or prepregs for high-end
composite applications. The research will be complemented with impregnated fibre
bundle tests and 3-point bending tests (3PBT).
5.2 Overview of different models for the prediction of composite mechanical

properties
The experimental stiffness and tensile strength results of aligned discontinuous

bamboo fibre-epoxy composites (UD-D) and continuous bamboo fibre-epoxy
composites (UD-C) will be compared with the predictions from several models. In
general, they have common assumptions such as the linear elastic behaviour of
fibres and matrix, the matrix is isotropic, and the fibres are either isotropic or
transversely isotropic. Also, the fibres have regular cross sectional area and the same
length, the fibres are well bonded to the matrix and fibre-matrix debonding is not
considered; there is a regular packing geometry of the fibres and perfect fibre
alignment. It should be noticed that the last assumptions, do not represent the real
conditions present in the studied composite samples. The next sections give an
overview of the models used to predict the properties of the bamboo fibre-epoxy
composites with different fibre configurations.
5.2.1 Prediction of longitudinal tensile stiffness
Rule of mixtures (ROM)
The rule of mixtures is the simplest mechanical model to estimate the properties of a
multiple component system. It estimates the composite material properties by taking
a volume weighed average of the corresponding properties of the individual
constituents. Concerning the prediction of the longitudinal stiffness of a composite,
this model assumes the presence of continuous aligned fibres through-out the entire
length of the composite and perfect bonding between fibre and matrix. Furthermore,
it assumes iso-strain behaviour for all the components during loading conditions.
Applying Hooke’s law and the previous assumptions yields Equation 5-1 that is used
to estimate the longitudinal stiffness of a composite:
104 Chapter 5b
Ec  V f E f  Vm Em (5-1)
Shear-lag theory
The shear-lag theory [11] considers a single cylindrical linear elastic and isotropic
fibre of finite length L and radius rf that is encased in a concentric cylindrical shell
of a linear elastic, isotropic matrix with radius Rm. A one-dimensional stress-state
situation is applied in which the matrix tensile strain becomes equal to the applied
strain (εappl) at a radial distance Rm from the fibre axis. Furthermore, the shear
transfer between matrix and fibre is thought to occur by interfacial shear stresses and
the tensile stress at the fibre ends is assumed to be zero. As a result, the maximum
tensile stress occurs in the middle of the fibre whereas the maximum shear stress
occurs in the interface at the ends of the fibre [11, 12]. Additional assumptions are
that the fibres are aligned and packed in an orderly manner, as it is also presumed
for the ROM. Also, both the fibre and the matrix are assumed to be perfectly linearly
elastic and isotropic, and the stress is transferred between the two constituents
without yielding (perfect bonding). In general, the shear-lag theory is a modified
ROM which incorporates a fibre-length correction factor ηl, that depends on the fibre
aspect ratio, a fibre packing factor (see Table 5-1), the matrix shear modulus and the
fibre’s Young’s modulus (Equations 5-2 to 5-5).
Ec  lV f E f  (1  V f ) Em (5-2)
tanh(nL / 2)
l  1 
nL / 2 (5-3)
2Gm
n2 
R  (5-4)
E f ln  m 
 rf
 
Rm KR
 (5-5)
rf Vf
There are some discussions in literature about the exact value of the radius of the
cylindrical matrix shell Rm. This value can be obtained as shown in Equation 5-5
and it will depend on the packing state of the fibres (KR). Table 5-1 gives an
overview of the different values of the KR constant according to different authors.
Fibre packing Value of KR Reference

Hexagonal, center-to-center distance 3.62 [11]
Hexagonal, half of the nearest neighbour distance 1.81 [13]

Composite cylinder 1 [14]
Square 0.78 [15]
Table 5-1. Overview of the different values found in literature for the packing dependent constant KR.
Halpin-Tsai equations
The Halpin-Tsai equations are a series of semi-empirical equations to predict the

longitudinal tensile Young´s modulus of composite materials. This model assumes
that the fibres are transversely isotropic and linear elastic, also that they are
percectly parallel and well bonded to their interface. A review of the entire
derivation of the model is given by Halpin and Kardos [16]. The Halpin-Tsai
equations were originally developed for continuous fibre composites and were
derived from the work of Hermans [17] and Hill [18]. Halpin and Tsai found that
three of Herman’s equations can be written in a common form as seen in Equations
5-6 and 5-7.
 1  V f 
Ec  Em   (5-6)
 1Vf
 
( E f / Em )  1
 (5-7)
( E f / Em )  
The parameter η is a function of the ratio of the fibre and matrix stiffness (Ef/Em) and
of the shape of the reinforcement. It was found in literature that the significance of
the ξ parameter for these equations had two interpretations. The first one considers it
as a fitting parameter, to adjust the Halpin-Tsai equations to the experimental data,
considering the packing arrangement and the geometry of the reinforcing fibres [19].
As a second approach, it has been suggested that ξ should vary from small values to
infinity as a function of the fibre aspect ratio L/d. By comparing model predictions
with available 2-D finite element results, it was found this parameter can be
determined by two times the aspect ratio of the fibres [13, 20], as shown in Equation
5-8. This was the approach followed in the present study.
ξ = 2(L/d) (5-8)
106 Chapter 5b
Mori-Tanaka model
The Mori-Tanaka (MT) model [21] is a commonly used micromechanics-based

model used to predict the elastic properties of a wide range of composites
(unidirectional, short fibre and particulate) [22]. The MT model defines explicit
expressions for the elastic constants of unidirectional anisotropic ellipsoidal short-
fibre-reinforced composites, which will be summarized next [23]:
The elastic stiffness matrix of composite, C-, can be expressed as in tensor notation
as:
C   C m  V f (C f  C m ) : B (5-9)
where Vf is the inclusion volume fraction, which is considered as the fibre volume
fraction for the fibre composite material system, Cm and Cf are the elastic stiffness
matrix of the matrix and fibre material, respectively. The fourth-order tensor B is
defined as:
1
B   I  (1  V f ) E : S m : (C f  C m )  (5-10)
where I is a fourth-order unit tensor, the compliance matrix of the matrix material
Sm = (Cm)-1, and E stand for Eshelby’s tensor, for which the analytical expressions
are listed clearly by Mura [24]. In this formulation, both fibre and matrix can be
generally anisotropic elastic, however, most commonly, only the reinforcing
inclusions is of anisotropic behavior. It is remarked that this model additionally
assumes that fibre and matrix are linearly elastic and perfectly bonded through the
entire deformation state, which might not be realistic, except at the earlier stage of
deformation.
The matrix is usually considered isotropic, whereas the fibres are modelled as
transversely isotropic. The Mori-Tanaka model does not depend on either the size of
the inclusion, nor on their positional coordinates, but it does depend on shape and
orientation of the inclusions, as well as their volume fraction [25]. The MT model
was implemented to estimate the elastic moduli of unidirectional kenaf epoxy
composites, where the prediction was 20% higher than the experimental results [23].
5.2.2 Prediction of longitudinal tensile strength
Rule of mixtures (ROM)
The rule of mixtures applies the iso-strain assumption, indicating that the component
with the lowest breakage strain will fail first. Since bamboo fibres, used in this
research, are more brittle than the epoxy matrix and since the produced composites
contain relatively high fibre volume fractions, it is expected that the composite will
break at the breakage strain of the fibres (see Equation 5-11). In this equation σ*f
indicates the longitudinal fibre tensile strength, σ´m the longitudinal tensile stress in
the matrix at fibre failure and Vf and Vm the volume fractions of fibre and matrix
respectively. This method further assumes that the stress concentration around one
broken fibre triggers the failure of other fibre failures (at the same strain) and that no
mechanisms of crack propagation deceleration are present.
σc= Vf σ*f + Vm σ´m (5-11)
Kelly-Tyson model
For the modelling of the strength of discontinuous fibre composites Kelly and Tyson
[26] extended the ROM for strength prediction of composites reinforced with fibres
aligned in loading direction. This model, it is assumed that no voids are present in
the composite. This model takes into account the interaction between axial tensile
stresses in the fibre and shear stresses at the interface of the fibre. The Kelly-Tyson
model further assumes interface failure to occur first and models the shear stress at
the interface as a constant in the debonded or yielded area [12]. By introducing the
critical aspect ratio Sc, defined as the smallest fibre aspect ratio at which the axial
fibre tensile stress can just reach the fibre strength and by assuming that the fracture
occurs when the axial fibre tensile stress reaches the fibre strength, the composite
strength is determined as stated in Equation 5-12.
  rf S c  Em   *f
 c   V f
*
1  
 m V  with Sc  (5-12)
f
  2L  E f  2 rz
where rf is the radius of the fibre, Sc is the critical aspect ratio, L is the fibre length
and τrz is the interfacial shear stress around the debonded fibre.
108 Chapter 5b
Global load sharing model (GLS)
The global load sharing model (GLS) introduced by Curtin [27, 28] assumes that the
stress from a broken fibre is redistributed globally across all the remaining intact
fibres in the cross section of the composite. An important characteristic of Curtin’s
model is that it accounts for load contributions of broken fibres and the matrix shear
loading recovers away from the fibre breakage. Moreover, it incorporates Weibull
statistics [29, 30] in the estimation of composite tensile strength (for more
information about Weibull’s model see paragraph 4.1.1). The ineffective fibre length
(δ) is calculated using the Kelly-Tyson approximations, leading to Equation 5-13 for
the longitudinal tensile strength of the composite.
1 1
 2  m1  m  1    0m L0 rz ,co  m 1 (5-13)

 c  V f 1     with 1   
 m2  m2  rf
 
where m and σ0 are the Weibull shape and scale parameters respectively and Lo
corresponds to the scale variable for the fibre length. More information about these
parameters can be found in section 4.1.1.
Local load sharing (LLS) model
The local load sharing (LLS) model [31-33] emerges from the intention to predict
the composite tensile strength in an analytical approach. This method is able to
predict the tensile strength of the discontinuous fibre composites with reasonably
good accuracy and low computational cost. An important difference in comparison
with the other models, is that it incorporates matrix shear loading to re-distribute the
applied load with only the adjacent fibres and therefore can account more accurately
for fibre breakage. A comprehensive description of this model is given by Taketa
[34] and is summarized as follows.
First, a spring element unit cell is designed using only dominant parameters for
composite strength. As shown in Figure 5-3a, one fibre is surrounded by six other
fibres. The unit cell considers the fibre represented by axial springs in the
longitudinal direction and the matrix by shear springs in transverse direction. Each
of the spring elements is assigned a stiffness matrix. Fibre elongation and matrix
shear are the mechanisms of load transfer through unidirectional composites. The
stiffness matrices of fibre spring element KL and matrix shear spring element KT are
defined in Equation 5-14, where l is the fibre spring length, Ef is fibre Young’s
modulus and G is the effective matrix shear modulus.
a. Unit cell b.
(hexagonal array)
Figure 5-3. a) Schematic of spring element unit cell and b) graphical visualization of the construction of a 3D-
volume by the stacking of hexagonal unit cells [31].
E f  rf2  1 1 Gm rf l  1 1

KL  and KT  (5-14)
l  1 1  3d  1 1 
A number of the unit cells are built up as three dimensionally illustrated in Figure 5-
3b. This enables a large enough simulation size (number of fibres and fibre length)
according to the case to ensure that the model is equivalent to the sample size for the
calculation of the tensile strength.
Strength variation is included through a Weibull distribution of the reinforcing

material. At each strain increment, the stresses in the springs are compared to the
fibre strength. If the fibre strength is exceeded, the spring element stiffness is
removed and a stress re-distribution is performed. At each step the normalized
composite stress is also calculated. This iterative scheme is repeated until the
relative difference in two subsequent composite stress values is larger than a pre-set
value. Monte Carlo simulation is then performed and the composite strength is taken
as the average of the simulated values. The LLS model code used in this PhD thesis
was developed at the University of Tokyo by Dr. T. Okabe, Dr. Nishikawa and Mr.
K. Ishii. The model was modified in order to incorporate all parameters and
restrictions needed for the present study, as it will be detailed further in section 5.5.7.
5.3. Materials
5.3.1 Bamboo fibres
Untreated bamboo fibres were used in this study. The location for the extraction of
the culms and the characteristics of the fibres and their extraction are described in
section 4.3.
110 Chapter 5b
5.3.2 Epoxy resin
Three different epoxy resins were used in this study. The resin used for the bamboo
fibre composites and which will be used for tensile testing is a commercial available
epoxy resin Epikote 828 LVEL and a DYTEK® DCH-99 Amine (1,2-
diaminocyclohexane) hardener. After mixing the two components in a mass ratio of
100/15.2 w/w the system is degassed in a vacuum oven at 1 bar for 15 minutes.
After the composite production the resin requires a curing temperature of 70°C for
an hour and an additional post-curing step at 150°C for 1 hour.
For the manufacturing of the impregnated bundles (IB), according to the standard
ISO 10618, it is recommended that the matrix has at least twice or triple the failure
strain of the reinforcement. For this reason epoxy Araldite® LY564 was chosen
together with amine Aradur® 3486 as a hardener with a mixing ratio of 100/34 w/w.
The same procedure for preparation and curing as described above was followed.
For 3-point bending test characterization, Epoxy HM 533 was used as the resin
component, supplied by Hexcel Composites S.A, with an areal weight of 250 g/m2
and a curing temperature of 125 °C. It is a film type of epoxy resin used to produce
prepregs, which makes it interesting to explore for composite manufacturing and to
determine the mechanical properties obtained with this type of resin. Table 5-2
shows the main properties of the epoxy resins used for the different test samples.
Epoxy resin Tensile Tensile Elongation Density Sample manufacturing

strength modulus at fracture (g/cm3)
(MPa) (GPa) (%)
Epikote 828 70 2.7 4.1 1.1 UD-C* and UD-D**
Araldite LY564 70 -74 2.8 – 3.0 4.6 – 5.0 1.2 Impregnated fibre bundle test
HM 533 50 2.9 4.0 1.2 3PBT
Table 5-2. Properties of the epoxy resins used as a matrix in this study. All data are obtained from data sheets.
*UD-C: unidirectional continuous bamboo fibre-epoxy composites. **UD-D: unidirectional discontinuous
bamboo fibre-epoxy composites.
5.4 Methods
5.4.1 Fibre patterns
In this study, different types of unidirectional discontinuous technical fibre patterns

(preforms) with a fibre length (Ls) of 50 mm were produced to manufacture UD-D’s,
in order to be compared with the mechanical behaviour of UD-C samples. A
schematic overview of the studied patterns vperarying a) the overlapping length of
the technical fibre ends (Lv) and b), the width of the fibre bundles is shown in Figure
5-4. The fibre bundle in this case, can be defined as a “cluster” of technical fibres
that are together and have the same Ls. The mentioned preforms were made to fit
into the cavity mould that had the final dimensions of the testing sample (25 mm
width (w) and 260 mm length). In general, the created patterns can be classified in
four different categories as follows:
1. Composites with fibre bundles, width half of w, Ls=50 mm and whose fibre
ends overlap (Lv)10, 30 and 50% of the adjacent fibre (see an schematic view
in Figure 5-4a, b and c);
2. Thinner fibre bundles, 1/5 of the sample width (w), Ls=50 mm and random Lv,
see Figure 5-4d;
3. Individual UD bamboo technical fibres with Ls= 50 mm whose fibre ends are
placed randomly along the composite (Lv= random overlap length between
individual fibres), see Figure 5-4e;
4. Finally, the fourth category belongs to the UD continuous disposition of the

fibres (UD-C and impregnated bundle (IB)) where the technical fibres have
the same length as the sample composite as is shown in Figure 5-4f.
Each composite, contained four layers of the preforms with the corresponding
patterns according to the case, having an inversion of symmetry between adjacent
layers (brick wall pattern), see Figure 5-5. No fibre ends were inserted in the
clamping area in order to reduce stress concentrations and to avoid premature failure
close to the grips. All the fibre bundles were weighed beforehand, after being
conditioned at room temperature (20° and 50% relative humidity for at least 72 h,
and dried again at 60 °C 24 h before they are used for composite manufacturing. The
fibre volume fraction (Vf) was recalculated through weight measurements after
composite production and normalized at 40%.
Finally, two different approaches can be considered for the determination of the
bamboo fibre’s aspect ratio (L/d), as input variable in some of the applied models. In
the first case, “d” represents the diameter of the entire fibre bundle which is defined
above as a “cluster” of technical fibres that are packed together and have the same
length Ls, as seen Figure 5-4 and 5-7. The second one proposes a more traditional
approach in which “d” represents the diameter of individual technical fibres. In this
case, an average fibre diameter of 160 µm was taken from 420 measurements. The
second approach was applied to calculate the fibre aspect ratio, taking into account
that every technical fibre will be surrounded by the resin and thus will acts as an
individual unit within the composite if a homogeneous fibre distribution is achieved
during the production of the composite plates.
112 Chapter 5b
Category 1 Category 2 Category 3 Category 4
a. b. c. d. e. f.
Lv=
Lv= random
random
Fibre
disconti-
nuities
w w w w w w
Figure 5-4. Schematic general view of the different fibre pattern configurations. Unidirectional discontinuous
composites (UD-D): a) 10%, b) 30% and c) 50% of fibre bundle overlapping, d) random fibre bundles and
e) full fibre randomization. Continuous unidirectional composites (UD-C and IB): f) continuous fibres. All
composites were normalized at 40% Vf. The fibre length (Ls) is 50 mm for all samples except UCD (f) and the
width (w) was 25 mm. The technical fibre ends (or discontinuities) can be seen in Figure 5-6.
a. First and third preform (layer) b.
Composite length
Fibre bundle
Thickness
Thickness
Second and fourth preform (layer)
Lv Ls Lv Ls
Figure 5-5. Schematic pattern for UD-D samples with fixed fibre bundle overlapping length. a) the upper view
showing the inversed symmetry that exists between adjacent layers of fibres and b), the side view presenting a
“brick wall” pattern assembly.
5.4.2 Elaboration of the prepregs
UD-D with fixed fibre bundle overlapping length
This configuration comprises Lv= 10, 30 and 50%, Ls= 50 mm and width ½ w. The
fibre bundles were manually introduced in series into the cavity mould in which
adjacent bundles were shifted 5, 15 and 25 mm, giving as a result three different
overlapping lengths (Lv) of 10, 30 and 50% respectively. The width of the fibre
bundles corresponded to half the width of the mould (½ w), see Figures 5-4 and 5-6.
The patterns were schematically presented in Figure 5-4a-c.
a. b.
4
1 Fibre
discontinuities 3
2 w=25 mm
Figure 5-6. Manual placement of the fibre bundles in the multicavity mould (260 x 25 mm). 1= placement of
the first layer of fibre bundle into to the mould at fixed overlapping length, 2= UD fibre bundle, 3= stacking
of several UD layers of fibre bundles, 4= upper mould. The fibre length (Ls) was set at 50 mm.
UD-D with random fibre bundle overlapping length
This fibre preform incorporates Lv= random, Ls= 50 mm and 1/5 w. In this case, the
effect of the fibre discontinuities was diminished by reducing the width of the
bundle to 5 mm, and by randomizing the overlapping length between them (see
Figure 5-4d). For the placement of the fibre bundles in order to make the preforms,
the next steps were followed:
- The cavity mould was virtually divided in five rows along the length (1/5 w),
in order to position fibre bundles of the same width in each row;
- Then, with the help of a random number generator (RNG), discrete numbers
for each row (between 0 and 260) were generated in order to position the
114 Chapter 5b
starting edge of the bundle according to this number. An example is given in

Figure 5-7a for rows 1 and 2 where the edges were positioned at distances
Xn+A and Xn+B respectively. In this way, the first bundles of each row are
positioned, see Figure 5-7a;
a.
Fibre bundle (Ls= 50 mm)
b.
Figure 5-7. Schematic representation of the procedure for the placement of the fibre bundles for the UD-D
samples with random fibre bundle overlapping length. a) Placement of the first fibre bundles in each row
(from 1 to 5) with the help of a random generation number (RNG). b) the positioning of the rest of the fibre
bundles.
- Then, the “free” space in each row was filled in series with more fibre
bundles. At the end the first layer is completed as depicted in Figure 5-7b
with Lv= random;
- The same operation is repeated 3 times more in order to have 4 layers in total
per sample. If by chance, a bundle discontinuity needed to be placed into the
gripping zone, this discontinuity was avoided. This was with the aim to
reduce stress concentrations and to avoid premature failure close to the grips.
UD-D with random single fibre overlapping length
Individual technical fibres (Lv= random, Ls= 50 mm) were placed manually directly
into the metal mould, in order to produce a UD discontinuous preform with
randomized fibre ends, as depicted in Figure 5-4e. A random number generator
helped for the positioning of the fibres (discrete numbers), where the maximum Lv is
50 mm. This operation gives as a result a different position of the fibre ends and
filling completely the length and width of the cavity mould (260 x 25 mm). This
operation is done for four preforms (contained in one sample), each one randomized
with different patterns.
Unidirectional continuous bamboo fibre-epoxy composites (UD-C and IB)
Figure 5-8. Unidirectional bamboo fibre preforms ready to be used for composite production.
For the unidirectional preform (Figure 5-4f) continuous bamboo fibres were used; as
for all the cases explained before, the fibres were weighed (being in room conditions
for at least 72 hours), carefully aligned and evenly spread by hand in order to have a
homogeneous thin layer of fibres (Figure 5-8). The fibre volume fraction of the final
composites was recalculated through weight measurements after composite
production and normalized at 40% for further comparisons. The same procedure was
followed for the preparation of the fibre preforms for the impregnated bundles (IB).
5.4.3 Composite production
5.4.3.1 Manufacturing of the samples for tensile testing
Light RTM was used to manufacture composite samples, see Figure 5-9. This
technique allowed keeping good alignment of the fibres and also a good surface
quality at both sides of the composite. The fibre preforms, described in the previous
sections, were placed into a mould with 5 cavities with a size of 260 x 25 mm, and
for the IB, another multicavity mould with smaller dimensions (250 x 10 mm) was
used. Both moulds had the corresponding upper moulds with final dimensions of the
samples for tensile testing, avoiding additional cutting of the composite plates. The
116 Chapter 5b
average dimensions of the samples are shown in Table 5-3. A fibre volume fraction
of 40% was targeted and verified by weight measurements after manufacturing.
Resin oulet Resin inlet
Mould
Upper
Vacuum bag mould
Bamboo fibres
Sealant tape Bottom plate
Hot plate
Figure 5-9. Light RTM process for the production of bamboo fibre-epoxy composites.
5.4.3.2 Sample production for 3-point bending test
Flexural three point bending tests (3PBT) were conducted with two different fibre
orientations, longitudinal and transverse, to evaluate the flexural properties of
bamboo fibre composites with a thermoset matrix (epoxy resin). The transverse fibre
orientation is chosen to evaluate the adhesion strength between fibre and matrix, as
the transverse mechanical properties of the composite are matrix and interface
dominated.
a.
b.
Figure 5-10. Schematic view of the mould used to produce 3PBT samples of bamboo fibre-epoxy composites.
For the composite manufacturing, several layers of UD fibres and epoxy film (HM
533) were alternated and laid up into the mould cavities of 60 x 40 mm, see Figure
5-10. Final consolidation is achieved by compression moulding applying pressure (3
bar) plus vacuum to minimize void content due to air entrapments in the material.
Samples were later on cut from the cured plates using a diamond saw and polished
appropriately to avoid side flaws that could act as crack initiators. Sample
dimensions are given in Table 5-3. The fibre volume fraction was measured through
weight measurements and the test results were normalized to 40%.
5.4.4 Sample testing
5.4.4.1 Tensile test
Tensile test samples were prepared according to the standard ASTM D3039. An
Instron 4467 machine with a load cell of 30 kN was used for the tests and a
crosshead speed of 2 mm/min was applied. The gauge length between the two
clamps was set at 150 mm and an extensometer with gauge length of 25 mm was
employed for measuring accurately the elongation of the composites. The use of end
tabs was not necessary because there was no failure close to the clamps, but instead
the samples were mechanically clamped using sand paper in the grips to prevent
slippage. The tensile test set up is shown in Figure 5-11. Before testing, all
specimens were conditioned at room conditions (21°C ± 2°C and 50 ± 2 %RH) for
at least 24 hours, see all specimen dimensions in Table 5-3.
Figure 5-11. Set-up for tensile test of bamboo fibre-epoxy composites.

118 Chapter 5b
Type of composite Length Width Thickness Span Tested

(mm) (mm) (mm) length samples
(mm)
Tensile test
UD-D fix fib. bundle overlap. length (Lv= 10%) 250 25 1.96 ± 0.11 150 5
UD-D random fibre bundle overlapping length 260 25 1.92 ± 0.09 150 5
UD-D random technical fibre overlapping length 260 25 1.84 ± 0.11 150 5
UD-C impregnated fibre bundle (IB) 250 10 0.76 ± 0.12 150 16
UD-C Continuous fibres 250 25 1.98 ± 0.3 150 5
3 point bending test
3PBT longitudinal 60 11 ±0.3 1.68 ± 0.03 35 6
3PBT transversal 60 5 ± 0.2 1.64 ± 0.04 35 5
Table 5-3. Dimensions of bamboo fibre – epoxy composites for tensile and 3PBT tests. For UD-D and UD-C
Ls= 50 mm.
Tensile test set up for the impregnated bundle test
Impregnated fibre bundle tests were performed according to ISO-10618. The same
set-up described above for the UD-C and UD-D samples was used, except for the
crosshead speed (1mm/min) and the extensometer (50 mm gauge length), as seen in
Figure 5-12. The fibre volume fraction was found to be 38 ± 4 % after weight
measurements. For all type of tensile samples the Young’s modulus was measured
between 0.1 and 0.3% of strain.
a. b. c.
Figure 5-12. a) fibre bundles ready to be tested, b) flexibility of the bundle due to its small thickness and c)
bundle during tensile test.
5.4.4.2 Flexural test
Flexural three-point bending tests (3PBT) were conducted with two types of fibre
disposition (longitudinal and transverse) and performed on a universal testing
machine (Instron 4426) according to ASTM D790M (see Figure 5-13). The bending
modulus of each sample was determined by calculating the slope of the stress-strain
curve between 0.1 and 0.3% of the strain. The crosshead speed was set at 1mm/min,
and a 1kN loadcell was used during the test. The load and the flexural displacement
are registered during the complete test. At least five samples were tested in each
configuration. See Table 5-3 for the sample dimensions.
Figure 5-13. Set-up for three-point bending test (3PBT) of bamboo fibre-epoxy composites.
5.4.5 Scanning electron microscopy (SEM) observations
Micrographs of fibres and composites were made by scanning electron microscopy

observations using secondary electrons using a voltage between 10 and 15 kV.
Due to some variations in the fibre volume fraction of the composites, an efficiency
factor and normalised values at 40% Vf for strength and Young’s modulus will be
used to allow comparison between the different specimens. Two values for the
efficiency factors are calculated; the first one corresponds to the ratio between the
experimental result and the calculation of the property using the rule of mixtures
(ROM), which assumes a perfect composite. For this, a fibre strength of 600 MPa
and Young’s modulus of 43 GPa were used for the calculation.
120
120 Chapter 5b Chapter 5
Lv Lv=
Lv Lv=
Lv random
random
UD-D with
UD-D with fixed UD-D with fixed UD-D with fixed UD-D with random
random fibre UD-C UD-C
fibre bundle fibre bundle fibre bundle fibre overlapping
bundle overlapping (impregnated (continuous
overlapping length overlapping length overlapping length length
length fibre bundle) fibres)
(Lv = 10%) (Lv = 30%) (Lv = 50%) ( Lv = random)
(Lv = random)
a a a a a a a
Eff. b Eff. b Eff. b Eff. b Eff. b Eff. b Eff. b
Average Average Average Average Average Average Average
Property Factor factor factor factor factor factor factor
value* value* value* value* value* value* value*
(%) (%) (%) (%) (%) (%) (%)
Tensile
86 16 ± 1 90 91 89 90 94 92
stiffness 17 ± 1 16 ± 2 17 ± 1 17 ± 1 19 ± 1 18 ± 1
84 89 84 89 89 - 94
(GPa)
Tensile
30 28 29 35 63 90 79
strength 85 ± 18 80 ± 12 81 ± 13 100 ± 18 191 ± 21 254 ± 18 222 ± 13
33 31 31 39 75 - 87
(MPa)
Strain at
breakage - 0.6 ± 0.11 - 0.5 ± 0.14 - 0.5 ± 0.1 - 0.7 ± 0.1 - 0.8 ± 0.03 - 1.3 ± 0.1 - 1.4 ± 0.2
(%)
Table 5-4. Tensile test results for unidirectional continuous (UD-C) and discontinuous patterns for bamboo fibre-epoxy composites (UD-D). The efficiency factor is
calculated as the ratio between the experimental result of the property and the calculated value using the rule of mixtures at the same fibre volume fraction. aThe first
value (up) corresponds to the efficiency factor calculated with the ROM, the second one (down) was obtained having as a reference the maximum composite strength
obtained with continuous fibres (254 MPa) that corresponds to the impregnated bundle test. bThe results are normalized at a fibre volume fraction of 40%. Lv =
overlapping length. For all UD-D samples the length of the fibres (Ls) was 50 mm.
The second efficiency factor is obtained as the ratio between the experimental result
and the maximum experimental composite strength value obtained with continuous
fibres, corresponding to the impregnated bundle test. All results of the tensile tests
performed on UD-D and UD-C samples, are presented in Table 5-4 and a typical
stress-strain curve is presented in Figure 5-14.
Figure 5-14: Typical tensile stress-strain curves for continuous (UD-C) and discontinuous (UD-D) bamboo
fibre epoxy composites.
5.5.1 Continuous fibre composites (UD-C)
Table 5-4 reveals that the experimental values for longitudinal tensile stiffness and
strength for UD-C respectively reach 92% and 79% of theoretical values found with
the ROM. This points out that a strong fibre-matrix interface is present between
fibres and matrix and that the resin impregnates the fibres very well and the good
alignment of the fibres. A visual inspection of the samples after failure indicated, in
general, a brittle fracture with a crack mostly propagating in one plane as shown in
Figure 5-15a. The fracture surfaces were also examined under SEM, showing good
resin impregnation and a quite clear fracture of the sample with relatively low
presence (~25-30 %) of fibre pull out (see also Figure 5-15b). This estimation is
based on the number of pulled out fibres relative to the total number of bamboo
fibres in the micrographs. Also, it was possible to observe a good dispersion of the
fibres because the layer-wise initial configuration of the composite is not visible in
the final material.
122 Chapter 5b
a. b.
25 mm
Figure 5-15. a) Composite fracture after testing and b) SEM observations of the fracture plane of the UD
continuous fibre composites after tensile testing.
Impregnated bundles
For UD-C impregnated bundles, the efficiency factor values (with respect to the
ROM) for longitudinal tensile stiffness and strength are 94% and 90% respectively.
The stiffness value (~19 GPa), is improved by ~5% in comparison with the standard
UD-C samples. This improvement can be attributed to the fact of having less fibres
in the sample; in practice, it is much easier to control the alignment of the fibres,
positively affecting the stiffness and strength of the composite material.
The tensile strength for this type of specimens was around 13% higher than for the
“standard” UD-C samples. This result can be expected if the volume of the sample is
considerably smaller, meaning less probability of presence of defects and improving
the strength properties. When the technical fibre strength was back calculated from
the impregnated bundle test, the values were found to be close to the strength
properties given by the dry fibre bundle test (DFB) presented in section 4.5.4 (see
Figure 5-16). Even though, the parameters vary between both tests, this can be an
indication that DFB test values are representative for the mechanical behaviour of
the fibre in the final composite material.
Single fibre test

Dry fibre bundle test
Back calculated (ROM)
fibre strength from the
impregnated bundle
test
40 mm gauge length 100 mm gauge length
Figure 5-16. Strength comparison between single fibre and dry fibre bundle test for technical bamboo fibres at
different gauge lengths. The single fibre strength data reported at 100 mm gauge length was extrapolated from
the experimental results at 40mm gauge length using Equation 4-5 (m=7.6), using the length of the technical
fibre (L) as a scale variable.
5.5.2 UD-D with fixed fibre bundle overlapping length
The results of the tensile tests, performed on UD-D samples (Lv= 10, 30 and 50%)
and Ls= 50 mm and ½ w) are given in Table 5-4. The introduction of discontinuities
at the technical fibre ends significantly reduces the strength of the composite
samples. The strength reduces to around one third with the insertion of these weak
points and the efficiency factor drops to approximately 30%. The strain to failure
also reduces from approximately 1.3% to only 0.5%. These discontinuities result in
the presence of a matrix rich area in the composite (Figure 5-17), that under tensile
load produces a strain magnification zone that creates shear stresses with the
adjacent fibre bundle, as seen in a schematic representation in Figure 5-18. These
shear stresses in turn, trigger a crack initiation that easily propagates due to the
limited toughness of the epoxy resin and causes the failure of the sample. The
Young’s modulus did not show a significant decrease and remained at ~17 GPa for
all the UD-D configurations, which represents a 6% of reduction in comparison with
the full UD continuous fibre composite (UD-C).
124 Chapter 5b
Tensile sample
Tensile
sample
Fibre
ends Resin-reach zone
Figure 5-17. As an example, a UD-D sample with LO=50%, where the introduction of the fibre ends
(discontinuities) results in the presence of resin-rich zones.
Technical
bamboo fibres
Shear
stress
UD-D
Crack
initiation
Strain
magnifi-
Resin-reach
Fibre cation
zone
disconti- zone
nuities
Figure 5-18. Schematic illustration of the fracture mechanism for UD-D samples.
5.5.3. UD-D with random fibre bundle overlapping length
The results of the tensile tests, performed on this type of composites (Lv= random,
Ls= 50 mm and 1/5 of the w) are given in Table 5-4. As compared to the composites
with fixed overlapping length, the reduction of the fibre bundle width and their
overlap randomization allowed a minor increase in the efficiency factor of ~5% for
the longitudinal tensile strength. SEM observations however pointed out, in some
cases, a similar fracture phenomenon as present in the UD-D samples with fixed
overlapping length. This indicates that the width of the fibre discontinuity is still
high and acts as a crack initiator leading to premature failure of the samples.
In industry short fibre tapes are emerging with a relatively new technology called
“fibre patch preforming” [35], which can be a suitable way to tackle the
discontinuity of the fibres. This technology allows short fibres to be positioned
according to nearly any specification, with a robot providing a full range of
possibilities in the fibres’ position and orientation. These preforms are cut from
continuous UD preforms of carbon fibre (in 60 x 20 mm patches) and then placed by
a robot directly into the mould in a defined pattern, being ready to use in standard
infiltration processes, and can be combined with other preform types. The range of
applications includes production of small and complex geometric parts including
medical prosthetics or high-performance sport equipment. This technology can be
potentially used for the current discontinuous bamboo fibres.
5.5.4. UD-D with individual (technical) fibres with random fibre ends
In this fibre configuration (Ls= 50 mm, Lv= random), the crack initiator effect of the
fibre discontinuities was removed due to the randomization of the fibre ends,
resulting in properties close to the UD-C properties, as seen in tensile tests (Table 5-
4). The introduction of randomized fibre discontinuities (individual fibres with Ls=
50 mm) leads to a preservation of 85% of the longitudinal tensile strength in
comparison with the UD-C. These results clearly show that randomization of the
discontinuities on individual bamboo technical fibres inside the composite, is
necessary to take advantage of the good mechanical properties of the fibres and to
translate them to the composite. By varying the fibre ends over the length of the
sample, the overall stress fields surrounding the discontinuities are expected to be
minimized, slowing down the initiation of cracks. After tensile testing, the
specimens exhibited a mixed mode of fracture, where in most of the cases the
fracture was scattered randomly along the cross-sectional plane of the samples,
indicating less contribution of the additive stress concentrations (due to the in plane
bundle discontinuities), and minimized by the presence of the random fibre ends.
5.5.5 Tensile fracture characteristics for UD-D samples
In general, for all UD-D samples with different overlapping lengths (Lv), the
discontinuities resulted in two different types of tensile fracture, as shown in Figure
5-19. The first consists of staggered fracture following the fibre bundle “brick wall”
construction inside the composite. The shear stresses created by the discontinuity of
the fibres initiate the crack and then it runs along the rich-resin zone created by the
fibre ends (discontinuities). Then, the crack continues growing preferably through
the fibre length parallel to the fibre bundles due to delamination between the
126 Chapter 5b
initiated crack and the adjacent fibre bundle which is “bridging” the fibre ends until
it connects to with another bundle discontinuity and causes the final failure of the
composite. This type of fracture occurs especially in composites with L v=10%
(Figure 5-19a), due to the proximity of the discontinuities; nevertheless this case is
also observed in some samples of the other configurations.
a. b.
c. d.
e. f.
Figure 5-19. Tensile fracture of UD-D samples. Specimens with a) Lv=10%, b) Lv=30%, c) Lv=50%, d)
random Lv with slits of one fifth of the composite width, e) UD-D with single technical fibres in random
disposition and f) their typical tensile failure (highlighted by a black line). For all samples Ls=50mm.
The second case consists of a rather straight failure across the UD-D sample. Here,
the crack starts as explained before due to a strain magnification, but instead of
being deflected by delamination along the fibre length to the next fibre discontinuity
(fibre bundle end), it continues all along the composite width. This situation is
presented in all fibre patterns but especially in configurations with Lv=30 and 50%
(Figures 5-19b and c).
In UD-D samples with randomized individual fibre ends, the failure mode can be
described as a long splitting along the gauge length mostly in the middle of the
specimen (SGM), see Figure 5-19f. The crack is initiated in the most critical defect
along the sample and running preferably in longitudinal direction of the fibres. All
different fibre patterns described above (Figure 5-4), did not affect the linear elastic
behaviour of the material as is shown in Figure 5-14.
5.5.6. Experimental stiffness compared with predicting models in UD bamboo

fibre-epoxy composites
Composite stiffness estimations from different models are compared in Table 5-5.
For this comparison the rule of mixtures (ROM), the shear lag theory, the Mori-
Tanaka model and the Halpin-Tsai equations were chosen. These models were
already described in the introductory part of this chapter in paragraph 5.2 and can be
compared with the experimental results in Table 5-4 and Figure 5-20.
Experimental
Mori- Halpin-
Rule of Shear- lag results
Property Tanaka Tsai
Mixtures theory UD-C UD-D
model equation
Longitudinal tensile
18.8 18.6 18.9 17.2 18-19 16-17
stiffness (GPa)
Table 5-5. Results for the longitudinal stiffness of bamboo fibre-epoxy composites estimated with several
models and compared with the experimental results (40% Vf).
The Young’s modulus of the bamboo technical fibre used in the models was 43 GPa
[36]. When comparing the experimental and theoretical results for the Young’s
modulus of the studied composites, the simplest model, the rule of mixtures (ROM),
shows a relatively large over-estimation for this property. This is due to the large
simplifications accounted for in this model, by assuming perfect bonding between
fibres and matrix giving an iso-strain condition under load, and by assuming
infinitely long and perfectly aligned fibres. Not only large simplifications, but also,
assumptions can lead to an inaccuracy for a model prediction as is the case for the
shear-lag theory and Mori-Tanaka model in this specific case of bamboo fibre-epoxy
128 Chapter 5b
composites. The results for the shear-lag theory, in which a square array was chosen
for the present study, show a slightly better approximation of the longitudinal
stiffness. It assumes, however, that the fibres can be modelled as perfect cylinders
and uses a hexagonal fibre packing. Bamboo fibres and their distribution inside the
composite do not satisfy these conditions.
Experimental results
Figure 5-20. Experimental results for the longitudinal Young’s modulus for bamboo fibre-epoxy composites
with different UD fibre configurations versus the estimation from different models.
The application of the Mori-Tanaka model requires knowledge of the different

engineering constants that set up the stiffness or compliance matrix. Considering the
anisotropic nature of the natural fibres, the required transverse stiffness and Poisson
ratio’s for bamboo fibres are still unknown. For these, a strong assumption needs to
be made, assuming that the fibre is an isotropic material with a fixed Poisson’s ratio
(0.33). This assumption was also made by Herrera-Franco et al. [37] when
modelling the tensile strength properties of UD henequen fibre and HDPE
composites, and by Sabeel et al. [38] with a value of 0.38 for the evaluation of jute-
glass fibre hybrid composites predicted by using the classical laminate theory.
Finally, the semi-empirical Halpin-Tsai equation gave a reasonably good estimation

of the longitudinal tensile stiffness for the bamboo fibre reinforced epoxy
composites. A possible reason is that this model not only takes into account the
geometry of the reinforcement, represented in the fibre aspect ratio (L/d) but also,
the fibre packing efficiency, instead of geometrical arrangement of the

reinforcement as is considered for the shear-lag model [39]. These considerations,
together with the good efficiency factor for the composite modulus reported
previously, suit better the “real” conditions of the composites and give better results.
In general, when comparing experimental results and theoretical predictions for the
Young’s modulus in natural fibre composites, controversial results were found in
literature. This poor agreement is attributed to the fact that the models do not take
into account the bonding between fibres and matrix [39]. However, for
unidirectional jute fibre reinforced composites with different fibre volume fraction,
the Halpin-Tsai model was successfully applied [40]. On the other hand, natural
fibre polymer-based composites analyzed through the same model, revealed that the
theoretical values for the Young’s modulus were higher than the experimentally
obtained values [41]. This can be again attributed to the poor interface between fibre
and matrix. It should be noted that this model is one of the most used material
property models, particularly in engineering design applications, due to its simple
universal form of expressions, and its wide applicability to a number of different
materials [42]. Nevertheless, Halpin–Tsai equations have been found to be
inaccurate at high Vf, because they do not take into account the limit in maximum
packing fraction in a real system [39] and give relatively low values for moderate to
high aspect ratios [13].
5.5.7 Experimental strength compared with predicting models in UD bamboo

fibre-epoxy composites
In this section, several models are applied to predict the tensile strength of the UD
discontinuous bamboo fibre-epoxy composites. The results of these predictions are
presented in Table 5-6 and Figure 5-21. The results of the continuous fibre
composites (UD-C) are given as a reference.
130 Chapter 5
130 Chapter 5b
Lv Lv Lv= Lv=
Lv
random random
UD-D with UD-D with UD-D with

UD-D with UD-D with
fixed fibre fixed fibre fixed fibre
random fibre random fibre UD-C UD-C
bundle bundle bundle
bundle overlapping overlapping (impregnated (continuous
overlapping overlapping overlapping
length length fibre bundle) fibres)
length length length
(Lv = random) ( Lv = random)
(Lv = 10%) (Lv = 30%) (Lv = 50%)
Strength Strength Strength Strength Strength Strength Strength

Model
(GPa) (GPa) (GPa) (GPa) (GPa) (GPa) (GPa)
Experimental* 85 ± 18 80 ± 12 81 ± 13 100 ± 18 191 ± 21 254 ± 18 222 ± 13
ROM 263 263 263 263 263 263 263
Kelly-Tyson 248 248 248 248 248 250 250
GLS 132 132 132 132 132 132 132
LLS 96 ± 12 98 ± 14 96 ± 14 126 ± 20 213 ± 16 219 ± 14 219 ± 14
Table 5-6. Overview of the different modelling attempts to predict the tensile strength of unidirectional continuous (UD-C) and discontinuous (UD-D) bamboo-epoxy
composites. The experimental results are also included as a comparison. The values of the experimental results were normalized at a fibre volume fraction of 40%. For
all UD-D samples the length of the fibres (Ls) was 50 mm. Lv= overlapping length.
Figure 5-21. Experimental results for the tensile strength for composites with different UD fibre
configurations versus the estimation from different models.
Local load sharing model LLS
The LLS model implements interactions between adjacent fibres that lead to a stress
redistribution through the application of matrix shear spring elements. With each
strain increment, the resulting tensile stresses in the axial fibre spring elements are
compared to the fibre strength (according to the probability of failure from the
Weibull distribution). If the fibre strength is exceeded, the fibre spring element is
removed from the system, simulating a broken fibre. After that, the stress
redistribution of the resultant load is recalculated [31, 32]. The original LLS model
model only allows the prediction of the tensile strength of continuous fibre
composites. After a modification, a UD-C can be modelled as a UD-D when
multiple axial fibre spring elements are removed from the system before starting the
simulation. This procedure allowed to implement the discontinuities already present
in the UD-D samples, as broken axial fibre spring elements under unloaded
conditions.
In order to implement the LLS model for this particular study with discontinuous
bamboo fibre-epoxy composites, several parameters needed to be set beforehand. As
132 Chapter 5b
it was stated before, the value of the matrix shear spring element stiffness (‫ )ז‬was
experimentally determined (32 GPa) by applying the model to the experimental
results of the unidirectional composites (UD-C). This parameter has no real physical
meaning and represents the behavior of the matrix and locally affected interface
when the fibre is broken. The same value was used to calculate the strength
predictions for the other UD-D samples. The numerical experiments were repeated
ten times in order to obtain reliable statistical variations. The list of introduced
parameters and their values required for the LLS model are presented in Table 5-7.
Parameter Value Parameter Value

Fibre Young’s modulus [Ef] (GPa) 43 Young’s modulus of the matrix [Em] (GPa) 2.7
Fibre radius [rf] (μm) 180 Fibre volume fraction [Vf] (%) 40
Weibull characteristic strength of
Composite length [Le] (mm) 150
life [σo] (MPa) 850
Weibull shape parameter [m] 7.6 Effective matrix shear yield stress [‫( ]ז‬GPa) 32
Weibull reference length for the
1
fibre [Lo] (mm)
Table 5-7. Overview of the different initial parameters set for the calculation of the tensile strength using the
local load sharing (LLS) model.
The modified LLS model was able to predict the UD-D composites tensile strengths
for the different pattern configurations tested. Since the model takes into account the
neighbour fibres which are surrounding the discontinuous fibres through the use of
shear spring elements, good and reliable predictions can be achieved. As explained
previously, the interactions between adjacent fibres allow the stress redistribution
with the application of matrix shear spring elements which are able to detect the
amount of continuous fibres around the fibre discontinuities that were introduced in
the model. A larger amount of continuous fibres surrounding the discontinuities
gives better bridging to these weak points, avoiding the formation of stress
concentrations and hence giving better strength properties. This effect can be clearly
seen in Figure 5-21 which shows that when the fibre discontinuities become
narrower, the strength of the composite is higher as is well predicted by the modified
LLS model (shown in the same figure). Figure 5-22 schematically represents this
situation for a typical transversal planes in the UD-D studied patterns (categories 1,
2 and 3 presented in section 5.4.1). This Figure shows the discontinuous fibres (▬)
and their neighbouring continuous fibres (▬) connected by the shear spring
elements, simulating the fibre distribution in the real composites according to each
case.
Large fibre discontinuity

a. UD-D
sample
Fibre
discontinuity
b. UD-D
sample
Cut fibre Continuos fibre
c. UD-D
sample
Figure 5-22. Schematic representation of the modified LLS model applied in this thesis (see section 5.2.2)
showing, in a transversal plane, the neighbouring fibres close to the fibre discontinuities according to the
different studied patterns: a) category 1, b) category 2 and c) category 3. For additional information about the
fibre patterns and categories, see section 5.4.1.
The difference between experimental and predicted values by the modified LLS
model were within a margin error of around 15%, with predictions systematically
higher than the experimental results. A possible explanation is that the strain field of
full continuous fibre composites (UD-C) under axial load is homogeneous, while
this is not the case for discontinuous composites (UD-D’s). When the fibre
discontinuities are introduced in the composite, the stress field becomes
inhomogeneous near the fibre edges, generating stress concentrations in the regions
adjacent to the pre-existing fibre discontinuities when loading the sample, and
producing premature failure. This situation cannot be accounted for in the LLS
model, preventing not only the incorporation of local stress concentrations around
the fibre discontinuities, but also the decrease in load transfer between adjacent
fibres upon interface debonding.
134 Chapter 5b
Other models
Although most of the studied analytical models manage to predict the longitudinal
tensile strength of continuous bamboo fibre epoxy composite within a reasonable
degree of accuracy, they fail to accurately predict the longitudinal tensile strength of
discontinuous samples. The application of the rule of mixtures gave an
overestimation of the composite strength of approximately 216%. The Kelly-Tyson
model gives only a slight improvement in the prediction of the composite tensile
strength by accounting for stress redistribution in the ineffective length of the fibre.
However, this model still presumes perfect plastic yielding behaviour of the matrix
and constant frictional shear stress at the interface to bring forth a linear axial fibre
stress in the ineffective length. Furthermore, it does not incorporate the statistical
influence of the fibre strength. The Kelly-Tyson model did not predict the tensile
strength of discontinuous composites with sufficient accuracy.
The GLS model does take into account the Weibull distribution of the fibre strength.
It treats the composite as a chain of bundles, and accounts for load contributions of
broken fibres. Stress recovery occurs in the ineffective length due to the presence of
a presumed constant interfacial shear stress. However, this model does not consider
the stress redistribution from the broken fibre to the adjacent fibres which has an
important effect on the final mechanical characteristics of the composites.
Furthermore, this model neglects the contribution of the matrix to the tensile
strength of the composite.
5.5.8 Properties evaluated in 3-point bending
Table 5-8 shows flexural properties for unidirectional bamboo fibre-epoxy

composites with longitudinal and transverse distribution of the fibres. For
longitudinal flexural strength, an efficiency factor of 82% was reached after
comparing the experimental value (265 MPa) with the theoretical value (323 MPa).
This indicates a good fibre/matrix adhesion, possibly enhanced due to the surface
fibre roughness that promotes mechanical interlocking. The composite stiffness,
with the same fibre disposition, reaches 95% of efficiency factor, with an
experimental value of 19 GPa, indicating a good alignment of the technical fibre in
UD disposition.
Fibre orientation Flexural strength Flexural stiffness Strain to failure

(MPa) (GPa) (%)
Longitudinal 265 ± 12 19.3 ± 0.8 2.4
Transversal 33 ± 2 2.7 ± 0.2 1.3
Table 5-8. Flexural properties with longitudinal and transversal disposition of the fibres. The fibre volume
fraction was normalized at 40%.
The good adhesion between fibre and matrix mentioned above is confirmed by the
experimental results in the transverse 3PBT. This test provides a direct estimation of
interface tensile strength. In the present study, the bamboo fibre-epoxy composites
reached a transverse strength of 33 MPa and a stiffness of 2.7 GPa with untreated
fibres (Table 5-8). Figure 5-23 shows the fracture surface after transversal 3PBT
with fibres covered by the polymer and good dispersion (wetting) of the resin
around the technical bamboo fibres. From these results, it is possible to affirm that
the interfacial fibre-matrix strength is reasonably good. For flax-epoxy composites
(Vf = 40 %), it was reported that the best value in transverse 3PB was 20 MPa after
1% of alkali treatment (20 min at RT) [43]. In this case, the treatment removed
impurities and waxy substances from the fibre surface and promoted the creation of
a rougher topography and enhanced the interface quality by mechanical interlocking.
Also, pineapple leaf fibres and PHBV resin composites (Vf = 28 %), achieved 32
MPa with untreated fibres [44].
200 µm
Figure 5-23. SEM images for untreated bamboo fibre – epoxy composites: (a) fibre surface covered by epoxy
resin after transversal 3PBT and b), a polished cross sectional area of the composite.
It was also noticed that the bending properties (strength and stiffness) are very close
to the values obtained with the impregnated bundle test when tested in tension as can
be seen by comparing Tables 5-4 and 5-8. This shows that the flexural strength is a
bit higher, which is a consistent result due to the fact that in 3PBT only the bottom
136 Chapter 5b
part of the sample is under tension. In the case of the impregnated bundle tensile test,
the whole sample volume is loaded and there is a higher probability of defects
present in the sample; hence, a lower strength can be expected.
a.
1. Bamboo + Epoxy (Vf: 40%)*

2. Flax + Epoxy (Vf: 40%)
3. Jute + Epoxy (Vf: 40%)
4. Jute + Vinylester (Vf: 35%)
5. Hemp + Epoxy (Vf: 35%)
6. Sisal + Epoxy (Vf: 37%)
7. Kenaf + Cashew nut Shell (Vf: 64%)
8. Bamboo + Polyester (Vf: 15%)
9. Kenaf + Polyester (Vf: 60%)
10. Hemp + Polyester (Vf: 60%)
11. Hemp + Cashew nut Shell (Vf: 65%)
b.
1. Bamboo + Epoxy (Vf: 40%)*

2. Flax + Epoxy (Vf: 40%)
3. Jute + Epoxy (Vf: 40%)
4. Jute + Vinylester (Vf: 35%)
5. Hemp + Epoxy (Vf: 35%)
6. Sisal + Epoxy (Vf: 37%)
7. Kenaf + Cashew nut Shell (Vf: 64%)
8. Kenaf + Polyester (Vf: 60%)
9. Hemp + Polyester (Vf: 60%)
10. Hemp + Cashew nut Shell (Vf: 65%)
Figure 5-24. Comparisons between a) flexural strength and b) flexural modulus for unidirectional natural
fibre-thermoset resin composites found in similar studies [2, 45-50]. This comparison is only indicative due to
the significant difference in Vf of the reported samples. *Normalized results from this study with untreated
bamboo fibres.
Figure 5-24 shows a comparison for flexural strength and flexural stiffness between
different natural fibres and thermoset matrices in 3-point bending testing with
longitudinal disposition of the reinforcement. In literature, a significant
improvement in properties for several natural fibre composites with thermoset
matrices was found after fibre alkali (NaOH) treatments at different concentrations,
temperature and time of exposure (see section 3.3.2, Table 3-3 and 3-4). For flexural
properties of bamboo fibre-epoxy composites with the same type of fibres used in
this study as well as the same manufacturing process, no improvement in flexural
strength after different fibre alkali treatments was observed (1,3, and 5% for 20 min
at RT) [36]. The fact that the flexural strength is the highest for the non-treated
fibres is an important advantage for future industrial uses, reducing both cost and
fibre preparation, as well as having less environmental impact for the entire
manufacturing process.
5.6 Conclusions
Unidirectional continuous (UD-C) and discontinuous (UD-D) bamboo fibre-epoxy

composites were successfully manufactured to evaluate the effect of different
unidirectional fibre patterns on their composite tensile properties and results were
compared with predictive models. Not only a UD fibre alignment procedure was
developed allowing the production of fibre-epoxy samples, but also a novel
discontinuous UD fibre randomization technique has been proposed at lab scale
(proof of concept), to produce continuous tape from discontinuous fibres.
In general, a high strength is found for the unidirectional continuous samples (UD-C)
tested in tension and bending. The results are close to what could be expected based
on single technical fibre properties. A reason for this good behaviour is the high
quality bamboo fibre obtained after the mechanical extraction process that conserves
their intrinsic good characteristics with a high fibre surface roughness that promotes
a good mechanical interlocking with the epoxy matrix. Moreover a good bonding is
present of the epoxy with the chemical groups present on the fibre surface,
according to the good results of the transverse flexural strength with untreated fibres.
It was found, in comparison with similar systems, that alkali treatment is not
necessary to enhance the strength of the composite. This fact reduces the production
cost and strengthens the environmental advantages of this natural fibre.
In this study, the full randomization of technical fibres of 50 mm length to make

composites was approached with positive results, reducing the longitudinal tensile
strength by only 15% in comparison with the UD-C (impregnated bundle). This
aspect opens the possibility for a number of industrial applications in which an
endless prepreg is needed. Also, in a real production process, the technical bamboo
fibres, will be longer, in its majority at least twice of the used fibre length,
generating less presence of discontinuities and diminishing their effect on the
composite strength. In this case, together with a standardized production process for
the discontinuous bamboo preform, it is expected that the difference between
138 Chapter 5b
UD-D’s and UD-C’s will be closer. The composite stiffness variations were not
significant for any of the studied UD fibre patterns.
For the strength predictions the best approximation was achieved by the local load
sharing model (LLS). This model was applied in a novel approach incorporating the
effect of the fibre discontinuities inside the bamboo fibre epoxy composites, and it
was able to predict the experimental tensile strength within an error margin of
around 14%. This positive result is based on a more realistic approach of the failure
of the material, represented in the probability of failure of the bamboo fibres,
Weibull distribution parameters found specifically for the studied fibres, and the
load redistribution to the neighbour fibres after fibre breakage.
Impregnated bamboo fibre bundle tests can be used as a good alternative to evaluate
the composite tensile properties. They present several advantages in comparison
with the “standard” composite such as less material needed and less preparation for
the specimen manufacturing. Also, because they use rovings (tow) or yarns, that do
not need to be woven into a fabric, stitched or delivered in a UD configuration in
order to have a preform (prepreg) to proceed with the composite manufacturing,
saving time and costs. Additionally, higher mechanical properties were obtained for
the impregnated bundles in comparison with the UD-C. This is due to the easier
control of the fibre alignment for the case of the stiffness, and reduced sample
volume (with the corresponding less probability of defects) that improves the
strength. These data can be interpreted as target values that can be achieved for the
composite material.
References
[1] Liese W. The Anatomy of Bamboo Culms (Inbar Technical Reports)1998.

[4] Ray AK, Das SK, Mondal S, Ramachandrarao P. Microstructural characterization of bamboo. J
Mater Sci. 2004;39(3):1055-60.
Mater. 1986;20(1):19-29.
[6] Thwe MM, Liao K. Environmental effects on bamboo-glass/polypropylene hybrid composites. J
Mater Sci. 2003;38(2):363-76.
[7] Londoño X. The American bamboos with emphasis in the genus guadua. Cali, Colombia:
Instituto Vallecaucano de Investigaciones Científicas INCIVA; 2004.
[9] Baets J, Plastria D, Ivens J, Verpoest I. Determination of the optimal flax fibre preparation for
use in unidirectional flax -epoxy composites. J Reinf Plast Compos. 2014;33(5):493-502.
[10] Goutianos S, Peijs T, Nystrom B, Skrifvars M. Development of flax fibre based textile
reinforcements for composite applications. Appl Compos Mater. 2006;13(4):199-215.
[11] Cox H. The elasticity and strength of paper and other fibrous materials. Br J Appl Phys.
1952;3:72-9.
[12] Lacroix T, Tilmans B, Keunings R, Desaeger M, Verpoest I. Modelling of critical fibre length
and interfacial debonding in the fragmentation testing of polymer composites. Compos Sci Technol.
1992;43(4):379-87.
[13] Tucker Iii CL, Liang E. Stiffness predictions for unidirectional short-fiber composites: Review
and evaluation. Compos Sci Technol. 1999;59(5):655-71.
[14] Rosen BW. Tensile failure of fibrous composites. AIAA J. 1964;2(11):1985-91.
[15] Robinson IM, Robinson JM. The effect of fibre aspect ratio on the stiffness of discontinuous
fibre-reinforced composites. Composites. 1994;25(7):499-503.
[16] Affdl JCH, Kardos JL. The Halpin-Tsai equations: A review. Polymer Engineering & Science.
1976;16(5):344-52.
[17] Hermans J. Elastic properties of fiber reinforced materials when fibers are aligned. Koninklijke
Nederlandse Akademie Van Weteschappen-Proceedings Series B-Physical Sciences. 1967;70(1):1-
9.
[18] Hill R. Theory of mechanical properties of fibre-strengthened materials: I. Elastic behaviour. J
Mech Phys Solids. 1964;12(4):199-212.
[19] Facca AG, Kortschot MT, Yan N. Predicting the elastic modulus of natural fibre reinforced
thermoplastics. Compos Part A: Appl Sci Manuf. 2006;37(10):1660-71.
[20] Halpin J. Stiffness and expansion estimates for oriented short fibre composites. J Compos
Mater. 1969;3:732-4.
[21] Mori T, Tanaka K. Average stress in matrix and average elastic energy of materials with
misfitting inclusions. Acta metallurgica. 1973;21(5):571-4.
[22] Gommers B, Verpoest I, Van Houtte P. The Mori-Tanaka method applied to textile composite
materials. Acta Mater. 1998;46(6):2223-35.
[23] Xue Y, Du Y, Elder S, Wang K, Zhang J. Temperature and loading rate effects on tensile
properties of kenaf bast fiber bundles and composites. Compos B: Eng. 2009;40(3):189-96.
[24] Mura T. Micromechanics of defects in solids. Second edition ed. The Netherlands: Kluwer
Academic Publishers; 1982.
[25] Eshelby JD. Elastic inclusions and inhomogeneities In: Sneddon I, Hill R, editors. Progress in
material mechanics. Amterdam1961. p. 89-1470.
[26] Kelly A, Tyson W. Tensile properties of fibre-reinforced metals: copper/tungsten and
copper/molybdenum. J Mech Phys Solids. 1965;13(6):329-50.
[27] Curtin WA. Theory of Mechanical Properties of Ceramic-Matrix Composites. J Am Ceram Soc.
1991;74(11):2837-45.
[28] Curtin WA. Tensile Strength of Fiber-Reinforced Composites: III. Beyond the Traditional
Weibull Model for Fiber Strengths. J Compos Mater. 2000;34(15):1301-32.
[29] Weibull W. A statistical theory of the strength of materials. Proceedings of the Royal Swedish
Institute of Engineering Research. 1939;151:1-45.
[30] Weibull W. A statistical distribution functions of wide applicability. J Appl Mech.
1951;18:293-7.
[31] Okabe T, Sekine H, Ishii K, Nishikawa M, Takeda N. Numerical method for failure simulation
of unidirectional fiber-reinforced composites with spring element model. Compos Sci Technol.
2005;65(6):921-33.
[32] Okabe T, Ishii K, Nishikawa M, Takeda N. Prediction of tensile strength of unidirectional
CFRP composites. Adv Compos Mater. 2010;19(3):229-41.
[33] Okabe T, Nishikawa M, Takeda N, Sekine H. Effect of matrix hardening on the tensile strength
of alumina fiber-reinforced aluminum matrix composites. Acta Mater. 2006;54(9):2557-66.
[34] Taketa I. Analysis of failure machanisms and hybrid effects in carbon fibre reniforced
thermoplastic comosites [Doctoral thesis]. Leuven: University of Leuven; 2001.
140 Chapter 5b
[35] Drechsler K. History, status and future potential of textile tecnologies for high performance
composites. Composites Week @ Leuven and TexComp-11 Conference. Leuven 2013.
[37] Herrera-Franco P, Valadez-Gonzalez A. Mechanical properties of continuous natural fibre-
reinforced polymer composites. Compos Part A: Appl Sci Manuf. 2004;35(3):339-45.
[38] Sabeel Ahmed K, Vijayarangan S. Elastic property evaluation of jute-glass fibre hybrid
composite using experimental and CLT approach. Indian journal of engineering & materials
sciences. 2006;13(5):435-42.
[39] Borges P, Godinho M, Figueirinhas J, De Pinho M, Belgacem M. All-cellulosic based
composites. In: Kalia S, Kaith B, Kaur I, editors. Cellulose fibres: Bio- and nano-polymer
composites Green chemistry and technology. New York (UK) 2011. p. 399-424.
[40] Bernal C, Cabral H, Vázquez A. Mechanical properties of jute-PP composites. In: Sol H,
Degrieck J, editors. Proceedings 10th European Conference on Composite Materials (ECCM-10).
Brugge, Belgium 2002.
[41] Satyanarayana K, et al. Dimensional analysis and surface morphology as selective criteria of
lignocellulosic fibres as reinforcement in polymeric matrices. In: Kalia S, Kaith B, Kaur I, editors.
Cellulose fibres: Bio- and nano-polymer composites Green chemestry and technology. New York
(UK)2011. p. 215-40.
[42] Hibbeler RC. Engineering Mechanics: Statics and Dynamics: Macmillan; 1992.
[43] Van de Weyenberg I. Flax fibres as a reinforcement for epoxy composites [Doctoral thesis].
Leuven: University of Leuven; 2005.
[44] Luo S, Netravali AN. Mechanical and thermal properties of environment-friendly “green”
composites made from pineapple leaf fibers and poly(hydroxybutyrate-co-valerate) resin. Polym
Compos. 1999;20(3):367-78.
Sci Technol. 2004;64(9):1219-30.
thermal properties of kenaf and hemp bast fibre composites: part 2 - cashew nut shell liquid matrix.
[47] Bledzki AK, Fink HP, Specht K. Unidirectional hemp and flax EP- and PP-composites:
Influence of defined fiber treatments. J Appl Polym Sci. 2004;93(5):2150-6.
[48] Gassan J, Bledzki AK. Possibilities for improving the mechanical properties of jute/epoxy
composites by alkali treatment of fibres. Compos Sci Technol. 1999;59(9):1303-9.
[49] Ray D, Sarkar BK, Rana AK, Bose NR. The mechanical properties of vinylester resin matrix
composites reinforced with alkali-treated jute fibres. Compos Part A: Appl Sci Manuf.
2001;32(1):119-27.
[50] Van de Weyenberg I, Chi Truong T, Vangrimde B, Verpoest I. Improving the properties of UD
flax fibre reinforced composites by applying an alkaline fibre treatment. Compos Part A: Appl Sci
Manuf. 2006;37(9):1368-76.
Thermal degradation in bamboo fibres and bamboo fibre polypropylene composites 141
Chapter 6
Thermal degradation in bamboo fibres and
bamboo fibre polypropylene composites
_______________________________________________________
6.1 Introduction
Thermoplastic matrices are a good alternative in comparison with thermosets to

produce bamboo fibre composites, they allow clean and fast processing, lower cost,
provide more impact resistant composites and allow mechanical downcycling of the
composite at the end-of-life (e.g. for injection moulding). For these reasons, recent
developments have been shifting to thermoplastic matrix composites [1]. However,
the degradation or thermal instability of natural fibres occur typically at
temperatures above 180°C due to the inherent organic composition of the natural
fibres, as shown previously in section 3.4.2. The overall advantages of bamboo
fibres have not yet been fully exploited for high-tech thermoplastic composite
applications because of the limitation of the low required melting temperature of the
polymer. This sets a serious limit to the processing conditions which narrows down
the thermoplastic choices [2].
High temperatures are applied mainly during the manufacturing (e.g. compression
molding), where the thermoplastic matrix needs to have reduced viscosity to
impregnate the fibre bundles and to shape the part. The exposure of the fibres to
high temperature, often in an oxidative environment (air), results in a decrease of the
mechanical properties accompanied with side effects such as discolouration and
unpleasant odour of the composites [3, 4], which means a significant disadvantage
for applications such as car interior applications.
142 Chapter 6b
Only polymers with a sufficiently low processing temperature, mostly polyolefines

such as polyethylene and polypropylene (PP), are considered suitable for natural
fibre reinforced parts [5]. Polypropylene is used as matrix for a number of reasons.
Firstly, it is easy to process and it is one of the cheapest polymers on the market, as
well as having a low processing temperature. On the other hand, a drawback is its
hydrophobic character (and this accounts for many thermoplastics) which is
incompatible with the typically more hydrophilic nature of plant fibres, creating the
need to modify either the surface of the fibre or the matrix. In the case of PP, maleic
anhydride modified polypropylene (MAPP) is commonly used to enhance chemical
adhesion with natural fibres, proving to be very effective in enhancing the
mechanical properties of the composite [6-8].
Several studies using TGA and DSC techniques on the thermal behaviour of
lignocellulosic (technical) fibres and their composites, such as hemp [9-11], sisal [12,
13], flax [14-16], kenaf [17, 18], jute [3, 14, 19] and bamboo fibres [7, 20, 21] are
available in literature. These studies have shown that the use of an inert atmosphere
during the thermal treatment can delay the thermal degradation in single natural
fibres [22, 23]. In spite of the available data, to the best of the author’s knowledge,
there are not studies that make a reliable connection between single fibre thermal
degradation, and its influence on the mechanical properties of fibres and composites
through mechanical testing. Also, there is a lack of information on the potential
benefit of using inert gas in thermal degradation of individual technical fibres or
during the manufacturing of the natural fibre composites. This is in spite of the good
TGA results under inert environment mentioned above.
It is therefore of practical importance to characterize the effect of high temperature

and exposure time during the processing on natural fibres and their composites. The
results will help to establish and/or adapt manufacturing parameters to prevent
excessive fibre degradation that can reduce the performance of the bamboo fibre-
thermoplastic composite. Also, this will open new options for thermoplastic matrices
and new industrial applications.
In this chapter, the results of the research on the thermal degradation of bamboo
technical fibres and bamboo fibre polypropylene composites (BFPP) are presented,
with the aim to characterize the loss in mechanical properties due to thermal
degradation. Individual technical fibres were thermally treated in different
environments (i.e. air and argon) at different temperature-time couples and
subsequently characterized mechanically using the single fibre tensile test. TGA
analysis (including isothermal TGA), was carried out at the same temperature-time
couples to establish a relationship between the mass loss and single technical fibre
strength.
Bamboo fibre polypropylene composites were produced at different consolidation

temperatures and in different environments (air and argon), in order to determine the
mechanical properties of the material. The samples were characterized in flexural 3
point bending tests (3PBT), with two fibre orientations (longitudinal and transverse).
The relation between the degradation due to thermal treatment and mechanical
properties of single technical fibres in final bamboo fibre PP composites (BFPP) will
be discussed in combination with the TGA results.
6.2 Materials
6.2.1 Bamboo fibres
Mechanically extracted technical bamboo fibres were used in this study. The
characteristics of the fibre and the extraction method are described in sections 4.3
and 4.5.1.
6.2.2 Polypropylene (PP) and maleic anhydride polypropylene (MAPP)
Polypropylene (PP) film with a density of 900 kg/m³ and 20 µm thickness was
supplied by Propex GmbH (Germany) and 0.3% maleic anhydride (MA) grafted
polypropylene (MAPP Bynel 50) was provided by Dupont (Switzerland). Table 6-1
shows the thermal and mechanical properties of these matrices.
Mechanical properties
Young’s
Tc Tm Density CTE Strength Strain to
Matrix modulus
(°C) (°C) (g/cm3) (10-6/K) (MPa) failure (%)
(GPa)
PP 115.7 160.6 0.9 62.7 – 73.2 1.6 – 1.8 55 - 65 >300
MAPP 100.3 147 0.89 112.2 – 175.8 0.3 18 475
Table 6-1. Thermal and mechanical properties of polypropylene (PP) and maleic anhydride polypropylene
(MAPP) used as a matrix, from manufacturer’s data sheets. DSC scans to determine Tm were conducted in a
temperature range from 40 to 600 °C, at a constant heating rate of 10 °C/min.
6.3 Methods
6.3.1 Thermal treatment of single bamboo fibres
The characterization of the thermal degradation of technical bamboo fibres was

carried out by measuring their tensile strength and stiffness after the treatment, at
144 Chapter 6b
different temperature-time couples in different environments (i.e. air and argon). The
starting temperature in the oven was 20 °C (time = 0); technical fibres were heated
at a rate of 5 °C/min until the target temperature (maximum temperature) and then
kept constant until the desired time. Before and after the thermal treatment, the
fibres were conditioned under standard environmental conditions (21°C± 2°C and
50±2 %RH) for 48 hours. Table 6-2 shows all temperature and time of exposure
combinations for the thermal treatment on individual technical fibres.
Time (min)
Max. temperature (°C) 25 50 60 70 80 100 120
180
200
220
250
Table 6-2. Overview of the temperature-time couples for thermal treatment of bamboo technical fibres in air
( ) and argon ( ) atmospheres. The treatment started at RT with a heating rate of 5 °C/min. These
combinations (round points) are also shown in Figures 6-9 and 6-10.
6.3.2 Thermogravimetric analysis (TGA)
This technique measures the mass change of a polymer as a function of temperature

in a controlled atmosphere (e.g. air or argon) as a function of increasing temperature
(with constant heating rate), or as a function of time (with constant temperature
and/or constant mass loss) [24]. TGA can characterize weight loss or gain due to
decomposition, oxidation or dehydratation, to determine the thermal stability and
oxidative stability of a material at temperatures up to 1000 °C. Usually the first
derivative of the TGA curve (DTG) is also calculated in order to identify the points
of greatest rate of change on the weight loss curve.
Thermogravimetric analysis (TGA) for bamboo technical fibres and PP samples

were carried out on a SDT Q600 T.A. Instruments, starting at room temperature
conditions. The experiments were carried out under air and inert atmosphere at a
flow rate of 20mL/min and a heating rate of 5 °C/min. For the TGA measurements,
samples between 11 and 16 mg were tested. For the TGA of bamboo fibres, it must
be mentioned that the fibres were conditioned at 50% HR for 72 hours before the
respective test. This is an important aspect because the moisture content of the fibre
directly depends on the environment. For the graphs and the analysis of the
information, TA Universal Analysis software (TA Instruments, USA) was used.
6.3.3 Single fibre tensile test
The specifications of the machine and the methodology followed for single fibre
tensile tests were described in section 4.4.1.3. For the determination of the technical
fibre’s cross sectional area, the samples were weighed before and after the thermal
treatment. The average diameter for all tested bamboo technical fibres was 181 ± 39
µm. For each batch at least 20 successful tests were carried out to measure the fibre
strength for each thermal treatment with a gauge length of 30 mm, and 10 single
fibre tests were conducted at a gauge length of 70 mm for the determination of the
Young’s modulus.
The fibre Young’s modulus was determined from the slope of the stress-strain curve
(between 0.1 and 0.3% of deformation) of single fibres tested at 70 mm gauge length.
With this long gauge length, it is expected that the machine compliance (e.g.
slippage at the grips) becomes negligible, giving an accurate measurement of the
elongation of the samples during the test, and thus, a reliable calculation for the
Young’s modulus. This is done for practical reasons, taking into account the
numerous tests to be performed for this study. To corroborate the reliability of this
methodology, the Young’s modulus was measured at different span lengths (5, 10,
25, 40 and 70 mm; 20 fibres for each length), and then compared with the values
obtained by Osorio et al [25] (43 GPa) for the same type of fibres. The results
showed that fibres tested at “long” span length (i.e. 70 mm) did not show a
significant difference with the mentioned reference value, see Figure 6-1.
Young’s modulus after machine compliance correction

measured by Osorio et al. [25]
5 10 25 40 70
Figure 6-1. Results from this study for the estimation of the Young’s modulus of bamboo fibres from the
stress-strain curve at different gauge lengths (5, 10, 25, 40 and 70 mm) and compared with the value found
by Osorio et al [25] after machine compliance correction (----).
146 Chapter 6b
In Osorio’s study [25], a theoretical correction for the machine compliance,

developed by Defoirdt et al [26], was applied for the same type of bamboo technical
fibres in order to determine the real elongation of the specimens. This methodology
consists of plotting the modulus versus 1/span length. The extrapolation to 1/span =
0 (infinite fibre length), provides the material modulus for which slip and machine
compliance may be ignored. Thus, when the real material modulus is known, an
estimation can be made for the machine compliance. Knowing this compliance,
strain values can be corrected and this calculation must be done for every single
experiment.
6.3.4 Scanning electron microscopy (SEM) observations
Micrographs of fibres and composites were made by Scanning Electron Microscopy

observations using a voltage between 10 and 15 kV.
6.3.5 Bamboo fibre – PP/MAPP composites
6.3.5.1 UD bamboo prepregs
For the preparation of bamboo fibre prepregs, special care was taken to accurately
align and evenly distribute the fibres in a unidirectional (UD) array. The bamboo
fibres were subsequently stabilized by mechanical clamping, as seen in Figure 6-2a.
Two layers of polymer film, polypropylene or maleic anhydride grafted
polypropylene (MAPP), were attached to the fibres on both sides using a hot iron as
seen in Figures 6-2b-c. The hot plate of the iron had a temperature of 172 °C and
was in contact with the fibres for 6 seconds on each side, enough time to pre-
impregnate the bamboo technical fibres with the polymer (Figure 6-2d). A Teflon®
sheet was placed in between the polymer and the hot plate during this operation to
avoid adhesion of the film to the iron. The average areal density of the fibres used in
the prepregs was 210 g/m2 ± 20 g/m2.
a. b. c. d.
Figure 6-2. Preparation of the UD bamboo fibre prepregs with PP and MAPP thermoplastic polymers.
6.3.5.2 Composite production
The prepregs were dried in an oven at 65°C for at least 72 h and subsequently placed
in a desiccator to avoid moisture absorption, prior to composite production.
a. b. c. d.
Figure 6-3. Preparation of bamboo fibre-PP composites: a) bamboo fibre – PP prepreg, b) preparation of the
stacking sequence of prepregs and PP films to target 45% V f, c) schematic view of compression moulding and
d), composite samples in longitudinal and transversal fibre direction ready to be tested in 3PBT.
Bamboo fibre thermoplastic composites with polypropylene (BFPP) and maleic

anhydride polypropylene (BFMA) were prepared by compression moulding (Pinette
hot press). Unidirectional bamboo prepregs (6 layers) and thermoplastic films of PP
or MAPP, where intercalated and placed into a cavity mould of 100 x 50 mm
(Figure 6-3a and b). Fibre volume fraction (Vf) was targeted at 45% by weight
measurements. Several temperatures were used for the consolidation of the
composites in air (CAI) and argon (CAR) environments, see Table 6-3.
Maximum
Atmosphere Pre-heating
consolidation
Composite during temperature (°C)
temperature (°C)
manufacturing (5 min)
(5 min)
CAI-175 Air 155 175
CAI-185 Air 165 185
CAI-200 Air 180 200
CAI-220 Air 200 220
CAI-230 Air 210 230
CAR-200 Argon 180 200
CAR-220 Argon 200 220
CAR-230 Argon 210 230
Table 6-3. Bamboo fibre – polypropylene/MAPP composites consolidated at different temperatures and
environments. The pre-heating temperature was chosen 20 °C below the consolidation temperature (see also
Figure 6-4), in order to evenly heat the mould.
The temperature profile of the consolidation experiments is shown in Figure 6-4.

The manufacturing procedure proceeded as follows:
- Starting at ambient temperature (20 °C)

- Heating rate of 5 °C/min up to the preheating temperature
- Dwell at preheating temperature for 5 minutes
148 Chapter 6b
- Heating rate of 5 °C/min up to the consolidation temperature

- Dwelling for 5 minutes
- Cooling down to room temperature
- A constant pressure of 15 bar was maintained through the whole process.
Figure 6-4. Temperature profile for the consolidation of bamboo-PP/MAPP composites. The pressure was
kept at 15 bar during the whole process
6.3.5.3 Composite production under inert atmosphere
Bamboo fibre-polypropylene composites (BFPP) were manufactured in argon inert

gas, to produce CAR composites (see Table 6-3). They followed the same
methodology for the preparation of BFPP in air, described in the previous section,
except for the final consolidation step. During the compression moulding, the inert
gas atmosphere was maintained by a bagging system, in order to maintain a
controlled atmosphere around the mould during the consolidation phase.
Thermalimide® film with a high working temperature (427 °C), supplied by Airtech
was used as bagging material. A schematic view of the set up is presented in Figure
6-5, and a general view plus details are shown in Figure 6-6.
Valves
Argon
Figure 6-5. Schematic representation of the set-up for the production of bamboo fibre-PP composites under
inert atmosphere.
A copper tube is connected on one side the bag (Figure 6-5 and 6-6c) with two
bypass valves giving the possibility to control the atmosphere inside of the bag (see
Figure 6-6c). The first valve is connected to a vacuum pump in order to evacuate the
air from the mould; for this purpose, vacuum (1 bar) was applied for at least 5
minutes to the mould. The second valve is used to regulate the inert gas flow inside
of the bag (see Figure 6-6d). The gas flow was maintained at least for 10 minutes
before the start of the compression moulding (Figure 6-6e), when the system started
to be compressed, the excess of argon was evacuated by the vacuum exit. A small
positive internal pressure of argon is maintained during all the process, controlled by
visual inspection, to avoid air into the bagging. The temperature was monitored
using a thermocouple attached directly to the mould, and a constant pressure of 15
bar was applied during the consolidation process. The temperature profiles followed
in these experiments are shown in Table 6-3.
a. b.
a. b.
Mould
Vacuum
Input Bag
inert gas
Metal
tape
c. Copper tube d. e.
Valves
Mould
Figure 6-6. Set-up for the production of bamboo BFPP under inert atmosphere. a) General view of the set-up
in the hot press, b) bagging system to contain the inert gas during the compression moulding, c) vacuum
applied to the mould to evacuate the air, d) injection of the inert gas (Ar) into the bagging system and e)
compression moulding under inert gas atmosphere.
6.3.6 Sample preparation and testing
After the production of the BFPP and BFMA composite plates (10 x 5 cm), samples
with average dimensions of 50, 14.4±0.2 and 1.85±0.1 mm, were cut using a low
speed bending saw machine (see Figure 6-7a). All specimens were kept at standard
150 Chapter 6b
room conditions (21°C ± 2°C and 50 ± 2 %RH) for at least 24 hours before testing.
Flexural three point bending tests (3PBTs) were performed according to the ASTM
D790-03 standard on a universal testing machine (Instron 4426), with a span length
of 32 mm (pure bending regime).
a. b.
c.
Figure 6-7. a) Three-point-bending test composite samples with longitudinal and transverse disposition of the
fibres; these orientations are represented schematically in b) and c) respectively.
The BFPP and BFMA samples had 0° and 90° fibre orientation (see Figure 6-7b and
c respectively), to evaluate longitudinal and transverse properties. The longitudinal
disposition was evaluated with the aim to determine the flexural properties such as
flexural strength and flexural stiffness. The transversal direction was carried out to
estimate the interface strength between the polymer and the thermoplastic matrix.
The Young’s modulus was calculated from the slope at the beginning of the stress-
strain curve, between 0.1 and 0.3% of deformation.
6.4.1 Thermogravimetric analysis (TGA)
TGA/DTG curves for bamboo fibres in air and argon environments are shown in
Figure 6-8. It is noticed that the first drop in mass loss, around 5.2%, at temperatures
around 100 °C is due to the moisture evaporation. The values found in literature for
natural fibres due to humidity release for the same range of temperature, ranged in
between 3 and 7 % [23, 27].
From 100 °C up to 193 °C there is no significant mass change. After this, the first
signs of thermal degradation started to take place with a slightly observable peak at
275 °C in air, and at 288 °C when treated in inert atmosphere. This mass loss is
attributed to the degradation of hemicellulose as reported by several authors [28, 29].
They reported that the decomposition of pure hemicellulose starts between 220 and
315 °C, with a maximum mass loss rate (0.95 wt.%/°C) at 268 °C [29]. For natural
fibres (e.g. sisal), this value started at 234 °C with a maximum at 297 °C [28]. The
hemicelluloses, are the most reactive component in the fibre (degrading at lower
temperature), due to their relatively low molecular weight [30].
Figure 6-8. TGA/DTG analysis for single bamboo fibres under different environments (air and argon).
At temperatures between 100 and 250 °C, some of the changes in physical properties
of the fibres can be explained in terms of alterations in either physical or chemical
structures such as depolymerization, hydrolysis, dehydration, decarboxylation and
recrystallization. Besides this, the formation of free radicals has been noticed [14].
They can contribute to accelerate the degradation process due to the formation of
hydroperoxide groups, responsible to a large extent for the depolymerization of
cellulose (bond scission). At higher temperatures, the cellulose thermal degradation
observed in the TGA analysis (Figure 6-8), for the bamboo technical fibres (310 °C),
is likely caused by the destruction of hydrogen bridges, the loss of water,
depolymerisation, glass transitions, changes in crystallinity, and some processes
mentioned before (i.e. the formation of free radicals, carbonyl groups, and carboxyl
groups) but at a larger scale (especially in air) [16]. The cellulose decomposition
under inert environment is shifted upwards by around 20-30 °C.
In air environment, the degradation peaks for hemicellulose (275 °C), cellulose
(310 °C) and lignin (444°C), were identified very clearly from the DTG analysis.
The same result was not found, for example for sisal fibres, where there were no
distinct peaks for each constituent, due to the presence of oxygen that accelerates the
152 Chapter 6b
degradation process, making not possible the identification of the peaks [28]. The
thermogravimetric behaviour of bamboo fibres has a direct correlation with their
chemical constituents: hemicellulose, cellulose and lignin. In fact, the TG/DTG
curves of common lignocellulosic fibres such as jute, sisal, wood and cotton display
similar aspects that could be correlated to the thermal decomposition of their main
constituents [31].
The lignin degradation is only visible for the fibres treated in air at around 444 °C.
Pure lignin was found to be the most difficult one to decompose, doing it slowly
over the whole temperature range from ambient to 900 °C [29]. Generally, the
contribution of hemicellulose and cellulose to the thermal degradation is of much
higher importance [23]. In general in literature, different peaks were found
depending on the type of environment where the analysis was carried out.
Decomposition in air was faster, more complete and proceeded at lower temperature
than in inert environment. A clear mass loss delay was observed when the fibre was
treated under inert atmosphere. This is because in air environment, the process of
decomposition of the cellulose occurs much quicker as a result of the reaction of free
radicals (R∙) with oxygen (oxidation process), during thermal degradation [32]. This
reaction produces a peroxy radical (ROO∙) which in turn removes an hydrogen atom
from another polymer molecule to form a hydroperoxide (ROOH) and so
regenerates another free radical through which the process can continue [33]. With
the presence of an inert atmosphere during the thermal treatment the reaction with
oxygen is limited to the minimum, attenuating the above reaction sequence.
In Figure 6-8, the onset temperature defined by the “shoulder” of the curve, between
100 and 190 °C remains almost invariable. But even if the mass loss is not evident in
this range, the tensile properties can be affected much earlier [28]. According to
Aziz et al [17], during this shoulder region there is still a gradual degradation,
including: depolymerisation, hydrolysis, oxidation, dehydration and decarboxylation,
mainly in the cellulose, and leading to a reduction of the fibre strength.
6.4.1.1 Isothermal TGA experiments
Isothermal TGA measurements for bamboo technical fibres were carried out over a
period of 170 minutes at different temperatures and in two different environments
(air and argon), see Figures 6-9 and 6-10. In both figures, an arrow indicates the
moment when the targeted temperature is reached and round points for air ( ) and
argon ( ) environments, show different combinations of temperature – exposure
time. These variables were used to accomplish thermal treatment on several batches
of single bamboo fibres for further tensile characterization.
160 °C
180 °C
160 °C
180 °C
200 °C
200 °C
220 °C
250 °C
220 °C
160 °C
180 °C 250 °C
200 °C
220 °C
250 °C
Figure 6-9. Isothermal TGA in air atmosphere. The arrow ( ) indicates the moment when the target
temperature is reached. The round points ( ) show different temperature-time- couples (see Table 6-2) for
further single (technical) fibre thermal characterization (see also Figure 6-11 and Table 6-4).
further single (technical) fibre thermal characterization (Figure 6-11).
180 °C
180 °C
200 °C 200 °C
220 °C 220 °C
250 °C
250 °C
180 °C
200 °C
220 °C
250 °C
Figure 6-10. Isothermal TGA in inert (argon) atmosphere. The arrow ( ) indicates the moment when the
target temperature is reached. The round points ( ) show selected different temperature-time couples (see
Table 6-2) for further single (technical) fibre thermal characterization (see also Figure 6-16 and Table 6-4).
154 Chapter 6b
The results show no apparent degradation at 160 °C at least after 140 min of
exposure (Figures 6-9 and 6-10). Above this temperature, the thermal stability of the
bamboo fibres starts to gradually decrease and becomes more critical when the
exposure time is longer as it can be seen in Figure 6-9. According to several authors
[23, 34, 35], first degradation typically occurs at temperatures above 180°C and
more severe damage due to thermal decomposition of most natural fibres occurs
within a temperature range between 215 and 310°C.
An initial mass loss due to moisture evaporation, reaching a “plateau”, at around 4-

5% was found. The total moisture content of a bamboo technical fibre at 47% RH, is
6% ± 0.5 measured by [36] in a previous study with the same type of fibres. The
moisture content of the fibres will be discussed in more detail in section 6.4.2.1. It is
clearly noticed that the mass loss in fibres exposed to an inert environment (Figure
6-10), is significantly delayed in comparison with air treatments; a difference in
mass loss of 3, 14 and 15 % can be observed after 150 minutes of exposure at 200,
220 and 250 °C respectively. These results indicate that an inert atmosphere protects
the fibre against thermal degradation.
6.4.2 Tensile properties for single fibre after thermal treatment
6.4.2.1 Fibre strength
For a better understanding of the mechanical behaviour of the composites, tensile

tests on single bamboo technical fibres were carried out after exposure at different
temperature-time couples. The results are presented in Figure 6-11 and Table 6-4;
they show that the fibre strength is evidently affected by the temperature and
decreases with increased exposure time.
This dependency has been also observed in other studies on natural fibres [16, 35],
where higher temperatures and longer exposure times result in higher decreases of
the mechanical properties. The decrease of the mechanical properties was not only
attributed to the thermal degradation. Studies on flax [37] and sisal [38] fibres at
elevated temperatures, attributed the decay in mechanical properties to differences in
coefficients of thermal expansion between cellulose, hemicellulose, lignin and
pectin. As a consequence, at elevated temperatures the mismatch in expansion
between structure forming components of the fibre creates internal stresses, resulting
in weakening of the fibre. Moreover, the thermal degradation of the fibre
components will intensify this problem.
Figure 6-11.Single bamboo technical fibre strength after thermal treatment at different temperature-time
couples in air atmosphere.
Mechanical properties
Temperature Time of Strength Young’s Strain to
(°C) exposure (min) (MPa) modulus (GPa) failure (%)
Air
Room temp. 0 733 ± 120 44 ± 2 1.8 ± 0.22
180 25 735 ± 100 39 ± 3 1.6 ± 0.16
180 60 564 ± 120 46 ± 4. 1.3 ± 0.12
180 120 462 ± 90 45 ± 2 1.1 ± 0.18
200 25 709 ± 120 45 ± 3 1.3 ± 0.31
200 50 487 ± 80 43 ± 4 0.67 ± 0.25
200 70 330 ± 50 41 ± 2 0.74 ± 0.16
200 90 312 ± 70 42 ± 4 0.50 ± 0.11
200 100 277 ± 80 41 ± 2 0.74 ± 0.10
220 25 591 ± 100 43 ± 3 1.4 ± 0.16
220 60 290 ± 70 36 ± 3 0.62 ± 0.10
220 70 226 ± 70 44 ± 3 0.52 ± 0.06
220 80 181 ± 70 37 ± 2 0.57 ± 0.18
250 50 149 ± 50 38 ± 2 0.37 ± 0.08
250 60 99 ± 60 37 ± 4 0.42 ± 0.13
Argon
200 25 680 ± 140 44 ±2 1.6 ± 0.17
200 50 591 ± 70 41 ± 2 1.1 ± 0.26
200 70 530 ± 90 45 ± 3 1.2 ± 0.25
250 50 271 ± 70 38 ± 2 0.8 ± 0.15
Table 6-4. Tensile mechanical properties for thermally treated bamboo technical fibres. The results for argon
treated fibres are also shown in Figures 6-15 and 6-16.
156 Chapter 6b
The reduction in the technical fibre strength after thermal exposure found in this
research, is in agreement with the study of Ochi et al [39], where temperatures
below 140°C do not significantly affect the mechanical properties of bamboo fibres.
However, in his study faster fibre degradation was found where at temperatures
around 160 °C, a gradual decrease of the tensile strength was observed after 30 min
of exposure, contrary to the present results. Moreover, at temperatures between
180°C and 200°C, the tensile strength dropped around 25% in the first 10 minutes
and tended to stabilize after 30 min of exposure.
Correlation between mass loss and fibre strength
In order to establish a correlation between the fibre mass loss and fibre strength after
thermal treatment, specimens for tensile testing were selected based on the
isothermal TGA results, specifically chosen from the round points ( ) shown in
Figure 6-9. Two criteria in the selection of these round points, based on literature
and previous experiences, were: a maximum 18% fibre mass loss and an exposure
time less than 120 minutes. This was done in order to establish the fibre strength
after certain time of exposure as shown in Figure 6-11, but also to determine the
fibre strength after certain mass loss.
A correlation was found between mass loss and strength properties of the technical
fibres in air atmosphere (Figure 6-12a). A distinction between mass loss due to both
moisture evaporation and thermal degradation is pointed out. The point P1 is the
reference corresponding to the fibre strength at room conditions (21°C± 2°C and
50±2 %RH), meaning zero mass loss. P2 shows fibre strength measurements after 25
minutes of thermal exposure that corresponds to the “plateau” region previously
shown in Figure 6-9. ANOVA (α= 0.05) revealed no statistically significant
differences between the average fibre strength corresponding to P1 and P2. The point
P3 marks the starting point of mass loss due to thermal degradation.
Figure 6-12b shows the P1, P2 and P3 points in an isothermal TGA in air at 200 °C.
This curve is representative for the other four temperatures analyzed with the same
technique (180, 220 and 250 °C). P2 is situated in the “plateau” region 25 minutes
after the beginning of the TGA experiment, at ~143 °C and a mass loss of around 4-
5%. At this point, the fibres have been exposed to a temperature above 100 °C
during 8.5 minutes. These conditions are often considered sufficient to release all
moisture present in the fibre, but this is not correct. To explain this, two concepts
need to be defined; “free water” and “bound water”. The first one is defined as the
“unbound” water in polymers and the second one, is the water chemically bonded
(strongly attached) to the hydroxyl groups of cellulose molecules [40].
a. P1
Mass loss due to thermal degradation
P2
P3
Mass loss due to

moisture evaporation
bb..
P2
Temperature vs time
Temperature (°C)
Mass (%)
P1
Mass loss vs time
P3
Mass loss due Mass loss due to thermal degradation

to moisture
evaporation
Figure 6-12. a) Correlation between the fibre strength and mass loss in air environment. P1 corresponds to the
fibre strength at starting room conditions. P2 refers to strength values, 25 minutes after starting the isothermal
treatment. This point corresponds to the TGA “plateau” region previously shown in Figure 6-9. P3
corresponds to the zone where the technical bamboo fibres start to lose mass due to thermal degradation (>
6% mass loss). b) Isothermal TGA at 200 °C (taken as an example for all isothermal experiments), showing
the same corresponding P1, P2 and P3 points explained above.
158 Chapter 6b
The plateau reached by the mass loss in Figures 6-9 and 6-12b in P2 is a consequence
of the evaporation of the “free water”. After a standard drying procedure, it is
expected that a natural fibre still has constitutional tightly bound water remaining in
its structure [37]. For the same type of bamboo technical fibres, the moisture content
was reported as 6% ± 0.5 [36] at 47% RH. At this point, it is expected that the free
water is released because even at prolonged heat exposure (100 °C for 16 h), no
further mass loss was observed. In Figure 6-12b, the mentioned 6% of mass loss is
reached at around 200 °C (where the fibres are already ~20 min above 100 °C).
After this, the mass loss started a quasi-linear decrease. Fibre thermal treatments of
manufacturing processes that gives more than ~7% of mass loss, start to
significantly affect the mechanical properties of the fibres (~33% of fibre strength
reduction), and hence the performance of the composites.
Figure 6-12a, shows a strong influence of the mass loss of the fibres due to thermal
degradation, on their strength properties. A simple quadratic curve fit (Equation 6-1)
describes this dependency with reasonable correlation (R2= 88 %), providing an
empirical model to estimate the fibre strength at certain mass loss of dry bamboo
technical fibres.
y= 3.1x2 – 106 x +1056 (6-1)
6.4.2.2 Fibre stiffness and strain to failure
A slight decrease in Young’s modulus of the technical fibre is observed after

thermal treatment, see Figure 6-13. The values ranged in between 36 and 45 GPa,
showing a rather constant tendency for all temperature-time couple combinations,
and remaining stable for low temperature treatments. Similar results are reported by
Ochi et al [39], where the Young’s modulus for the bamboo fibres was found to be
almost independent of the applied thermal treatment. The TGA results (Figure 6-9)
and the Young’s modulus measured on thermally treated fibres at specific
temperature-time couples in air (Table 6-2) were used to establish a correlation
between mass loss and the stiffness of the fibre (Figure 6-14). This relationship can
be described with a regression of the obtained experimental data, by a simple
quadratic equation (Equation 6-2) with reasonable correlation of the effect of the
thermal treatment (R2 = 0.81).
y= 0,0408x2 – 1.58x + 53.22 (6-2)

Figure 6-13. Young’s modulus for single bamboo technical fibres at different time-temperature couples in air
environment.
Mass loss due to

moisture evaporation Mass loss due to thermal degradation
Figure 6-14. Correlation between Young’s modulus and mass loss in single bamboo technical fibres in air.
This relationship given by equation 6-2 (R2 = 0.81), becomes also an empirical model to describe the stiffness
in function of the mass loss for technical bamboo fibres.
160 Chapter 6b
Thermal treatment shows a significant effect on the strain to failure of the bamboo
technical fibre. Exposure times of 50 minutes or more at temperatures of 200 °C or
higher, show a dramatic drop in failure strain to less than 50% of the original value,
see Table 6-4. Changes taking place in molecular structure of the fibre due to
decomposition of the long polymers cause a decrease in elasticity of the fibre and
the fibre becomes more fragile [30]. Furthermore, Gassan and Bledzky [14] reported
a significant reduction of the tenacity in natural fibres when treated at low
temperatures (i.e. 150 °C) and longer exposure times (up to 4h). Other studies report
up to a factor or 3 in reduction of the failure strain after thermal treatment [15]. The
strain at maximum strength vs mass loss is shown in Figure 6-15, showing the high
sensitivity of this property due to a thermal treatment.
Mass loss due to Mass loss due to thermal degradation

moisture evaporation
Figure 6-15. Correlation between strain at maximum strength and mass loss in single bamboo technical fibres
after thermal treatment in air (R2=0.68).
6.4.2.3 Influence of an inert atmosphere on the mechanical properties of

thermally treated bamboo technical fibres
A clear positive effect on fibre strength is observed when the thermal treatment is
carried out under inert atmosphere, showing less fibre strength reduction amounting
up to 37% and 45% for treatments at 200 °C for 70 minutes and 250 °C for 50
minutes respectively (see Figure 6-16). Higher thermal resistance in flax [15], jute [3]
and hemp [22] technical fibres, had been found when treating these fibres under
inert atmosphere.
Fibre stiffness did not show a significant response when thermally treating technical
bamboo fibres at 200 °C under different environments (i.e. air and argon) and
several times of exposure (25, 50 and 75 min), see Figure 6-17. For the same
treatment combinations, the strain at maximum strength showed less reduction under
inert atmosphere (Figure 6-17).
18%
37%
45%
Figure 6-16. Fibre strength comparison for thermally treated bamboo technical fibres in air and argon
atmospheres at different temperatures.
Figure 6-17. Young’s modulus comparison for thermally treated bamboo technical fibres in air and argon
atmospheres at 200°C.
162 Chapter 6b
6.4.2.4 Thermally treated bamboo fibres under SEM
SEM micrographs were made of the fibre fracture surface after tensile testing. For
untreated fibres, mainly three types of fracture were identified: (a) a straight crack
right through the elementary fibres, (b) a crack proceeding along the primary layer
that surrounds the elementary fibres and (c), a combination of the two. There was
not found a regular pattern of failure types (see also section 4.3.1 for more
information about the type of failure in individual untreated fibres).
For thermally treated fibres, two main characteristics were observed in the type of
fracture. The first one corresponds to a significant structural damage of the technical
fibre with a limited presence of pull-out of elementary fibres. This is largely
observed for treatments up to 200 °C and exposure times between 30 and 50 minutes
(see Figure 6-18a). This phenomenon can be attributed to the degradation of the
hemicellulose which tends to degrade first at moderate temperatures. The second
characteristic corresponds to a relatively clean fracture of the fibres, caused by the
damage on the structure and resulting in a higher brittleness of the fibres. This is
observed mostly for severe thermal treatments (i.e. 220 and 250 °C for exposure
times longer than 60 minutes), as shown in Figure 6-18b.
In parallel to these phenomena, a smoothening of the fibre surface was observed for
some of the fibres (see Figure 6-18c). The more regular surface can be attributed to
the softening of the lignin layer considered to be a thermoplastic polymer [22, 41],
exhibiting a glass transition temperature of around 90° and a melting temperature of
around 170 -200 °C [42, 43]. Studies on bamboo technical fibres [44], indicated that
the surface is homogeneously covered with a lignin layer instead of cellulose.
Furthermore, AFM observations reported by Fuentes et al. [45], show that autoclave
treatment (3 bar at 150 °C) applied to bamboo technical fibres, reduced the surface
irregularities by smoothening of the lignin layer present on the fibres.
XPS analyses were carried out on thermally treated fibres in presence of argon and
air. The results indicated that the thermal treatment of technical bamboo fibres in
any of the environments did not change the surface chemical components of the
fibres, in comparison with the untreated samples (see Appendix 2). Also, after SEM
observations for the different batches, there was no evidence for any of the time-
temperature couples including air and inert atmospheres, for splitting or visible
damage to the fibre surface as a consequence of exposure to the high temperatures.
These two aspects where reported by Prasad et al. [22], where the splitting of hemp
fibres was present in both air and nitrogen environment at 220 °C, but fibre surface
damage happened only in air.
a.
b.
c.
Figure 6-18. SEM images of the failure modes after tensile testing of thermally treated bamboo technical
fibres showing a) structural damage of the bamboo technical fibre with a limited presence of pull-out, b) a
relatively clear fracture and c) a smoothening of the fibre surface observed for some of the fibres.
164 Chapter 6b
6.4.3 Bamboo fibre polypropylene composites (BFPP)
6.4.3.1 Fibre degradation during composite processing
The fibre mass loss after following the same temperature profile used to
manufacture the BFPP composites (CAI and CAR profiles) was characterized by
TGA. Three different manufacturing patterns were analyzed in air (CAI-200, CAI-
220 and CAI-230) and argon (CAR-200, CAR- 220 and CAR-230) atmospheres.
Figure 6-19 shows a TGA profile for the CAI-200 composite, following all the steps
that were applied for the manufacturing of the BFPP (i.e. starting temperature
conditions, heating rate of 5 °C/min, dwell and maximum consolidation
temperatures). In this analysis, the total mass loss of the fibre at the end of the
process was 6.7% in air conditions and 5.8% when the treatment was performed in
inert atmosphere. The same difference in mass loss (~1%) between these two
environments was observed for the other two studied patterns (CAI/CAR-220 and
CAI/CAR-230), as can be seen in Table 6-5. The mass loss measured by the TGA
for technical bamboo fibres after the corresponding manufacturing process (as it was
done for CAI/CAR-200 in Figure 6-19), was used in Equations 6-1 and 6-2, to
obtain the estimated fibre strength and stiffness (see Table 6-6). These values were
used to recalculate the properties of the BFPP and will be shown as the “theoretical
composite strength” in section 6.4.3.4.
The estimated fibre strength after thermal treatments CAR-200, CAR-220 and CAR-
230 (Table 6-5), was calculated to be respectively 12, 17 and 18% higher than for
fibres treated in air environment with the same temperature profile (CAI’s). This
improvement, theoretically, will give an improvement in strength properties in the
final composites between 10 and 13%.
These results suggest not only that it is advantageous to use an inert atmosphere for
the manufacturing of BFPP, but also reveal the magnitude of the fibre strength
reduction, as a consequence of the composite manufacturing. It is common to
consider the original fibre strength in virgin state as a reference for the estimation of
the final composite properties. In the case of untreated bamboo technical fibres, this
value is around 733 MPa, and the theoretical strength of the unidirectional BFPP
(with the rule of mixtures) will be ~360 MPa (45% Vf). However, the theoretical
composite strength taking into account the strength reduction of the fibre after
processing will be evidently lower (e.g. almost a reduction by half for the composite
strength for CAI-230).
Figure 6-19. TGA analysis for bamboo fibres carried out in air and argon, following the same temperature
profile for the manufacturing of the composite (CAI-200 and CAR-200).
Estimated properties after

thermal treatment
Maximum Fibre
Process Mass Fibre
processing Young’s
Composite duration loss** strength***
temperature modulus***
(min)* (%) (MPa)
(°C) (GPa)
Air
CAI-200 200 45.2 6.7 485 44
CAI-220 220 49.6 9.0 351 42
CAI-230 230 51.1 10.9 269 41
Inert (argon)
CAR-200 200 45.2 5.8 545 44
CAR-220 220 49.6 7.9 412 43
CAR-230 230 51.1 9.7 319 41
Table 6-5. Fibre mass loss, and estimated technical fibre properties after different thermal treatments
following the same temperature profiles used in the manufacturing of BFPP composites. The theoretical
composite strength according to the estimated fibre strength for each manufacturing process is also shown.
*Starting at room conditions. ** Estimated after TGA measurements following the composite temperature
profile. ***Estimated properties obtained by Equations 6-1 and 6-2 respectively.
166 Chapter 6b
6.4.3.2 Matrix degradation
The matrix is also exposed to thermal degradation during composite manufacturing.

Figure 6-20 shows the TGA analysis for the polypropylene used in this study in air
and argon environments. According to this test, the onset degradation temperature in
air is taking place at around 250 °C while in inert atmosphere (argon) it is around
293 °C. This result shows that the maximum consolidation temperature used in this
research (230 °C, see Table 6-3), is low enough not to expect a drop in properties of
the PP matrix.
These results are in agreement with similar studies [3], where it was also found that
the degradation of polypropylene occurs about 65 °C lower in air atmosphere than in
inert atmosphere. It was found that in argon, PP starts to degrade at around 300°C
with a maximum weight loss at 431 °C. This degradation is initiated primarily by
thermal scissions of C–C chain bonds accompanied by a transfer of hydrogen at the
site of scission [46]. In air environment, the thermal stability of PP is significantly
reduced by oxidative dehydrogenation accompanied by hydrogen abstraction [47].
6.4.3.3 TGA analysis for bamboo fibre-polypropylene composites
A TGA comparison between bamboo fibres, neat PP and BFPP is shown in Figure
6-20. The BFPP’s present better performance than the individual constituents (i.e.
neat PP and bamboo fibres), especially in air. Nevertheless, it was noticed that both
single technical fibres and BFPP, have equal starting mass loss at 193 °C as can be
seen in the detail in Figure 6-20. The main contribution of the initial mass loss in the
composite is caused by the fibre, nevertheless the degradation appears be slower in
the composite, meaning that the fibres are protected by the polymer.
The overall better performance of BFPP in comparison with the neat PP and the
single fibre is in accordance with the results of Doan et al. [3], attributing the better
thermal behaviour of the composite to the interfacial adhesion between fibre and
matrix. According to the results obtained by Martin et al. [28] on DSC and TGA
analyses in sisal fibres, the maximum temperatures to produce composites
reinforced by sisal fibre should be under 185 °C due to the degradation of
hemicellulose and cellulose components, that starts around 180 °C; which is lower
than what was observed in bamboo fibres. The starting temperature degradation for
the hemicellulose in bamboo technical fibres was found to be around 200 °C in
combination with a mass loss around 6% based on the TGA (section 6.4.1, Figure 6-
8). Martin et al [1] also reported a significant decrease in the strength properties after
6% mass loss as it was shown in this study in Figures 6-12a and b. Regarding the
degradation of the fibre in terms of mechanical properties, the results indicate that it
is more relevant to consider the mass loss instead of the temperature and time of
exposure, to assess the fibre strength degradation and hence the properties of the
composite.
Initial DTG curves for single

bamboo fibre and BFPP
samples
Figure 6-20. TGA for technical bamboo fibres, neat PP and BFPP in air and argon environments; inset shows
DTG detail for single bamboo fibre and BFPP, showing similar onset of degradation at 193°C.
Influence of the fibre volume fraction on the thermal degradation of the BFPP
When the matrix and reinforcement are analyzed with the TGA technique (Figure 6-
20), it is observed that the thermal stability of the neat PP in air is considerably
lower than for the bamboo fibres (this is again assigned to the degradation of PP by
thermal scission of C–C chain bonds [3]), and the opposite is found when they are
analyzed under inert atmosphere. In a similar study, the neat PP had a higher thermal
resistance in comparison with jute fibres in nitrogen [48]. As a consequence, the
overall thermal resistance of PP-jute composites decreased with increasing fibre
content in nitrogen atmosphere, but increased in air environment due to the lower
thermal stability of PP compared to jute. George et al. [49] also reported a decrease
of the thermal stability of the whole composite when increasing the fibre volume
fraction of jute fibre in the system. Even though experiments with different fibre
168 Chapter 6b
volume fractions were not performed in this study, a similar situation can be
expected for the BFPP’s in inert environment.
The degradation of the individual components, i.e. polymer and fibres, characterized
by TGA, can help to understand the degradation mechanism of the entire natural
fibre composite. Its threshold decomposition temperature indicates the limit for the
composite manufacturing temperature, but might be accompanied by other
complementary techniques such as isothermal TGA and thermal single fibre testing
to have a clearer picture of the degradation phenomena.
6.4.3.4 Flexural strength of bamboo fibre-polypropylene composites
The flexural strength of bamboo fibre-polypropylene composites (BFPP) and their

correspondent efficiency factors at different consolidation temperature profiles are
shown in Table 6-6 and Figure 6-21. The flexural strength of UD bamboo fibre
composites is affected by two factors. First, the fibre strength is reduced by thermal
degradation due to the exposure to relatively high temperatures during the
manufacturing process (see section 6.4.2). Second, the effective fibre/matrix
adhesion determines the ability to transfer the load between the two constituents.
Maximum Experimental Strain at Theoretical

Efficiency
processing flexural maximum composite
Composite factor*
temperature strength strength strength
(%)
(°C) (MPa) (%) (MPa)**
Air
CAI-175 175 178 ± 15 2.4 ± 0.15 55 360
CAI-185 185 192 ± 8 1.9 ± 0.12 53 345
CAI-200 200 169 ± 2 2,1 ± 0,08 67 251
CAI-220 220 148 ± 3 1,1 ± 0,03 78 190
CAI-230 230 90 ± 15 0,8 ± 0,2 58 154
Argon
CAR-200 200 173 ± 5 2,3 ± 0,30 61 278
CAR-220 220 157 ± 4 1,5 ± 0,08 72 218
CAR-230 230 126 ± 9 0,9 ± 0,06 71 177
Table 6-6. Flexural properties of BFPP consolidated under different environments. *The efficiency factors
and the theoretical composite strength were calculated with the correspondent fibre strength after thermal
treatment according to Table 6-5, see also Figure 6-26. ** Calculated with the ROM, considering the
properties of the fibre after thermal degradation due to the manufacturing process and a Vf = 45%.
Figure 6-21. Longitudinal flexural strength for BFPP consolidated in air and argon at different temperatures,
the values were normalized at Vf= 45%. The theoretical composite strength (obtained with the ROM), was
calculated with the correspondent fibre strength after thermal treatment (according to Table 6-5) for air and
argon treatments. The efficiency factors are shown in Figure 6-26.
Experimental flexural strength results for BFPP show no statistically significant

difference according to the Student’s t-test (α= 0.05), between CAI-175 and CAI-
200 and between CAI-175 and CAI-185. This shows an unclear tendency between
these relatively low consolidation temperatures. Furthermore, they exhibited a
continuous deformation (plateau), without a clear fracture when tested in 3PBT (see
CAI-200 stress-strain curve in Figure 6-22). This continuous deformation is an
indication that either the processing temperature is not enough to melt and properly
consolidate the layers of fibres and polypropylene, creating bad impregnation (see
e.g. fracture surface of CAI-185 in Figure 6-23), or that the fibre-matrix interface is
weak. The bad or weak fibre/matrix interface generates an extensive pull-out during
the high strain of the sample. The higher the strain, the higher is the pull-out length
[50]. For higher consolidation temperatures, e.g. CAI-220, the matrix flows better
around and between the fibres and the composites showed a clearer fracture surface
(Figure 6-22). The reduced fibre strength will also play a role in this behaviour.
The good thermal resistance for single fibres under inert environment is effectively
translated to the flexural strength of the composites, where an improvement of 6 and
28% at 220 and 230 °C respectively is obtained in comparison with the composites
produced in air (Table 6-6). This shows that the composite production under inert
170 Chapter 6b
gas is protecting the fibre delaying the reduction in strength properties. The
composites are becoming more brittle with increasing processing temperature with a
better strain to failure behaviour in argon environment (e.g. ~25% more when
comparing CAI and CAR at 220 °C. In the SEM images, shown in Figure 6-24, it
can be observed that the pull-out length of the BFPP’s decreases with increasing
temperature, which is a sign of embrittlement and shift of interface failure to fibre
failure.
CAI-200
CAI-220
Figure 6-22. Typical flexural stress-strain curve for CAI-200 and CAI-220.
Figure 6-23. SEM images of fractured BFPP samples after consolidation in air at 185 °C (CAI-185) showing
bad impregnation and large scale pull-out.
temperature: 200 °C Air environement Inert environment

Max. Consolidation
temperature: 220 °C
Max. Consolidation
temperature: 230 °C
Max. Consolidation
Figure 6-24. SEM images of fractured BFPP samples after consolidation at different temperatures: a)
CAI/CAR-200, b) CAI/CAR-220 and c) CAI/CAR-230 after processing in air and inert conditions.
Figure 6-25 shows when comparing the strength - strain-to-failure results for all
produced BFPP’s, that there is a narrow process window to obtain a good balance
between good impregnation and low fibre degradation, where the inert atmosphere
can enlarge the window to obtain better properties. This is in agreement with what is
found in literature (based in TGA analysis), regarding the benefit of using inert
atmosphere with natural fibres to obtain better mechanical properties. Wielage et al
[35] stated that in air atmosphere, thermal decomposition of technical flax fibres
started significantly earlier than in inert atmosphere. The results, show that the use
172 Chapter 6b
of inert gas increases the resistance of the fibres to a relatively high temperature.
This increases the possibility to use higher processing temperatures or longer cycle
times, opening new alternatives for thermoplastic matrices with better properties,
such as PVDF and PA6 with shaping temperatures of 190-230 and 240 °C
respectively.
Figure 6-2xx. Stress at maximum strain of BFPP consolidated in air and argon atmospheres.
Figure 6-25. Flexural strength vs strain at maximum strength of BFPP at different consolidation temperatures
in air and argon environments. 1= full impregnation, low fibre pull-out and high fibre degradation. 2= bad
impregnation and high fibre pull-out and 3= partial impregnation and still presence of fibre pull-out.
Finally, as it was demonstrated, the fibre damage after thermal exposure will vary
according to the severity of the thermal treatment. For this reason, an estimation of
realistic theoretical composite properties and efficiency factors (EF) will depend on
the fibre strength values that are chosen. Figure 6-26 shows significant differences
between the EF’s calculated with the fibre strength of an untreated fibre (730 MPa);
in comparison with EF’s obtained with the correspondent fibre strength after the
manufacturing process (Table 6-5), for CAI and CAR samples.
Figure 6-26. Efficiency factors calculated with two different values, the first one with the original fibre
strength of the bamboo technical fibre (~730 MPa) and the second one with the fibre strength after each
manufacturing process (see Table 6-6), in air and argon atmospheres.
In the first case (Figure 6-21), when EF is calculated with the strength of an
untreated fibre, the values present a constant decrease (Figure 6-26) because of the
fibre strength reduction due to the thermal exposure. For the CAR samples, the EF’s
are higher than the CAI because the fibre is protected by the argon as show in Table
6-5. For the second case, when the EF is calculated according to the fibre strength
corresponding to each manufacturing process, is possible to observe a peak value for
CAI-220. This EF improvement can be explained by the fact that there is a balance
between moderate fibre degradation and good wetting, for CAI-230 samples, the
fibre degradation is too high and the EF decreased. When the process is carried out
in an inert atmosphere, the EF’s present an improvement for the CAR-220/230. In
this case, there is good wetting, due to the relative high temperatures, and at the
same time the strength of the fibre is better conserved (reaching values close to
80%). These results proves that the loss in properties of the fibre using high-
temperature manufacturing have to be taken into account, because it points out the
importance of an efficient fibre/matrix consolidation manufacturing process as the
heat exposure can affect the intrinsic properties of the fibres.
174 Chapter 6b
6.4.3.5 Bamboo fibre-maleic anhydride polypropylene composites (BFMA)
BFMA samples in longitudinal direction CAI-175, CAI-185 and CAI-200 had a

maximum flexural strength of 163, 179 and 152 MPa respectively. These values did
not present an improvement in properties in comparison with the BFPP consolidated
at the same temperatures shown in Table 6-6. Although an improvement in interface
strength was found (see Table 6-7 further-on), this shows that the behaviour in the
UD composite is still largely determined by fibre strength. Although MAPP has a
better adhesion to natural fibres, it is also a weaker polymer than PP. In all, flexural
composite strength is not affected much.
Even if epoxy resins give good efficiency factors for bamboo fibre composites of 82
and 95 % for strength and stiffness respectively, the BFPP composite strength for PP
and MAPP remains relatively low. This reflects a combination of insufficient bond
strength and weak polymer mechanical properties. In Figure 6-27 flexural properties
of bamboo fibre thermoplastic composites are compared with similar natural fibre
composites. This graph is indicative because fibre volume fractions vary strongly.
Chattopadhyay et al. [7] reported a general increase of 81% in the flexural strength,
150% in the flexural modulus, 105% in tensile strength and 191% in tensile modulus
for short bamboo fibres (1-6 mm) reinforced MAPP compared to the PP benchmark.
Also Chen et al [8], demonstrated that a content that the MA in bamboo PP
composites (short fibre), resulted in higher tensile strength. This positive effect is
attributed to the stronger interaction between the fibre and matrix due to the reaction
between acid groups of the maleic anhydride groups and hydrophilic groups on the
fibre surface [51]. In the case of flax, good results have been obtained with MAPP
[52]. It is clear that compatibility is the key in these systems. For more data on
flexural and tensile properties of natural fibre thermoplastic composites, see Tables
3-5 and 3-6.
When the unidirectional composite is tested in transverse direction, the fibre-matrix

interface will often dominate the final composite properties, and the interface quality
of the composite can be characterized. Transverse properties of BFMA are shown in
Table 6-7. From these results a small improvement in interface strength for BFMA
is apparent, but transverse strength will be limited due to the relatively low strength
of MAPP (see Table 6-1). Furthermore, there are indications that the surface of the
current studied bamboo fibres is covered with lignin [45], reducing its capability to
form covalent bonds with MAPP. Similar results on transverse 3 point bending
values for BFPP and BFMA of 15 and 18 MPa respectively, are reported in
literature [45].
Figure 6-27. Flexural properties for unidirectional natural fibre/thermoplastic composites [52-54]. 1,2Results
from this study.
Matrix Transverse flexural strength (MPa)
175 °C 185 °C 200 °C*

Polypropylene 18 ± 2.2 17 ± 1 15 ± 0.6
MAPP 18 ± 0.8 20 ± 0.3 18 ± 0.7
Table 6-7. Results for transverse flexural strength of BFPP and BFMA. *Values from [45]. The
manufacturing procedure was the same as for CAI-175, CAI-185 and CAI-200 explained in section 6.3.5.1.
The low improvement of the MAPP on the mechanical properties in the present
study can be attributed mainly to several reasons. First, the adhesion to bamboo fibre
may not be optimal yet. It has been stated that the MAPP increased the reaction
between the anhydride functions and the hydroxyl functions on cellulose, and as a
result, the mechanical properties of the composites are enhanced [8, 52]. The surface
constituents for the analyzed bamboo technical fibre are close to the reference
materials for lignin (see Figure A2.2, Appendix 2). This indicates that the technical
fibre is still cover by a layer of lignin and possibly other molecules but not cellulose
after mechanical extraction reducing its capability to form covalent bonds with
MAPP. This finding is in agreement with the studies of Fuentes et al. [45], in the
same bamboo fibre samples used in this study. Secondly, in this study the MAPP
employed, with relatively poor mechanical properties as compared to examples in
literature [55], MA was added in a low percentage as compatibilizer. A third reason
176 Chapter 6b
is that possibly not enough temperature was applied to the BFMA for the activation
of the copolymer by heating (170 °C) a is stated by Bledsky et al. [56]. Further
experiments varying the content of maleic anhydride and with higher consolidation
temperatures will be necessary in order to explore the possibility to obtain the
maximum performance of the material.
6.5 Conclusions
The aim of this study was to characterize the loss in mechanical properties of
individual technical fibres and BFPP and BFMA due to thermal degradation. The
fibres were exposed at different temperature-time couples and the composites were
manufactured at different maximum consolidation temperatures, both under two
environments i.e. air and argon.
TGA results show that technical bamboo fibres are not affected at temperatures
below 160 °C but show a significant weight loss upon exposure to temperatures
above 200 °C. The mass loss in the fibre was characterized mainly by two stages,
the first one, common to all treatments, corresponded to a rapid decrease of the mass
(~4-5 %), due to moisture evaporation present in the structure of the fibre and
followed by a plateau. In a second stage, a mass loss above ~6% is due to thermal
degradation of the fibre and becoming more severe according to the temperature and
the time of exposure.
The decrease of the technical fibre strength in air environment was in agreement
with what is found in other studies and the strength decreased according to the
severity of the treatment (temperature and time of exposure). The Young’s modulus
was not significantly influenced by the different thermal treatments applied in this
study. The results show that the inert atmosphere had a positive effect on the tensile
strength for individual fibres; this is observed especially in a temperature range
between 220 and 250 °C with improvements of 37% and 45% respectively in
comparison with the air treated samples.
Through the results it was possible to establish the maximum time and/or
temperature allowed during the processing of bamboo fibre thermoplastic
composites to avoid excessive loss in properties. Even if the thermal degradation in
fibres is a complex process, in this study it was possible to link the composite
properties after a specific manufacturing process, with the strength and the Young’s
modulus of the fibre at the end of that process. For this, an empirical model based on
the correlation between single fibre experimental results and TGA measurements
was developed allowing estimating the fibre properties as a function of the mass loss
(above 200 °C) after thermal treatment or after a composite manufacturing process.
This procedure can be also applied to other natural fibres, and is described as
follows:
- First, it is necessary to obtain the full correlation between mass loss and fibre
strength and modulus. For this, the procedure described in paragraphs 6.3.1,
6.4.2.1 and 6.4.2.2 can be followed, yielding equations like 6-1 and 6-2;
- Define the desired manufacturing temperature and time profile for the production
of the composite (including heating rate(s) and dwell time(s));
- Measure the mass loss of the fibres by TGA following the same steps of the
manufacturing profile defined above. The initial conditions of the fibres (e.g.
moisture content) should be the same as the ones for the composite production;
- The fibre strength or fibre stiffness after the fibre treatment or manufacturing
process can be found by filling in the mass loss (x value) in Equations 6-1 and 6-
2.
With this novel approach it is proposed to rather consider the link between fibre
properties and fibre mass loss as a key phenomenon, instead of the traditional
connection between the fibre properties and the time-temperature variables, as
usually employed in literature. Most of the encountered studies on the thermal
degradation of natural fibres in literature are exclusively based on TGA-DSC
analysis, excluding the measurement of the fibre properties. The proposed procedure
is applicable for other natural fibres, taking into account that the structure and
chemical composition between them are quite similar. But, all the corresponding
measurements such as isothermal TGA, and fibre strength measurements, should be
performed for each particular fibre.
Unidirectional BFPP consolidated at different temperatures were characterized by 3

point bending tests. To reach the potential of bamboo fibre thermoplastic composites,
further work is needed to improve the compatibility between fibre and matrix.
Experiments carried out with MAPP did not show improvement in properties in
comparison with PP matrix. This is attributed to a lack of interaction caused by a
lignin layer that is surrounding the fibres reducing its capability to form covalent
bonds and the low mechanical properties of the MAPP. Other contributing factors
may be the low content of MA and the need of a higher temperature for the
activation of the copolymer.
178 Chapter 6b
The flexural strength for bamboo fibre/ polypropylene composites consolidated

under inert atmosphere indicated a significant improvement (30%) on the flexural
strength, especially for relatively high processing temperatures, in comparison with
production in a normal air environment. This process can be easily adopted for the
processing of natural fibres and thermoplastic matrices; nevertheless it needs to be
combined with an optimized manufacturing process in order to avoid thermal over-
exposure of the fibres.
References
[1] Saheb DN, Jog J. Natural fiber polymer composites: a review. Adv Polym Technol.
1999;18(4):351-63.
[2] Sena Neto AR, Araujo MAM, Souza FVD, Mattoso LHC, Marconcini JM. Characterization and
comparative evaluation of thermal, structural, chemical, mechanical and morphological properties
of six pineapple leaf fiber varieties for use in composites. Industrial Crops and Products.
2013;43(0):529-37.
[3] Doan T, Brodowsky H, Mader E. Jute fibre/polypropylene composites II. Thermal,
hydrothermal and dynamic mechanical behaviour. Compos Sci Technol. 2007;67(13):2707-14.
[4] Yao F, Wu Q, Lei Y, Guo W, Xu Y. Thermal decomposition kinetics of natural fibers:
Activation energy with dynamic thermogravimetric analysis. Polym Degradation Stab.
2008;93(1):90-8.
[5] Bledzki A, Gassan J. Composites reinforced with cellulose based fibres. Prog Polym Sci.
1999;24(2):221-74.
[6] Mohanty S, Nayak S, Verma S, Tripathy S. Effect of MAPP as a coupling agent on the
performance of jute–PP composites. Reinforced Plastics and Composites. 2004;23(6):625-37.
[7] Chattopadhyay SK, Khandal RK, Uppaluri R, Ghoshal AK. Bamboo fiber reinforced
polypropylene composites and their mechanical, thermal, and morphological properties. J Appl
Polym Sci.119(3):1619-26.
[8] Chen X, Guo Q, Mi Y. Bamboo fiber-reinforced polypropylene composites: A study of the
mechanical properties. J Appl Polym Sci. 1998;69(10):1891-9.
[9] Beckermann GW, Pickering KL. Engineering and evaluation of hemp fibre reinforced
polypropylene composites: Fibre treatment and matrix modification. Compos Part A: Appl Sci
Manuf. 2008;39(6):979-88.
[10] Pracella M, Chionna D, Anguillesi I, Kulinski Z, Piorkowska E. Functionalization,
compatibilization and properties of polypropylene composites with Hemp fibres. Compos Sci
Technol. 2006;66(13):2218-30.
[11] Elkhaoulani A, Arrakhiz FZ, Benmoussa K, Bouhfid R, Qaiss A. Mechanical and thermal
properties of polymer composite based on natural fibers: Moroccan hemp fibers/polypropylene.
Materials & Design. 2013;49(0):203-8.
[12] Joseph PV, Rabello MS, Mattoso LHC, Joseph K, Thomas S. Environmental effects on the
degradation behaviour of sisal fibre reinforced polypropylene composites. Compos Sci Technol.
2002;62(10-11):1357-72.
[13] Gañan P, Garbizu S, Llano-Ponte R, Mondragon I. Surface modification of sisal fibers: Effects
on the mechanical and thermal properties of their epoxy composites. Polym Compos.
2005;26(2):121-7.
[14] Gassan J, Bledzki AK. Thermal degradation of flax and jute fibers. J Appl Polym Sci.
2001;82(6):1417-22.
[15] Van de Velde K, Baetens E. Thermal and Mechanical Properties of Flax Fibres as Potential
Composite Reinforcement. Macromolecular Mater Eng. 2001;286(6):342-9.
[16] Van De Velde K, Kiekens P. Thermal degradation of flax: The determination of kinetic
parameters with thermogravimetric analysis. J Appl Polym Sci. 2002;83(12):2634-43.
Sci Technol. 2004;64(9):1219-30.
[18] Sgriccia N, Hawley MC. Thermal, morphological, and electrical characterization of microwave
processed natural fiber composites. Compos Sci Technol. 2007;67(9):1986-91.
[19] Ray D, Sarkar BK, Basak RK, Rana AK. Study of the thermal behavior of alkali-treated jute
fibers. J Appl Polym Sci. 2002;85(12):2594-9.
[20] Lee S, Wang S. Biodegradable polymer/bamboo fiber biocomposite with bio-based coupling
agent. Composites Part A. 2006;37:80-91.
[21] Shih Y-F. Mechanical and thermal properties of waste water bamboo husk fiber reinforced
epoxy composites. Materials Science and Engineering: A. 2007;445-446(0):289-95.
[22] Prasad B, Sain M. Mechanical properties of thermally treated hemp fibers in inert atmosphere
for potential composite reinforcement. Mater Res Innov. 2003;7(4):231-8.
[23] Rachini A, Le Troedec M, Peyratout C, Smith A. Comparison of the thermal degradation of
natural, alkali-treated and silane-treated hemp fibers under air and an inert atmosphere. J Appl
Polym Sci. 2009;112(1):226-34.
[24] Prime R, Bair H, Vyazovkin S, Gallaher P, Riga A. Thermogravimetric analysis (TGA). In:
Menczel J, Prime R, editors. Thermal analysis of polymers: fundamentals and applications: John
Wiley & sons, Inc; 2009.
2010;41(5):588-95.
[27] Ouajai S, Shanks RA. Composition, structure and thermal degradation of hemp cellulose after
chemical treatments. Polym Degradation Stab. 2005;89(2):327-35.
[28] Martin A, Martins M, da Silva O, Mattoso L. Studies on the thermal properties of sisal fiber
and its constituents. Thermochim Acta. 2010;506:14-9.
[29] Yang H, Yan R, Chen H, Lee DH, Zheng C. Characteristics of hemicellulose, cellulose and
lignin pyrolysis. Fuel. 2007;86:1781-8.
[30] Kocaefe D, Poncsak S, Boluk Y. Effec of thermal treatment of the chemical composition and
mechanical properties of birch and aspen. BioResources. 2008;3(2):517-37.
[31] Monteiro S, Calado V, Rodriguez R, Margem F. Thermogravimetric behavior of natural fibers
reinforced polymer composites: An overview. Materials Science and Engineering: A.
2012;557(0):17-28.
[32] Shafizadeh F, Bradbury AGW. Thermal degradation of cellulose in air and nitrogen at low
temperatures. J Appl Polym Sci. 1979;23(5):1431-42.
[33] Peterson JD, Vyazovkin S, Wight CA. Kinetics of the thermal and thermo-oxidative
degradation of polystyrene, polyethylene and poly (propylene). Macromol Chem Phys.
2001;202(6):775-84.
2009;40(8):1111-8.
[35] Wielage B, Lampke T, Marx G, Nestler K, Starke D. Thermogravimetric and differential
scanning calorimetric analysis of natural fibres and polypropylene. Thermochim Acta.
1999;337:169-77.
[36] Verheyden S. Karakterisering van vochtabsorptie van bamboe vezel composiet [Master thesis].
Leuven: University of Leuven, Faculteit Ingenieurswetenchappen, 80 p.; 2010.
180 Chapter 6b
[37] Hornsby PR, Hinrichsen E, Tarverdi K. Preparation and properties of polypropylene

composites reinforced with wheat and flax straw fibres: Part I Fibre characterization. J Mater Sci.
1997;32(2):443-9.
[38] Chand N, Hashmi SAR. Mechanical properties of sisal fibre at elevated temperatures. J Mater
Sci. 1993;28(24):6724-8.
[39] Ochi S, Takagi H, Niki R. Mechanical properties of heated-treated natural fibers. In: Brebbia C,
De Wilde W, editors. High performance Structures and Composites. Southampton: WIT press; 2002.
p. 117-23.
[40] Nakamura K, Hatakeyama T, Hatakeyama H. Studies on Bound Water of Cellulose by
Differential Scanning Calorimetry. Textile Research Journal. 1981;51(9):607-13.
[41] Strom G, Carlsson G. Wettability of kraft pulps-effect of surface composition and oxygen
plasma treatment. J Adhes Sci Technol. 1992;6(6):745-61.
[42] Prasad M, Sain M, Roy N. Structure Property Correlation of Thermally Treated Hemp Fiber.
Macromolecular Mater Eng. 2004;289(6):581-92.
[43] John MJ, Anandjiwala RD. Recent developments in chemical modification and
characterization of natural fiber-reinforced composites. Polym Compos. 2008;29(2):187-207.
[44] Fuentes CA, Tran LQN, Dupont-Gillain C, Vanderlinden W, De Feyter S, Van Vuure AW, et
al. Wetting behaviour and surface properties of technical bamboo fibres. Colloids Surf A:
and Bioenergy. 2012;6(4):456-62.
[46] Tripathy D. Practical guide to polypropylene. UK2002.
[47] Zanetti M, Camino G, Reichert P, Mülhaupt R. Thermal Behaviour of Poly(propylene) Layered
Silicate Nanocomposites. Macromol Rapid Commun. 2001;22(3):176-80.
[48] Doan T, Gao S, Mader E. Jute/polypropylene composites I. Effect of matrix modification.
Compos Sci Technol. 2006 66(7-8):952-63.
[49] George G, Joseph K, Nagarajan ER, Tomlal Jose E, Skrifvars M. Thermal, calorimetric and
crystallisation behaviour of polypropylene/jute yarn bio-composites fabricated by commingling
technique. Compos Part A: Appl Sci Manuf. 2013;48(0):110-20.
[50] Sydenstricker THD, Mochnaz S, Amico SC. Pull-out and other evaluations in sisal-reinforced
polyester biocomposites. Polym Test. 2003;22(4):375-80.
[51] Araújo J, Waldman W, De Paoli M. Thermal properties of high density polyethylene
composites with natural fibres: Coupling agent effect. Polym Degradation Stab. 2008;93(10):1770-5.
[52] Van de Velde K, Kiekens P. Effect of material and process parameters on the mechanical
properties of unidirectional and multidirectional flax/polypropylene composites. Compos Struct.
2003;62:443-8.
[53] Khondker O, Ishiaku U, Nakai A, Hamada H. Fabrication mechanical properties of
unidirectional jute/PP composites using jute yarns by film stacking method. J Polym Environ.
2005;13(2):115-26.
[54] Khondker OA, Ishiaku US, Nakai A, Hamada H. A novel processing technique for
thermoplastic manufacturing of unidirectional composites reinforced with jute yarns. Compos Part
A: Appl Sci Manuf. 2006;37(12):2274-84.
[55] Oksman K, Clemons C. Mechanical Properties Polypropylene-Wood. J Appl Polym Sci.
1998;67:1503-13.
[56] Bledzki AK, Reihmane S, Gassan J. Properties and modification methods for vegetable fibers
for natural fiber composites. J Appl Polym Sci. 1996;59(8):1329-36.
Technology assessment and application potential of bamboo fibre composites 181
Chapter 7
Technology assessment and application potential of bamboo
fibre composites
_______________________________________________________
7.1 Introduction
Previous chapters were dedicated to the characterization of bamboo technical fibres

(BF) and bamboo fibre composites (BFC) using different testing methodologies and
conditions. The main goal of this chapter is to assess the potential industrialisation
possibilities of BFC from a technological and economical point of view, in
comparison to other natural fibres and glass fibres. A successful implementation of a
new technology such as bamboo fibre prepreg, its commercialization and its
potential extensive use in the composites industry depends on different factors such
as:
 The availability of the resource;

 Environmental benefits and sustainability of the resource;
 The extraction process and the effect on the ecological impact and fibre cost;
 Manufacturing and environmental impact of the prepreg production;
 Competitive cost of the prepreg;
 Performance of the final composites and potential applications (market);
 Durability and recycling of the parts (end of life possibilities).
An overview of these factors will be presented in the following sections,

nevertheless, it has to be considered that it is a new technology and not all
information is available yet.
182 Chapter 7
b
7.2 Availability of the resources
Regarding the availability of the resource, approximately 36 million hectares is

covered worldwide with bamboo forest: 24 million hectares in Asia and Oceania, 10
million hectares in Latin America and 2.7 million hectares in Africa, distributed over
the tropical and subtropical climate zones. Around 1200 bamboo species have been
identified, of which 64% are in Asia, 33% in Latin American and the rest in Africa
and Oceania [1, 2].
Not all bamboo species are adequate to be used as a source of fibre due to the culm
size, specifically the height and thickness of the culm wall, and consequent low
volumes of fibre yield. For this reason, only “giant” bamboos will be considered for
this purpose. Table 7-1 shows a list of bamboo species that have been identified as a
potential source of fibre due to their anatomical characteristics and their common
use in traditional construction. In tropical zones the most prevalent ones are the
Bambusa, Dendrocalamus, Gigantochloa, Guadua and Phyllostachys. According to
their characteristics, it can be assumed that they are a suitable fibre source in relation
to the fibre yield and the extraction processing feasibility.
Type of bamboo Height Diameter Origin

(m) (cm)
Bambusa balcooa 12 – 20 8 - 15 Southeast Asia
Bambusa chungii 10 - 15 6 China
Bambusa polymorpha 27 15 China
Bambusa vulgaris 18 10 Asia, Americas
Bambusa bambos (L.) Voss 30 15 - 18 Southeast Asia
Bambusa oldhamii 6 - 12 3 - 12 Taiwan
Dendrocalamus edulis 20 16 China
Dendrocalamus giganteus 25 - 35 30 India, Sri Lanka and Thailand
Dendrocalamus asper 25 20 Malaysia
Dendrocalamus latiflorus 20 20 Taiwan and Southern China
Dendrocalamus strictus 8 - 20 8 India
Gigantochloa atroviolacea 10 - 12 10 Indonesia
Gigantochloa atter 25 10 Indonesia, Malaysia
Gigantochloa apus 16 10 Malaysia and Indonesia
Gigantochloa levis 16 10 - 15 Philippines
Guadua aculeata 25 12 Mexico and Panamá
Guadua angustifolia Kunth 15 - 30 11 Colombia, Ecuador, Perú and Brazil
Guadua chacoensis 20 8 - 12 Argentina and Bolivia
Guadua magna 12 - 23 6 - 12 Brazil
Guadua lynnclarkiae 18 15 Perú
Guadua superb Huber 20 9 - 12 Bolivia
Phyllostachys heterocycla var. 12 17 China, introduced to United States
pubenses (Moso) and Japan
Table 7-1. Some eligible “giant “bamboo species as a source of fibres [3-5].
In general, the total amount of bamboo plantations are well defined, but information
about the number of hectares for a specific type of species (e.g. the ones defined in
Table 7-1), is not well specified, outdated or even not reported in an official
inventory. This makes it difficult to have a global estimation of the potential of the
resource for fibre extraction in the world. Nevertheless, according to the collected
data, China has bamboo plantations of approximately 5 million hectares, in which
the bamboo species Phyllostachys hetrocycla var. pubenses (Moso) occupies around
70% [2]. In India the area of bamboo growths exceeds 11.4 million hectares of
which 67% corresponds to the species Dendrocalamus strictus and 13% corresponds
to the specie Bambusa bambos [1]. In Colombia, Guadua angustifolia covers
approximately 28000 ha [6]. For the same species, in a very conservative estimation,
Ecuador reports 15000 ha and Mexico 4000 ha. Brazil has a total surface close to 18
million hectares of bamboo plantations where the most dominant genius is Guadua
plus around 30000 ha of the species Bambusa vulgaris [2]. In total it can be
estimated that more than 15 million hectare of giant bamboo is available worldwide
and a rough estimation of more than millions of 480 million tons of raw fibre
available per year. For more information about the distribution of the bamboo
resources per country or by continent see Londoño [7] and Lobovikov et al.[2].
The sustainability of the bamboo resource is assured by several factors, one of the
most important is the fast growth and the possibility to be harvested annually
without depletion and deterioration of the soil. The species Guadua angustifolia
grows at its peak 21 cm per day and in one month reaches 80% of its maximum
height; it then completes in an additional 5 months reaching on average 23.6 meters
[8]. The productivity is between 1200 and 1350 culms per hectare per year [3].
Bamboo can grow on slopes and other areas where foresting of wood and agriculture
is not possible [9], and it does not need to use land used for food (as is the case of
sugar cane for bio-fuels). The total demand for flax and hemp fibres for material
users is expected to represent 190% of the current surface allocated for the growing
of such fibres in 2015 and 350 % in 2030 [10]. This means that the agricultural
surfaces dedicated to flax and hemp fibres for the material users will have to double
by 2015 to face the demand. This situation can lead to problems concerning the
allocation of fields for food crops and non food crops, making BFs a good
alternative to supply the future demand of high quality fibres.
184 Chapter 7
b
7.3 Environmental benefits
As mentioned in section 3.5, one of the most important environmental benefits of

the bamboo is the large capturing of CO2. It was stated by Cruz [11] that Guadua
angustifolia captures 150 tons of CO2 per hectare in the first 7 years after planting,
with an average of 21 tons/ha/year). A natural growth of guadua plantation (without
technical management), with a density of 5755 culms/hectare does capture a total of
133 tonnes of carbon. These values are rather consistent with Arango and Camargo
[12], in their study on bamboo plantations of 7-8 years old (7700 culms / hectare)
which captured 25 tons of CO2 /ha/year. After 6 – 7 years the carbon absorption
stabilizes due to the fact that all vegetative development is completed. With an
adequate management and harvesting, the self-regeneration of the bamboo
plantation provides a permanent CO2 absorption, which does not happen with other
species [3]. This high carbon rate sequestration can be translated into carbon credits1
in the voluntary carbon market, meaning that managed bamboo forests can generate
extra benefits for both farmers and the environment at the same time. For example,
several Chinese companies have already pre-ordered more than 8000 tonnes of
carbon credits. The money they pay supports the planting of new bamboo forests in
China. Now, companies in China that want to offset their carbon emissions can buy
bamboo carbon credits on the voluntary market through the China Green Carbon
Foundation (CGCF) 2 [13]. In Colombia the accessibility to apply to carbon
voluntary market schemes is still not available but there are already good conditions
to meet REDD+ 3 initiatives, in terms of well established political and legal
frameworks that will facilitate actions in the near future [12].
Besides the CO2 capture, bamboo cultivation has other environmental benefits.
Thanks to the dense network of bamboo roots they help for the reduction of soil
erosion. Also, one hectare of Guadua angustifolia can retain 30000 liters of water
helping with the regulation of the hydraulic flow, conserving the water in the rainy
season and using it later in the dry season [3], and finally the number of employment
1
A carbon credit for greenhouse gas reductions is achieved by one party that can be purchased and used to
compensate (offset) the emissions of another party. Carbon offsets are typically measured in tonnes of CO2
equivalents (or CO2e) and are bought and sold through a number of international brokers and trading
platforms.
2
CGCF is the first nation-wide non-profit public funding foundation dedicated to combating climate change
by increasing the carbon sink in China. It was recognized as a valid mechanism to mitigate climate change
during the Bali conference in 2007.
3
Reducing Emissions from Deforestation and Forest Degradation (REDD+) is a global initiative designed to
pay groups or countries for protecting their forests and reducing emissions of greenhouse gas pollutants,
especially CO₂.
that this activity generates in rural areas. Therefore, local authorities stimulate the
planting and harvesting of bamboo by giving subsidies to the farmers (e.g. 50% of
the establishment and maintenance costs of the plantation up to the 5th year in
Colombia).
One bamboo hectare can produce between 495 and 600 tons of biomass after 6 years
of the maturation of the plantation, being the weight of the culm the part of the plant
that produce most biomass followed by the rizome [11, 14]. One hectare of adult
guadua bamboo can also produce 5.8 times more biomass compared to most other
forest species [15]. This biomass can be a source of renewable energy, according to
the UK Department of Energy and Climate Change 4 ; switching to biomass from
fossil fuel alternatives reduces carbon emissions by 70%. Biomass also emits
significantly lower levels of ash, nitrogen, sulfur, mercury and other heavy metals
that are harmful to the environment. Part of this biomass (e.g. branches, leaves, etc.)
can be used for alternative uses aside of the fibre production, e.g. the extraction of
baseline chemicals for fertilizers, etc.
7.4 The extraction process and the effect on the ecological impact and fibre cost
The newly developed extraction process does not use pressure, chemicals or high
temperature, and all the operations involved were designed to be an in-line process
avoiding batches during the extraction and limited use of water which can be
recovered. For this reason it is expected that the price of the fibre extraction can be
competitive regarding other natural fibres and glass fibres. It might be noticed that
this good mechanical behaviour of BF or its composites reported in Chapters 4 and 5
were achieved without any fibre treatment (e.g. chemical treatment) being an
important environmental and cost aspect issues that strength the use of the BFC.
The relative cost advantage of natural fibres over glass fibres will disappear if
expensive chemical treatments are needed to improve the composite performance
[16]. These additional treatments are expensive operations that have to be done
during the fibre preparation (e.g. soaking in alkali, rinsing the fibres, drying, etc.)
and which can have negative environmental impact due to the use of chemicals and
water in the process.
An estimation of the environmental impact of such treatment process, e.g. the

calculation for the energy consumption, in this early stage (prototype machines) will
not be accurate due to the scaling effect of the production at industrial volumes. For
4
https://www.gov.uk. Press release, August 2013.
186 Chapter 7
b
this reason, data from the flax industry will be used as reference for some
comparisons with glass fibres, assuming that certain parts of the flax extraction
process, e.g. scutching, can be similar in terms of volume of the machinery and
energy consumption. According to Le Duigou et al. [17], the energy consumption
increases from 4.4 MJ/Kg for flax scutched fibres to 11.6 MJ/kg for hackled fibres
due to the additional energy consumed by the hackling process (0.55 kWh/kg of
fibres hackled). The scutching step requires 0.116 kWh/kg of fibres. In general, the
production of flax fibres appears to be an environmentally attractive alternative to
glass fibres in terms of non-renewable energy consumption. The energy
consumption for the production of flax fibres and mats is around 73% and 82%
lower than used for the manufacturing of glass fibres and mats respectively (see
Table 7-2).
Type Non renewable energy requirements Reference

(MJ/kg fibres)
Flax fibre (scutched) 4.4 [17]
Flax fibre (hackled) 11.7 [17]
Flax fibre (hackled) 12.4 [18]
Flax fibre mat 9.5 [19]
Glass fibre 45 [17]
Glass fibre mat 54.7 [19]
Table 7-2. Non renewable energy requirements for flax and glass fibres and their mats.
Most of the environmental aspects used are favourable to flax fibres, with the
exception of the eutrophication indicator. For flax fibres, this remains high, mainly
due to the use and production of fertilizers [17]. The usage of fertilizers, herbicides
and other pesticides plays an important role in the environmental impact values.
Their optimal use in the case of the bamboo plantations requires performing a soil
test to check deficiencies, but in general, the NPK fertilizers composed of three main
elements: Nitrogen, Phosphor and Potassium, are the most used ones. As an
alternative, natural fertilizers can be used as a complement to the chemical ones in
order to reduce the impact on the eutrophication indicator. They can be coffee pulp
residues (e.g. in the Coffee Region), already used extensively for this purpose in
agricultural soils. This last option is available thanks to the currently waste related
policies formulated by the Colombian government which have focused its effort on
the solid organic residues management from the organic municipal waste generated
by the population (70% of the total residues production 5 ), producing natural
fertilizers such as compost.
5
www.asocompost.org
During the operations for the extraction of long BFs (> 50 mm), several side stream
products are produced (see Figure 7-1) and can be valorized, they are:
 Biomass (e.g. chemicals, fertilizers, coal, etc.);

 Dust and particles;
 Short fibres (>2mm);
 Medium length fibres (5 – 50 mm).
Forestry
Harvesting
Operations
Cutting and preparation
Extraction of the fibres / cleaning

Raw materials
Biomass Dust Short fibres Medium Long fibres

(branches & (< 5 mm) length fibres (>50 mm)
& leaves) particles (5-50 mm)
Manufacturing
Chemical Separa- Extrusion Random UD

transfor- tion & positioning discont.
mation pelletizing randomiz.
Intermediate
Baseline Polymer Pellets for Mats, SMC, UD

products
chemicals fillers injection BMC, LFT prepregs

(e.g. moulding
fertilizers)
Figure 7-1. Raw materials extracted from the bamboo after different operations and their corresponding
intermediate products that can be offered to the market. SMC= sheet moulding compound, BMC= bulk
moulding compound and LFT= long fibre reinforced thermoplastic.
A concise overview of each of these side products will be presented a continuation

to explain their potential market (main uses) and profit opportunity.
188 Chapter 7
b
7.5 Manufacturing and performance of the final composites; potential

applications
Long bamboo fibres
The manufacturing of high performance composites will be possible with the

unidirectional disposition of the fibres into the composite. This fibre configuration
can reach the highest mechanical performance (see Figure 7-2) due to the axial
alignment of the fibres to the loading direction and the absence of crimp optimizing
and strengthening the composite with the added advantage of reducing the amount
of resin, thereby minimizing the weight of the final composite.
Tensile strength (MPa)
Tensile modulus (GPa)

Figure 7-2. Ashby-plot, tensile modulus vs. tensile strength, for various processing techniques used in the
processing of composite materials including natural fibre composites. In this plot the processing technique is
linked with the type of fibre reinforcement [20].
For this purpose, it is necessary to stabilize the extracted long bamboo fibres (>50
mm) in a continuous unidirectional pattern where the fibre ends are randomly
distributed. This is possible holding the fibres together with a binder or stitching to
have a preform that can be used in RTM and autoclaving. Another alternative is by
adding the exact amount of matrix (thermoset or thermoplastic) necessary to achieve
a suitable fibre volume fraction to obtain a pre-impregnated material (“prepreg”).
A configuration of narrow strips makes possible to use different existing

technologies such as automated fibre placement and filament winding. Also, it opens
the possibility to apply both weaving and braiding to produce textile fibre
architectures ready to be used in composite production. A wider bamboo tape would
allow the use of manufacturing technologies such as: automated tape laying,
thermoforming and matched die moulding. These scenarios will open several
possibilities for future applications for the UD bamboo fibre prepreg (tape) or
preform, adding the highest value to the long bamboo fibres as an intermediate
product that will be offered to the composites industry.
In terms of the performance, the strength of the BFC (Vf=40%) with epoxy matrix
in tensile and flexural tests was found to be comparatively high with respect to other
natural fibres as shown in section 5.6 (see Tables 3-3 to 3-6). Furthermore, the BFC
are suitable to be used in applications where the stiffness is the most important
performance characteristic since it has been shown in section 5.4.1 that the
composite stiffness is not affected by the different fibre end patterns that were
investigated, benchmarked and compared to fully unidirectional fibre composite.
The specific properties of the bamboo fibres, when normalized to the material’s
density, are in average 595 MPa*cm3/g (σ/δ) and 31 GPa*cm3/g (E/δ) for the
strength and Young’s modulus respectively. These values are comparable to the
specific properties of E-glass which are in average 610 MPa*cm3/g and 30 GPa
*cm3/g. The specific stiffness formula E/δ actually only applies to materials subject
to a tensile stress. For applications that are subject to bending stresses (“beams”), the
square root of the stiffness (E1/2/δ) has to be introduced into the formula, which
increases the relative importance of the specific weight. And for materials loaded in
compressive stress or for plates in bending, the cubic root of the stiffness has to be
used (E1/3/δ) for further comparisons. A study revealed also that the compressive
properties of BFC were 150 MPa for the strength and 16 GPa for the modulus.
These values correspond to the 67% and 82% of the maximum tensile strength and
tensile modulus respectively and being slightly better than flax-epoxy composites at
the same fibre volume fraction (Vf =40%) [21]. The specific stiffness (E/δ) and the
specific stiffness for bending (E1/3/δ) in flax and hemp (as it is also the case for
bamboo fibres), are higher that the values of E-glass [22], which make them ideal
for stiff and light composites.
Medium and short bamboo fibres and particles
Medium and short bamboo fibres are obtained during the extraction and cleaning
processes. They can be easily separated to produce random mats or pellets for the
production of low cost materials, because they do not require intermediate
190 Chapter 7
b
operations such as fibre alignment, as is the case for the UD prepreg. Long fibre
reinforced thermoplastics (LFT), sheet moulding compound (SMC), bulk moulding
compound (BMC) and injection moulding are visible manufacturing processes to
improve the mechanical properties of the polymer and to improve dimensional
stability of the parts. The particles (e.g. dust) also obtained during the extraction
process, can be used as filler for polymers to cheapen and enhance properties of the
end products. In general bamboo fibres, similar to flax and hemp, are less abrasive
in comparison to glass fibres and present less fibre breakage (shortening) during
compounding and extrusion processes.
Potential market
In 2010, the global natural fibre composites market reached €1.6 Billion. Lucintel6
forecasts the natural fibre composite market to to have a high growth increasing to
€2.9 Billion in 2016 and to grow with a good CAGR (Compound Annual Growth
Rate) of 11.2%7 during the 2014-2019 period. The automotive market will remain
the largest segment in terms of both value and volume during this period. The major
drivers of using natural fibre composites are to take advantage of the natural fibres
properties such as: lightweight, possibilities of recycling, sound insulation, acoustics
and governmental regulations that encourage the protection of the human health and
the environment. Also, an increasing demand from the building and construction
industries as well as from the electrical and electronic industries is leading the
growth or the natural fibre composites market. According to these predictions a
range of different applications will continue having a big impact in the natural fibres
business7:
 Automotive sector: door panels, seat backs, headliners, dash boards;

 Electrical and electronics: mobile cases, laptop cases;
 Sporting goods: tennis rackets, bicycles, frames, snowboards;
 Building and construction: door panels, decking, railing, window frames.
6
Lucintel is a global management consulting and market research firm (http://www.lucintel.com).
Opportunities in natural fibre composites.
7
Lucintel (http://www.prnewswire.com/news-releases). Global natural fibre composites market 2014-2019:
Trends, forecast, and opportunity analysis.
7.6 Durability and recyclability
There is little knowledge about the durability and the BFC long term behaviour
(creep, fatigue, etc.) and they still need to be investigated. Furthermore, moisture
uptake trough time is a major concern for the industry players. Some studies on
tensile testing of individual fibres conditioned to different humidity levels show an
increase in failure strain with increasing humidity level. Damage analysis of broken
fibres showed that this is due because of the plasticization of the fibres. Moreover,
tests performed on unidirectional bamboo epoxy composite samples show a
statistically relevant increase in both failure strain and strength with increase in
atmospheric humidity level [23]. This good behaviour at high humidity levels is
expected because the bamboo fibre tends to be less moisture sensitive (e.g. in
comparison with flax and hemp), because of the higher lignin content on the surface
on the fibre as it was demonstrated in Appendix 2. Also, some studies in fatigue
behaviour with flax fibres reported promising results at high cycling fatigue loads
that may potentially be expected also for BFC. Shah [24], found that the fatigue
strength degradation rates are lower in natural fibres reinforced plastics than in glass
fibre reinforced plastics. Moreover, according to Bensadoun et al. [25], it was
observed that at normalized stress by density, the fatigue life of flax-epoxy plain
weave composite have comparable behaviour to glass plain weave epoxy composite,
which validates the possibility of replacing glass fibre by flax in the near future for
certain applications.
The recyclability is another important parameter that will affect the end-of-life
option for the BFC. The first aspect to consider is the type of matrix used, because
the alternatives for recycling will depend directly on it. There are nowadays several
techniques that can be applied to natural fibre composites. The most common one is
the mechanical recycling, especially for thermoplastic composites applying a
downgrade recyclability and giving as a result pieces with short fibres that later on
can be added to an extrusion machine to be compounded and pelletized [26]. Other
options are the incineration for energy recovery and the thermal recycling. For the
first, it has been shown that the calorific value of natural fibres are superior to
synthetic ones leading to an increased energy recovery once burned and leads to a
positive carbon credits and lower warming effect [19]. All natural fibre composites
can be incinerated for energy recovery where between 50% and 70% of the heat can
be recuperated as energy [20]. In this case, an advantage is that in the case of natural
fibre composites, not only the matrix material but also the fibres can be combusted;
the composite can thus be considered a delayed biomass, after having served a
lifetime as a light-weight component. In the case of glass fibre composites, only the
matrix burns and the fibres tend to remain as residue, although usable as raw
material for cement production [27]. The second mainly consist of the degradation
192 Chapter 7
b
of the matrix trough temperature in order to recover the reinforcing fibres, but due to
the low degradation temperature of natural fibres >200 oC, this is not an option.
However, it is possible for the thermoplastic composite structure to increase the
temperature close to the melting point of the polymer matrix in order to re-shape it,
and to obtain a new product. Several layers of new material can be added for
reinforcement and for better looking of the component. The concern with this
method is the degradation to the fibre due to the thermal treatment needed during the
process. Chemical recycling of natural fibre composites seems to be not feasible due
to unavoidable damage to the fibres when dissolving the matrix; moreover chemical
treatments are expensive and not environment-friendly.
7.7 Conclusions
The availability of the bamboo resource is assured due to the high amount of
available hectares of giant species around the world making it an ideal source of
fibre for reinforcing composites. Several governmental incentives (e.g. subsidies)
stimulate the planting and harvesting bamboo in a continued manner due to its
environmental and social benefits. After an overview of the ecological,
environmental, technological and performance issues concerning the production and
use of BFs and BFCs, they appear to be competitive with similar natural fibres
already implemented in the composite market. Also, they present superior
environmental advantages in comparison to glass fibres, and in terms of mechanical
properties they may replace them in certain applications. The market forecast for the
natural fibre composites seems to be very positive with a regular annual growth led
by more stringent environmental regulations, new recycling possibilities and product
design and innovation. Nevertheless there are several aspects for the bamboo fibres
and their composites that still need to be investigated in the future: Some are
summarized in the future research segment in section 8.4.
References
[1] Chaowana P. Bamboo: An Alternative Raw Material for Wood and Wood-Based Composites.
Journal of Materials Science Research. 2013;2(2):p90.
[2] Lobovikov M, Paudel M, Piazza M, Ren H, Wu J. World bamboo resources: a thematic study
prepared in the framework of the Global Forest Resources Assessment. Non wood fores products
18 -World bamboo resources, FAO, 2005. p. 1-18.
[3] Minke G. Building with Bamboo: Design and Technology of a Sustainable Architecture: De
Gruyter; 2012.
[4] Londoño X. A decade of observations of a "Guadua angustifolia" plantation in Colombia. The
Journal of the American Bamboo Society. 1998;12:37-42.
[5] Londoño X. Dos nuevas especies de guadua para el Perú (Poaceae: Bambusoidae: Bambuseae:
Guaduinae). Journal of the Botanical Research Institute of Texas. 2013;7(1).
[6] Kleinn C, Morales-Hidalgo D. An inventory of Guadua (Guadua angustifolia) bamboo in the
Coffee Region of Colombia. European Journal of Forest Research. 2006;125(4):361-8.
[7] Londoño X. La Guadua un bambú importante de América. First Bamboo Seminar. Guayaquil,
Ecuador 2001. p. 12-8 [Spanish].
[9] Vogtlander J, Van der Lugt P, Brezet H. The sustainability of bamboo products for local and
Western European applications. LCAs and land-use. J Clean Prod. 2010;18(13):1260-9.
[10] Bewa H. Fibres végétales: une disponibilité régulée de la ressource. JEC Composites
conferences. Paris2012.
[11] Cruz H. Bambú - Guadua. Bosques naturales en Colombia y plantaciones comerciales en
Mexico: Colmex; 2009.
[12] Arango A, Camargo J. Bosques de guadua del Eje Cafetero de Colombia: oportunidades para
su inclusión en el mercado voluntario de carbono y en el Programa REDD+. Recursos Naturales y
Ambiente. 2011;61:77-85.
[13] Evans K. Bamboo carbon credits now on sale in China: INBAR. Forest news A blog by the
Center for International Forestry Research (http://blogcifororg/13245/bamboo-carbon-credits-now-
on-sale-in-china#U8v4_fmSzng), 2012.
[14] Riaño N, Londoño X, López Y, J G. Plant growth and biomass distribution on Guadua
angustifolia Kunth in relation to ageing in the Valle del Cauca - Colombia. J Am Bamboo Soc.
2002;16(1):43-51.
[15] Environmental impact of bamboo. http://wwwguaduabamboocom/blog/environmental-impact-
of-bamboo.
[16] Thomason J. Why are natural fibres failing to deliver on composite performance? 17th
International Conference on Composite Materials, ICCM172009.
[17] Le Duigou A, Davies P, Baley C. Environmental impact analysis of the production of flax
fibres to be used as composite material reinforcement. Journal of biobased materials and bioenergy.
2011;5(1):153-65.
[18] González-García S, Hospido A, Feijoo G, Moreira M. Life cycle assessment of raw materials
for non-wood pulp mills: Hemp and flax. Resour, Conserv Recycl. 2010;54(11):923-30.
[19] Joshi SV, Drzal LT, Mohanty AK, Arora S. Are natural fiber composites environmentally
superior to glass fiber reinforced composites? Composite Part A: Appl Sci Manuf. 2004;35(3):371-6.
[20] Shah DU. Developing plant fibre composites for structural applications by optimising
composite parameters: a critical review. J Mater Sci. 2013;48(18):6083-107.
[21] Wouters K. Compressive properties of natural fibre reinforced composites. Master thesis:
Leuven University; 2014.
[22] Confédération européenne du lin et du chanvre. Book: Flax and Hemp Fibres: A Natural
Solution for the Composite Industry: JEC Composites, editor Reux F.; 2012.
[23] Verheyden S. Karakterisering van vochtabsorptie van bamboe vezel composiet [Master thesis].
Leuven: University of Leuven, Faculteit Ingenieurswetenchappen, 80 p.; 2010.
[24] Shah DU. Characterization and optimization of the mechanical performance of plant fibre
composites for structural applications: PhD thesis, Univerisity of Nottingham; 2013.
[25] Bensadoun F, Depuydt D, Baets J, Vuure V, Willem A, Verpoest I. Influence of fibre
architecture on impact and fatigue behaviour of flax fibre-based composites. The 19th International
Conference on Composite Materials (ICCM), Montreal, Canada 2013.
[26] Summerscales J, Dissanayake N, Virk A, Hall W. A review of bast fibres and their composites.
Part 2 - Composites. Compos Part A: Appl Sci Manuf.41(10):1336-44.
[27] Vanderfeesten B. Environmental impact of newly developed composites with thermoplastic
matrix for high volume applications. Master thesis. Leuven: University of Leuven; 2014.
194 Chapter 7
b
General conclusions 195
Chapter 8
General conclusions
____________________________________________________________________
This PhD thesis has presented a study on the mechanical properties of long bamboo
technical fibres and their composites with thermoset and thermoplastic matrices, and
includes recent developments in the fibre preparation and manufacturing techniques
in order to obtain high quality bamboo composites.
The main contribution of this research work to the state of the art is the development
of a unidirectional bamboo fibre composite with high mechanical properties. This
was possible thanks to a novel mechanical extraction process, applied during this
PhD research work that proved to be an effective method to produce long high-
quality bamboo fibres. These unique long high quality fibres are an innovative and
renewable alternative for other natural fibres and particularly for synthetic fibres (e.g.
glass) used in composites. The conclusions of this research are summarized as
follows:
8.1 Bamboo technical fibre
As a first stage, a new environment friendly process was developed to extract

bamboo fibres. The effectiveness of this process was verified by measurements of
tensile strength, Young’s modulus and strain to failure, all confirming the good fibre
quality. The modified length-based Weibull model, accurately describes the fibre
strength distribution of the bamboo technical fibres at various gauge lengths. The
Weibull shape parameter (m) revealed low variation for the fibre strength,
demonstrating the good condition of the fibres produced by the new mechanical
extraction method. It was possible to establish a close relation between the
196 Chapter 8
b
extraction process and the quality and morphology of the bamboo fibres. For this
reason, fibre properties cannot be generalized and should be linked with the
correspondent extraction method in order to state fair comparisons and statements.
Moreover, an adequate statistical treatment of the data is necessary to compare,
describe and predict the mechanical properties of the fibres and hence its composites.
Single technical fibre characterization requires sample pre-selection, where weaker

fibres can break during separation. Also, for statistical reasons a large number of
single fibre tests are needed. Therefore, as an alternative for fibre characterization,
the dry fibre bundle test was proposed to obtain reliable values for the mechanical
properties and Weibull parameters for a larger population of non-impregnated fibres,
with a relatively small amount of experiments. In comparison with the single fibre
test, the bundle test systematically gives lower fibre strength and higher variability
due to the fact that no pre-selection of the fibres is needed. The Young’ modulus,
however, remains almost unchanged. The developed methodology is generic enough
to be used for other natural fibres as well, avoiding intensive manipulation of
individual single fibre sampling and testing.
8.2 Unidirectional continuous and discontinuous bamboo fibre epoxy

composites
Unidirectional long bamboo fibre/epoxy composites were produced showing in

general the high potential of bamboo fibres for high-end composite applications.
Experiments show that the mechanical properties of bamboo fibres are transferred
to the composite properties, reaching 92% for stiffness and 79% for strength of the
ideal bamboo fibre composite properties. This points out the strong fibre-matrix
interface present between fibres and epoxy matrix, promoted by the high fibre
surface roughness, and confirmed by a relatively high transverse flexural strength
(33 MPa).
The limited (<25 cm) length of technical bamboo fibres inevitably results in fibre
ends and consequent stress discontinuities in real size composites. An innovative
approach for industrialization of bamboo fibres in high-performance composite
materials is the development of an up-scalable method to randomize the fibre ends
for the production of a continuous bamboo fibre prepreg (tape). This fibre
configuration will allow the future commercialization at industrial scale and the use
of existing technology for the composite manufacturers.
In order to simulate the effect of the inherent fibre ends on the composite
mechanical properties, six discrete patterns of fibre discontinuity distributions were
designed and tested. The results show that a highly randomized distribution is
necessary to preserve the composite properties: 85% of the longitudinal tensile
strength of a fully continuous bamboo fibre composite is obtained. The composite
stiffness variations were not significant for any of the studied fibre-end patterns. The
results clearly show that full randomization of the fibre ends in UD composites is
necessary to fully exploit the good mechanical properties of the fibres. The
experimental composite strength was predicted by a modified local load sharing
model (LLS) within a reasonable margin of around 14%.
It was found that the impregnated fibre bundle test is a good option to evaluate the
composite tensile properties, with some advantages over the “standard” composite
tensile test: less material is needed and easier specimen preparation saves time and
cost. It gives easier control of the fibre alignment for the measurement of the
modulus, and the reduced sample volume (with the corresponding lower probability
of defects) will improve the composite strength.
8.3 Thermal degradation in single technical fibres and bamboo fibre

polypropylene composites
In literature, the thermal degradation of natural fibres is only scarcely linked with
the loss in mechanical properties of single fibres (i.e. strength, stiffness and strain to
failure) and composites (e.g. fibre degradation during the manufacturing process).
This makes it difficult for a process engineer to design a manufacturing process with
natural fibres and thermoplastic matrices, where sufficient time at elevated
temperature is needed to achieve full impregnation and consolidation.
Single technical bamboo fibres were exposed at different temperatures and times of
exposure in presence of air and argon. The results show that the fibres start to be
affected at temperatures above 160 °C and that thermal treatments or manufacturing
processes that induce more than ~7% of mass loss, start to considerably affect the
mechanical properties of the fibres. The use of an inert argon atmosphere for thermal
treatment of technical fibres results in an improvement of the strength properties
compared to exposure to air, of 45 % at 250 °C respectively (for 50 min of
treatment). The strength results of single thermally treated fibres show that the mass
loss due to the thermal degradation is dependent on the temperature and becomes
more critical as the time of exposure increases. The Young’s modulus was not
significantly influenced by any of the applied thermal treatment cases. Furthermore,
198 Chapter 8
b
XPS analyses show no changes in the surface components for untreated or thermally
treated fibres.
Results show that a correlation exists between the mass loss and the fibre strength. If
the mass loss of the technical fibres for a given set of process parameters (i.e.
temperature and time of exposure) is known after TGA measurements, the
correlation graph allows the prediction of fibre and composite strength. This study
served not only to have a better understanding of the final composite behaviour, but
also to have an approach for the process parameters selection, assuring no damage
occurs in the internal structure of the fibre (caused by too high temperatures during
the production of bamboo fibre/thermoplastic composites).
Unidirectional bamboo fibre polypropylene composites consolidated at relatively

high temperatures in an inert atmosphere, exhibit 30% higher flexural strength
compared to composites produced in air. The use of inert atmosphere during
manufacturing of natural fibre thermoplastic composites is a promising method that
can easily be adapted. Nevertheless, it is necessary to consider the costs associated
to future mass production and nitrogen might be a cheaper alternative to argon.
Matrix modification using MAPP showed no significant improvement of the
composite properties in comparison with the unmodified PP samples.
8.4 Potential use of bamboo fibre composites
In the literature, it is well established that natural certain fibres are suitable for use as
reinforcement in polymeric matrices and they have the potential to replace synthetic
fibres such as glass in several applications due to a number of advantageous
characteristics. Some of them are good specific mechanical properties, good acoustic
insulation and damping, renewability, reduced environmental footprint as well as
giving better possibilities of recycling for the composites.
Bamboo fibres in particular present several ecological, environmental, technical and

mechanical performance aspects concerning the production and use of bamboo
technical fibres and their composites. These make them suitable to fulfil the
requirements of the new worldwide requirements toward the use of renewable and
sustainable materials at competitive price. Nevertheless, as important as the
technical and environmental aspects associated with the use of natural fibres, is their
relevance in the local economies and the societal impacts, especially in the
developing countries where the bamboo grows.
Bamboo grows massively in equatorial countries through four continents and this
aspect makes the bamboo fibres an excellent alternative to the already existing high
quality natural fibres used in composites such as flax and hemp. This is under the
premise that for low cost and commercial exploitation, the fibre production should
be close to where plants grow. They should not be seen as a competence itself,
conversely, they together become strong to face the challenge of developing and
maintain markets in which they can compete effectively with synthetic fibres (e.g.
glass fibres).
8.5 Suggestions for future research
Bamboo fibre epoxy composites have good mechanical properties, but epoxy is a
relatively expensive matrix and non-biodegradable fossil fuel-based resin. It is
suggested for future work to study various thermoset and thermoplastic matrix
systems, preferably bio-based, to meet the market expectations. The following
studies can be considered:
- Bio-based matrices like GreenPoxy 55® resin (whose molecular structure

is originated from plants), polyamide 6 (PA 6) and Solanyl® (made from
leftovers of potato starch), are very interesting options in terms of
properties and biodegradability;
- Regarding thermoplastic matrices, Polyoxymethylene (POM) presents

good mechanical properties and acceptable production temperature; it has
also shown good adhesion to other natural fibres.
The tensile strength of UD discontinuous bamboo fibre epoxy composites is 15%

lower than for the UD fully continuous composite. Additional work is needed to
explain and to improve this performance. It can be hypothesized that a tougher
thermoplastic matrix will be less sensitive to the discontinuities at the fibre ends.
Further improvement in the standardization of the manufacturing of the continuous
fibre preform (tape) is required by enhancing all variables involved in the
randomization process (proof of concept machine) to obtain the desired results.
To reach the full potential of bamboo fibre thermoplastic composites, further work is
needed to improve the compatibility. Experiments carried out in this thesis with
MAPP did not show improvement in properties in contrast with data on other natural
fibre composites presented in literature. The pure MAPP used in this study did have
low intrinsic mechanical properties. Further studies with blends of MAPP with
higher performance polymers and by varying and controlling the content of maleic
200 Chapter 8
b
anhydride or other compatibilizers, can improve the actual mechanical properties of

the bamboo fibre composites. Furthermore, treatments made on the dry fibres such
as cold plasma can also be considered.
To achieve better properties for thermoplastic matrices, further research to optimise

the manufacturing process to avoid fibre degradation by minimizing the exposure of
the fibre to high temperature is needed. The use of inert atmosphere can help to
improve the mechanical properties of the final composite.
In-service properties of bamboo fibre reinforced plastics (BFRP) must be evaluated

to determine the long term reliability:
- Fatigue behaviour;
- Creep behaviour;
- Interlaminar fracture toughness, impact behaviour and properties after
impact;
- Effects of moisture absorption, swelling and wet-dry cycling on the decay
in properties.
Bamboo fibre composites are expected to be applied in many high performance

industrial applications, such as sporting goods, building and construction and the
transportation sector, etc. The characterization of long term BFRP behaviour is very
important to achieve long service life and for a good performance in a LCA
assessment, especially in the use phase. The light-weight natural fiber composites
improve fuel efficiency and reduce emissions in the use phase of the composite
component, especially in transport applications1. This advantage will dominate the
results and favour the BFRP.
1
Joshi et al. Composites Part A 35 (2004), 371-376.
Appendix 1 201
Appendices
202 Appendix 1
Appendix 1
A1.1 Calculation of the Weibull parameters by linear regression
Different approaches available in the literature can be used to estimate the σ0 and m
parameters of the Weibull distribution from a set of experimentally measured
fracture stresses. The most common methodologies are both the maximum
likelihood (ML) and linear regression (LR), known as the graphical method. It has
been stated that ML method leads to the most precise estimation of the Weibull
modulus (m) [1]. However, several studies [2, 3] concluded that this method yielded
similar values as LR for m and σ0. Also, it was shown that ML tends to overestimate
the Weibull modulus leading in a lower safety factor in reliability predictions. For
these reasons, the LR method is preferred from the engineering point of view [4, 5].
In natural fibres both methods are used as is show in Table 4-1.
To obtain the two parameter Weibull by LR it is necessary to rearrange the

distribution (Equation 4-6) by taking twice the natural logarithm, and rewriting it in
a linear form y= mx+C:

 L m  1
ln(ln(1/1  P))  m ln( )  m ln( 0 )  ln    1   (A1-1)
 L0   m
A straight line should be found when ln(σ) is plotted against ln(ln(1/1-P)), the
Weibull modulus (m) corresponds to the slope of the line. The characteristic strength
(σ0) can be calculated based on the intersection with y-axis via linear regression and
m, as follows:
 L
C  m ln  0  ln   (A1-2)
 L0 
thus:
  L 
 C  ln   
 0  exp  m 0  
L
(A1-3)
 
 
The p-value (P) is estimated by a function known as the probability index. These are
the most commonly used probability estimators (Equations A1-4 to A1-7) found in
literature [6-12]:
Appendix 1 203
i  0.5
P (A1-4)
n
i
P (A1-5)
n 1
i  0.3 (A1-6)
P
n  0.4
i  (3 / 8) (A1-7)
P
n  0.25
Where n is the size of the sample and i is the rank of ith data point. In order to
calculate P from the experimental values, the data are sorted in ascending order and
given number as a counter (i). Among this commonly used probability estimators, it
is reported that estimator A1-4 gives the least-biased estimation of the Weibull
modulus for n ≥20, followed by estimators A1-7 and A1-6, and it is considered the
best probability estimator by many authors [4, 13, 14]. Furthermore, it was found
that after statistical treatment of the data, the estimator A1-4 yielded systematically
the highest m value, while estimator A1-5 yielded the lowest number and highest σ0
[6]. Figure A1-1 shows the Weibull plot based on Equation A1-1 for all bamboo
fibres correspondent to the 30 mm span length (430 samples) in which the Weibull
parameter is determined by different probability estimators described above.
○ Estimator A1-5 m = 8.0, σo = 790 MPa, R²= 0,973

Δ Estimator A1-6 m = 8.2, σo = 790 MPa, R²= 0,976
□ Estimator A1-7 m = 8.3, σo = 790 MPa, R²= 0,974
◊ Estimator A1-4 m = 8.3, σo = 789 MPa, R²= 0,981
Figure A1-1. Weibull plot for fibres tested at 30 mm span length using LR method and different probability
estimators (P).
It is observed that there were not significant differences when using the different
estimators, nevertheless the estimator A1-4 gave slightly better fitting with a R2
204 Appendix 1
equal to 0.98. This estimator was used to estimate the Weibull parameters for the
different gauge lengths as shown in Figure A1-2. The experimental data fit a straight
line reasonably well for each span length, which demonstrates good agreement with
the Weibull statistics. Deviations from Weibull distribution could be mainly
attributed to the weakest fibres, probably damaged during the preparation process
[15] or another external factor.
a. b.
c. d.
e. f.
Figure A1-2. Weibull plots for the strength to failure of bamboo fibres at different gauge lengths: a) 1 mm, b)
2 mm, c) 5 mm, d) 10 mm, e) 20 mm, f) 40 mm. Fitting of the data to the straight line demonstrates good
agreement with the Weibull statistics.
Appendix 1 205
As discussed in section 4.5.3 it was assumed that during the tensile test at these short
span lengths, some elementary fibres are clamped; but at the same time there is
presence of fibre ends randomly distributed along the gauge length. All this
contribute to have a “miscellany system” that when pulled at tension could have an
influence on the fibre fracture. An indication of that is when the data plot under
Weibull coordinates shows some deviations and sharp corners. These characteristics
suggest a presence of a mixture of failure modes [16], as it seems to be the case for
gauge lengths of 1 and 2 mm (see Figure A1-2a and b) where the linear regression
fits with lower R2. When the gauge length became longer, a more consistent type of
fibre failure occurred represented in the straightness of the data plot.
Shape Scale (σ0)

Tested gauge
parameter parameter
length (mm)
(m) (MPa)
1 12.1 984
2 9.4 960
5 8.8 880
10 8.0 872
20 12.8 785
30 8.3 789
40 7.8 796
All 7.6 850
Table A1-1. Weibull parameters obtained by linear regression (LR) at different span lengths with the length
(L) as a scale variable.
c.
References
[1] Khalili A, Kromp K. Statistical properties of Weibull estimators. J Mater Sci. 1991;26(24):6741-
52.
[2] Patankar SN. Weibull distribution as applied to ceramic fibres. Journal of Materials Science
Letters. 1991;10(20):1176-81.
[3] Doremus R. Fracture statistics: A comparison of the normal, Weibull, and Type I extreme value
distributions. Journal of applied Physics. 1983;54:193-9.
[4] Wu D, Zhou J, Li Y. Methods for estimating Weibull parameters for brittle materials. J Mater
Sci. 2006;41(17):5630-8.
[5] Wu D, Lu G, Jiang H, Li Y. Improved estimation of Weibull parameters with the linear
regression method. J Am Ceram Soc. 2004;87:1799-802.
[6] Zafeiropoulos NE, Baillie CA. A study of the effect of surface treatments on the tensile strength
of flax fibres: Part II. Application of Weibull statistics. Composite Part A: Appl Sci Manuf.
2007;38(2):629-38.
[7] Cao Y, Wu Y. Evaluation of statistical strength of bamboo fiber and mechanical properties of
fiber reinforced green composites. J Cent S Univ Technol. 2008;15(0):564-7.
[8] Defoirdt N, Biswas S, De Vrise L, Tran N, Van Acker J, Ahsan Q, et al. Assessment of the
2010;41(5):588-95.
[9] Kulkani A, Satyanarayana K, Rohatgi P. Weibull analysis of coir fibres. Fibre Sci Technol.
1983;19:59-76.
206 Appendix 1
[10] Xia ZP, Yu JY, Cheng LD, Liu LF, Wang WM. Study on the breaking strength of jute fibres
using modified Weibull distribution. Composite Part A: Appl Sci Manuf. 2009;40(1):54-9.
[11] Placet V. Characterization of the thermo-mechanical behaviour of Hemp fibres intended for the
2009;40(8):1111-8.
[13] Wu D, Zhou J, Li Y. Unbiased estimation of Weibull parameters with the linear regression
method. Journal of the European Ceramic Society. 2006;26(7):1099-105.
[14] Bergman B. On the estimation of the Weibull modulus. Journal of Materials Science Letters.
1984;3(8):689-92.
[15] Andersons J, Sparnins E, Joffe R, Wallstrom L. Strength distribution of elementary flax fibres.
Compos Sci Technol. 2005;65:693-702.
[16] Rinne H. The Weibull distribution: a handbook: CRC Press; 2010.
Appendix 2 207
Appendix 2
A2.1 X-ray photoelectron spectroscopy (XPS) methodology
The XPS analyses were carried out at the division of Bio- and Soft Matter (BSMA)
of the Institute of Condensed Matter and Nanosciences (IMCN) of the University of
Louvain (UCLouvain). Bamboo fibres were fixed overhanging on a flat stainless
steel trough with an insulating double sided adhesive tape. The analysed zone was
situated on the overhanging part of the fibre, which totally prevents any contribution
of the sample support to the recorded signal. The signal was carefully optimized in
X, Y and Z. Analyses were repeated in triplicate. XPS analyses were performed on a
Kratos Axis Ultra spectrometer (Kratos Analytical – Manchester – UK) equipped
with a monochromatized aluminium X-ray source (powered at 10 mA and 15 kV)
and an eight channeltrons detector. The spectrometer was interfaced with a HP PC
workstation. Instrument control and data acquisition were performed with the
Vision2 program.
The pressure in the analysis chamber was about 10-6 Pa. The angle between the
normal to the sample surface and the direction of photoelectrons collection was
about 0°. Analysis was performed in the hybrid lens mode (a combination of
magnetic and electrostatic lenses) with the slot aperture and the iris drive position
set at 0.5, the resulting analysed area was 700 µm x 300 µm. The pass energy of the
hemispherical analyser was set at 160 eV for the survey scan and 40 eV for narrow
scans. In the latter conditions, the full width at half maximum (FWHM) of the Ag
3d5/2 peak of a standard silver sample was about 0.9 eV. Charge stabilisation was
achieved by using an electron source mounted co-axially to the electrostatic lens
column and a charge balance plate used to reflect slow electrons back towards the
sample. The magnetic field of the immersion lens placed below the sample acts as a
guide path for the low energy electrons returning to the sample. The electron source
was operated at 0.16 A filament current and a bias of –1.2 eV. The charge balance
plate was set at 4 eV.
Powdered samples were deposited on a double side adhesive tape fixed on a piece of
polymer allowing complete insulation of the samples. The following sequence of
spectra was recorded: survey spectrum, C 1s, O 1s, N 1s, Ca 2p, Zn 2p, Si 2p, P 2p
and C 1s again to check for charge stability as a function of time and the absence of
degradation of the sample during the analyses. To set the binding energy scale, the
C-(C,H) component of the C1s peak of carbon has been fixed to 284.8 eV.Spectra
were decomposed with the CasaXPS program (Casa Software Ltd., UK) with a
208 Appendix 2
Gaussian/Lorentzian (70/30) product function after subtraction of a linear baseline.

Molar fractions (%) were calculated using peak areas normalised on the basis of
acquisition parameters, experimental sensitivity factors and transmission factors
(depending on kinetic energy, analyser pass energy and lens combination) provided
by the manufacturer.
A2.1.1 Thermal treatment for samples analyzed with XPS
Untreated and treated technical bamboo fibres in air, nitrogen and argon
environments were analyzed with the XPS technique. Thermal treatments were
performed in a horizontal tube furnace, starting at room standard conditions. When
using argon, and in order to assure the absence of air in the cavity, a constant flow of
gas was passed through the furnace for 30 minutes before starting from room
temperature up to 220°C, at a constant temperature ramp of 5 °C/min, where it is
held during 40 min. The temperature was monitored with a thermocouple placed
close to the sample holder.
A2.2 Results
A2.2.1 XPS analysis
The XPS technique was used to characterize the surface chemistry of bamboo of
untreated and thermally treated bamboo fibres in two different environments (i.e. air
and argon). Table A2-1 presents the surface components of the treated and untreated
fibres detected by XPS, consisting of relative atomic percentages of the elements,
oxygen-carbon ratio and the decomposition of the C 1s peak into four sub-peaks C1-
C4. According to the number of oxygen atoms bonded to C, those peaks represent
[1-3]: carbon solely linked to carbon involved in ester or hydrogen C-C or C-H (C1),
carbon singly bound to oxygen or nitrogen C-O or C-N (C2), carbon doubly bound
to oxygen O-C-O or C=O (C3) and carbon involved in ester or carboxylic acid
functions O=C-O (C4). The decomposed C1s spectra for untreated and thermally
treated bamboo technical fibres in three different environments are shown in Figure
A2.1.
In Figure A2-2 shows the results of the surface chemical constituents of untreated
and treated (air and argon environments) bamboo technical fibres obtained from the
decomposition of the high resolution carbon 1s spectrum for each fibre. The content
of lignin on the surface was analysed by the O/C (oxygen to carbon ratio) and C1
(the relative concentration of the C1 component). As noticed, the atomic ratios for
cellulose and lignin are different and also pure cellulose lacks of C1 bonds.
Appendix
Appendix2 2 209 209
Thermal Number C O N Si O/C Binding energy (eV)

treatment of
atmosphere samples
(%) (%) (%) (%) (%) 284.8 286.3 287.5 288.8
C1 (%) C2 (%) C3 (%) C4 (%)
(C-C/C-H) (C-O) C=O/O-C-O O-C=O
Untreated fibre* 10 74.3± 1.5 22.9 ± 0.3 1.8 ± 0.7 0.6 ± 0.4 0.31 ± 0.02 58.0 ± 3.07 28.8 ± 2.3 7.6 ± 1.3 5.6 ± 0.4
Bamboo (air) 3 79.1 ± 1.6 16.7 ± 0.8 2.1 ± 0.4 1.7 ± 0.7 0.21 ± 0.07 53.3 ± 3.7 31.2 ± 3.8 10.5 ± 1.9 4.9 ± 0.4
Bamboo (Ar) 3 79.3 ± 0.5 16.7 ± 0.6 2.5 ± 0.5 1.0 ± 0.8 0.21 ± 0.08 52.5 ± 3.5 34.3 ± 2.1 9.3 ± 0.9 3.7 ± 0.6
Table A2-1. Relative atomic percentages (excluding hydrogen), O/C ratio and decomposition of C 1s peaks for untreated and thermally treated bamboo fibres in air and
argon atmospheres obtained by XPS. *Values obtained by Fuentes et al [4].
1_B_Argon_1 : C 1s start/6 1_B_Argon_2 : C 1s start/6

1_B_Air_1 : C 1s start/6 1_B_Air_2 : C 1s start/6 1_B_Air_3 : C 1s start/6
90 70
45
50
70
80
45 40 60
60
70 40
Intensity (102 counts/s)

35

50
60 35 50
30
40
30
50 40 25
25
20 30
40
30
20
30 15
20
15 20
10
20
10
10 10
5
10 5
296 292 288 284 280 296 292 288 284 280
296 292 288 284 280 296 292 288 284
Binding 280rg y (eV)
Ene 296 292 288 284
Binding 280
Ene rg y (eV)
Binding Ene rg y (eV) Binding Ene rg y (eV) Binding Ene rg y (eV)
Untreated fibre Treated in air Treated in argon

Figure A2-1. Typical C1s spectra comprising of the decomposition into four components C1 – C4 for a) untreated bamboo fibre [5], b) thermally treated bamboo fibre
in air and c) in argon (Ar).
210 Appendix 2
The ratios between the total amount of oxygen and carbon (O/C) of 0.21 and 0.21
for fibres treated in air and argon respectively, are close to the ones reported
previously by Fuentes et al. [5], ratio O/C of 0.31, for untreated bamboo technical
fibres. These values, clearly lower than 1, could reveal the abundant presence of C-
O-C ether functions which are characteristic of the presence of lignin at the surface
of the fibers [6].
Hypothetically, if only cellulose and lignin are present on the fibre surface, the
results show that the surface constituents for the analyzed samples are close to the
reference materials for lignin (see Figure A2.2). This indicates that the technical
fibre is cover by lignin and possibly other molecules but not cellulose. This is in
agreement with the studies of Fuentes et al. [4, 5] in which is stated that the surface
of the technical bamboo fibres (same as used in this study), is composed by lignin
instead of cellulose or hemi-cellulose. This is supported by the value of the ratio
O/C of 0.21 for both treated fibres and the untreated fibre (0.31) (see Table A2.1).
This value is close to the value ranges reported for lignin (0.24 – 0.36) [6, 7] and far
different from the O/C ratio of pure cellulose 0.83 [8]. Finally, the results obtained
from this XPS study indicate that the thermal treatment for technical bamboo fibres
in air and argon does not change the superficial components of the fibres, in
comparison with the untreated samples.
Figure A2.2. A correlation graph between the percentage of C1/C ratio versus O/C ratio for chemical groups
at the surface of untreated and thermally treated bamboo technical fibres (in air and argon), lignin form Granit
and theoretical values for cellulose and lignin according to Shchukarev [8].
Appendix 2 211
References
[1] Inari GN, Petrissans M, Lambert J, Ehrhardt JJ, Gérardin P. XPS characterization of wood
chemical composition after heat-treatment. Surf Interface Anal. 2006;38(10):1336-42.
[2] Kazayawoko M, Balatinecz J, Sodhi R. X-ray photoelectron spectroscopy of maleated
polypropylene treated wood fibers in a high-intensity thermokinetic mixer. Wood Science and
Technology. 1999;33(5):359-72.
[3] Shen Q, Mikkola P, Rosenholm JB. Quantitative characterization of the subsurface acidâ€“base
properties of wood by XPS and Fowkes theory. Colloids and Surfaces A: Physicochemical and
Engineering Aspects. 1998;145(1):235-41.
and Bioenergy. 2012;6(4):456-62.
[5] Fuentes CA, Tran LQN, Dupont-Gillain C, Vanderlinden W, De Feyter S, Van Vuure AW, et al.
Wetting behaviour and surface properties of technical bamboo fibres. Colloids Surf A: Physicochem
Eng Aspects. 2011;380:89-99.
[6] Johansson L-S, Campbell J, Koljonen K, Stenius P. Evaluation of surface lignin on cellulose
fibers with XPS. Appl Surf Sci. 1999;144:92-5.
[7] Tran LQN, Fuentes CA, Dupont-Gillain C, Van Vuure AW, Verpoest I. Wetting analysis and
surface characterisation of coir fibres used as reinforcement for composites. Colloids Surf A:
[8] Shchukarev A, Sundberg Br, Mellerowicz E, Persson P. XPS study of living tree. Surf Interface
Anal. 2002;34(1):284-8.
Curriculum Vitae
Personal data
Name: Efraín Eduardo Trujillo De Los Ríos
Nationality Colombian
Address: Ijzerenmolenstraat 22/301
3001 Heverlee
Belgium
e-mail: eduardo.trujillo@mtm.kuleuven.be
etrujillod@hotmail.com
Education
2009 - present PhD., Materials Engineering, Katholieke Universiteit Leuven,
Belgium.
Dissertation: Polymer composite materials based on bamboo fibres.
2005 - 2007 M.Sc. Materials Engineering: Polymers and Composites (EUPOCO
program), Katholieke Universiteit Leuven, Belgium.
Thesis: Bamboo fibre composites.
1999 - 2004 B.Sc., Industrial Engineering, Universidad Nacional de Colombia.
Thesis: Physical-chemical characterization of the fibre of Guadua
angustifolia Kunth.
Awards
2014 Colciencias, Locomotora de la innovación para el apoyo del
desarrollo tecnológico (call 642). Development of a bamboo fibre
reinforcement to be used in high-tech composite applications,
Colombia.
2011 “Winner of the award for the best exploitation plan”. Doctoral
school training course “Exploitation of Research – Technology &
Knowledge Transfer”, Belgium.
2010 "JEC - SAMPE Student Award" for the best student selected on
subject environmental aspects of composites, France.
2005 “Development and Co-operation Prize 2005”, financed by the
Belgian Development Cooperation for young scientists, based on a
scientific work of high relevance for development, Belgium.
2004 Laureate bachelor’s thesis of Industrial Engineering at Universidad
Nacional de Colombia, Manizales.
Publications
Journal papers
1. E. Trujillo, M. Moesen, L. Osorio, A.W. Van Vuure, J. Ivens, I. Verpoest. Bamboo fibres
for reinforcement in composite materials: Strength Weibull analysis. Compos Part A:
Appl Sci Manuf. 2014;61:115-25.
2. E. Trujillo, M. Moesen, L. Osorio, A.W. Van Vuure, J. Ivens, I. Verpoest. Statistical
analysis of the mechanical properties bamboo fibres obtained by the dry fibre bundle
test, (2014). (To be submitted).
3. L. Osorio, E. Trujillo, A.W. Van Vuure, F. Lens, J. Ivens, I. Verpoest. Composite
micromechanics explain the behaviour of bamboo fibres, (2014). (To be sumitted).
4. L. Osorio, E. Trujillo, A.W. Van Vuure, J. Ivens, I. Verpoest. Morphological aspects and
mechanical properties of single bamboo fibres and flexural characterization of
bamboo/epoxy composite. Journal of Reinforced Plastics and Composites 2011; 30 (5):
396 –408.
Conference proceedings
1. E. Trujillo, D. Perremans L. Osorio, A.W. Van Vuure, J. Ivens, I. Verpoest, (2014).
Characterization of unidirectional discontinuous bamboo fibre/epoxy composites. ECCM
16 European Conference on Composite Materials, Seville, Spain.
2. E. Trujillo, J. Vertommen, L. Osorio, A.W. Van Vuure, J. Ivens, I. Verpoest, (2013).
Investigating the flexural properties of bamboo fibre – PP composites consolidated under
inert atmosphere. ICCM-19 International Conference on Composite Materials. Montreal,
Canada.
3. E. Trujillo, L. Osorio, A.W. Van Vuure, J. Ivens, I. Verpoest, (2013). Assessment of the
properties of bamboo fibre and bamboo fibre composites. In Lomov,
S. (Ed.), Composites Week @ Leuven. Composites Week @ Leuven. Leuven, Belgium.
4. E. Trujillo, L. Osorio, A.W. Van Vuure, J. Ivens, I. Verpoest. Bamboo (Guadua
angustifolia) fibres for storng-light composite materials. 9th World Bamboo Congress.
Antwerp, Belgium.
5. E. Trujillo, M. Moesen, L. Osorio, A.W. Van Vuure, J. Ivens,
I. Verpoest, (2012). Weibull statistics of bamboo fibre bundles: methodology for tensile
testing of natural fibres. In Quaresimin, M. (Ed.), 15th European Conference on
Composite Materials. European Conference on Composite Materials. Venice, Italy.
6. E. Trujillo, L. Osorio, A.W. Van Vuure, J. Ivens, I. Verpoest, (2010). Characterisation
of polymer composite materials based on bamboo fibres. Proceedings ECCM-14, 14th
European Conference on Composite Materials. European Conference on Composite
Materials. Budapest, Hungary.
7. E. Trujillo, L. Osorio, C.A. Fuentes, A.W. Van Vuure, I. Verpoest, (2010). Bamboo fibre
thermoplastic composites for transport applications. Society for the Advancement of
Material and Process Engineering, SAMPE Europe. Paris. France.

Thesis Bamboo Fibres E Trujllo 24082014

Cargado por

Información del documento

Derechos de autor

Formatos disponibles

Compartir este documento

Compartir o incrustar documentos

Opciones para compartir

¿Le pareció útil este documento?

¿Este contenido es inapropiado?

Copyright:

Formatos disponibles

Thesis Bamboo Fibres E Trujllo 24082014

Cargado por

Copyright:

Formatos disponibles

POLYMER COMPOSITE MATERIALS

BASED ON BAMBOO FIBRES

Eduardo TRUJILLO DE LOS RÍOS

Members of the Examination

Chapter 1: General introduction………………………………………………………... 1

3.6.2 Mechanical extraction………………………………………………………... 44

Chapter 5: Unidirectional continuous and discontinuous bamboo fibre

6.3.2 Thermogravimetric analysis (TGA)....…………...………………………….. 144

Chapter 8: General conclusions……….……………………...…………………………. 195

As an example, transportation by road contributes in average about one-fifth of the

The present research “polymer composite materials based on bamboo fibres”

them, such as variability in fibre mechanical properties, the thermal degradation of

 To develop operating prototype machines incorporating environmentally

For the thermal characterization, the effect of typical composite manufacturing

In Chapter 6, bamboo fibre thermoplastic (TP) matrix composites using PP and

3.1 Natural fibres

3.1.1 Fibre microstructure

3.1.2 Chemical composition of natural fibres

Cellulose Hemicellulose Lignin

Cellulose is a natural polymer whose elementary macromoleule unit is the anhydro-

Figure 3-3. Cellulose structure [7].

The third main constituent is lignin, with a complex three-dimensional copolymer of

3.2 Mechanical properties of plant fibres

Table 3-2 presents some characteristics and results in mechanical properties of

Microfibril Elongation Tensile Young’s

The industrial adoption of natural resources for reinforcing composites is an active

3.3 Natural fibre composites and their advantages

Nowadays, a substantial amount of research is carried out to analyse the possibility

The composite material performance in the final application, depends on several

Before treatment After treatment

UD Flax + epoxy 3% NaOH, 20 min at 20 °C 40 218 18 283 22 30/22 [37]

Before treatment After treatment

System Process parameters Vf (%) % MA Flexural properties Ref.

UD Jute – PP 140, 150 and 160°C 49 - 101 11 - [44]

UD Jute yarns-PP 180°C 10 min 21 - 122 9 - [45]

UD Flax-PP 12 min pre-heating, 27 - 129±19 16± 2 1.2 [46]

UD Flax-MAPP 12 min pre-heating, 28 0.6 164±56 14± 6 2.0 [46]

UD Flax-MAPP 12 min pre-heating, 32 0.6 212±14 23± 2 1.4 [46]

UD Glass-MAPP 200°C 45min 58 10 1037 32 3.3 [47]

UD Glass-PP 200°C 45min 58 - 774 38 2.1 [47]

System Process parameters Vf (%) % MA Tensile properties Ref.

UD Hemp-MAPP 180 °C 3min 30 10 37 12 0.5 [48]

UD Hemp-PP Film stacking 35 - 125 - - [17]

UD Hemp-MAPP Film stacking 35 0.5 122 - - [17]

3.3.1 Environmental impact

 Carbon dioxide fixation of natural fibres is an important issue in the reduction of

The development of bio-composites based on biopolymers and natural fibres as

3.3.2 Coupling agents

Studies on composite materials have shown that the properties of composite

3.3.2.1 Alkali treatment

Alkali treatment in single fibres

Alkalization cannot be considered as an authentic physical modification since the

Fibre – OH + NaOH  Fibre – O- – Na+ + H2O

The important modification is the removal of hydroxyl groups in the network

Before treatment After treatment

Sisal - 500 16 810 27 16%, 48h at RT [18]

Curauá 94 913 30.3 736 27 5%, 1h at RT [70]

Palm 180 - 208 233 7 366 5 6%, 3h at 95 °C [71]

Effect of the alkali treated fibres on natural fibre composite properties

3.3.2.2 Maleic anhydride

According to Chattopadhyay et al. [78], short bamboo fibre reinforced

3.4 Limitations in the use of natural fibre composites

3.4.1 Moisture absorption

The effect of moisture on the mechanical properties of natural fibre composites