CHARACTERISTICS AND BEHAVIOUR OF GOMUTI FIBRE COMPOSITES WITH THERMOSET RESIN

Transcription

1 CHARACTERISTICS AND BEHAVIOUR OF GOMUTI FIBRE COMPOSITES WITH THERMOSET RESIN A dissertation submitted by Adriana Nanow Everce Ticoalu BEng (Civil), MEng (Civil) For the award of Doctor of Philosophy Principal supervisor Prof Thiru Aravinthan Associate supervisor Dr Francisco Cardona School of Civil Engineering and Surveying Faculty of Health, Engineering and Sciences University of Southern Queensland Australia 2013

2 ABSTRACT Gomuti fibre, as a natural fibre obtained from the Arenga pinnata tree, is a versatile fibre that has been used conventionally for various purposes such as thatch roofing, water filters, brooms, rope and insulation. Its usage has diminished over time due to the introduction of some more durable, installation-easy and aesthetically pleasing roofing materials such as metal, plastic and concrete. With the current reemerging interest in natural fibres and their composites as potential building and construction materials, there exists the opportunity to investigate the potential of gomuti fibre. The main objective of this study is to evaluate the characteristics and behaviour of gomuti fibre composites with polymer thermoset resins, and investigate its feasibility to be developed as a building and construction material. The experimental study covers firstly, the investigation on the strength, stiffness and thermal stability of gomuti fibre itself. Secondly, the strength and stiffness of gomuti fibre composites and thirdly, the impact strength and behaviour of gomuti fibre composites. The experimental results demonstrated the effects of fibre treatment and the use of different thermoset resin matrices on the properties of gomuti fibre. Lastly, a parametric study on the mechanical properties of a flat roof tile was performed to evaluate the feasibility of developing gomuti fibre composites into a building and construction material. Investigation on the characteristics of gomuti fibre has revealed that gomuti fibre has an average single fibre tensile strength of MPa, Young s modulus of 3.85 GPa, and elongation of 12.8 %. The results are particularly comparable with coir fibre. Fibre treatment with sodium hydroxide resulted in increased fibre density, strength, modulus and thermal stability, whilst reducing the diameter. Furthermore, 5 A b s t r a c t i

3 % treatment was found to give the optimal properties although there were no significant variations between the results of 5 % and 10 % treatment amount. Improvement of the properties was observed by treating the fibre with sodium hydroxide however, the results indicated that a higher concentration of sodium hydroxide may damage the fibre causing a reduction in properties. With variations of fibre treatment and different thermoset resin (polyester, vinylester and epoxy), overall, the tensile strength of gomuti fibre composites was found to be between 28.4 MPa and 48.5 MPa, and Young s modulus between 4.32 GPa and 6.25 GPa. Compressive strengths of the composites were found to be approximately 2 4 times the tensile strengths; the flexural strengths approximately 2 times the tensile strengths, while shear strengths were approximately times the tensile strengths. The use of different resins resulted in significant variations in the mechanical properties while fibre treatment caused minor variations. The composites with treated fibre show various responses but, overall, improvement was seen on composites with 5 % alkali treated fibre. Results of strength and stiffness evaluation confirm that while fibre treatment has obvious effects to the composite properties, the difference between 5 % and 10 % of NaOH was not significant. This testing result relates to the strength of single fibre. Furthermore, results from mechanical testing indicated that gomuti fibre composites may offer more significant benefits when used for products with low to moderate strength requirements. A similar inference was found on the results of impact testing. A study of impact strength characteristics of gomuti fibre composites found that, overall, the average impact strength varies from 15 kj/m 2 to 71.7 kj/m 2. Samples with untreated fibre exhibited higher impact strength compared to samples with treated fibres, except for 10% NaOH treated fibre with vinylester. Samples with epoxy resin A b s t r a c t ii

4 exhibit lower impact strength compared to samples with polyester and vinylester, except for untreated fibre with epoxy which exhibited higher value compared to samples with vinylester. The higher impact strength relates to inferior fibre-matrix bonding, while lower impact strength can be attributed to better fibre-matrix bonding. It was found that although strength and stiffness of gomuti fibre composites are below than that of glass fibre, they are comparable to similar natural fibre composites such as coir fibre composites. The functionality of gomuti fibre is considered enhanced when processed into natural fibre polymer composites. After evaluating the mechanical properties of gomuti fibre composite samples which has provided important assessments of the strength, stiffness and impact strength of the composites, the feasibility of developing gomuti fibre composites into a flat roof tile was studied by parametric design. The results indicated that, with a tile dimension of 330 mm x 420 mm and under the transverse bending loading conditions, the stiffness of tile is the major criteria in the design of a roof tile. A graph was constructed where a design thickness of tile can be obtained based on the value of modulus, or to obtain the values of modulus and design strength based on thickness. By using the graph, it was found that for gomuti fibre composites with a range of modulus from 4.3 to 6.3 GPa, the required minimum thickness is approximately from 13.5 to 15.4 mm and strength of 6.5 to 8.4 MPa. For the design of gomuti fibre composite roof tile, the stiffness was found to be the governing criteria. While strength criteria is not a governing factors in the design, when the design strength incorporate a safety factor, strength of the composites will become an important criteria. In special cases when a material exhibits high modulus and low strength, design based on the strength is likely to become major criteria. A b s t r a c t iii

5 CERTIFICATION OF DISSERTATION I certify that the ideas, experimental works, results, analyses, software and conclusions reported in this dissertation are entirely my own effort, except where otherwise acknowledged. I also certify that the work is original and has not been previously submitted for any other award, except where otherwise acknowledged. Signature of Candidate Date ENDORSEMENT Signature of Principal Supervisor Date Signature of Associate Supervisor Date

6 ACKNOWLEDGEMENTS I would like to convey sincere gratitude to my principal supervisor, Professor Thiru Aravinthan, for the guidance, critical suggestions and consistent support during this PhD study. The ideas, help and discussion with my associate supervisor, Dr. Francisco Cardona are sincerely acknowledged. I am thankful for the financial support of the Indonesian Government for my PhD study through Directorate General of Higher Degree (DGHE) Scholarship. For the generous advice and help from Dr Allan Manalo, I am very grateful. I thank Mazed Kabir for advice and help in some experimental works, and Ernesto Guades for the helpful advice on impact test data analysis. The kind assistance in experimental works and administrative support from Wayne Crowell, Mohan Trada, Adrian Blokland, Martin Geach and Juanita Ryan are greatly appreciated. The assistance in proof-reading of the thesis manuscript provided by Ms. Sandra Cochrane is hereby acknowledged. For the sincere friendship and beneficial discussions with the staff and fellow postgraduate students in the CEEFC, I feel really blessed. I would like to thank my mum, dad, my brother Arthur and Dr Julius Pontoh from Sam Ratulangi University Manado, for the help in making sure a sufficient amount of gomuti fibre was prepared for my research. For the love, prayers and support in numerous ways from my church family, Rangeville Community Church (RCC) Toowoomba and from the Thursday night home-group where I share my ups and downs, I am forever grateful. My vast gratitude is for the love and cheering from my husband, Mario and our daughter, Kearney, that have made this study a challenge worthy to prevail. I wish to express my sincere gratitude to all the people I have failed to mention here but who have been part of this study in various ways. Lest I boast in myself, my utmost praise is to the chief Cornerstone, the everlasting Foundation; the Way, the Truth and the Life; the living and wonderful Redeemer; who has made the impossible possible. A c k n o w l e d g e m e n t s v

7 ASSOCIATED PUBLICATIONS Journal papers 1. Ticoalu, A, Aravinthan, T & Cardona, F 2013, A review on the characteristics of gomuti fibre and its composites with thermoset resins, Journal of Reinforced Plastics and Composites, vol. 32, no. 2, pp Ticoalu, A, Aravinthan, T & Cardona, F 2013, A study into the characteristics of gomuti (Arenga pinnata) fibre for usage as natural fibre composites, Journal of Reinforced Plastics and Composites, published online 8 October 2013, DOI: / Ticoalu, A, Aravinthan, T & Cardona, F 2013, Mechanical properties of gomuti fibre composites: the influence of fibre treatment and different thermoset resins, (Under review for publication in journal Composite Structures). 4. Ticoalu, A, Aravinthan, T & Cardona, F 2013, An experimental study on the impact characteristics of gomuti fibre composites (Under review for publication in journal Composite Structures). Peer-reviewed conference papers 1. Ticoalu, A, Aravinthan, T & Cardona, F 2013, A review of current development in natural fibre composites for structural and infrastructure applications, Proceedings of the 2010 Southern Region Engineering Conference (SREC 2010), Nov 2010, Toowoomba, Queensland, Australia. 2. Ticoalu, A, Aravinthan, T & Cardona, F 2013, Experimental investigation into gomuti fibres/polyester composites, Proceedings of the 21st Australasian Conference on the Mechanics of Structures and Materials (ACMSM 21), 7-10 December 2010, Melbourne, Australia. 3. Ticoalu, A, Aravinthan, T & Cardona, F 2013, Properties and behaviour of gomuti fibre composites under tensile and compressive load, Proceedings of the 22nd Australasian Conference on the Mechanics of Structures and Materials (ACMSM 22), December 2012, Sydney, Australia. vi

8 TABLE OF CONTENTS Abstract Certification of dissertation Acknowledgements Associated Publications Table of contents List of figures List of tables i iv v vi vii xii xvi CHAPTER 1 Introduction General Background Objectives Scope Structure of dissertation Summary 9 CHAPTER 2 Literature review on the characteristics of gomuti fibre, its composites and potential applications Introduction Traditional and existing usage of gomuti fibre Characteristics of gomuti fibre Physical characteristics Morphology of gomuti fibre Diameter of gomuti fibre Density of gomuti fibre Mechanical properties of single gomuti fibre Chemical composition and thermal stability of gomuti fibre 18 T a b l e o f c o n t e n t s vii

9 2.4 Characteristics of gomuti fibre composites with polymer thermoset resins Characteristics of thermoset resins Fibre treatment and composite fabrication Tensile strength and stiffness characteristics Prediction of tensile properties of gomuti fibre composites Flexural properties Impact properties Potential applications of gomuti fibre composites Summary 35 CHAPTER 3 Investigation into the characteristics of gomuti fibre Introduction Experimental methods Fibre material Fibre treatment Observation of fibre morphology and diameter measurement Measurement of fibre density Single fibre tensile test Measurement of fibre thermal stability Experimental results Morphology of gomuti fibre Diameter of gomuti fibre Density of gomuti fibre Single fibre tensile properties Thermal stability of gomuti fibre Discussions Effects of treatment to the morphology of gomuti fibre Effects of treatment to the diameter of gomuti fibre Effects of treatment to the density of gomuti fibre Effects of treatment to the tensile properties of single gomuti fibre Effects of treatment to the thermal stability of gomuti fibre 67 T a b l e o f c o n t e n t s viii

10 3.5 Conclusions 68 CHAPTER 4 Investigation on the strength and stiffness of gomuti fibre composites with thermoset resins Introduction Experimental methods Materials Fibre treatment Preparation and fabrication of composite panels Testing methods Tensile test Compressive test Flexural test Shear test Experimental results Tensile properties and load-extension behaviour Compressive properties and load-extension behaviour Flexural properties and load-displacement behaviour Shear properties and load-displacement behaviour Discussions Effects of fibre treatment and resin variation Comparison of theoretical prediction of mechanical properties based on Rule of Mixture (ROM) methods and experimental results Failure of gomuti fibre composites Conclusions 108 CHAPTER 5 Investigation on the impact strength and behaviour of gomuti fibre composites Introduction Materials and methods 116 T a b l e o f c o n t e n t s ix

11 5.2.1 Material Specimen preparation Drop-weight impact test Experimental results Damage characteristics Impact force-time characteristics Absorbed energy-time characteristics Impact force-deflection behaviour Discussions Impact strength Effects of fibre treatment Effects of different thermoset resins Conclusions 136 CHAPTER 6 Strength and stiffness design of gomuti fibre composites roof tile Introduction Overview of roofing materials and roof tiles Performance requirement of roof tile Design of gomuti fibre composites roof tiles Behaviour of gomuti fibre composites roof tile Gomuti fibre composites roof tile design guidelines Conclusions 161 CHAPTER 7 Conclusions General Characteristics of gomuti fibre and the effects of alkali treatment Strength and stiffness characteristics of gomuti fibre composites with thermoset polymer resins Impact strength and behaviour characteristics of gomuti fibre composites with thermoset polymer resins 165 T a b l e o f c o n t e n t s x

12 7.5 Gomuti fibre composite roof tile Recommendation for future research 167 LIST OF REFERENCES 170 APPENDICES Appendix A: Overview of natural fibres and their composites Appendix B: Fibre characterisation details Appendix C: Mechanical properties testing details Appendix D: Impact test data Appendix E: Simulation of gomuti fibre composites roof tiles A1 B1 C1 D1 E1 T a b l e o f c o n t e n t s xi

13 LIST OF FIGURES Figure 1.1 Simplified illustration of the research flow Figure 2.1 Figure 2.2 Figure 2.3 Figure 2.4. Figure 2.5. Arenga pinnata tree with gomuti covering the tree trunk Gomuti fibre: (a) Layers of trimmed intermingled long fibre, and (b) chopped fibre. Gomuti thatch roof in Bali, Indonesia (Original photo caption: Shrines in the inner sanctuary of Tirta Empul, courtesy of Timothy Tye, used with permission) Surface morphology of gomuti fibre under optical microscope. Illustration of stress-strain of gomuti fibre and thermoset resin. Figure 3.1 Figure 3.2 Single fibre observation: (a) schematic of specimen arrangement for microscope observation; and (b) schematic of specimen mounting tab for single fibre tensile test (SFTT) using dynamic mechanical analyser (DMA) machine. Microscope observation of fibre cross-section. Figure 3.3 Surface morphology of untreated (UN) and treated (5N and 10N) gomuti fibre. Figure 3.4 Figure 3.5 Figure 3.6 Figure 3.7 Figure 3.8 Figure 3.9 Figure 3.10 Figure 3.11 Figure 3.12 Figure 3.13 Figure 3.14 Cross-section of gomuti fibre. Fibre diameter measurement. Example of stress-strain curves of single-fibre-tensile tested specimens. TGA plots of untreated gomuti fibre. TGA plots of 5% NaOH treated gomuti fibre. TGA plots of 10% NaOH treated gomuti fibre. DSC plot of untreated and treated gomuti specimens. Boxplot of gomuti fibre diameter data. Boxplot of the tensile strength of gomuti fibre. Boxplot of the Young s modulus of gomuti fibre. Boxplot of the strain of gomuti fibre. L i s t of f i g u r e s xii

14 Figure 3.15 Figure 3.16 Figure 3.17 Figure 3.18 Weibull probability plot of the tensile strength of untreated and alkali-treated gomuti fibre. Weibull probability plot of the tensile modulus of untreated and alkali-treated gomuti fibre. Weibull probability density function (pdf) of the tensile strength of untreated and alkali-treated gomuti fibre. Weibull probability density function (pdf) of the tensile modulus of untreated and alkali-treated gomuti fibre. Figure 4.1 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 Figure 4.6 Figure 4.7 Figure 4.8 Figure 4.9 Figure 4.10 Figure 4.11 Figure 4.12 Figure 4.13 Figure 4.14 Figure 4.15 Figure 4.16 Figure 4.17 Figure Unidirectional aligned gomuti fibre. Test set-up: (a) tensile; (b) compressive; (c) flexural and (d) shear. Schematic of tensile specimen and idealised loading condition. Schematic of compressive specimen and idealised loading condition. Schematic of flexural specimen and idealised loading condition. Schematic of shear specimen and idealised loading condition. Comparison of tensile strength Comparison of tensile modulus. Boxplot of tensile strength. Load-extension plots of typical tensile-tested specimen: (a) specimens with polyester resin, (b) specimens with epoxy resin, and (c) specimens with vinylester resin. Comparison of compressive strength Comparison of compressive modulus Boxplot of compressive stress Load-extension plots of typical compressive-tested specimen: (a) specimens with polyester resin; (b) specimens with epoxy resin; and (c) specimens with vinylester resin. Comparison of flexural strength Comparison of flexural modulus Boxplot of flexural stress. Load-displacement plots of typical flexure-tested specimen: (a) specimens with polyester resin; (b) specimens with epoxy resin; and (c) specimens with vinylester resin. L i s t of f i g u r e s xiii

15 Figure 4.19 Figure 4.20 Figure 4.21 Figure 4.22 Figure 4.23 Figure 4.24 Figure 4.25 Figure 4.26 Figure 4.27 Figure 4.28 Comparison of shear strength. Comparison of shear modulus. Boxplot of shear stress Load-displacement plot of typical shear-tested specimen: (a) specimens with polyester resin; (b) specimens with epoxy resin; (c) specimens with vinylester resin. Comparison of predicted and test result of tensile strength Comparison of predicted and test result of tensile modulus. Failure of tensile specimen. Failure of compressive specimen. Failure of flexural specimen. Failure of shear specimen. Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 Figure 5.7 Figure 5.8 Figure 5.9 Dimension of specimen for impact test. Drop-weight impact test set-up: (a) securing specimen to the test machine; (b) mounted specimen before impact; and (c) specimen after impact. Example pictures of the impacted specimens. Damage mode of a typically impact-tested specimen taken from the impacted point (mid-span). Example of force vs time and energy vs time plots of impacted gomuti fibre composites specimen. Force/area vs time plots of typical impact-tested specimen: (a) polyester specimen; (b) epoxy specimen; and (c) vinylester specimen. Energy/area-time plots of typical impact-tested specimen: (a) polyester specimen; (b) epoxy specimen; and (c) vinylester specimen. Force-deflection curve of typical impact-tested gomuti fibre composites. Force/area-deflection plots of typical impact-tested specimen: (a) polyester specimen; (b) epoxy specimen; and (c) vinylester specimen. L i s t of f i g u r e s xiv

16 Figure 5.10 Figure 5.11 Figure 5.12 Percentage comparison of average impact strength. Comparison of the impact strength of gomuti fibre composite samples: effects of fibre treatment. Comparison of the impact strength of gomuti fibre composite samples: effects of different thermoset resin. Figure 6.1 Figure 6.2 Figure 6.3 Figure 6.4. Figure 6.5. Figure 6.6 Figure 6.7 Figure 6.8 Figure 6.9 Figure 6.10 Figure 6.11 Figure 6.12 Various types of roof tiles Schematic of roof tile arrangement (front view) Schematic of roof tile arrangement (top view) Transverse bending strength test configuration Flat roof tile profile Deflection along the length of tile for different modulus and different thicknesses. E-tile vs deflection plot for the thickness variations E-tile vs deflection plot for the thickness variations Behaviour of a roof tile with variations of elastic moduli P-max calculated based on strength of the tile and thickness variations. Plate thickness vs plate stress Roof tile design curves L i s t of f i g u r e s xv

17 LIST OF TABLES Table 2.1 Table 2.2 Table 2.3 Table 2.4 Table 2.5 Table 2.6 Density and diameter of natural fibres compared to glass fibre. Single fibre properties of natural fibres compared to glass fibre. Chemical compositions of various natural fibres compared to gomuti fibre. Typical tensile and flexural properties of some common thermoset polymers. Tensile properties of gomuti fibre composites with thermoset resins. Flexural properties of gomuti fibre composites with thermoset resins. Table 3.1 Table 3.2 Table 3.3 Table 3.4 Table 3.5. Table 3.6 Descriptive statistics of diameter of untreated and alkali treated gomuti fibre (n = 25). Density of untreated and treated gomuti fibre Average tensile strength, Weibull parameters and goodness-of-fit parameter of gomuti fibre samples. Average Young s modulus, Weibull parameters and goodness-of-fit parameter of gomuti fibre samples. Weibull modulus (b) for tensile strength of different fibres. Properties of gomuti fibre. Table 4.1 Table 4.2 Table 4.3 Table 4.4 Table 4.5 Table 4.6 Table 4.7 Table 4.8 Table 4.9 Designated sample names. Average tensile strength of gomuti fibre composites. Average Young s modulus of gomuti fibre composites. Average compressive strength of gomuti fibre composites. Average compressive modulus of gomuti fibre composites. Average flexural strength of gomuti fibre composites. Average flexural modulus of gomuti fibre composites. Average shear strength of gomuti fibre composites. Calculated shear modulus. L i s t o f t a b l e s xvi

18 Table 4.10 Theoretical prediction of tensile properties of gomuti fibre composites Table 4.11 Percentage tensile strength of the composite based on fibre tensile strength Table 5.1 Table 5.2 Table 5.3 Table 5.4 Table 5.5 Studies on impact strength of natural fibre composites. Designated names of the composite samples. Average dimension of the specimens. Average maximum impact force and energy absorbed by gomuti fibre composites. Average impact strength of gomuti fibre composites. Table 6.1 Table 6.2 Table 6.3 Table 6.4 Table 6.5 Table 6.6 Table 6.7 Roofing materials and their respective attributes. Typical dimensions and weight of roof tiles. Roof tile profile under consideration. Range of mechanical properties of gomuti fibre composites for roof tile design. Determination of minimum required thickness based on the properties of material and loading condition. Design variables for gomuti fibre composites roof tile. Examples of roof tile design L i s t o f t a b l e s xvii

19 CHAPTER 1 INTRODUCTION 1.1 General Natural fibre polymer composites (NFPC) are one type of composites that incorporates the use of naturally occurring components, such as natural plant fibres with polymer resins. During the past few decades, research into NFPC has been extensive due to the recognised benefits of the use of environmentally friendly materials. Significant number of studies into NFPC have encompassed investigations into the micro-characteristics and macro-characteristics. Topics of these studies range from micromechanical characterisation to potential applications in various area such as automotive, merchandise, construction and infrastructure uses (O Donnell 2004; Alms et al. 2009; Dweib et al. 2006; Yu et al. 2008). As composites are developed from two or more components to obtain a material with the required properties and performance, it is essential to evaluate the characteristics of the composite in relation to the characteristics of the constituents. The potential of natural fibres as composite materials is justified when their properties and behaviour can be carefully examined and understood. Environmental benefits have been the main motivation of research into NFPC. Being natural resources, natural fibres can be categorised as a renewable material. Moreover, the use of natural fibre offers a more environmentally Adriana N E Ticoalu C h a p t e r 1 1

20 sustainable supply of material and promotes less synthetic waste. The main attractions of natural fibre composites are their tailored ability, lightweight, and environmental benefits. Furthermore, natural fibres are available in various forms and types and in large quantities in most parts of the world, making them globally sustainable renewable resources. Common examples of natural fibres are cotton, coir, straw, hemp and jute. The fact that natural fibres are sourced from plants which, unlike synthetic fibres, can be easily biodegraded has encouraged increasing research into the field. Therefore, modern knowledge of composites has looked into incorporating natural fibres for environmental benefits purposes. The challenge of working with NFPC is the large variation in properties and characteristics. The properties of NFPC gave to a large extent, influenced by the type of fibres, the environmental conditions where the plant fibres are sourced and the type of fibre treatments employed. However, with their unique and wide range of variability, natural fibre composites could emerge as a new alternative engineering material which can substitute the use of synthetic fibre composites. Unlike their synthetic counterparts, the incorporation of natural fibres often results in a large variability of composites properties. This variability is influenced by natural conditions such as temperature, humidity, moisture, climate, harvesting process used and by the manufacturing process used. However, the exploitation of natural fibres for use in composites offer advantages in terms of friendly processing, reduced risk of environmental damage due to fibre production and fibre decomposable issues, cost competitiveness and economical assistance to rural communities in developing countries. Friendly processing means that, compared to Adriana N E Ticoalu C h a p t e r 1 2

21 the processing or handling of synthetic fibres, there is minimal risk of skin and respiratory irritation to people handling natural fibres. Although variability in properties is a major challenge, in civil engineering applications, natural fibre composites may find their place in the development and production of non-structural elements or elements that require lower to moderate strength such as roofing, decking or slats, wall cladding and insulation panels. The suitability of a natural fibre composite in a certain engineering application essentially depends on the characteristics and behaviour of the composite itself. As an engineered material, similar to synthetic fibre polymer composites, the properties of NFPC can be tailored to meet the desired requirements. Although their benefits are many, NFPCs have their drawbacks. Investigation into the characteristics of NFPC is, therefore, essential to ensuring proper knowledge of the properties and behaviour which will be beneficial in determining potential applications. With regard to fibre composite material, especially concerning natural fibres, obtaining the properties of the fibre itself is necessary. Properties and behaviour of a material can be measured and assessed with mechanical tests. Basic mechanical property tests include tensile, flexure, compressive and shear tests. These tests are performed to obtain the important basic mechanical properties of natural fibre composites. 1.2 Background Natural fibre composites generally imply that at least one of the constituents is naturally sourced. Either the fibre, the matrix or both. Early utilisation of natural Adriana N E Ticoalu C h a p t e r 1 3

22 fibres includes textile for clothing, brick reinforcement, roof covering and wall insulation. In rural developing countries, natural fibres have been used for roofing material, multipurpose rope, wall insulation, bags, brooms and other non-structural applications. Natural fibres have been utilised since Ancient time for various applications such as for textile and construction material. Combining natural fibre with polymer resin has wide and varied potential. Not only can many different fibres be used to create new composite materials, these materials have a large variety of applications. It is, therefore, important to investigate and understand the properties and behaviour of natural fibre for the purpose of developing natural fibre composites. As has been revealed by many studies, substantial variations exist in the properties of natural fibres which, unlike its synthetic counterparts, are dependent on the type of fibres, cultivation conditions, etc. Among many types of natural fibres available in Indonesia, gomuti fibre is one of the most versatile and durable. It has been used traditionally by Indonesian people for various applications such as rooves, brooms, water filters and ships rope. Overtime, however, the use of gomuti fibre has decreased due to the introduction of many other materials such as metal roofing, plastic roofing, and synthetic or plastic ropes and brooms. The production and utilisation of gomuti which actually supported local economies continues to diminish as it is replaced by other materials. Gomuti has failed to compete with other products for reasons of aesthetics and convenience, rather than performance. Therefore, a study into the characteristics of gomuti fibre will provide important information on its usefulness and viability as a natural fibre composites material and justify further potential uses. Adriana N E Ticoalu C h a p t e r 1 4

23 1.3 Objectives An assessment of the characteristics of natural fibre composites, based on the characteristics of its constituents, is essential in determining the composite s viability as a functional engineering material. This research aims to investigate the characteristics and behaviour of natural fibre composites using gomuti fibre and thermoset resins, and to evaluate the feasibility of possible applications in building and construction. Figure 1.1 provides a simple illustration of the research objectives and flow: a) Investigate the characteristics and behaviour of gomuti fibre b) Investigate the strength, stiffness and impact resistance properties of gomuti-fibre/thermoset resin composites c) Determine the effect of alkali treatment and the use of different resins to the strength, stiffness and impact resistance of gomuti fibre composites d) Evaluate the potential of developing gomuti fibre composites roof tiles. Gomuti fibre existing usage: roof, filter, rope Fibre characterisation Strength Stiffness Elongation at break Density Diameter Thermal stability Gomuti fibre composites (Variables: alkali treatment & thermoset resins) Strength & stiffness: Tensile Compressive Flexural Shear Gomuti fibre composites (Variables: alkali treatment & thermoset resins) Impact resistance characteristics and behaviour Potential applications Roof tiles Figure 1.1 Simplified illustration of the research flow Adriana N E Ticoalu C h a p t e r 1 5

24 1.4 Scope This study focuses on investigating the characteristics of gomuti fibre and the properties of its composites with thermoset resins. The study comprises experimental investigation on individual gomuti fibre and on the composites made of gomuti fibre with three different thermoset resins to evaluate the feasibility of development into a construction material. The study specifically covers the following: 1. Reviews of the characteristics of gomuti fibre and its composites with thermoset resins: polyester, vinylester and epoxy. 2. Characterisation of gomuti fibre covers: observation of surface morphology with optical microscope, measurement of fibre diameter from fibre cross-section, measurement of fibre density, single fibre tensile test and measurement of fibre thermal stability. 3. Testing of unidirectional gomuti fibre composites to obtain mechanical properties (tensile, flexural, compressive and shear). 4. Instrumented drop-weight charpy impact testing. 5. Strength and stiffness of gomuti fibre composites roof tile (evaluation based on the transverse bending strength test). 6. Statistical analysis with ANOVA to evaluate the effects of fibre treatment and different resins. The following issues are beyond the scope of this study: Adriana N E Ticoalu C h a p t e r 1 6

25 1. The use of phenolic types of resin 2. Determination of fibre chemical composition 3. Wind uplift resistance of roof tile 4. Long-term behaviour such as creep and fatigue 1.5 Structure of dissertation This research study is presented in seven chapters which generally comprise an introduction, review of related publications, investigations, findings and a conclusion. Further outline of each chapter is as follows: Chapter 1 presents a general introduction to natural fibres, natural fibre composites, background to idea of gomuti fibre composites and the objectives of this study. This chapter sets the direction of the research study and provides an overall general outline of the dissertation. Chapter 2 provides a detailed literature review on the characteristics of gomuti fibre and its composites and the potential applications. Existing research on gomuti fibre composites is studied as basis for the characterisation of gomuti fibre and its composites. Chapter 3 presents the experimental investigation into the properties of gomuti fibre, which covers the physical, mechanical and thermal properties of the fibre, and the effect of alkali treatment on its properties. This investigation will be a basis for evaluating the characteristics of the composites, which will be discussed in the next chapter. Adriana N E Ticoalu C h a p t e r 1 7

26 After investigation of the characteristics of gomuti fibre in the previous chapter, Chapter 4 focuses on the composites of gomuti fibre and thermoset resins, and the experimental investigation into their mechanical properties. This chapter covers the evaluation into the strength and stiffness of gomuti fibre composites by performing tensile, flexure, compressive and shear tests on nine composite samples to study the variations in the properties affected by alkali treatment to the fibre, and different thermoset resins. Although fibre composites provides strength and stiffness properties, impact strength is also major concern. Therefore, the next chapters will evaluate the impact characteristics of gomuti fibre composites. Chapter 5 presents the experimental investigation into drop-weight charpy impact properties of gomuti fibre composites and the evaluation of variations in the impact properties due to alkali treatment and different thermoset resins. Impact resistance characteristics together with strength and stiffness characteristics of gomuti fibre composites are important factors in designing the fibre composite element for applications in building and construction, which will be covered in the next chapter. Chapter 6 presents the parametric and simulation studies on the design of the gomuti fibre composites roof tile. The study covers a flat type of roof tile with variations in thickness, modulus of elasticity and strength. This chapter gives insight into the potential use of gomuti fibre composites and provides recommendation on its use as a roof tile. Conclusion of the findings from experimental results and analysis are presented in Chapter 7 together with recommendations for future research in this area. Adriana N E Ticoalu C h a p t e r 1 8

27 1.6 Summary Natural fibre composites are promising materials which combine low cost, acceptable performance and environmental friendliness. Reducing the production of synthetic fibres for low to moderate strength and stiffness requirements may be one of the many alternatives for promoting environmentally responsible practice in engineering. Domestic use of natural fibre is particularly beneficial to the community and the environment. However, strength, stiffness and compatibility need to be investigated in order to confirm the viability of the material. This chapter has provided a general introduction into natural fibre composite material, their potential and issues related to their research and development. The following chapters discuss, in more detail, the investigation of gomuti fibre and its composites as a potential natural fibre composite material. Adriana N E Ticoalu C h a p t e r 1 9

28 CHAPTER 2 LITERATURE REVIEW ON THE CHARACTERISTICS OF GOMUTI FIBRE, ITS COMPOSITES AND POTENTIAL APPLICATIONS 2.1. Introduction Many types of natural fibre are available around the world and many of them have been subjected to extensive research for their viability as natural fibre reinforced polymer composites (NFPC). Examples are jute, kenaf, hemp, cotton, coir, banana leaf, sisal, pineapple leaf and sugar-cane bagasse. These fibres can be grouped into bast fibres, leaf fibres, seed fibres, core fibres, grass and reed, and other fibres (Rowell 2008, cited in Pickering 2008). The use of these natural fibres in composites is particularly beneficial when they are developed in their native countries where there is a continuous and adequate supply. The use of natural fibre as composite material presents minimal risk of skin and respiratory irritation compared with handling synthetic fibres. Further, it offers reduced risk of environmental damage from fibre production and fibre decomposable issues. On the financial side, natural fibre can be cost competitive and provides economic assistance to rural communities in developing countries. The major challenge in utilising natural fibre is, compared to synthetic fibres, the large variability in the properties. Their properties are largely dependent on natural conditions such as temperature, humidity, moisture, climate, harvesting process and Adriana N E Ticoalu C h a p t e r 2 10

29 manufacturing process. Therefore, evaluation of the characteristics and properties of the fibre is necessary before attempting to develop its composites. One of the many natural fibres available in Indonesia is gomuti fibre. Mogea et al. (1991) informed that the fibre has been traditionally used as rope, filter and roofing material. Gomuti is the black coloured fibre that covers the trunk and the pinnate-leaf bases of the Arenga pinnata or Arenga Saccharifera tree as illustrated in Figure 2.1. Other names associated with gomuti are sugar-palm fibre, Arenga pinnata fibre, ijuk, injuk, gomutu, gomutus and black fibre. The tree is one of the main sugar yielding palms (Mogea et al. 1991) and is native to the South-East Asia region, India, Sri Lanka and Papua New Guinea (Mogea et al. 1991; Jones 2000). The trees can be found on most Indonesian islands and is particularly known for its production of palm sugar (Mogea et al. 1991; Smits 2010), vinegar, a distilled alcoholic drink and its sweet sap. Generally, in Indonesia, gomuti can be purchased for a relatively low price, approximately less than US$1/kg. Another positive aspect of using gomuti fibres is that when obtaining the fibres, there is no need to cut or harvest the whole tree, hence preserve the plantation of the Arenga pinnata. Figure 2.1. Arenga pinnata tree with gomuti covering the tree trunk. Adriana N E Ticoalu C h a p t e r 2 11

30 As introduced in Chapter 1, the key interest in research on natural fibre composites is their natural characteristics which promote environmentally friendly processing and usage. Large variability is, however, a major drawback of NFPC. This chapter aims to provide an overview of the characteristics of gomuti fibre and its composites with thermoset resins, based on the existing information in the literature and to formulate the necessary approach for evaluating the feasibility of gomuti fibre for use as a natural fibre composite material Traditional and existing usage of gomuti fibre When layers of gomuti fibre are taken from the tree, they are in the form of randomly intermingled long fibre as shown in Figure 2.2(a). Figure 2.2(b) shows the chopped gomuti fibre. Figure 2.2. Gomuti fibre: (a) Layers of trimmed intermingled long fibre, and (b) chopped fibre. Traditionally, to make rope, the fibre bunch is manually spun. While to make broom, the fibre layers are trimmed and brushed. For roofing, it is common practice to use gomuti in its original fibre orientation. Gomuti roof sheets are made by Adriana N E Ticoalu C h a p t e r 2 12

31 stacking, aligning and trimming layers of gomuti into a rectangular shape with dimensions of mm by mm or according to consumer specifications. Layers of gomuti sheets are then fixed into a bamboo clamp. Several of the rectangular gomuti roof sheets are then laid onto timber or bamboo purlins until the whole house or building is covered. Another practice is to lay layers of brushed fibres as thatch roof as shown in Figure 2.3. Figure 2.3. Gomuti thatch roof in Bali, Indonesia (Original photo caption: Shrines in the inner sanctuary of Tirta Empul, courtesy of Timothy Tye, used with permission). Gomuti is a durable natural fibre as proven by its usage for traditional roofing (thatch roof), water filters, ship s rope, brooms and reinforcement to traditional cement-mortar walls. It is one of the cheapest options for traditional roofing which can sustain tropical climate for a number of years. Furthermore, the fibre has also been used in road construction, the base of sport course and as shelters for fish Adriana N E Ticoalu C h a p t e r 2 13

32 breeding (Mogea et al. 1991). The fibre is also durable when exposed to seawater (Mogea et al. 1991) Characteristics of gomuti fibre To assess the viability of a fibre for usage in NFPC, a knowledge of the fibre s characteristics is essential. Observing the features of gomuti fibre such as physical characteristics, single fibre tensile properties, chemical composition and thermal stability are important in assisting the development of gomuti fibre composites and evaluating their behaviour. This section will look at the characteristics of gomuti fibre which include the physical, mechanical and thermal properties Physical characteristics Physically, gomuti is a black-coloured, coarse fibre that covers the trunk and the pinnate-leaf bases of Arenga pinnata or Arenga Saccharifera tree. In this section, the physical characteristics related to surface morphology, diameter, shape and density are presented Morphology of gomuti fibre A preliminary observation, under optical microscope, the surface of gomuti fibre reveals parallel lines along the length of the fibre, and that on the fibre surface, are pore-like spots present in almost regular spaces. This is quite similar to the surface of coir. The picture of gomuti fibre under optical microscope is shown in Figure 2.4. In the case of coir, the spots are referred to as tyloses and cover the pits on the cell walls (Prasad et al. 1983) while the parallel lines may be referred to as Adriana N E Ticoalu C h a p t e r 2 14

33 microfibrils. In a picture of the Scanning Electron Microscopy (SEM) of coconut fibre after removal of the waxy layer, tyloses and pits can be clearly seen (Brahmakumar et al. 2005). Figure 2.4. Surface morphology of gomuti fibre under optical microscope The surface morphology of natural fibre affects the fibre-matrix adhesion in a composite. If the surface of the fibre is waxy, the resin finds it difficult to penetrate and grip the fibre. This occurrence can be described as the fibre and the matrix not bonding properly. Low fibre-matrix adhesion leads to the poor performance of a composite Diameter of gomuti fibre Different studies report different values for the diameter range of gomuti fibre. Suriani et al. (2007) reported a range of mm, Bachtiar et al. (2010) reported mm, while Razak and Ferdiansyah (2005) reported mm. Table 2.1 presents a compilation of the diameter and density of several natural fibres, in comparison with the diameter and density of glass fibre (Wonderly et al. 2005). As observed in the Table, the diameter range of gomuti is similar to coir and Adriana N E Ticoalu C h a p t e r 2 15

34 sisal, which is about times that of cotton, flax, hemp, palmyrah and glass fibre. Table 2.1. Density and diameter of natural fibres compared to glass fibre. Type of fibre Density Fibre diameter (g/cm 3 ) (µm) Coir Cotton Flax Hemp Jute Palmyrah Sisal Glass fibre Source: 1 (Satyanarayana & Wypych 2007); 2 (Wonderly et al. 2005) Density of gomuti fibre Even though coir and sisal have a larger diameter especially when compared to glass fibre, their densities are lower (Table 2.1). This means that they typically have lower weight. Gomuti exhibits similar characteristics, having larger diameter with lower density. Bachtiar et al. (2010) reported that the density of gomuti fibre was 1.29 g/cm3, while Razak and Ferdiansyah (2005) reported 1.05 g/cm Mechanical properties of single gomuti fibre Strength, modulus of elasticity and elongation of single fibre can be obtained from a Single Fibre Tensile Test (SFTT). The properties of single fibre are important for evaluating the fibre s behaviour as an individual and in a composite. Adriana N E Ticoalu C h a p t e r 2 16

35 Various methods can be employed to obtain the single fibre properties, two of which are ASTM and ASTM Principally, the technique is to elongate the single fibre under tensile load until it fails or breaks. Variation of properties from the same type of fibre is common in the case of natural fibre. Many factors such as environmental conditions, temperature, humidity, climate, cultivation, harvesting, fibre separation process, the fibre s chemical compositions and moisture content contribute to this variation. Overall, it is observable that natural fibres generally have lower strength compared to glass fibre. In the case of stiffness, several natural fibres such as flax and hemp have comparable modulus with glass fibre. Gomuti fibre has similar features to coir: lower density, lower strength, lower modulus but higher elongation. Tensile strength of single natural fibres is in the range of 100 MPa 1500 MPa, as shown in Table 2.2. In contrast, the typical strength of glass fibre is 2400 MPa (Wonderly et al. 2005). Hence, the strength of natural fibres is approximately 4% 60% of the strength of single glass fibre. Table 2.2. Single fibre properties of natural fibres compared to glass fibre. Type of fibre Tensile strength Young s modulus Elongation at break (MPa) (GPa) (%) Coir Cotton Flax Hemp Jute Palmyrah Sisal Glass fibre Source: 1 (Satyanarayana & Wypych 2007); 2 (Wonderly et al. 2005) Adriana N E Ticoalu C h a p t e r 2 17

36 Various results of the properties of single gomuti fibre have been reported in several studies. Bacthiar et al. (2010) reported that the average stress at break of single fibre was MPa while Ishak et al. (2011) reported the value between MPa (for fibres taken from 7m, 9m, 11m, 13m and 15m of the tree) and less than 150 MPa (for fibres taken from 1m, 3m and 5 m of the tree). Meanwhile, Razak & Ferdiansyah (2005) informed that the yield strength of gomuti fibre is in the range of MPa. Modulus of elasticity and elongation at break for fibres with the strength between MPa are GPa and %, respectively (Ishak et al. 2011). These features are similar to that of coir and palmyrah fibres as shown in Table 2.2. Moreover, the modulus of natural fibres is approximately 5 % 38 % of the modulus of glass fibre, except in the case of flax and hemp. Elongation at break of several fibres such as flax, hemp, jute and sisal is close to the elongation of glass fibre (Table 2.2). Coir and palmyrah fibres, however, have a distinctive difference because the values of elongation are higher Chemical composition and thermal stability of gomuti fibre An investigation of the chemical composition of natural fibres reveals the amount of cellulose, lignin, ash and other naturally occurring chemicals in the fibres. Trindade et al. (2005) outlined the process for determining the chemical composition of natural fibres which include determination of klason lignin content by TAPPI T13M-54, holocellulose content by TAPPI T19m-54 and the determination of alpha-cellulose content. Adriana N E Ticoalu C h a p t e r 2 18

37 The chemical composition of various types of natural fibres is found in many papers and there is a broad list found in Satyanarayana and Wypych (2007) and Rowell (2008). Table 2.3 is a compilation of the chemical composition of several natural fibres (Satyanarayana & Wypych 2007) in comparison with that of gomuti fibre (Ishak et al. 2011; Dandi cited in Suriani 2007). It can be observed that gomuti fibre has a similar amount of alpha cellulose to palmyrah, and a similar amount of lignin to both coir and palmyrah. Fibres with a higher amount of lignin generally exhibit higher elongation characteristics, and their composites tend to have lower tensile strength and modulus compared to fibres with high cellulose content. Table 2.3. Chemical compositions of various natural fibres compared to gomuti fibre. Fiber alphacellulose (%) hemicellulose (%) Lignin (%) Ash (%) Extracts (%) Coir n/a n/a Cotton Flax Hemp Jute Palmyrah Sisal Sugar-palm (moisture) Sugar-palm (cellulose) (moisture) Source: 1 (Satyanarayana & Wypych 2007); 2 (Ishak et al. 2011); 3 (Dandi et al. cited in Suriani et al. 2007) The main chemical components (hemicellulose, alpha-cellulose and lignin) present in natural fibre degrade at a particular temperature. The profile of the fibre s thermal degradation can be measured as a weight changes versus temperature using Thermogravimetric Analysis (TGA) technique. During TGA scans of the fibre, the Adriana N E Ticoalu C h a p t e r 2 19

38 temperature range where there is no significant weight loss of the fibre is the thermal range where the fibre is considered thermally stable. The corresponding thermograph (change in weight vs temperature) is recorded. The TGA analysis requires placing a small amount of finely chopped fibres (approximately mg) into the TGA furnace. The fibre specimen is then heated to a maximum temperature of 600 o C. In the case of jute fibre (Ray et al. 2002), first weight change which indicates the loss of moisture occurred before the temperature reaches 100 o C. The second weight change at 297 o C was referred to the degradation of hemicelluloses and the third weight change at o C to the degradation of alpha-cellulose (Ray et al. 2002). Degradation of lignin happens slowly during the TGA process (De Rosa et al. 2010), after which the remaining weight is considered as the percentage of ash or residue. For typical natural fibres, at temperature between o C, the fibres can be considered as thermally stable. After 200 o C, a decline of the fibre s thermal stability can be observed along with a significant reduction in weight. Results of TGA may vary according to the type of natural fibres and also the condition of the fibres itself. It is reasonable to expect different results even from the same type of natural fibres. Any treatment to the fibre will also affect the TGA results Characteristics of gomuti fibre composites with thermoset resins On its own, gomuti fibre can be considered as a resilient and durable fibre because it has been used as roofing material which resists a tropical climate and it is known as seawater-resistant fibre (Mogea et al. 1991). However, its potential and Adriana N E Ticoalu C h a p t e r 2 20

39 application are limited by its original form and physical appearance. As with many other natural fibres, there is an opportunity to improve the application and usage of gomuti fibre by developing natural fibre composites with polymer resins. The polymer resin acts as a matrix material to provide shape and bind the fibre filaments so that, together, they can carry the design load. Examples of natural fibre composites with polymer resins are coir/polyester (Prasad et al. 1983), jute/vinylester (Ray et al. 2002), cotton/epoxy (Müssig 2008), sisal/epoxy (Oksman et al. 2002), flax/phenolic (Hepworth 2000), hemp/polypropylene (Beckermann & Pickering 2008), kenaf/poly-lactic-acid (Ochi 2008), and hemp/cashew-nut-shellliquid (Mwaikambo & Ansell 2003). Choosing a matrix material can be largely influenced by the type of fibres used, intended applications, fabrication method and cost consideration. Research into the feasibility of gomuti fibre for natural fibre composites have been reported in several papers. Thermoset resin, particularly epoxy has been used in most of the published literature on gomuti fibre composites (Suriani et al. 2007; Bachtiar et al. 2008; Bachtiar et al. 2009; Ishak et al. 2009; Leman et al. 2008; Leman et al. 2008; Sastra et al. 2006; Sastra et al. 2005). The use of epoxy may have been favourable due to its excellent properties and good bonding features. Therefore, the following overview on the characteristics of gomuti fibre composites will be dominated by the composites with epoxy resin. Two publications were found on the mechanical properties of gomuti fibre composites with polyester resin (Ticoalu, Aravinthan & Cardona 2010; Misri et al. 2010). However, the studies are limited to 10% (Ticoalu, Aravinthan & Cardona 2010) and 18% (Misri et al. 2010) fibre fraction by weight of untreated fibre. Adriana N E Ticoalu C h a p t e r 2 21

40 Moreover, publications on gomuti/vinylester or gomuti/phenolics are difficult to find. Therefore there is a need for more comprehensive studies on gomuti fibre thermoset composites with polyester, vinylester and phenolics matrices. These gaps are valuable benchmarks for further investigations into gomuti fibre composites with thermoset resins. With any type of matrix, in order to evaluate its usefulness and viability as composite material, it is essential to evaluate the characteristics of the composites which include the mechanical properties and behaviour, and how they relate to the characteristics of the constituents Characteristics of thermoset resins Choice of matrix can make a significant difference in term of performance, cost and compatibility with the fibre, especially in the case of natural fibre. A suitable type of matrix will also assists in the fabrication process. The two broad categories of polymer resins are thermosets and thermoplastics. Polymer thermoset resin is the type of polymer that will harden when cured and, unlike thermoplastic resins, do not liquefy when heated above their curing temperature. Therefore, thermosets cannot be remoulded. Examples are unsaturated polyester (UPE, a thermoset form of polyester), epoxy,vinylester and phenolics. UPE offers the advantages of lower price, acceptable properties and easy processing despite its poor temperature and UV-light tolerance (Åström 1997). Epoxy is suitable when a combination of economical value and excellent properties is sought. It is more expensive, but provides advantages such as low viscosity, low shrinkage and good bonding (Åström 1997). Properties of vinylester are between Adriana N E Ticoalu C h a p t e r 2 22

41 those of polyester and epoxy. Vinylester is produced by the reaction of an epoxy with an unsaturated acid, and was developed to have fast, simple crosslinking of UPE with mechanical and thermal properties of epoxies (Åström 1997). Phenolics are resins made of formaldehyde. Except for their brittle characteristic, they provide excellent fire resistant, high service temperatures, good electrical properties, high tensile modulus, excellent chemical resistance, and low-cost (Aranguren & Reboredo 2007). Table 2.4 presents the typical properties of common thermoset polymers. Table 2.4. Typical tensile and flexural properties of some common thermoset polymers. Polymers Tensile strength (MPa) Tensile modulus (GPa) Flexural strength (MPa) Flexural modulus (GPa) Epoxy Vinylester Polyester (Unsaturated) Phenolic Source: (Aranguren & Reboredo 2007) Fibre treatment and composite fabrication Fabrication processes of natural fibre composites can generally adapt methods used for synthetic fibre composites such as hand lay-up, vacuum assisted processes, compression moulding and pultrusion. Hand lay-up is the simplest but most labour-intensive method of fabricating composites. The other methods require specialised equipment and a specifically trained operator as initial investments, but large production is feasible with less time and labour. Adriana N E Ticoalu C h a p t e r 2 23

42 Fabricating gomuti fibre composites can be done with hand lay-up as well as with other methods. Existing published studies have attempted the hand lay-up method (Bachtiar et al. 2008; Bachtiar et al. 2009; Ishak et al. 2009; Leman et al. 2008; Ticoalu, Aravinthan & Cardona 2010), compression moulding (Sahari et al. 2011) and hot press moulding (Leman et al. 2008). However, it was acknowledged in Leman et al. (2008) that producing gomuti composites to the expected surface morphology was unsuccessful when using hot press moulding or using aluminium moulding because many voids were observed in the resulted composites. Similarly, a trial fabrication using a vacuum infusion process by the author has produced a panel with poor surface because of voids. However, higher fibre content (up to 50% wt.) was achievable which means that less resin was used and therefore, lighter and thinner panels can be produced. Void content affects the quality and strength of the composites (O Donnell et al. 2004), therefore, when fabricating a composite test panel, it is necessary to eliminate as many voids as possible. However, the complete removal of voids is not an easy task, not only with vacuum infusion but with hand lamination as well. In the vacuum infusion process, it is recommended to dry the fibres and to degas the resin (Brouwer 2000). In hand lamination, because the density of gomuti is about the same as the resins, when resin is applied to the fibre, the fibre tends to drift up from the resin. Therefore, one possible technique is to put weight or pressure on top of the uncured panel. To remove the air voids, careful attention is needed in wetting the fibre with a roller or squeeze. Leman et al. (2008) advised that the best employed method in fabricating gomuti fibre composite with the hand lay-up process was by using the heat resistant transparent plastic or Mylar. Adriana N E Ticoalu C h a p t e r 2 24

43 The extent of wetting during the manufacturing of composite affects the adhesion between fibre and matrix (Brouwer 2000). Many studies in natural fibre composites have attempted to improve the bonding characteristics between fibre and matrix by means of treatments. Different treatments for natural fibres are found in the literature, such as alkali treatment (Ray et al. 2002, Pickering et al. 2007, Aziz & Ansell 2004), hot water treatment (Delft University of Technology 2003) and silane treatment (Valadez- Gonzalez 1999). The correct amount and duration of alkali treatment assisted in the removal of lignin, pectin and hemicelluloses, therefore improve the fibre s properties (Pickering et al. 2007). An incorrect treatment, however, may degrade the fibre. For example, over-treatment with alkali will reduce fibre properties (Pickering et al. 2007). Effects of alkaline treatment, freshwater and seawater treatment to the gomuti-epoxy interface and to the mechanical properties of gomuti fibre composites are discussed in several papers (Bachtiar et al. 2008; Bachtiar et al. 2009; Ishak et al. 2009; Leman et al. 2008). There is no exact matching of treatments for every natural fibre. Treatment effects are mostly studied experimentally. Experimental trials have so far been beneficial in determining the optimum treatment time, type and concentration Tensile strength and stiffness characteristics Tensile properties such as tensile strength and modulus are some of the basic information of mechanical properties that allow the evaluation of strength and stiffness of a material. Based on the published literature, Table 2.5 presents the compilation of tensile properties of gomuti thermoset composites. To date, Adriana N E Ticoalu C h a p t e r 2 25

44 published studies on gomuti thermoset composites have mainly been on composites with epoxy, and a few studies with polyester. Fibre treatment Table 2.5. Tensile properties of gomuti fibre composites with thermoset resins. Thermoset resin Fibre fraction (% wt.) Fibre form Tensile strength (MPa) Tensile modulus (GPa) - Neat epoxy Untreated Epoxy 10 chopped random Untreated Epoxy 10 chopped random Untreated Epoxy 10 long random Untreated Epoxy 10 long random Untreated Epoxy 10 woven Untreated Epoxy 10 woven Untreated Epoxy 15 chopped random Untreated Epoxy 15 chopped random Untreated Epoxy 15 chopped random Untreated Epoxy 20 chopped random Untreated Epoxy 20 chopped random Untreated Epoxy 20 long random Untreated Epoxy 20 long random Untreated Epoxy 10 unknown NaOH 0.25M 1h Epoxy 10 unknown NaOH 0.5M 8h Epoxy 10 unknown Freshwater 30-days Epoxy 15 chopped random Seawater 30-days Epoxy 15 chopped random Untreated Polyester 10 random original Untreated Polyester 10 Chopped Untreated Polyester 10 Unidirectional Untreated Polyester 10 Woven mat Untreated Polyester 18 Woven Source: (Suriani et al. 2007; Sastra et al. 2006; Leman et al. 2008; Bachtiar et al. 2008; Ticoalu, Aravinthan & Cardona 2010; Misri et al. 2010). A combination of gomuti with polyester offered lower strength (approximately 9-25 MPa) compared to gomuti with epoxy. Variations in fibre Adriana N E Ticoalu C h a p t e r 2 26

45 forms or fibre orientation, as can be seen from Table 2.5, show that composites with chopped fibres generally have a lower tensile strength. In the case of gomuti/epoxy, composites with woven fibres achieved the highest average tensile strength. For gomuti polyester, composites with unidirectional fibre achieved the highest average tensile strength However, there is limited study on composites made of unidirectional and woven gomuti. For woven fibre, there is also limited information on the detail of weave pattern. Furthermore, Sastra et al. (2006) informed that the availability of woven form of gomuti is limited because it needs to be pre-ordered. Chopped fibre in a composite gives lower strength, but it is simple to process. Chopped gomuti polyester composites obtained an average tensile strength of MPa (Ticoalu, Aravinthan & Cardona 2010). Meanwhile, as shown in Table 2.5, the value of tensile strength of chopped gomuti/epoxy is approximately MPa. There is no significant variation shown in tensile strength for chopped gomuti epoxy even with weight fraction differences, except in one study that reported MPa (Leman et al. 2008). Moreover for gomuti epoxy composites, distinction between the results of chopped random and long random fibre was insignificant when the composites contained 20% fibre by weight. Based on the tensile properties summarised in Table 2.5, it can be inferred that gomuti fibre functioned as tensile reinforcement in epoxy composites with 10% long random and 10% woven fibres. The tensile strength of these composites is higher than the neat epoxy resin reported in Sastra et al. (2006). Gomuti polyester composites with 10% fibre by weight, as reported in Ticoalu, Aravinthan and Cardona (2010), obtained lower tensile strength, approximately 8 70% of the strength of the typical neat polyester resin (Aranguren & Reboredo 2007). Adriana N E Ticoalu C h a p t e r 2 27

46 Similarly, Misri et al. (2010) reported that the strength of gomuti/polyester with 18% fibre fraction by weight was MPa. When the strength of the composites is lower than that of the neat resin, several explanations are likely such as ineffective bonding between fibre and matrix and/or the presence of voids. When gomuti is treated with sodium hydroxide (NaOH), the composites containing fibres that have been soaked for 1 hour in 0.25M NaOH showed an increased tensile strength while other variations showed only increasing of tensile moduli (Bachtiar, Sapuan & Hamdan 2008). This advises the need for a more comprehensive study on variations of alkaline treatments to gomuti fibre. Meanwhile, increasing tensile strengths are recorded for the composites with gomuti fibres that have been treated by soaking the fibres in either freshwater or seawater (Leman et al. 2008). The tensile strength increased by the longer soaking time, except for the 24-days-freshwater specimen (Leman et al. 2008). Comparing the values of gomuti/epoxy composites in Table 2.5 to the reported neat epoxy in Sastra et al. (2006), it can be inferred that the best composites out of gomuti fibre are made when long random or woven fibre is used with 10% fibre weight. Increasing the amount of fibre to 15% and 20% by weight decreased the tensile strength. Based on the compilation in Table 2.5, there is a need for thorough research which evaluates the properties consistently with a controlled experimental plan including testing of the properties of neat resin. Adriana N E Ticoalu C h a p t e r 2 28

47 2.4.4 Prediction of tensile strength and modulus of gomuti fibre composites Strength and modulus of composites can be predicted by analytical models. Several models have been established for composites made of synthetic fibres. Efforts to predict the strength and modulus of natural fibre composites have led to modification of the models. Facca et al. (2006) have evaluated various models such as rule of mixtures (ROM), inverse rule of mixtures (IROM), Halpin-Tsai equation, Nairn s generalised shear-lag analysis and Mendels et al. models, to predict the elastic modulus of hemp fibre, rice hulls and oak wood flour short-fibres in highdensity polyethylene (HDPE), with consideration of the changes in moisture content and fibre density. The Rule of Mixtures (ROM) is a prediction model used for unidirectional continuous-fibre composites (Chou 1992) which is commonly used due to its simplicity. The following formula is the general ROM formula for the determination of elastic modulus: E c = E f V f + E m (1 V f ) (2.1) where Ec is the modulus of the composite, Ef is the modulus of fibre, Vf is the volume fraction of fibre and Em is the modulus of matrix. In the practice of fabricating natural fibre composites, fibre fraction is usually determined by weight fraction. The fibre and matrix are weighed and determined before fabrication. Therefore, modulus of the composite can be calculated by substituting the volume of fibre with the ratio of weight fraction and density of fibre: Adriana N E Ticoalu C h a p t e r 2 29

48 E c = E f w f ρ f + E m (1 w f ρ f ) (2.2) For strength prediction, the following is the ROM model to calculate the stress of the composite under uniaxial loading (Chou 1992): σ c = σ f V f + σ m (1 V f ) (2.3) where σ c is the strength of the composite, σ f is the strength of fibre, V f is the volume fraction of fibre and σ m is the strength of matrix. Similarly, the ratio of weight fraction and density of fibre can be used in place of volume fraction: σ c = σ f w f ρ f + σ m (1 w f ρ f ) (2.4) Prediction of the strength and modulus of gomuti fibre composites based on the properties of gomuti fibre and resin requires the consideration on the fact that, unlike some other natural fibres, gomuti fibre has a higher strain value whereas, the ROM formula was initially developed assuming similar strain of the fibre and resin (Chou 1992). Therefore, when thermosets such as epoxy, polyester or vinylester are used with gomuti, the relationship falls into the category of ductile fibre/brittle matrix. This category for gomuti is illustrated in Figure 2.5. Theoretically, using ROM prediction of strength and stiffness, the strength and stiffness of a composite will have a value in between the value range of the constituents. For natural fibre composites such as gomuti fibre composites, Adriana N E Ticoalu C h a p t e r 2 30

49 adjustments are needed in the prediction of strength and modulus. The adjustments shall incorporate factors that represent the experimental investigations. Figure 2.5. Illustration of stress-strain of gomuti fibre and thermoset resin Flexural properties Flexural properties of gomuti fibre/epoxy composites exhibit similar tendencies to the tensile properties. Composite specimens with chopped fibres have lower values while specimens with woven fibres have the highest values of flexural strength. When the value of neat epoxy from Sastra et al. (2005) is used as comparison, from all the listed gomuti epoxy composites in Table 2.6, the best combinations are 10% long random, 10% woven, 15% long random and 15% chopped. Overall from Table 2.6, it is shown that the flexural strength of gomuti epoxy is between 50 MPa to 108 MPa while the flexural modulus is between 3 GPa and 4.5 GPa (Ishak et al. 2009; Sastra et al. 2005). For both long random and chopped fibres, the strength and modulus increased when the weight fractions Adriana N E Ticoalu C h a p t e r 2 31

50 increased from 10% to 15%. However, when the weight fractions were further increased to 20%, there was a decrease in flexural strength and modulus. Flexural strength of gomuti/polyester composites exhibit a different trend from their tensile strength (Ticoalu, Aravinthan & Cardona 2010). The composites containing woven fibre had the highest flexural strength, followed by the composites with unidirectional, random and chopped fibre. However, the gomuti polyester composite specimen with chopped fibre had the highest flexural modulus. Overall, flexural strength of untreated 10% gomuti polyester composites was in the range of MPa for all four different fibre forms, and the flexural modulus was between 2.9 GPa and 3.5 GPa. Table 2.6. Flexural properties of gomuti fibre composites with thermoset resin. Fibre treatment Resin Fibre fraction (% wt.) Fibre form Flexural strength (MPa) Flexural modulus (GPa) - Neat epoxy Untreated Epoxy 10 long random Untreated Epoxy 10 chopped Untreated Epoxy 10 woven Untreated Epoxy 15 long random Untreated Epoxy 15 chopped Untreated Epoxy 20 long random Untreated Epoxy 20 chopped Untreated Epoxy 20 chopped Untreated Epoxy 30 chopped Seawater 30-days Epoxy 20 chopped Seawater 30-days Epoxy 30 chopped Untreated Polyester 10 random original Untreated Polyester 10 Chopped Untreated Polyester 10 Unidirectional Untreated Polyester 10 Woven mat Source: (Ishak et al. 2009; Sastra et al. 2005; Ticoalu, Aravinthan & Cardona 2010) Adriana N E Ticoalu C h a p t e r 2 32

51 Contrasting results of flexure tests are reported in Ishak et al. (2009) where different trends can be observed for the variations in weight fractions and fibre treatment. For the composite specimens containing 20% fibre, the flexural strength of the sea-water treated specimen is lower than the specimen with untreated fibre (Ishak et al. 2009). On the contrary, the flexural strength of the sea-water treated specimen is higher than the untreated specimen when composites are made of 30% fibre (Ishak et al. 2009). Possible clarifications of the contrasting results may relate to the fabrication process of chopped fibres Impact properties A number of studies have been done on the impact properties of gomuti fibre composites, mainly on the composites using epoxy resin. Bachtiar, Sapuan and Hamdan (2009) found that alkali treatment to gomuti fibre has resulted in an increase of the (izod) impact strength of composites containing 10% fibre and epoxy matrix. This increase was reported as the result of better bonding between the fibre and matrix due to fibre treatment. Similarly, Ishak and Leman (2009) reported an increase of impact strength after fibre treatment with seawater. Another study by Ali, Sanuddin and Ezzeddin (2010) reported the impact strength of gomuti fibre and epoxy composites, and found that the composite samples with accelerated ageing exhibited slightly lower impact strength. Although direct comparison with the value of impact strength from this study is not suitable due to different method (drop weight or pendulum) and specimen type (Izod or Charpy), it can be observed from the study that the impact strength was increased for the treated specimen. It shows that a higher concentration of alkali provides increased impact strength. Adriana N E Ticoalu C h a p t e r 2 33

52 2.5. Potential applications of gomuti fibre composites Reviewing the characteristics of gomuti fibre and its composites with thermoset polymer resin matrices, it is found that the tensile strength of gomuti/epoxy is in the range of MPa (Suriani et al. 2007; Bachtiar et al. 2008; Leman et al. 2008; Sastra et al. 2006) and that the flexural strength is around MPa (Ishak et al. 2009, Sastra et al. 2005). When polyester resin is used, the tensile strength is in the range of 9 25 MPa and flexural strength approximately MPa (Ticoalu, Aravinthan & Cardona 2010). Gomuti fibres exhibit similar characteristics (morphology, strength, stiffness, chemical composition and elongation) to coir fibre. The tensile strength of single gomuti fibre is in the range of MPa, Young s Modulus is approximately 2 4 GPa and elongation is approximately 10 30% (Bachtiar et al. 2010; Ishak et al. 2011). These characteristics are particularly different from other natural fibres such as hemp, flax and kenaf. Finding a suitable end use of natural fibre composites that match the properties is an essential factor in developing a beneficial and economical product. Existing applications of natural fibre composites include car rooves and interiors (Rijswijk et al. 2003), fibre-boards and flush doors (John et al. 2007), natural fibre composite panels (John et al. 2007; Van Erp & Rogers 2008), rooves (Bisanda 1993; Dweib et al. 2006; Satyanarayana et al. 1986), furniture (O Donnell, Dweib & Wool 2004), fluid containers (Rijswijk, Brouwer & Beukers 2003), bone and tissue repair (Cheung et al. 2009; Hutmacher & Woodruff 2007), small boats, and the structural rehabilitation of underground drain pipes and water pipes (Yu et al. 2008). Adriana N E Ticoalu C h a p t e r 2 34

53 Apart from the traditional applications such as roofing, broom, water filter and ship cordage, existing development of gomuti fibre composites have looked at the application as hybrid fibre for manufacturing a small boat (Misri et al. 2010). Based on its characteristics, untreated gomuti fibre exhibits similar properties to untreated coir. Newman (2008) suggested that fibres from fruit or seed such as coir and cotton may be more ideal for insulation rather than reinforcements, due to their low bulk density. Therefore, it can be assumed that gomuti can also be used as insulation material or in applications where lower strength is acceptable such as low-cost roofing panels, insulating fibre boards and wall cladding Summary Several studies have reported that gomuti fibre from the Arenga pinnata tree has properties and characteristics similar to coir fibre and comparable to other natural fibres. Compared with other common natural fibres such as flax, hemp and jute, gomuti fibre exhibited lower strength, lower density, higher elongation and higher amount of lignin. In combination with polymer thermosetting matrices such as epoxy, gomuti fibre composites can typically offer tensile strength of up to MPa and flexural strength up to 100 MPa. Lower tensile strength (less than 25 MPa) and flexural strength (less than 50 MPa), however, are obtained when gomuti is combined with polyester. Chemical, water or saltwater treatment of gomuti fibre alters the properties of the composites. However, a more thorough investigation is needed and further verification in this area must be devised to provide more references of gomuti fibre polymer thermoset composites. Adriana N E Ticoalu C h a p t e r 2 35

54 Based on the existing publications on gomuti fibre composites, it is deemed essential to undertake further investigation into the characteristics of gomuti fibre to confirm the properties and further, to evaluate the effects of fibre treatment to gomuti fibre composites by firstly evaluating the effects on the fibre. This investigation is covered in Chapter 3. In Chapter 4, investigation is continued towards manufacturing gomuti fibre composite samples to evaluate the effects of fibre treatment and the use of different polymer thermoset resins. Strength, stiffness and impact characteristics are studied to assess the samples characteristics and behaviour. Chapter 5 covers the study of the impact strength characteristics of gomuti fibre composites. Feasibility of gomuti fibre composites as roof tile is evaluated afterwards in Chapter 6. In order to observe the large variability in natural fibres and their composites, which is the key issue in developing natural fibre composites, this study has limited the fibre content to 20 % fibre fraction by weight. This amount of fibre is relatively suitable for hand lay-up manufacturing of natural fibre composites. Therefore, the variables in this study will be the fibre treatment and different resin type. The use of gomuti fibre as a composite ingredient can be explored with greater emphasise on usage for low-cost and low to moderate load bearing elements. While the strength of gomuti fibres epoxy composites showed the need for improvement, further study into this kind of fibres is feasible due to its environmental and cost benefit. To extend the potentiality of gomuti fibre composites from laboratory scale to practical larger scale applications, especially in construction industries, there exists a number of options such as vacuum infusion, vacuum bagging and compression moulding method. These methods are likely to produce larger number Adriana N E Ticoalu C h a p t e r 2 36

55 of products even though they require higher initial cost for the equipment and specialised training for the operator. The use of these processing will promote even lesser usage of resin matrix, therefore it is expected that the performance of the composites will be more dependent on the natural fibre itself. These processing, however, require lengthy preparation of the materials and significant initial cost. Therefore, the use of the above-mentioned manufacturing methods will be more covered in future study The following Chapter 3 provides an investigation into the physical, mechanical and thermal stability of gomuti fibre with a view to develop natural fibre composites with thermoset resin. Adriana N E Ticoalu C h a p t e r 2 37

56 CHAPTER 3 INVESTIGATION INTO THE CHARACTERISTICS OF GOMUTI FIBRE 3.1. Introduction The utilisation of natural plant fibres for various applications is undoubtedly supported by their availability in bulky amount in many countries and the fact that they are obtained from natural resources. For certain applications such as those which require low to moderate strength, the use of natural fibres offers competitive benefits in comparison to synthetic fibres which have to be fabricated and are difficult to degrade naturally. When natural fibres are developed into composite materials, they promote less production of synthetic fibres therefore reducing synthetic waste and the risk of environmental damage. Handling natural fibre also offers a reduced risk of health hazards compared to the handling of synthetic fibres. Additionally, the development of natural fibre composites may provide financial benefits to local communities where the fibres are sourced. The obvious challenge in using natural fibres as composite materials is related to their widely diverse types and properties. Each type of natural fibre has different characteristics and typical properties. In a comprehensive review of natural fibres for biocomposites, Faruk et al. (2012) has pointed out similar facts regarding the variation in the properties of natural fibres. A relatively large variation may also exist within each similar variety of fibres, and further, within one batch of a fibre Adriana N E Ticoalu C h a p t e r 3 38

57 sample. Observation of the characteristics of natural fibre is therefore essential to assess its viability as natural fibre composite materials. This chapter aims to study the characteristics of the gomuti fibre for the purpose of evaluating its viability as a natural fibre composite material. The characteristics under assessment are the physical characteristics, the single fibre tensile properties and the fibre s thermal stability. Furthermore, the effects of sodium hydroxide treatment of the fibre will also be assessed. A limited number of studies have examined the characteristics of gomuti fibre. Bachtiar et al. (2010) reported that the density of Arenga pinnata fibre was 1.29 g/cm3, average single fibre tensile strength was MPa, average tensile modulus was 3.69 GPa, average strain was 19.6 % and diameters vary between µm. A similar range of diameters, µm, was reported by Suriani et al. (2007), while Razak and Ferdiansyah (2005) reported slightly different range of diameters, from 300 to 500 µm. Compared to other natural fibres, the properties of the gomuti fibre exhibit similarity to coir (coconut husk fibre), having lower strength, larger diameter and distinctively higher elongation-at-break. As presented and discussed in Chapter 2, in term of strength, most of the listed natural fibres have lower values compared to glass fibre. By observing the density values of natural fibres, which are approximately % of the density of glass fibre, it is obvious that natural fibres may offer competitive strength-to-weight ratio. For use as a composite material, it is important that the fibre adhere properly to the matrix. Fibre-matrix bonding can be improved by certain treatments of fibres. In the case of natural fibres, alkali treatment with sodium hydroxide (NaOH) is one Adriana N E Ticoalu C h a p t e r 3 39

58 of the most commonly studied. Sodium hydroxide treatments to natural fibre provide cleansing and removal of the impurities, wax, oils, hemicelluloses and a parts of lignin from the surface of the natural fibres (Amico, Sydenstricker & da Silva 2005; El-Shekeil et al. 2012; Faruk et al. 2012; Mwaikambo & Ansell 2002; Symington et al. 2009) Therefore, improvements of the fibre s properties by alkali treatment are possible (Mwaikambo & Ansell 2002). Faruk et al. (2012) reported that the major modification in the fibre when treated with alkali is the increased fibre surface roughness due to the disruption of the hydrogen bonding in the network structure. Since the chemical compositions of natural fibres are different, the treatment type, amount and duration will provide different results. A suitable treatment will assist in enhancing the fibre properties and behaviour, which will be of further benefit to the composite. Gomuti fibre has a similar chemical composition to coir fibre (Dandi, Sapuan & Hamdan, cited in Suriani 2007; Satyanarayana & Wypych 2007; Ticoalu, Aravinthan & Cardona 2013), therefore similar treatment may become an initial estimation to observe the influence of the treatment to the properties of the fibre. Alkali treatment with different percentage, duration and results have been reported in various literature. Increased tensile properties were shown for coir fibre treated with 5 % NaOH (Prasad, Pavithran & Rohatgi 1983; Silva et al. 2000), sisal fibre treated with 2 %, 9 % and 18 % of NaOH for 24 h (Padmavathi, Naidu & Rao 2012), date-palm fibre treated with 2 % and 5 % NaOH for 2 h and 4 h (Taha, Steuernagel & Ziegmann 2007), and sisal fibre treated with 3 % NaOH for 10 minutes (Symington et al. 2009). Improvements in the mechanical properties of Adriana N E Ticoalu C h a p t e r 3 40

59 composites containing alkali-treated natural fibre have also been reported. For example, the mechanical properties of coir/polyester composites were improved when the fibre was treated with 2 %, 5 % and 10 % NaOH for 1 hour (Rout et al. 2001) Another example is the improvement in sisal/epoxy composites after fibre treatment with 2 %, 9 % and 18 % NaOH for 24 h (Padmavathi, Naidu & Rao 2012). On the other hand, reductions in fibre tensile properties have been reported for 10 % NaOH treatment to sisal fibre (Amico, Sydenstricker & da Silva 2005), 3 % NaOH treatment (10, 20 and 30 minutes) to flax and abaca (Symington et al. 2009), 3 % NaOH treatment of sisal fibre (20 and 30 minutes) (Symington et al. 2009), and 5 % NaOH treatment (30 minutes) to flax, linen and bamboo single-strand yarns (Yan, Chouw & Yuan 2012). Furthermore, NaOH treatment of kenaf fibre resulted in reduced tensile, flexural and impact strengths of kenaf-polyurethane (thermoplastic polyurethane) composites (El-Shekeil & Sapuan 2012; El-Shekeil et al. 2012). Based on the information and findings from these studies, it can be inferred that different fibres behave differently in response to the alkali treatment. Therefore, to obtain an initial estimation of the effect of alkali treatment to gomuti fibre, treatment with 5 % and 10 % NaOH is considered adequate at this stage Experimental methods All experimental works on fibre characterisation which include fibre treatment, microscopic observation, determination of fibre density, single fibre tensile test and thermal analysis were conducted at the testing facilities in the Faculty of Health, Engineering and Sciences and in the Centre of Excellence in Adriana N E Ticoalu C h a p t e r 3 41

60 Engineered Fibre Composites (CEEFC), University of Southern Queensland (USQ), Toowoomba, Australia Fibre material Gomuti fibre from North Sulawesi, Indonesia, was used. Layers of gomuti fibre were extracted manually from the Arenga pinnata tree. Afterwards, the fibres were manually cleaned and rinsed with cold water before being air-dried thoroughly Fibre treatment Three batches of gomuti fibre, each weighing approximately 10 grams, were taken randomly from the bunch, and placed in different containers. One batch was not treated (UN), one batch was treated with 5 % sodium hydroxide (5N) and one batch was treated with 10 % sodium hydroxide (10N). Sodium hydroxide (NaOH) pellets analytical grade was used for the treatment. The samples were soaked in water-diluted NaOH for approximately 2 h. Afterwards, they were rinsed several times with water to remove excess NaOH. The samples were then strained and airdried just to remove excessive moisture before drying in an oven at 80 o C for approximately 4 hours Observation of fibre morphology and diameter measurement Surface morphology of gomuti fibre was observed by an optical microscope. The microscope with its built-in software (analysis docu 5.0 (Build 1200), Olympus Soft Imaging Solutions GmbH) was also used to measure the cross section area and to calculate the diameter based on the area s equivalent circular diameter (ECD). As it is with most natural fibres, gomuti presents non-uniformity along the Adriana N E Ticoalu C h a p t e r 3 42

61 cross-section. Therefore, an estimation of the fibre diameter is made by taking an average from ECD measurements from the two ends of every single fibre in a sample. Figure 3.1 shows the schematic of specimen arrangement for measuring the fibre diameter at the two ends, while Figure 3.2 shows the actual measurement with an optical microscope. Figure 3.1. Single fibre observation: (a) schematic of specimen arrangement for microscope observation; and (b) schematic of specimen mounting tab for single fibre tensile test (SFTT) using dynamic mechanical analyser (DMA) machine. Adriana N E Ticoalu C h a p t e r 3 43

62 Figure 3.2. Microscope observation of fibre cross-section Measurement of fibre density A helium pycnometer was used to measure the volume (V) of fibre samples and a microbalance was used to obtain the mass (m). Five random measurements of density were taken for each sample. Density (ρ) was then determined as: ρρ = mm VV (3.1) Single fibre tensile test Single fibre tensile tests were carried out using a Dynamic Mechanical Analyser (DMA) machine Q800 v5.1, manufactured by TA Instruments. ASTM and ASTM were used as reference standards. Specimens were prepared by taking randomly from each of the three groups of specimen (UN, 5N and 10N). Single fibres were attached to thick papers, with configuration as shown Adriana N E Ticoalu C h a p t e r 3 44

63 in Figure 3.1. After microscope observation of fibre diameter, the paper pieces were cut along the dashed lines to separate each single fibre for tensile testing. Gauge lengths of the specimens, which were measured by a micrometer, were approximately mm. The DMA machine was set to controlled force for stress/strain type of test. Specimen type of cylindrical was selected, and the values of diameter and gauge length of each specimen were entered. Testing temperature was held isothermal at 28 o C and the force rate was 0.5 N/min. Each specimen was mounted carefully by holding the upper paper tab using tweezers. Firstly, the upper paper tab was clamped by the fixed-top-grip. Next, after proper tightening of the upper part, the bottom paper tab was aligned to be clamped to the movable, spring loaded grip clamp. For each specimen, the microscopemeasured diameter and the micrometer-measured gauge length were entered into the analysis software for stress calculation. The value of diameter entered for each specimen was the average of two ECDs measured from the two ends of the respective specimen. Results were then recorded with the built-in Universal Analysis software. A minimum of 25 specimens were tested for each fibre group. Slipped specimens or those that were loose from the grip during testing were disqualified. For analysis purpose, 25 specimens were used. The 25 number of specimens were considered practical at this stage in order to make an initial evaluation of the properties of gomuti fibre. A similar number of specimens (20-25 single fibres) in a sample of natural fibre has been reported in literature (Bachtiar et al. 2010; De Rosa, et al. 2010; Joffe, Andersons & Wallstrom 2003; Symington et al. 2009). Adriana N E Ticoalu C h a p t e r 3 45

64 Distribution of the results data was evaluated using descriptive statistics and Weibull distribution. Weibull distribution is employed to obtain an understanding of the characteristics of the fibre. The general equation for Weibull probability density function is: yy = bb aa xx aa bb 1 ee xx aa bb (3.2) where: x is data, a is Weibull scale parameter or characteristics strength, b is the shape parameters or Weibull modulus. In this study, the Weibull parameters (a and b) were obtained by using MATLAB Statistics Toolbox (The Mathworks, Inc. R2011b) which uses the Maximum Likelihood Estimation (MLE). Furthermore, to evaluate the significance of the effects of the fibre treatment, inference statistics with one-way analysis of variance (ANOVA) was used. Romhany, Karger-Kocsis and Czigany (2003) reported the use of the Kolmogorov-Smirnov goodness of fit test to judge the appropriateness of the use of Weibull distribution in evaluating distribution of data. A similar approach is applied in this study. The basic principle is to calculate the maximum difference (dmax) from the resulting Weibull distribution and the empirical distribution of the data. A Weibull Distribution is considered appropriate to be used when dmax < dcrit (Romhany, Karger-Kocsis & Czigany 2003). In this study, the number of specimen was 25, therefore n is 25. For n = 25 and significant level of 0.05, dcrit = (Makhnin, 2010). Data analysis in this study was performed with the aid of Microsoft Excel and MATLAB Statistics Toolbox (The Mathworks, Inc., R2011b). Adriana N E Ticoalu C h a p t e r 3 46

65 3.2.6 Measurement of fibre thermal stability Thermogravimetric Analysis (TGA), using TA Instruments Q500 v6.1, was performed to observe the thermal stability of the gomuti fibre. The analysis was done to obtain data of weight changes in the fibre under a controlled condition (ramp 5 o C/min to 600 o C). The changes were measured as a function of temperature and time. A small amount of finely chopped fibres, approximately mg was placed in a small aluminium pan that goes inside the TGA furnace where the fibres were heated in a nitrogen atmosphere to a maximum temperature of 600 o C with a heating rate of 5 o C/min. The corresponding thermograph (change in weight vs temperature) was recorded. Differential Scanning Calorimetry (DSC), using TA Instruments Q100 v8.1, was also performed providing the details of changes in the specimen by temperature changes, time and heat flows Experimental results Morphology of gomuti fibre Gomuti fibre is black-coloured and its physical appearance is coarse. Under optical microscope, as shown in Figure 3.3, the surface of an untreated gomuti fibre (UN) reveals very fine parallel lines which are known as microfibrils. There are also small circular shaped pores containing white substances. A similar surface was observed in coir fibre where the white substances were identified as tyloses and the pores as pits on the cell walls (Brahmakumar, Pavithran & Pillai 2005). For Figure 3.3, the scale magnification used was 10x eyepiece magnification with a further 40x camera software magnification. Adriana N E Ticoalu C h a p t e r 3 47

66 Figure 3.3. Surface morphology of untreated (UN) and treated (5N and 10N) gomuti fibre. Adriana N E Ticoalu C h a p t e r 3 48

67 A similar microscopic pictures with magnification of (5x eyepiece, 40x software magnification) has been added to the Appendix B. Figure 3.4 shows a cross-section of gomuti fibre showing the fibre defect common in natural fibre. Cross-sectional shapes of gomuti fibre are generally arbitrary but the cross-sectional area can be estimated by fitting either circular or elliptical shapes which will be discussed further in the next section. Figure 3.4. Cross-section of gomuti fibre Diameter of gomuti fibre The built-in microscope software provides the calculation of the cross-section area and the equivalent circular diameter. From 25 specimens, it was found that the average diameter of untreated gomuti was 168 µm and the minimum to maximum values were 81 to 313 µm. This range is similar to the reported value of the diameter Adriana N E Ticoalu C h a p t e r 3 49

68 of gomuti (sugar-palm) fibre as reported by Bachtiar et al. (2010). Figure 3.5 shows the fibre diameter measurement. Figure 3.5. Fibre diameter measurement. Table 3.1 presents the descriptive statistics of diameter data, where it is shown that the average diameter of untreated gomuti fibre is slightly larger compared to the averages of the alkali treated fibres. The detailed results of diameter data is presented in Appendix B. Table 3.1. Descriptive statistics of diameter of untreated and alkali treated gomuti fibre (n = 25). Sample Diameter (µm) Mean S.D. UN N N Density of gomuti fibre The results of density measurement are presented in Table 3.2. The average value of 5 specimens is presented with the standard deviation. Adriana N E Ticoalu C h a p t e r 3 50

69 Table 3.2 Density of untreated and treated gomuti fibre Sample Density (g/cm 3 ) Average S.D. UN N N Single fibre tensile properties Figure 3.6 provides examples of stress-strain plots from the recorded results for UN, 5N and 10N. It is to be noted, however, that the values used in the analysis were the average results from 25 observed specimens for each sample. The three plots in Figure 3.6 show the initial linear elastic region before the stress reached approximately 50 MPa. After the linear elastic region, the plots started to show nonlinear behaviour before ultimate stress and failure. This behaviour could be attributed as a typical stress-strain behaviour of gomuti fibre which exhibits large elongation before fracture. A similar plot shape was observed from the stress-strain curve of coir fibre as reported in Tomczak, Sydenstricker and Satyanarayana (2007) and Silva et al. (2000) while okra fibre (De Rosa, et al. 2010) and flax fibre (Baley 2002) showed different stress-strain behaviour. Nechwatal, Mieck and Reußmann (2003) explained that natural fibres show a different initial shape of stress-strain curve. Based on the similarities exhibited by gomuti and coir, it can therefore be implied that natural fibres with higher elongation are more likely to produce the stress-strain curves as shown in Figure 3.6. Adriana N E Ticoalu C h a p t e r 3 51

70 Figure 3.6. Example of stress-strain curves of single-fibre-tensile tested specimens. The Modulus of Elasticity of a single gomuti fibre was calculated from the elastic linear region of the stress-strain curves of each specimen. The average modulus of gomuti fibre was 3847 MPa and the average strain at failure or elongation at break was 12.8 %. The results are different from the results from the existing study on gomuti fibre by Bachtiar et al. (2010) that reported the value of 3690 MPa and 19.6 % strain. Such variation is common in the study of natural fibre. As the literature review revealed, the elongation at break of gomuti fibre is distinctively different compared to glass fibre and other natural fibres, with the exception of coir Thermal stability of gomuti fibre Figures 3.7, 3.8 and 3.9 show the TGA curves of untreated and treated gomuti fibre. For the UN sample, the first peak of weight change occurred at around 48.1 o C (<100 o C) which relates to the release of moisture. The weight was reduced by Adriana N E Ticoalu C h a p t e r 3 52

71 approximately 4.1 %. The next weight change occurred at around o C (> 200 o C) when the weight has been reduced by approximately 17.5 %. The third weight change, which occurred at around o C, is a major weight change because the weight of the sample has been reduced by almost half (44.1 %). The onset of thermal degradation (degradation temperature) of the UN was found to be approximately o C. It is the temperature when untreated gomuti fibre can be considered as starting to degrade. Figure 3.7 TGA plots of untreated gomuti fibre. Adriana N E Ticoalu C h a p t e r 3 53

72 Figure 3.8 TGA plots of 5% NaOH treated gomuti fibre. Figure 3.9 TGA plots of 10% NaOH treated gomuti fibre. DSC results show that the first exothermic peaks in DSC plots, which relates to fibre decomposition (Mwaikambo & Ansell 2002), occurred at different Adriana N E Ticoalu C h a p t e r 3 54

73 temperatures for untreated and treated samples. Figure 3.10 shows the DSC plot of untreated and treated gomuti specimens. It can be observed from the Figure that after an endothermic peak, which relates to the heat absorbed during the release of moisture present in the sample, the UN curve displays two exothermic peaks, the 5N curve displays one broad peak, and the 10N displays one peak after a shouldered curve. Endothermic peaks for all samples occurred before 100 o C. The first exothermic peaks were found at approximately o C, 319 o C and 339 o C, respectively for UN, 5N and 10N. The finding from DSC results is in agreement with the TGA results and indicates that alkali treatment has resulted in a higher decomposition temperature, which further implies an improved thermal stability. Figure 3.10 DSC plot of untreated and treated gomuti specimens. Adriana N E Ticoalu C h a p t e r 3 55

74 3.4. Discussions Effects of treatment to the morphology of gomuti fibre Compared to other natural fibres, the surface morphology of gomuti is different to that of jute, kenaf, flax and hemp while closely resembling the surface of coir. The alkali treated specimens (5N and 10N) exhibit a reduced number of tyloses on the fibre surface. The surface of 10N was almost clear of tyloses, which infers that the treatment has removed the tyloses and other impurities. Cross-sectional shapes of gomuti fibre are generally arbitrary but the crosssectional area can be estimated by fitting either circular or elliptical shapes. The built-in microscope software provides the calculation of cross-section area and the equivalent circular diameter. From 25 specimens, it was found that the average diameter of untreated gomuti was 168 µm and the minimum to maximum values were 81 to 313 µm. This range is similar to the diameter value of gomuti fibre as reported by Bachtiar et al. (2010). The density of gomuti fibre was found to be approximately 1.40 g/cm 3. This value is higher compared to the reported value found in Bachtiar et al. (2010), which was 1.29 g/cm 3. Such variation is anticipated in the case of natural fibres. Moreover, the measured density is within the range of densities of natural fibres as reported in Chapter Effects of treatment to the diameter of gomuti fibre Table 3.1 presents the descriptive statistics of diameter data, where it is shown that the average diameter of untreated gomuti fibre is slightly larger compared to the averages of the alkali treated fibres. The difference in data can also be observed Adriana N E Ticoalu C h a p t e r 3 56

75 from the boxplot in Figure From the boxplot, it is shown that fibres treated with 10 % NaOH (10N) have a slightly smaller diameter compared to untreated fibres, but slightly higher compared to 5N sample. This finding suggests the idea that alkali treatment to gomuti has reduced the fibre s diameter. The idea is a reasonable one because the treatment would have removed the alkali prone substances from the surface of the fibre. Similar findings have been reported in existing literature on natural fibre (Taha et al. 2007; Sawpan 2009). To further evaluate the differences between the samples, an analysis of variance (ANOVA) simulation was conducted which returned a p-value of The value is more than the significance level of 0.05, which indicates that there is no significant difference between the diameter of untreated and treated gomuti fibre. Despite the ANOVA result showing that there is no clear difference between the diameters of the fibre samples. The summary in Table 3.1 and the boxplot in Figure 3.11 clearly show that the diameters of the treated samples are smaller compared to the untreated sample. The smaller average diameter of the treated samples is most likely caused by the elimination of the waxy layers, impurities and hemicelluloses from the fibre surface. This finding corresponds to the TGA result. Adriana N E Ticoalu C h a p t e r 3 57

76 0.5 Fibre diameter (mm) UN 5N 10N Figure 3.11 Boxplot of gomuti fibre diameter data Effects of treatment to the density of gomuti fibre Densities of the treated fibres were 1.48 g/cm 3 for 5N and 1.44 g/cm 3 for 10N, higher than the density of UN (refer to Table 3.2). Significant difference is observed between untreated and treated fibres, which is confirmed with a p-value of (<0.05) from ANOVA. Treated fibres, particularly 5N, have a higher density compared to untreated fibre. Similar finding has been reported in the literature. Sawpan (2009) reported that the density of hemp fibre increased after 5 % NaOH treatment for 30 minutes. It was suggested that, because the treatment removed the impurities from the fibre, the fibre cell wall became more compact, hence the density increased (Sawpan 2009) Effects of treatment to the tensile properties of single gomuti fibre Overall, the alkali-treated samples exhibit higher tensile strength. Compared to UN, the average tensile strength of 5N is 34.3 % higher and 10N is 29.6 % higher, Adriana N E Ticoalu C h a p t e r 3 58

77 while the modulus of elasticity of 5N is 20.8 % higher and 10N is 18.4 % higher. The increased tensile strength has been suggested as a result of a better restructuring of cellulose chains in the fibre after the treatment (Sawpan 2009; Taha, Steuernagel & Ziegmann 2007). The ultimate tensile strength of 5N and 10N exhibit higher values compared to UN. Observing the tensile capacity (peak load) of each individual fibre, as presented in Tables B.9, B.10 and B.11, the values are clearly varied, while the mean capacity between untreated and treated fibres show slight variation. Similar tendency is also shown by the calculated sectional stiffness (EA). The mean tensile capacities of untreated and treated gomuti fibre suggests that NaOH treatment has decreased the capacity and sectional stiffness. Therefore, to obtain a clear comparison of each single fibre, in this study, the ultimate strength of single fibre is measured based on its respective diameter. Based on the mean tensile strength of each fibre sample, it can be observed that the treated fibre samples exhibit higher strength and modulus compared to untreated fibre sample. Furthermore, NaOH treatment to gomuti fibre has provided advantages such as removal of dirt and waxy layer, which provides better fibre-matrix adhesion. Data distribution resulting from the tensile test is illustrated in boxplot as shown in Figures 3.12, 3.13 and It is shown that the 5N sample has the highest median value and a smaller range of data for tensile strength, modulus and elongation at break. The properties exhibit a slightly lower value when the treatment was increased to 10 %. The higher concentration of alkali may have caused damage to the fibres hence reducing the properties, as was observed in Pickering et al. (2007). Adriana N E Ticoalu C h a p t e r 3 59

78 Average tensile elongation-at-break of the treated samples was 17.3 % for 5N and 17.1 % for 10N. The results show that the untreated sample has a lower average strain compared to the treated specimen. A larger scatter is observed from strain data of UN, as shown in Figure It can be implied, therefore, that the treated fibre samples have larger elongation and less scatter of strain data. To further evaluate the data, data of tensile strength and modulus are plotted into the Weibull probability plot as shown in Figures 3.15 and Weibull parameters (a and b) for each sample are presented in Tables 3.3 and 3.4, respectively for the strength and modulus data. The use of a Weibull distribution is validated by the results from the Kolmogorov-Smirnov goodness-of-fit test. The maximum differences (dmax) are presented in Tables 3.3 and 3.4, respectively for strength and modulus data. It was found that dmax for all data are less than dcrit, therefore Weibull distribution can be used to present the data Tensile Strength (MPa) UN 5N 10N Figure 3.12 Boxplot of the tensile strength of gomuti fibre. Adriana N E Ticoalu C h a p t e r 3 60

79 Young's Modulus (MPa) UN 5N 10N Figure 3.13 Boxplot of the Young s Modulus of gomuti fibre Strain (%) UN 5N 10N Figure 3.14 Boxplot of the strain of gomuti fibre. Table 3.3 Average tensile strength, Weibull parameters and goodness-of-fit parameter of gomuti fibre samples. Fibre samples Tensile strength (MPa) Weibull parameters UN N N a b dmax Adriana N E Ticoalu C h a p t e r 3 61

80 Table 3.4 Average Young s Modulus, Weibull parameters and goodness-of-fit parameter of gomuti fibre samples. Fibre samples Tensile strength (MPa) Weibull parameters UN N N a b dmax Table 3.5. Weibull modulus (b) for tensile strength of different fibres. Type of fibres b Reference Glass fibre Chawla, 1998 Carbon fibre 5-6 Chawla, 1998 Brown coir 8 Defoirdt et al., 2010 Jute 2.7 Defoirdt et al., 2010 Sisal 1.94 Amico et al., 2005 Untreated gomuti 4.59 Current study 5% NaOH treated gomuti 9.32 Current study 10% NaOH treated gomuti 5.12 Current study In the Weibull probability plot, as in Figures 3.15 and 3.16, data of all three samples can be fitted reasonably well with straight lines. The lines show that there is a clear difference between untreated and treated fibres both for the tensile strength and modulus. Further, the lines show that the strength and modulus of treated fibres are higher compared to UN. It is interesting to observe that the lines of 5N and 10N for both tensile strength and modulus intersect. This suggests that there may not be a clear difference between the influence of 5 % and 10 % NaOH treatment to the strength and modulus of gomuti fibre. The fitted lines also show that 5N data have a steeper slope compared to UN and 10N. Steeper slope relates to smaller variation and more predictable results (Abernethy 1996). Adriana N E Ticoalu C h a p t e r 3 62

81 A steeper slope of the fitted line in a Weibull probability plot means a bigger value of the Weibull modulus (b). For the purpose of comparison, Table 3.5 outlines the Weibull modulus of several types of fibres that have been recorded in the literature (Amico, Sydenstricker & da Silva 2005; Chawla 1998; Defoirdt et al. 2010) and from the analysis from this current study. Lesser variation in the strength of glass fibre is shown by higher b value. As shown in Table 3.5, jute and sisal have b values under 4.0, while the other listed fibres have b values over 4.0. All samples of gomuti fibre in this study exhibit b over 4.0. This indicates relatively smaller variation compared to the other samples. Further, it indicates the old age or rapid wear out category which is a typical failure mode for material properties based on Weibull plot (Abernethy 1996) N Probability UN 5N Tensile strength (MPa) Figure 3.15 Weibull probability plot of the tensile strength of untreated and alkali-treated gomuti fibre. Adriana N E Ticoalu C h a p t e r 3 63

82 N 0.5 UN Probability N Young's modulus (MPa) Figure 3.16 Weibull probability plot of the tensile modulus of untreated and alkali-treated gomuti fibre. Figures 3.17 and 3.18 shows the Weibull probability density function for data of tensile strength and modulus, which further illustrates the distribution of data. In Figure 3.17, lesser variation in data is shown by a slimmer curve. Data of 5N exhibits less variation although the location of peak is almost similar to 10N. The locations of peaks are observably different between the untreated and treated samples. A similar trend is observed for the modulus. Evaluating the data with ANOVA for all three samples returned a p-value of for strength and for modulus, which confirm that there are differences between UN, 5N and 10N data. This result implies that the NaOH treatment has resulted in different tensile properties of the fibre samples. Adriana N E Ticoalu C h a p t e r 3 64

83 While it is clear that treatment with NaOH has made a difference in single fibre tensile properties of the gomuti fibre, it is of interest to evaluate whether there is an actual difference between 5N and 10N. An ANOVA simulation was performed on the data of 5N and 10N, which returned p-values of 0.5 and 0.657, respectively for tensile strength and modulus. These values strongly suggest that there is no significant difference in 5N and 10N data. The alkali treatments have been shown to affect the tensile properties of gomuti fibre, however the difference in the tensile properties due to treatment amount is shown to be minor. This finding will be a base consideration for further work with gomuti fibre UN (a:191.10, b:4.59) 5N (a:246.98, b:9.32) Weibull pdf N (a:245.26, b:5.12) Tensile strength (MPa) Figure 3.17 Weibull probability density function (pdf) of the tensile strength of untreated and alkali-treated gomuti fibre. Adriana N E Ticoalu C h a p t e r 3 65

84 UN (a: , b:7.57) 5N (a: , b:8.03) Weibull pdf N (a: , b:5.82) Young's modulus (MPa) Figure 3.18 Weibull probability density function (pdf) of the tensile modulus of untreated and alkali-treated gomuti fibre. Results from the testing shows that gomuti fibre has an average tensile strength of MPa with a standard deviation of 46.9 MPa or approximately 27 % of the average value. Such deviation shows a relatively large scatter of data and is commonly observed in the properties of natural fibres. Table 3.6 summarises the diameter, density and tensile properties of gomuti fibre. Material Density (g/cm 3 ) Table 3.6 Properties of gomuti fibre. Tensile strength (MPa) Young s modulus (GPa) Strain-at-failure (%) UN N N Adriana N E Ticoalu C h a p t e r 3 66

85 3.4.5 Effects of treatment to the thermal stability of gomuti fibre As shown in Figures 3.7, 3.8 and 3.9, the treated samples (5N and 10N) do not exhibit a minor second weight change peak which occurred in the untreated specimen. This may indicate the elimination of waxy layers and the absence of hemicelluloses in the treated specimens. The degradation temperature of 5N was found to be approximately o C, and 10N approximately o C. Comparing the degradation temperatures (the onset of thermal degradation) of gomuti fibre, it was found that the untreated sample started to degrade at lower temperature compared to the treated samples even though the major weight change of UN occurred at higher temperature. DSC results show that the first exothermic peaks in DSC plots, which relate to fibre decomposition (Mwaikambo & Ansell 2002) occurring at different temperatures for untreated and treated samples. The previous Figure 3.10 shows the DSC plot of untreated and treated gomuti specimens. It can be observed from the Figure that after an endothermic peak, which relates to the heat absorbed during the release of moisture present in the sample, the UN curve displays two exothermic peaks, 5N curve displays one broad peak and 10N displays one peak after a shouldered curve. Endothermic peaks for all samples occurred before 100 o C. The first exothermic peaks were found approximately o C, 319 o C and 339 o C, respectively for UN, 5N and 10N. The finding from DSC results is in agreement with the TGA results and indicates that alkali treatment has resulted in higher decomposition temperature of gomuti fibre, which further implies an improved thermal stability. Adriana N E Ticoalu C h a p t e r 3 67

86 3.5. Conclusions The characteristics of gomuti fibre have been studied. Under microscope observations, gomuti fibre shows the physical appearance similar to coir fibre. Microfibrils and tyloses can be observed from the surface of a single gomuti fibre filament. The arbitrary shape cross-section of gomuti can be considered elliptical or circular shape with diameters ranging from µm. The density of untreated gomuti was 1.40 g/cm 3. The average strength of a single gomuti fibre was MPa, the average modulus of elasticity was 3.85 MPa and the average of elongation-at-break was 12.8 %. Alkali treatment with 5 % and 10 % NaOH to gomuti fibre resulted in altered properties of the fibre. A lesser number of tyloses on the surface of treated gomuti fibres and the smaller average diameter of the treated samples support the indication that waxy layers, impurities and hemicellulose have been removed by the treatment. Meanwhile, the average densities and tensile properties of 5N and 10N are higher compared to UN. Since relatively large scatters were observed in the diameter and tensile properties data with respect to the number of specimens, the assessment on the influence of alkali treatment is therefore limited to inferring the tendency exhibited by the results. The results show that the average diameter tends to decrease after treatment while the density and tensile properties tend to increase after treatment. In addition, by observing the amount of NaOH, 5 % treatment to gomuti fibre has resulted in better results compared to 10 %. Evaluating a larger amount of fibre and different NaOH amount is therefore recommended. Adriana N E Ticoalu C h a p t e r 3 68

87 TGA results indicate that the untreated sample began to degrade at a lower temperature (222.8 o C) compared to the treated gomuti fibre samples (236.1 o C for 5N and o C for 10N). The TGA results correspond to the results from DSC and provide the indication that the alkali treatment resulted in a higher degradation temperature. Compared to the commonly used natural fibre for fibre composites such as coir, the characteristics of gomuti fibre is in the acceptable range for use as a natural fibre composite material. Gomuti fibre, therefore, can be considered to have the potential to be developed into natural fibre composites. The strength and stiffness characteristics of the composites of gomuti fibre with polymer thermoset resins will be discussed in Chapter 4, while impact strength characteristics of the composites will be the focus of Chapter 5. Adriana N E Ticoalu C h a p t e r 3 69

88 CHAPTER 4 INVESTIGATION ON THE STRENGTH AND STIFFNESS OF GOMUTI FIBRE COMPOSITES WITH THERMOSET RESINS 4.1 Introduction With gomuti fibre's characteristics of low strength, low modulus and high elongation, as investigated and reported in Chapter 3, the composite incorporating gomuti fibre is considered most likely to exhibit similar characteristics. To investigate the properties and obtain knowledge of the behaviour of a material, mechanical tests are essential. Tensile, compressive, flexural and shear tests are the mechanical tests that can be performed to evaluate basic mechanical properties. This chapter aims to investigate the mechanical properties of gomuti fibre composites. Tensile, compressive, flexural and shear tests are performed to observe the response of the composite specimens to the respective loading conditions and to investigate the strength and stiffness capacity of the composites. Furthermore, the influence of fibre treatment and different thermoset resins to the mechanical properties of the composites will be determined. It must be acknowledged that, unlike utilising synthetic fibres, utilising natural fibres presents more challenges due to their widely varied characteristics. The obvious challenge in incorporating natural fibres lies on their naturally diverse types, forms and individual fibre properties. In a composite form, the variations are broadened with several more factors such as fibre orientation, fibre/matrix Adriana N E Ticoalu C h a p t e r 4 70

89 composition, fibre/matrix compatibility, matrix selection and fabrication method. It is important, therefore, to investigate the mechanical properties of natural fibre composites to determine their typical characteristics. As previously discussed in Chapter 2, compared to other natural fibres such as jute, hemp, kenaf and flax, individual gomuti fibre possess lower strength, lower density but higher elongation and larger diameter. When gomuti fibres were processed into composites using thermoset resins, the tensile strength was expected to be around 9 52 MPa, depending on the type of resin, fibre fraction, fibre form and fibre treatment, while a modulus of ppraoximately GPa was expected (Bachtiar, Sapuan & Hamdan 2008; Leman & Sapuan 2008; Sastra & Siregar 2006; Suriani et al. 2007; Ticoalu, Aravinthan & Cardona 2011; Ticoalu, Aravinthan & Cardona 2013). The range of tensile strength of gomuti fibre composites is comparable to the tensile strength of several natural fibre composites such as coir/epoxy (Biswas, Kindo & Patnaik 2011; Harish et al. 2009), hemp/polyester (Rouison, Sain & Couturier 2006), and sisal/polyester (Bisanda 1991). Existing studies reveals that the maximum values of tensile strength of gomuti/epoxy were reached by composite specimens containing 10% fibre fraction by weight of untreated woven fibres and untreated long random fibres (Sastra et al. 2006). The combination of gomuti fibre with polyester resin obtained lower tensile strength (Ticoalu, Aravinthan & Cardona 2011). For flexural strength, 10% by weight of untreated woven fibres with epoxy specimens have the maximum flexural strength followed by 15% by weight untreated long random fibres (Sastra et al. 2005). Untreated-gomuti/polyester composite specimens with 10% fibre by weight have lower average flexural strength compared to the specimens with epoxy Adriana N E Ticoalu C h a p t e r 4 71

90 (Ticoalu, Aravinthan & Cardona 2011). These studies on gomuti fibre composites provide a strong indication that 10% fibre fraction and the use of epoxy will produce optimum strength for gomuti fibre composites. As is commonly reported regarding natural fibre composites, fibre-matrix bonding or adhesion is a complex issue. Bonding between fibre and matrix is an essential platform which determines the properties of a composite. However, perfect bonding is often difficult if not impossible, particularly when natural fibres are included. Inadequate bonding, on the other hand, is not desirable as it will lead to a composite with poor properties. Fibre-matrix bonding can be improved by means of fibre treatment and or matrix selection. In selecting the suitable type of matrix, it is important to consider the type of fibres, intended applications, fabrication method and cost. Epoxy, polyester and vinylester are three thermoset resins which have their own characteristics. Generally, thermosetting-polyester exhibits faster curing compared to epoxy and offers minimum cost. However, epoxy has better tensile and flexural properties. Epoxy is more expensive but it offers better bonding to other materials. Vinylester is cured similar to polyester but it offers better strength. The existing studies on gomuti fibre composites have reported variations in the tensile, flexural and compressive properties based on fibre form and fibre fractions. Limited publications were found on the evaluations of mechanical properties of gomuti fibre composites with combinations of different resins and fibre treatment. It is therefore important to evaluate the mechanical properties and behaviour of gomuti fibre composites with different resins and fibre treatment. Adriana N E Ticoalu C h a p t e r 4 72

91 This study is focused on unidirectional-aligned gomuti fibre because it was observed from the literature that using woven, long random and unidirectional gomuti fibres have resulted in higher tensile and flexural strength. The woven form of gomuti composites requires more time and effort to process, therefore it is not included. Furthermore, even though it was found that 10% fibre by weight offers the best results, this study aims to reduce the amount of resin without compromising the workability of composite panels produced by the hand lay-up method. 4.2 Experimental methods Gomuti fibre composite panels processing and testing were conducted in the Centre of Excellence in Engineered Fibre Composites (CEEFC), University of Southern Queensland (USQ), Toowoomba, Australia Material Gomuti fibres were obtained from Minahasa region in North Sulawesi, Indonesia. The fibres were manually cleaned, rinsed with water and sun-dried. Unidirectional-aligned fibres, as shown in Figure 4.1, were prepared by manual brushing. Subsequently, the fibres were trimmed to approximately 400 mm long for composite fabrication. Adriana N E Ticoalu C h a p t e r 4 73

92 Figure 4.1 Unidirectional aligned gomuti fibre. Three thermoset resins, namely polyester, vinylester and epoxy were used. Polyester and vinylester were prepared with 1.5% catalyst while epoxy was prepared with 4:1 ratio of resin and hardener Fibre treatment Fibre treatment was done by soaking two sets of fibres in water-dissolved sodium hydroxide (NaOH) pellets for approximately 2 h. One set in 5% NaOH and one in 10% NaOH. Two NaOH concentrations were chosen to observe the influence of fibre treatment on the mechanical properties of the gomuti fibre composites. The 2 hour soaking time was performed to ensure proper wetting of the solution to the fibres. Afterwards, the fibres were rinsed several times with water to remove excess NaOH solutions. The fibres were then strained and thoroughly air-dried to remove excessive moisture before drying in an oven for approximately 1 h at 80 o C, with the intention to dry the fibres without causing damage or degradation. Adriana N E Ticoalu C h a p t e r 4 74

93 4.2.3 Preparation and fabrication of composite panels Composite panels were prepared by the manual hand lay-up method in CEEFC s processing laboratory. Each group of fibre, approximately 200 gr, was placed evenly on top of a sheet of heat resistant plastic. The plastic sheet was placed on top of a clean and smooth metal base. When the pack of fibre was ready, resin was poured onto the fibre. Wetting of the fibre was assisted by using a roller and plastic squeeze. Soon afterwards, another sheet of heat resistant plastic was placed on top of the mixture, making sure to squeeze gently to remove apparent air voids. A clean and smooth-surfaced piece of metal or timber was placed on top of the plastic-covered uncured panel. The panels were room-temperature cured for a minimum 24 h before being post-cured for 4 h at 80 o C. In this study, approximately 20 % fibre fraction by weight was achievable with a hand lay-up method targeted at spending the least resin possible while retaining good workability. After a series of experimental trials, it was decided to process the composite by trying to squeeze out as much resin as possible without leaving too many voids. This was achieved with 20 % fibre fraction. An experimental study with 10 % fibre fraction has been done by the authors and has been published in a conference paper (Ticoalu, A, Aravinthan, T & Cardona, F, 2011), which provides information that the average tensile strength of 1 0% gomuti/polyester composites was MPa and the flexural strength was MPa. Working with 10 % fibre fraction has provided better workability as there are more resin, whereas working with more fibre (> 20 %) has resulted in poor specimens. Pictures showing comparison of composites panel with different fibre content (10 %, 20 %, > 20%) are presented in Appendix C. Adriana N E Ticoalu C h a p t e r 4 75

94 Composites specimens were cut and prepared for tensile, compressive, flexure and shear tests in longitudinal direction. Minimum of five specimens were tested for each variation. Table 4.1 outlines the designated sample names. Table 4.1 Designated sample names. Fibre treatment Resin Designated sample name Polyester UN.PE Untreated Epoxy UN.E Vinylester UN.VE Polyester 5N.PE 5% NaOH Epoxy 5N.E Vinylester 5N.VE Polyester 10N.PE 10% NaOH Epoxy 10N.E Vinylester 10N.VE Testing methods Tensile test Tensile tests were carried out using a MTS 100 kn machine. The dimensions of tensile specimen were t (thickness) x 25 mm (width) x 250 mm (length). Tensile tests were performed based on the Australian Standard AS (Standards Australia 2001). An extensometer was used for strain reading. When the strain reached 0.3%, the extensometer was removed and the specimen was loaded until failure. The loading rate for tensile specimens was 2 mm/min. Figure 4.2(a) shows the tensile test set-up and Figure 4.3 shows the schematic of the tensile specimen and the idealised loading condition. Adriana N E Ticoalu C h a p t e r 4 76

95 Figure 4.2 Test set-up: (a) tensile; (b) compressive; (c) flexural and (d) shear. Figure 4.3 Schematic of tensile specimen and idealised loading condition Compressive test Compressive specimens were cut in the shape of a rectangular prism with the dimensions of t (thickness) x 2t (width) x 2t (height). Compressive tests were carried out based on ASTM D (ASTM International 2010) by using a MTS Adriana N E Ticoalu C h a p t e r 4 77

96 100 kn machine with a loading rate of 1.3 mm/min. The specimens were loaded until failure. Figure 4.2(b) shows the compressive test set-up and Figure 4.4 shows the schematic of the compressive specimen and idealised loading condition. Figure 4.4 Schematic of compressive specimen and idealised loading condition Flexural test Flexural specimens were tested using a three-point-bending test, with specimen dimensions of t (thickness) x 15 mm (width) x 20t (length), and the span of 16t. The tests were performed following ISO 14125:1998 (International Organization for Standardization 1998) by using a MTS 10 kn machine. Loading rates were varied, depending on the average thickness of the specimens, and were calculated based on the formula in the standard. Load and displacement were recorded until the specimen failed. Figure 4.2(c) shows the flexural test set-up and Figure 4.5 shows the schematic of the flexural specimen and the idealised loading condition. Adriana N E Ticoalu C h a p t e r 4 78

97 Figure 4.5 Schematic of flexural specimen and idealised loading condition Shear test Shear tests were performed following ASTM D5379M-05 (ASTM International 2005). Specimens with dimensions of t (thickness) x 20 mm (width) x 76 mm (length) were V-notched symmetrically at the centre as shown in Figure 4.6. Shear tests were carried out using a MTS 10 kn machine and Iosipescu test fixture with test speed of 2 mm/min. The specimens were tested until fracture. Figure 4.2(d) shows the shear test set-up and Figure 4.6 presents the schematic of shear specimen and idealised loading conditions. Figure 4.6 Schematic of shear specimen and idealised loading condition. Adriana N E Ticoalu C h a p t e r 4 79

98 4.3 Experimental results Tensile properties and load-extension behaviour The average tensile strength of the gomuti fibre composite samples is presented in Table 4.2. The symbols of UN, 5N and 10N denote the treatment variables which are untreated, 5% NaOH treated and 10% NaOH treated, respectively. Whereas the symbols of PE, E and VE denote the type of resin which are polyester, epoxy and vinylester. These symbols will be used in the dissertation hereafter. In the Table, it can be observed that overall, the average tensile strength varies from 28.4 MPa to 48.5 MPa. Sample 10N.PE exhibits the lowest average tensile strength while 10N.VE exhibits the highest tensile strength. Further, from Table 4.2 and Figure 4.7, it can be observed that the samples with polyester resin exhibited lower tensile strength compared to samples with epoxy and vinylester. Similar findings were reported in the case of kenaf fibre reinforced laminates (Rassmann, Paskaramoorthy & Reid 2011) and flax fibre (Joffe, Wallstrom & Berglund 2001). Table 4.2 Average tensile strength of gomuti fibre composites. Tensile strength (MPa) PE E VE UN 35.2 ± ± ± 3.8 5N 41.5 ± ± ± N 28.4 ± ± ± 2.1 Note: UN = Untreated, 5N = 5% NaOH treated, 10N = 10% NaOH treated, PE = Polyester, E = Epoxy, VE = Vinylester. Adriana N E Ticoalu C h a p t e r 4 80

99 Figure 4.7 Comparison of tensile strength For specimen with vinylester resin, 5 % NaOH treatment of the fibre resulted in a decrease in tensile strength and 10% treatment resulted in a slight increase in tensile strength. For the specimen with polyester and epoxy, 5% treatment has resulted in an increase in tensile strength. A higher concentration of NaOH has resulted in a decreased tensile strength. Average tensile strength of the samples with polyester were notably lower compared to samples with epoxy and vinylester. This result provides an indication that resin type has been dominant in affecting the tensile strength of the specimens. Furthermore, the results can be attributed to the typical tensile strength trend of neat resins where polyester has the lowest tensile strength and vinylester and epoxy has higher strength. Table 4.3 presents the average Young s modulus of the composite samples. It shows that the modulus varies from 4.32 GPa to 6.25 GPa. Overall, samples with untreated fibres exhibit lowest Young s modulus. From Figure 4.8 it is clear that samples with polyester exhibit higher modulus compared to samples with epoxy and vinylester. Adriana N E Ticoalu C h a p t e r 4 81

100 Table 4.3 Average Young s modulus of gomuti fibre composites. Young s modulus (GPa) PE E VE UN 5.43 ± ± ± N 6.25 ± ± ± N 5.54 ± ± ± 0.41 Figure 4.8 Comparison of tensile modulus.. Variation in the data of tensile strength is shown by the standard deviation and also in a form of boxplot as shown in Figure 4.9. Tensile strengths of specimens with polyester and epoxy are more varied compared to specimens with vinylester. This suggests that gomuti with vinylester may have sufficient fibre-matrix compatibility, hence reduces the variations. Adriana N E Ticoalu C h a p t e r 4 82

101 Tensile stress (MPa) UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Figure 4.9 Boxplot of tensile strength. Load-extension curves of the typical tensile-tested specimens are presented in Figure It is shown that the failure of the specimens have a similar pattern, regardless of the treatment and resin type. The specimens exhibit brittle failure when subjected to tensile load. Failure occurred abruptly following peak stress. Before reaching peak stress, the specimens exhibit linear elastic curves with extension up to 3 5 mm. This indicates that binding gomuti fibre with thermoset resins have altered the fibre s characteristic of large elongation. Adriana N E Ticoalu C h a p t e r 4 83

102 Figure 4.10 Load-extension plots of typical tensile-tested specimen: (a) specimens with polyester resin, (b) specimens with epoxy resin, and (c) specimens with vinylester resin. Adriana N E Ticoalu C h a p t e r 4 84

103 4.3.2 Compressive properties and load-extension behaviour Overall, as presented in Tables 4.4 and 4.5, compressive strength of the specimens is between 81.4 MPa and MPa, and the compressive modulus is between 1.28 to 2.37 GPa. Composite specimens 5N.VE exhibit the highest average compressive strength and compressive modulus, followed closely by 5N.PE. Meanwhile, 10N.E has the lowest compressive strength and modulus. As can be observed from the Table, specimens with treated fibres display a slightly higher compressive strength. For each resin type, 5N-specimens have the highest strength. For each treatment group, specimens with vinylester show the highest strength followed by specimens with polyester. Specimens with epoxy show a distinctively lower strength and modulus. Table 4.4 Average compressive strength of gomuti fibre composites. Compressive strength (MPa) PE E VE UN 104 ± ± ± N 113 ± ± ± N 108 ± ± ± 2.91 Table 4.5 Average compressive modulus of gomuti fibre composites. Compressive modulus (GPa) PE E VE UN 2.05 ± ± ± N 2.26 ± ± ± N 1.91 ± ± ± 0.15 Adriana N E Ticoalu C h a p t e r 4 85

104 A graphical plot of compressive strength and compressive modulus of the tested specimens are presented in Figure 4.11 and Figure These results provide an indication that fibre treatment and different type of resin affect the compressive properties. Similar trend can be observed from the curves with results of samples with polyester and vinylester are similar, while samples with epoxy show significantly lower compressive strength. Furthermore, samples with 5 % treated fibres exhibit best compressive strength. This indicates a good fibre-matrix bonding. Comparison of the compressive modulus as illustrated in Figure 4.12 shows that for polyester and vinylester specimens, 5 % treatment has improved the moduli. Different trend, however, is shown by epoxy specimens where the untreated samples have the highest modulus compared to the treated samples. This could be due to defects or excessive voids in the samples. Figure 4.11 Comparison of compressive strength Adriana N E Ticoalu C h a p t e r 4 86

105 Figure 4.12 Comparison of compressive modulus Figure 4.13 shows compressive strength data variation for each composite specimen in boxplot chart. As shown in the Figure, specimens with epoxy have significantly lower strength and specimens with vinylester have the highest strength. Data are varied within a range of approximately 5 15 MPa. 120 Compressive Stress (MPa) UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Figure 4.13 Boxplot of compressive stress Adriana N E Ticoalu C h a p t e r 4 87

106 Figure 4.14 shows the load-extension plots of a typical compressive-tested specimen. Generally, the load-extension curves of compressive specimens exhibit a linear elastic region followed by a plastic region before ultimate strength and failure. Curves of some samples such as the curves of UN.PE, 5N.PE, 10N.PE, UN.E and 5N.E exhibit a toe region before the elastic region. In the calculation of compressive properties, the toe region has to be compensated, as specified in ASTM D (ASTM International 2010). After displaying a linear elastic region and plastic region, the specimens reached the ultimate strength at approximately 1 2 mm of extension, followed by declining load capacity. For the specimen with polyester, minor matrix cracking before failure can be observed from the slight distortion in the graph. However, for the specimen with epoxy and vinylester, failures were observed from the declining strength capacity without prior noticeable cracking. Adriana N E Ticoalu C h a p t e r 4 88

107 Figure 4.14 Load-extension plots of typical compressive-tested specimen: (a) specimens with polyester resin; (b) specimens with epoxy resin; and (c) specimens with vinylester resin. Adriana N E Ticoalu C h a p t e r 4 89

108 4.3.3 Flexural properties and load-displacement behaviour The average flexural strength and modulus of the gomuti fibre composite samples are presented in Table 4.6 and Table 4.7, respectively. Details on test report of flexural strength are presented in Appendix C. The flexural strength is calculated based on the following formula: σσ ff = 3FFFF 2bbh 2 (4.1) where F = force/load (N), L = span length (mm), b = width (mm), h = thickness (mm). As presented in Tables 4.6 and 4.7, the flexural strength of the specimens varies around MPa and the moduli vary around GPa. Composite specimen 5N.E has the highest average flexural strength followed by 10N.VE and 10N.E, while 10N.PE has the lowest value. The trend of the flexural strength result is relatively similar to the result for tensile strength. Table 4.6 Average flexural strength of gomuti fibre composites. Flexural strength (MPa) PE E VE UN 72.2 ± ± ± 8.9 5N 77.8 ± ± ± 20 10N 64.1 ± ± ± 11 Adriana N E Ticoalu C h a p t e r 4 90

109 Table 4.7 Average flexural modulus of gomuti fibre composites. Flexural modulus (GPa) PE E VE UN 4.11 ± ± ± N 5.06 ± ± ± N 4.33 ± ± ± 0.21 A graphical plot of flexural strength and flexural modulus of the tested specimens are shown in Figures 4.15 and They demonstrated that samples with polyester exhibit lower flexural strength while epoxy and vinylester samples exhibit higher values. For 5N.VE, however, the strength slightly decreased. From all the vinylester samples, it can be inferred that treatment to the fibre only provides marginal differences. Similarly, in the case of flexural modulus, the vinylester samples exhibit minor variations. Figure 4.15 Comparison of flexural strength Adriana N E Ticoalu C h a p t e r 4 91

110 Figure 4.16 Comparison of flexural modulus Flexural strength results, however, have larger dispersion of data as can be observed from the boxplot in Figure Data are dispersed with a range of approximately MPa. The largest dispersion is observed for the 5N.VE while UN.VE has minimal variation. Unlike the results of tensile strength which shows that specimens with polyester and epoxy have larger variations, for flexural test results, the specimens with vinylester have larger variations Flexural Strength (MPa) UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Figure 4.17 Boxplot of flexural stress. Adriana N E Ticoalu C h a p t e r 4 92

111 Furthermore, from the boxplot, it can also be observed that for untreated and 10N specimens, specimens with polyester have the lowest strength and vinylester has the highest, while for 5N specimen, specimens with epoxy have the highest strength. These results can be attributed to the typical trend of the strength of neat resins where polyester has the lowest strength, while epoxy and vinylester exhibit higher flexural strength. Load-displacement curves from flexural tests are presented in Figure 4.18, which shows that specimens with polyester exhibit particularly different shapes of the curves. As can be observed from Figure 4.18, the specimens exhibit an elastic linear region up to approximately 4 5 mm of displacement before minor matrix cracking occurred, which can be seen as the first distortion in the curves. After the first cracks, gradual increments in load and displacement can be observed until ultimate strength, followed by gradual ruptures. For specimens with epoxy and vinylester, failure occurred suddenly after peak load without prior notification of cracking. Adriana N E Ticoalu C h a p t e r 4 93

112 Figure Load-displacement plots of typical flexure-tested specimen: (a) specimens with polyester resin; (b) specimens with epoxy resin; and (c) specimens with vinylester resin. Adriana N E Ticoalu C h a p t e r 4 94

113 4.3.4 Shear properties and load-displacement behaviour The average shear strength of gomuti fibre composite samples is presented in Table 4.8, while Table 4.9 shows the calculated shear modulus. It can be observed, from Table 4.8, that sample 10N.E has the highest shear strength followed closely by 5N.E while sample UN.PE exhibits the lowest shear strength. The use of epoxy and treated fibre shows better results of shear strength. The boxplot in Figure 4.21 shows that the data are varied, and that the largest variation is observed from 10N.E sample, and the smallest variation is observed from UN.VE. Table 4.8 Average shear strength of gomuti fibre composites. Sample Shear strength (MPa) Polyester (PE) Epoxy (E) Vinylester (VE) UN 28.0 ± ± ± N 37.0 ± ± ± N 28.7 ± ± ± 3.23 Table 4.9 Calculated shear modulus. Sample Shear Modulus (GPa) Polyester (PE) Epoxy (E) Vinylester (VE) UN N N Overall, from the results of the shear test, shear strength can be expected to be between 28 and 39 MPa, with specimens containing treated fibres exhibiting higher shear strength. Furthermore, specimens with epoxy show a significantly higher shear strength compared to the specimens with polyester and vinylester, as Adriana N E Ticoalu C h a p t e r 4 95

114 illustrated in Figure In the case of the calculated shear modulus, as shown in Figure 4.20, samples with polyester show higher shear modulus compared samples with epoxy and vinylester. Figure 4.19 Comparison of shear strength. Figure 4.20 Comparison of shear modulus. Adriana N E Ticoalu C h a p t e r 4 96

115 The boxplot in Figure 4.21 shows that the data is varied and largest variation is observed from 10N.E sample and smallest variation is observed from UN.VE Shear stress (MPa) UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Figure 4.21 Boxplot of shear stress Figure 4.22 shows load-displacement curves of a typical shear-tested specimen. For the specimen with polyester and UN.VE, initial cracks were followed by a gradual increase of load and displacement up to peak load, followed by a gradual decline of capacity until failure. The initial cracks, as shown as the first distortion in the graphs, were observed as cracks initiated from the notch root. Other specimens (5N.E, 5N.VE, 10N.E and 10N.VE) exhibit sudden brittle ruptures with failure initiated from notch root as well. Overall, failure can be expected at around mm of displacement. Adriana N E Ticoalu C h a p t e r 4 97

116 Figure 4.22 Load-displacement plot of typical shear-tested specimen: (a) specimens with polyester resin; (b) specimens with epoxy resin; (c) specimens with vinylester resin. Adriana N E Ticoalu C h a p t e r 4 98

117 4.4 Discussions Effects of fibre treatment and resin variation Studies on the effects of alkali treatment with sodium hydroxide on the mechanical properties of natural fibres have reported various results. Some reported improvement while some others reported declining values or no effects. The results are dependent on many factors, such as the type of fibre, fibre chemical composition, amount of treatment and the duration of treatment. The findings from the investigation reported earlier in Chapter 3 regarding the characteristics of gomuti fibre, indicate that alkali treatment to gomuti fibre increases the fibre s tensile strength. Therefore, it is expected that the incorporation of treated fibre in the composites will result in improved mechanical properties. The effect of NaOH treatment on the tensile strength of gomuti fibre composites can be observed from Figure 4.7. For specimens with polyester, 5N.PE has higher tensile strength followed by UN.PE then 10N.PE. Similarly, for specimens with epoxy, the composites with 5%-treated fibres provide slightly higher tensile strength compared to the other specimens. The results confirm that NaOH treatment has influenced the composites properties. For composite samples with different resin, the treatment has provided different effects. This relates to fibre-matrix interfacial adhesion that has been affected by fibre treatment. A correct and optimum amount and duration of NaOH treatment offers removal of impurities from the fibre and improve the fibre-matrix adhesion. In addition to fibre treatment, the tensile, flexural, compressive, and shear strength of gomuti fibre composites are also influenced by other factors such as the type of resin, manufacturing and curing conditions, and also defects and voids in Adriana N E Ticoalu C h a p t e r 4 99

118 the composites. Improper or prolonged NaOH treatment may result in adverse effects such as fibre damage. Changes in ductility of the composite specimens are observed from the compressive test curves (Figure 4.14). This can be more attributed to the compressive behaviour of the resins. Comparing the tensile strength of a single fibre to the tensile strength of a composite, it is clear that fibre treatment affects the properties. The treated fibre samples showed improved properties in comparison to the untreated, with 10N showing slightly lower properties compared to 5N. In the composite form, this behaviour of a single fibre is represented by the samples with epoxy resin. To evaluate the significance of the differences between tensile strength data based on fibre treatment and resin type, a two-way analysis of variance (ANOVA) was performed using MATLAB Statistical Toolbox (MATLAB R2011b). The outcome of the two-way ANOVA indicates that fibre treatment has not produced significant differences in the result of tensile strength, but it does indicate that the use of different types of resin has resulted in significant differences in tensile strength. This confirms that variations in tensile strengths were more influenced by type of resin, which may also refer to fibre-matrix compatibility, rather than fibre treatment. However, both fibre treatment and type of resin have influenced the tensile moduli. When two-way ANOVA was performed for compressive strength and modulus data, the outcome confirms that fibre treatments and different type of resins have caused significant differences in compressive properties. Overall, treatment with 5 % NaOH has resulted in higher compressive properties. Adriana N E Ticoalu C h a p t e r 4 100

119 Additionally, higher values were recorded for specimens with vinylester and polyester. Effects of fibre treatment on the flexural test results are similar to that for the tensile test. For specimens with polyester, treatment of the fibre has resulted in higher tensile strength for 5N but significantly lower for 10N. For specimens with epoxy, treatment of the fibre has slight effect on higher tensile strength. Two-way ANOVA confirms that fibre treatment has not produced clear differences between the flexural strengths of each composite sample. By varying the resins, significant differences were detected, however, both the fibre treatment and the type of resin have an influence on the results of flexural modulus. Two-way ANOVA of the data confirms that fibre treatment and resin type have influenced the shear strength. There are significant differences in the results that can be observed by ANOVA. Overall, fibre treatment and the use of epoxy have resulted in higher shear strength Comparison of theoretical prediction of mechanical properties based on Rule of Mixture (ROM) method and experimental results Micromechanical prediction of natural fibre composites can be evaluated using the Rule of Mixture (ROM) methods. However, the prediction is limited due to assumption made when using this method. A separate further study is needed in micromechanical modelling of natural fibre composites. Table 4.10 was completed using Equations 2.2 and 2.4, to obtain the predicted strength and modulus. Adriana N E Ticoalu C h a p t e r 4 101

120 Table 4.10 Theoretical prediction of tensile properties of gomuti fibre composites. Prediction Tested composite Fibre Fibre fraction Fibre properties Resin properties Composite results properties density sample Weight Volume (g/cm 3 ) Strength Modulus Strength Modulus σσ cc.pp Ec.p σσ cc Ec fraction (Wf) fraction (Vf) UN.PE $ 3160 $ UN.E # 4883 # UN.VE N.PE $ 3160 $ N.E # 4883 # N.VE N.PE $ 3160 $ N.E # 4883 # N.VE Source: #) Manthey, et al. 2013; $) Kabir, Adriana N E Ticoalu C h a p t e r 4 102

121 Comparison of the result from ROM prediction and test results for tensile strength and modulus is presented in Figures 4.23 and 4.24, respectively. Figure 4.23 Comparison of predicted and test result of tensile strength. Figure 4.24 Comparison of predicted and test result of tensile modulus. Adriana N E Ticoalu C h a p t e r 4 103

122 From Table 4.10, Figure 4.23 and Figure 4.24, it is seen that the predicted strength values are higher compared to the test results, while the predicted modulus value are lower compared to the test results. This means that the prediction model is less accurate when predicting the properties of gomuti fibre composites. Furthermore, the prediction values are largely influenced by the properties of the resin. Table 4.11 presents the percentage tensile strength of gomuti fibre composites when compared with the fibre s tensile strength. This is done to evaluate the contribution of fibre strength to the strength of the composites. As shown in Table 4.11, the percentage ranges from 12.6 % to 26.6 %. The percentage tend to follow the pattern of resin s strength, except for 5N.VE sample. This may be due to defects or voids in the specimens. Composite sample Table 4.11 Percentage tensile strength of the composite based on fibre tensile strength fibre density fibre weight fraction fibre volume fraction Fibre strength Resin strength Tested composite properties Percentage composite tensile strength from fibre σσ ff σσ mm σσ cc UN.PE UN.E UN.VE N.PE N.E N.VE N.PE N.E N.VE Based on the test results of gomuti fibre composites, more resin in the composite is likely to produce lower tensile and flexural capacity. Comparison can Adriana N E Ticoalu C h a p t e r 4 104

123 be observed from the result of 10 % and 20 % fibre weight fraction. Gomuti fibre/polyester composites with 10 % fibre fraction has resulted in average tensile strength of MPa and flexural strength of MPa (Ticoalu, Aravinthan & Cardona, 2011), while gomuti fibre/polyester composites with 20 % fibre has resulted in average tensile strength of 35.2 MPa and flexural strength of 72.2 MPa. Furthermore, the composite tensile properties seems to follow the tensile properties of resin, with the exception of 5N samples. The better compressive capacity of the composites compared to tensile capacity reflects the properties of the resin Failure of gomuti fibre composites Typical failure of the tensile specimens for each sample group is shown in Figure Distinct differences are shown by the specimen with polyester. For specimens with polyester, matrix cracking was followed by a de-bonding of fibres from matrix before fibre ruptures. Specimens with vinylester and epoxy show sudden brittle rupture where it is clear that there was no excessive fibre pull-out indicating good fibre-matrix bonding. An example of compression specimen failure is shown in Figure The specimens with polyester clearly show the weak fibre-matrix bonding which has resulted in brittle and shattering failure of the matrix. Specimens with polyester exhibit matrix cracking followed by fibre-matrix de-bonding before the ultimate failure, while failure of other specimens was started with fibre kinking and fibrematrix de-bonding. Generally, ruptures occurred in 1 or 2 diagonal directions. For flexural specimens, the typical failures are shown in Figure Failure of the specimens was initiated from the tension area. Matrix cracking was noticed Adriana N E Ticoalu C h a p t e r 4 105

124 as the first crack followed by fibre-matrix pull-out and then fibre breakage. For polyester specimens, weak fibre-matrix debonding is obvious at mid-span. Figure 4.25 Failure of tensile specimen. Adriana N E Ticoalu C h a p t e r 4 106

125 Figure 4.26 Failure of compressive specimen. Figure 4.27 Failure of flexural specimen. Adriana N E Ticoalu C h a p t e r 4 107

126 Figure 4.28 shows the typical failure of the shear specimens, where it can be seen that the cracks developed at the notch root and propagated to the opposite direction. Figure 4.28 Failure of shear specimen. 4.5 Conclusions This chapter has studied and discussed the strength and stiffness of gomuti fibre composites under tensile, compressive, flexure and shear loading. The mechanical properties were influenced by 5% and 10% sodium hydroxide treatments, and the use of different thermoset resins were investigated. Composite specimens were prepared from unidirectional-aligned gomuti fibre combined with Adriana N E Ticoalu C h a p t e r 4 108

127 polyester, epoxy or vinylester resins. Results indicated that compressive strengths of the composites are approximately 2 4 times the tensile strengths while shear strengths are approximately times the tensile strength. The use of different resins has resulted in significant variations in the mechanical properties, while fibre treatment provides minor variations. It was found that gomuti fibre composites may offer more significant benefits when used for products with low to moderate strength requirements. Results of mechanical tests show that unidirectional-aligned gomuti fibre composites with approximately 20% fibre fraction by weight, exhibit tensile strengths and moduli between 28.4 and 48.5 MPa and GPa, respectively; flexural strengths and moduli between 64.1 and 97.0 MPa and GPa, respectively; compressive strengths and moduli between MPa and GPa, respectively; and shear strengths between MPa. As observed from the results, compressive strengths of the composites are higher, approximately 2 4 times the tensile strength while shear strength of the specimens are approximately times the tensile strength. The combinations of untreated or treated gomuti fibre with different resins have revealed different effects on the mechanical properties of the composites. The 10N.VE combination has reached the highest average tensile strength, 5N.PE and 5N.VE have reached the highest compressive strength, 5N.E has obtained highest flexural strength, and 5N.E and 10N.E have reached best results for shear strength. Highest tensile, flexural and shear moduli were reached by 5N.PE while highest compressive modulus by 5N.VE. Overall, samples treated with 5 % NaOH demonstrated the best results. The effects of different treatment amount of 5 % and Adriana N E Ticoalu C h a p t e r 4 109

128 10 % NaOH to the strength of the composites were minor as confirmed by ANOVA. Meanwhile, the use of different resins has made significant differences in the mechanical properties. Samples with polyester showed poor tensile strength and flexural strength, and better compressive strength compared to epoxy samples. Samples with epoxy showed highest shear strength, while samples with vinylester showed highest compressive strength. Observing these properties and behaviour, the best combination of treatment and effects can be attributed to composites with 5% treated fibres and composites with vinylester resin, which offers better overall performance. Large variation in the mechanical properties was observed from the results. The variation is more significantly affected by the type of resin used. NaOH treatment to the fibre provides better fibre-matrix bonding and improved mechanical properties. However, the amount of NaOH of 5% and 10% shows only minor differences. Therefore, treatment with only low amount of NaOH can be deemed sufficient to provide cleansing and promote fibre-matrix adhesion. The use of natural fibre as composite ingredients with polymer thermoset resins may not necessarily increase the tensile strength of the resin, but it reduces the amount of resin used with acceptable properties. This study into the properties and behaviour indicates that gomuti fibre composites with thermoset resins may offer more significant benefits when used for products with low to moderate strength requirements. While strength and stiffness are important assessments of a material s performance and behaviour, evaluating the impact strength of the composites is important because some applications of the material itself may relate to a condition Adriana N E Ticoalu C h a p t e r 4 110

129 under impact loading, and that composites have a higher susceptibility when exposed to impact loading. Impact resistance of gomuti fibre composites will be discussed in Chapter 5. Adriana N E Ticoalu C h a p t e r 4 111

130 CHAPTER 5 INVESTIGATION ON THE IMPACT STRENGTH AND BEHAVIOUR OF GOMUTI FIBRE COMPOSITES 5.1 Introduction The impact strength of a material relates to its capacity to sustain shock or high-velocity sudden loading. These types of loading can be prompted by actions such as a punching action, nail or screw driving, objects falling on the material, striking and high blast conditions. Dhakal (2007) reported that impact loading occurred due to smashing, collisions, striking and or falling objects and debris. For a building construction element, impact loading may appear in the form of hail stones, falling debris and other damaging actions. Nail or screw driving during installation and maintenance are also potential source of impact loading. The previous Chapter 4 has discussed the strength and stiffness of gomuti fibre composites under tensile, compressive, flexural and shear loading. The results indicate that the composites samples exhibit moderate strength and stiffness when compared to various existing natural fibre composites and display close similarity with coir fibre composites. Coir fibre composites are well-known for their impact resistance. Since the impact resistance of a composite material is also an important measure of its performance, this chapter aims to study the impact strength and behaviour of gomuti fibre composites. Adriana N E Ticoalu C h a p t e r 5 112

131 Resistance to the impact loading of a material can be measured by its impact strength and behaviour when subjected to a sudden and high velocity load. Impact loading can be a single sudden action (single impact) or repetitive actions (repeated impact). The impact resistance of a material can be measured by performing an impact test which will measure the energy absorption capacity and characteristics, and the extent of damage to the impacted specimen. Impact properties of natural fibre composites have been reported in several publications. Natural fibre composite materials are known to be relatively prone to impact damage (Sparnins 2009; Oksman 2000; Garkhail, Heijenrath & Peijs 2000). Table 5.1 presents several studies into the impact strength of natural fibre composites. As observed from Table 5.1, sources of variations in the measurement of impact strength are, the type of fibre and resin, the test type (pendulum or drop weight), specimen type (Izod or Charpy; notched or unnotched), fibre fraction and fibre treatment. Direct comparison of the results of different tests are not possible. Adriana N E Ticoalu C h a p t e r 5 113

132 Table 5.1 Studies on impact strength of natural fibre composites. Fibre Resin Fibre fraction (%) Fibre treatment Type of test Impact strength (untreated) Fibre treatment effects Flax 1 Epoxy 22 (vol) 1 % NaOH 1 h Charpy kj/m 2 25 % increase Coir 2 Epoxy 30 (wt) - Pendulum/Notched 17.5 kj/m 2 - Coir 3 Epoxy - - Izod kj/m 2 - Coir 4 Polyester 17 (wt) 5 % NaOH 1 h Izod/Notched 433 J/m 46 % increase Sugar-palm 5 Epoxy 10 (vol) 0.5 M NaOH 8 h Izod/Notched J/m 13 % increase Palmyra 6 Polyester 30 (wt) 5% NaOH 30 min Izod/Notched 14 kj/m 2 42 % increase Flax 7 Epoxy 55 (vol) 5% NaOH 30 min Izod/Unnotched kj/m 2 7.3% reduction Linen 7 Epoxy 55 (vol) 5% NaOH 30 min Izod/Unnotched kj/m % reduction References: 1) George, Ivens and Verpoest n.d.; 2) Biswas, Kindo and Patnaik 2011; 3) Harish et al. 2009; 4) Rout et al. 2001; 5) Bachtiar, Sapuan and Hamdan 2009; 6) Thiruchitrambalam and Shanmugam; 7) Yan 2013 Adriana N E Ticoalu C h a p t e r 5 114

133 The effects of several parameters, however, can be observed. Alkali treatment, for example, has resulted in increased impact strength (George, Ivens & Verpoest n.d.; Rout et al. 2001; Bachtiar, Sapuan & Hamdan 2009; Thiruchitrambalam & Shanmugam 2012) and decreased impact strength (Yan 2013). In relation to the characteristics of the fibre constituents, coir fibre composites exhibited higher impact strength compared to flax and several other fibres. This high impact strength is attributed to the high elongation of coir (Wambua, Ivens & Verpoest 2003). In relation to the characteristics of the matrix constituents, among common thermoset resins, composites with polyester are seen to give higher impact strength compared to epoxy (Rassmann, Paskaramoorthy & Reid 2011). This is because, having a lower bonding capacity, composites with polyester dissipate the impact energy through fibre pull-out. This is also the case with the impact strength of neat resin (Rassmann, Paskaramoorthy and Reid 2011). In the case of composites with natural fibres, there are two major characteristics that can be observed regarding impact strength. Fibres that have lower elongation (approximately 2 5%) are shown to exhibit lower impact strength. On the other hand, fibres that have higher elongation (approximately more than 10%) exhibit higher impact strength. One example is coir fibre composites which generally exhibit lower tensile and flexural strength but higher impact strength. Wambua, Ivens & Verpoest (2003) reported that coir polypropylene (PP) composites exhibit better charpy impact strength compared to jute PP and kenaf PP composites, which was suggested to be due to the higher strain-to-failure of coir fibres. The composites with coir fibre exhibit an increased impact strength compared to the matrix only. This is deemed to be due to the toughness of coir fibre Adriana N E Ticoalu C h a p t e r 5 115

134 (Rout et al. 2001). Better impact properties were exhibited by cotton-epoxy compared to ramie-epoxy (Mussig 2008). It was found gomuti fibre composites exhibit higher elongation compare to other natural fibres but lower compared to coir. It is therefore expected that impact strength of gomuti fibre composites will also show high value. One of the major problems with natural fibre composites is fibre-matrix compatibility and adhesion. Wambua, Ivens and Verpoest (2003) stated that, for fibre composites, the impact response is greatly influenced by fibre-matrix interfacial bond strength and the properties of matrix and fibre. A poor fibre-matrix bonding or adhesion was found to cause higher energy absorption (Bax & Mussig 2008; Thomason & Vlug 1997). Therefore, in this study, fibre treatment and different matrix variables are evaluated to assess the impact characteristics of the variations. The focus of this chapter is to evaluate the impact properties of gomuti fibre composites with variations in fibre treatment and different thermoset resins. Variations of fibre treatment and different resin are further investigated to obtain comparison with the strength and stiffness under tensile, compressive, flexure and shear loading. 5.2 Materials and methods Material The natural fibre in this study is gomuti fibres which were obtained from Minahasa region in North Sulawesi, Indonesia. Bunch of gomuti fibres were Adriana N E Ticoalu C h a p t e r 5 116

135 manually removed from an Arenga pinnata tree. The fibres were manually cleaned from coarse impurities, rinsed with water and sun-dried afterwards. The fibres were brushed to obtain unidirectionally aligned fibres. Next, the fibres were trimmed to a length of approximately 400 mm to produce a composite. Three thermoset resins, namely polyester, vinylester and epoxy were used. Polyester and vinylester were prepared with a 1.5% catalyst while epoxy was prepared with a 4:1 ratio of resin and hardener. Fibre treatment aims to enhancing the fibre-matrix adhesion. In this study, fibre treatment was done by soaking two sets of fibres in water-dissolved sodium hydroxide (NaOH) pellets for approximately 2 h. One set in 5% NaOH and one in 10% NaOH. Two NaOH concentrations were chosen to observe the influence of fibre treatment to the mechanical properties of the gomuti fibre composites. The 2 h soaking time was performed to ensure proper intrusion of the solution to the fibres. Afterwards, the fibres were rinsed several times with water to remove any excess NaOH solution. The fibres were then strained and thoroughly air-dried just to remove excessive moisture before drying in an oven prior to composite fabrication, for approximately 1 h at 80 0 C, with the intention of drying the fibres without causing damage or degradation Specimen preparation Composite panels with approximately 20 % fibre weight fraction were prepared by the manual hand lay-up method, cured at room-temperature for a minimum of 24 h before being post-cured for 4 h at 80 o C. After curing, composite specimens were prepared as for charpy impact test with a width of 20 mm and Adriana N E Ticoalu C h a p t e r 5 117

136 length of 160 mm. The specimens were unnotched. Figure 5.1 shows the dimensions of specimen for impact test. Average thickness of each sample was recorded. Five specimens were tested for each sample. The specimens were prepared for testing in longitudinal direction. ISO 179-1:2010 is used as reference standard for the impact testing. Figure 5.1 Dimension of specimen for impact test Drop-weight impact test An instrumented drop-weight impact test was performed using Instron Dynatup The test is started by mounting each specimen horizontally, supported and secured at both ends to avoid slipping. The specimens were struck once (single impact) at mid span with a vertically falling mass of 2.84 kg. Table 5.2 outlines the designated sample names of the impact-tested specimens, and Table 5.3 presents the dimensions of the impact specimen. The test set-up and pictures of the impact specimen before and after impact are presented in Figure 5.2. The corresponding time, impact force, velocity and deflection of the specimens were recorded. Adriana N E Ticoalu C h a p t e r 5 118

137 Table 5.2 Designated names of the composite samples. Fibre treatment Thermoset resins Polyester (PE) Epoxy (E) Vinylester (VE) Untreated (UN) UN.PE UN.E UN.VE 5% NaOH (5N) 5N.PE 5N.E 5N.VE 10% NaOH (10N) 10N.PE 10N.E 10N.VE Table 5.3 Average dimension of the specimens. Sample name Width (mm) Length (mm) Thickness (mm) UN.PE UN.E UN.VE N.PE N.E N.VE N.PE N.E N.VE Impact strength is calculated using the following formula based on ISO 179-1:2010 for an unnotched charpy impact type of specimen: a cu = E c hb 103 (5.1) where a cu is the charpy impact strength (kj/m 2 ); Ec is the corrected energy absorbed by breaking the test specimen (J); h is the thickness of the specimen (mm); and b is the width of the test specimen. Adriana N E Ticoalu C h a p t e r 5 119

138 Figure 5.2 Drop-weight impact test set-up: (a) securing specimen to the test machine; (b) mounted specimen before impact; and (c) specimen after impact. 5.3 Experimental results Damage characteristics Damage to fibre composites from impact force may occur in different forms. These forms can be distinguished as matrix cracking, fibre-matrix debonding, fibre breakage and fibre pull-out. Dhakal et al. (2007) stated that there are four major failure modes of impact damage in fibre reinforced composites which are matrix cracking, delamination, fibre breakage and penetration of the impacted surface. Adriana N E Ticoalu C h a p t e r 5 120

139 Similarly, Wambua, Ivens and Verpoest (2003) reported that impact energy is dissipated by debonding, fibre and/or matrix fracture and fibre pull out. According to Wambua, Ivens and Verpoest (2003), under impact loading, fibre fracture dissipates less energy compared to fibre pull out. Therefore, when fibre fracture is dominant in the failure of composite, it is expected that the specimen exhibits less absorbed energy compared to specimens with fibre pull out as the main cause of failure. Figure 5.3 provides an example picture of the impacted specimen. The damage was typically initiated at the impacted point of the specimen. All of the specimens exhibited total damage at the point of impact and breakage on the area of the supports. A comparison of damage characteristics for each sample at the impacted point are presented in Figure 5.4. Figure 5.3 Example pictures of the impacted specimens. Adriana N E Ticoalu C h a p t e r 5 121

140 Figure 5.4 Damage mode of a typically impact-tested specimen taken from the impacted point (mid-span). As a result of the impact test, overall, the specimens exhibited total breakage in which the specimens were parted at the point of impact load. This infer that the energy is thoroughly absorbed and dissipated by the specimen. Overall, as can be observed, damage to the specimens vary between the samples but generally includes fibre-matrix debonding, fibre pull-out and fibre rupture. Distinct damage Adriana N E Ticoalu C h a p t e r 5 122

141 mode is shown by samples with polyester. It can be observed that samples with polyester exhibit excessive matrix chippings and fibre pull-out, especially for UN.PE, which can be attributed to a very low bonding capacity between untreated fibre and polyester. Furthermore, debonding and fibre pull-out followed by fibre ruptures are evident from the figures of specimens with polyester. Similarly, debonding and fibre pull-out are also observable from the pictures of UN.E, UN.VE and 10N.VE samples. This can be attributed to the processing of the specimen themselves. Poor processing causes poor fibre-matrix bonding, hence affect the properties of the composites. On the other hand, better fibre-matrix bonding is obvious from the ruptured surface of samples 5N.E, 5N.VE and 10N.E, as shown in Figure 5.3. It can be inferred therefore, that fibre treatment with 5 % NaOH has provided better bonding, which confirms the findings on the mechanical properties in Chapter 4. The use of epoxy resin also provides better bonding. Kim and Mai (1991, cited in Dhakal et al. 2007) indicate that the strong fibre-matrix interface bonding results in a brittle fracture mode with lower energy absorption while a weak bonding will results in a multiple shear mode with high energy absorption. Similar failure damage characteristics was shown by coir-polypropylene composites (Wambua 2001), where the main failure mode reported was fibre pullout due to weak fibre-matrix bonding. Furthermore, in Rassmann, Paskaramoorthy and Reid (2011), SEM micrographs of impact failure surfaces of composites with 22.5% fibre volume fraction shows that the kenaf-polyester laminates failed by fibre pull-out, the kenaf-epoxy laminates fail by fibre fracture and the kenafvinylester laminates fail by a combination of both fibre pull-out and fibre fracture. Adriana N E Ticoalu C h a p t e r 5 123

142 The kenaf-epoxy laminates exhibit a cleaner failure surface compared to kenafpolyester and kenaf-vinylester resins, which indicate short fibre pull-out lengths and goof fibre-matrix bonding, resulting in poor impact strength (Rassmann, Paskaramoorthy & Reid 2011) Impact force-time characteristics Figure 5.4 shows the typical force-time and energy-time plots of the impact tested gomuti fibre composites demonstrating the characteristic peak force and absorbed energy. As indicated by the curve, the incident energy was totally absorbed by the specimen through complete damage. No rebound energy was registered in the case of gomuti fibre composite samples. This energy versus time plot shows that the specimen has absorbed most of the energy in a form of damage (Guades et al. 2013). In this case, the energy is not stored but dissipated through the collapse of the specimen. Total absorption by the impacted specimen is shown by total breakage of the specimen. A similar energy-time plot characteristic was also shown by impacted hemp polyester composites (Dhakal et al. 2007). Bax and Mussig (2008) reported that during an impact test, there occurred three mechanisms of energy absorption that are fibre-matrix debonding, fibre pullout and fibre fracture. In the case when the absorbed energy is very small, the impactor bounces back and when most of the energy is absorbed through various modes of damage, there will be no rebound energy and at high energy level, perforation of the specimens happen (Guades et al., 2012). Adriana N E Ticoalu C h a p t e r 5 124

143 Figure 5.5 Example of force vs time and energy vs time plots of impacted gomuti fibre composites specimen. From Figure 5.5 and Table 5.4, it can be observed that the average time taken to reach maximum impact force is between 0.9 ms and 2.2 ms, while the average total time to complete failure is approximately 20 ms. The average maximum impact force needed to break gomuti fibre composite specimens is between 0.61 kn and 1.31 kn, while the average total absorbed energy is between 2.2 J and 8.3 J. Table 5.4 Average maximum impact force and energy absorbed by gomuti fibre composites. Sample Time to max. force (ms) Max. force (kn) Total absorbed energy (J) UN.PE 2.2 ± ± ± 1.4 UN.E 1.5 ± ± ± 0.6 UN.VE 1.3 ± ± ± 1.0 5N.PE 0.9 ± ± ± 0.9 5N.E 1.0 ± ± ± 0.4 5N.VE 1.1 ± ± ± N.PE 1.3 ± ± ± N.E 1.1 ± ± ± N.VE 1.2 ± ± ± 3.0 Adriana N E Ticoalu C h a p t e r 5 125

144 As observed from the load-time curves, it can be inferred that when subjected to impact loading, gomuti fibre composites exhibit tough break, that is according to ISO 179-2:1997(E): yielding followed by stable cracking, resulting in a force at the deflection limit less than or equal to 5 % of the maximum force. Figure 5.6 shows the representative force/area-time plots of gomuti fibre composites. In Figure 5.6(a), it can be observed that for the samples with polyester resin, the treated samples exhibit a higher impact load capacity. Further, it is shown that the untreated specimen exhibited different load-time characteristics while between the two treatments, the difference is marginal. Similar behaviour is shown by the samples with epoxy, as shown in Figure 5.6(b), while for samples with vinylester, the behaviour is varied. Adriana N E Ticoalu C h a p t e r 5 126

145 Figure 5.6 Force/area vs time plots of typical impact-tested specimen: (a) polyester specimen; (b) epoxy specimen; and (c) vinylester specimen. Adriana N E Ticoalu C h a p t e r 5 127

146 5.3.3 Absorbed energy-time characteristics Figure 5.6 presents the energy-time plots of gomuti fibre composites, grouped according to resins type. Although the plots show a similar pattern, they also show that there is a great variation in impact energy values between the samples, which could be attributed to the variability in treatment and different resin. These differences may have resulted in variations in interfacial bonding and rupture characteristics. Some potential sources of difference may also inevitably exist due to the manufacturing and machining of specimens. All composite samples in this study exhibit a similar typical energy versus time plot where there was no rebound energy and complete damage was observed for all specimens. Figure 5.7 clearly shows that specimen with untreated fibres absorbed more energy compared to the treated samples, except for samples with vinylester. By taking into account that higher impact energy is related to poor fibre-matrix adhesion, this difference can be attributed to poor matrix-adhesion due to composite preparation or the large variability existing in the properties of natural fibre composites. Adriana N E Ticoalu C h a p t e r 5 128

147 Figure 5.7 Energy/area-time plots of typical impact-tested specimen: (a) polyester specimen; (b) epoxy specimen; and (c) vinylester specimen. Adriana N E Ticoalu C h a p t e r 5 129

148 5.3.4 Impact force-deflection behaviour Figure 5.8 presents the force-deflection curve of typical impact-tested gomuti fibre composites showing the position of maximum impact force and the corresponding deflection, the failure mode of the specimen and the deflection limit. As shown by the curve, the specimen exhibited tough break which is anticipated for fibre composite material. Figure 5.9 shows the load-deflection plot of typical impact-tested specimens, grouped by the type of resin. It shows, that in general, the deflection at peak load is less than 4 mm for every sample, except UN.PE which showed a deflection at peak load of over 6 mm. Figure 5.8 Force-deflection curve of typical impact-tested gomuti fibre composites. Adriana N E Ticoalu C h a p t e r 5 130

149 Figure 5.9 Force/area-deflection plots of typical impact-tested specimen: (a) polyester specimen; (b) epoxy specimen; and (c) vinylester specimen. Adriana N E Ticoalu C h a p t e r 5 131

150 5.4 Discussion Impact strength Table 5.5 presents the average of the calculated impact strength of gomuti fibre composites, based on Equation 5.1. Overall, samples with polyester showed higher impact strength while samples with epoxy showed lower impact strength. As shown in the table, UN.PE has the highest impact strength while 10N.E exhibits lowest impact strength. Using the value of UN.PE as the basis for comparison, Figure 5.10 presents the percentage comparison of the average impact strength. Table 5.5 Average impact strength of gomuti fibre composites. Sample Impact strength (kj/m2) PE E VE UN 71.7 ± ± ± N 43.8 ± ± ± N 49.4 ± ± ± Figure 5.10 Percentage comparison of average impact strength. Adriana N E Ticoalu C h a p t e r 5 132

151 As it can be observed from Table 5.5 and Figure 5.10, all the samples exhibit a reduction in impact strength capacity when compared to UN.PE. It can be inferred therefore that the use of untreated fibre and the use of polyester resin has resulted in extremely poor fibre-matrix compatibility. A reduction of impact strength exhibited by all the samples compared to UN.PE is between approximately 29.4 % %. This shows that a relatively large data variation has resulted from fibre treatment and resin type variables. This is generally common for natural fibre composites Effects of fibre treatment Generally, fibre treatment is used to enhance fibre-matrix adhesion. Better fibre-matrix adhesion is sought to improve the mechanical properties of a composite material. In the case of impact strength, contrast findings have been reported. As previously presented in Table 5.1, several studies have reported that due to good fibre-matrix bonding, the impact strength of the treated specimens is increased, while other studies indicated the opposite findings that good fibre-matrix bonding has resulted in decreased impact strength. Results in this study correspond to the latter finding. As was observed in Table 5.5 and Figure 5.10, treatment to the fibre has generally resulted in a lower impact strength, except for 10N.VE. To further evaluate the effects of fibre treatment to gomuti fibre composites samples, a graphical illustration of the comparison of the impact strength is presented in Figure Adriana N E Ticoalu C h a p t e r 5 133

152 Figure 5.11 Comparison of the impact strength of gomuti fibre composite samples: effects of fibre treatment. Varied results of impact strength are obvious from Figure Overall, alkali treatment to gomuti fibre has resulted in altered impact strength. For polyester and epoxy specimens, fibre treatment has caused a reduction in impact strength although the difference between the amount of treatment (5 % and 10 %) is marginal. These findings generally indicate that alkali treatment of the fibre has provided improved fibre-matrix bonding. The literature reports that higher impact strength relates to inferior fibre-matrix bonding, while lower impact strength can be attributed to better fibre-matrix bonding (Bax & Mussig 2008; Thomason & Vlug 1997; Wambua, 2001; Sarkar, 2002; Canche-Escamilla 2002). In the case of the vinylester samples, as seen in Figure 5.8, 5% treatment has resulted in lower impact strength compared to the untreated sample and the 10% treated sample. However, unlike 10N.PE and 10N.E, the 10N.VE sample exhibited higher impact strength compared to the untreated and 5 % treated vinylester samples. Furthermore, the impact strength value is significantly different from 5N.VE. This may indicate that the 10N.VE sample has relatively poor fibre-matrix Adriana N E Ticoalu C h a p t e r 5 134

153 bonding which may have occurred during processing. Fibre defects are also a possible explanation. Comparing the results from the testing of mechanical properties in Chapter 4, where 5 % treatment and the use of epoxy have provided better fibre-matrix bonding and relatively improved properties compared to untreated samples, it can be inferred that the effects of fibre treatment to the impact properties is confirmed Effects of different thermoset resin The choice of resin determines the fibre-matrix compatibility. Epoxy is known to provide better bonding, while polyester provide less bonding and vinylester bonding fell between polyester and epoxy. By observing the bar chart in Figure 5.12, it is clear that overall gomuti fibre composite samples with polyester resin exhibit highest impact strength, especially for the untreated and 5 % NaOH treated samples. For the 10% NaOH treated samples, the use of polyester and vinylester only resulted in almost negligible difference. For the treated samples (5N and 10N), the use of epoxy resulted in lowest value of impact strength. Similar findings were reported in a study by Rassmann, Paskaramoorthy and Reid (2011) on mechanical properties of kenaf fibre with three different resin (polyester, vinylester and epoxy). The study reported that the type of resin significantly affect the impact energy. It was found that kenaf with polyester laminates exhibit the highest absorbed energy before failure while kenaf with epoxy laminates exhibit the lowest absorbed energy for all fibre volume fractions. Their study concluded that kenaf with epoxy laminates have better fibre-matrix bonding Adriana N E Ticoalu C h a p t e r 5 135

154 compared to kenaf-polyester and kenaf-vinylester, which resulted in lower absorbed energy (Rassmann, Paskaramoorthy & Reid 2011). Figure 5.12 Comparison of the impact strength of gomuti fibre composite samples: effects of different thermoset resin. 5.5 Conclusions The impact strength characteristics of gomuti fibre composites have been evaluated using an instrumented drop-weight charpy impact test. Effects of variations in fibre treatment and resin type were evaluated. It was found that, overall, samples of gomuti fibre composites in this study exhibit impact strength between kj/m 2 and 71.7 kj/m 2. The use of epoxy has resulted in a lower impact strength which can be attributed to better fibre-matrix bonding. The use of polyester has the opposite effect, while vinylester shows varied results in between these two. Adriana N E Ticoalu C h a p t e r 5 136

155 Treatment of the fibre has resulted in lower impact strength which again confirms improved fibre-matrix bonding. It was observed that different amounts of fibre treatment resulted in small differences in the impact strength. It was observed from the tests that the incorporation of gomuti fibre into the resin matrices, resulted in the composites displaying a tough break failure mode, which is confirmed by the load-deflection plot. Pictures of the specimen after impact tests shows that the specimen with the highest impact strength (UN.PE) show excessive matrix cracking, fibre pull-out and fibre-matrix debonding at the impacted area, while specimens with lower impact strengths show relatively flat splits. Strength and behaviour of a material under impact loading are important properties to identify, especially when assessing the feasibility of the material for the use in building and construction. With the impact strength performance of approximately kj/m 2, gomuti fibre composites are suited for applications such as roof, decking and wall cladding. The next chapter will discuss about a study into the potential of gomuti fibre composites for roof tile applications. Adriana N E Ticoalu C h a p t e r 5 137

156 CHAPTER 6 INVESTIGATION ON THE MECHANICAL PROPERTIES OF GOMUTI FIBRE COMPOSITES FOR ROOF TILE APPLICATIONS 6.1 Introduction The previous chapters (Chapter 4 and Chapter 5) have highlighted the study of the mechanical properties and impact behaviour of gomuti fibre composites when combined with thermoset resins. With their strength, stiffness and impact resistance, gomuti fibre composites possess characteristics within the acceptable range for elements requiring low to moderate properties. Rijswijk, Brouwer and Beukers (2001) suggested that natural fibres can be potentially developed into products such as roofing panels, fluid containers, bridges and small boats. Since gomuti fibre itself has been used as roofing and insulation material, it is of particular advantage to examine the viability of developing gomuti fibre composites roofing. Binding the fibre with a matrix will provide not only a more aesthetic shape, but also enhance the usage of the fibre. The use of fibre only as thatch roof exposes the fibre into susceptibility to weather conditions. Combination with resin matrix provides longer-lasting element and better protection from weather. Furthermore, the application of natural fibre composites as roof material would encourage a reduction in production of synthetic fibres. It was reported in Marsh (2000) that the roofing sector has the second largest percentage (after Adriana N E Ticoalu C h a p t e r 6 138

157 electricity) of glass fibre usage in Europe s building and construction industry. These sectors where moderate strength is required and high demand is shown, offer a significant opportunity for the introduction of natural fibre composites. This chapter aims to investigate the mechanical properties of gomuti fibre composites for use as roof tiles. The investigation involves a parametric study on the strength and stiffness of gomuti fibre composites as a roof tile, based on the investigated mechanical properties. Results from this study will be beneficial in the development of natural fibre composites for building and construction purposes. As there is no specific standard for natural fibre composites roof tiles, this study will adapt existing standards for roofing material, terracotta (clay tile) and concrete roof tiles. Furthermore, there has been relatively limited literature assessing the performance of natural fibre polymer composites roof tile, and the existing installation of roof tiles is typically based on conventional practices or respective company-based research. 6.2 Overview of roofing materials and roof tiles Traditionally, gomuti fibre roof is prepared manually, either by stacking brushed fibres directly onto the roof frame or by using bamboo clamps. Other plantroofing or thatch roof are typically installed in similar way. Incorporating a resin matrix with the fibre offers more opportunity for various shapes, strength and stiffness to suit needs. Roofing made of natural fibre composites has been investigated in several studies. Satyanarayana et al. (1986) reported the use of coir-polyester composites to develop various end products including roofing material. Sisal-CNSL (Bisanda Adriana N E Ticoalu C h a p t e r 6 139

158 1993) and recycled paper-aeso (Dweib et al. 2006) have been developed into roof material. O Donnell, Dweib and Wool (2004) developed natural fibre composites from recycled paper-aeso and suggested that the composites could be suitable for applications in housing construction materials, furniture and automotive parts. Common materials for roofing are straw or vegetable (thatch), clay tiles (terracotta), metal, concrete, fibrous composites, bituminous felt (asphalt shingle), timber (shake), etc. Table 6.1 outlines several roofing materials with their attributes. Thatch roof from straw or other plants offers low cost, but it is extremely susceptible to fire danger. Clay and concrete tiles are exceptionally durable, however, unit weight of the tiles is a major issue as lightweight material is required for roofing. Table 6.1 Roofing materials and their respective attributes. Roof type Materials Positive attributes Negative attributes Thatch Straw, other plants Low cost Fire hazard Metal Aluminium, Zinc Low weight Moderate to excellent resistance Asphalt shingles Fibreglass shingles Clay tile (Terracotta) Organic felt mat: asphalt saturated wood fibres or recycled paper Non-woven fibreglass mat and phenolic resin to impact from hail Better tear resistance Resistance to nail pull-through More pliable Easier to work with in cold weather Lightweight Not fireproof Not waterproof Premature failure Clay Low cost, durable Weight Concrete tile Concrete Durable Weight Timber slate Timber Durable, good elasticity Difficulty sourcing Composite tile Various Design to suit needs No standard Fibre-cement tile Portland cement and wood fibres (or previously, asbestos) Light-weight (Adapted from Bliss 2006) Processing Adriana N E Ticoalu C h a p t e r 6 140

159 Furthermore, as outlined in Table 6.1, fibre composite roof tiles incorporating fibre glass and phenolic resin is lightweight, and considering the strength of glass fibre, the tile should have been strong. However, premature failure of fibre glass shingles have been reported (Bliss 2006). The failure related to the brittle nature of the phenolic resin that was used. Timber slate is cost efficient, aesthetically pleasing and has good elasticity. However, today sourcing of timber is becoming major issue. The use of asbestos is not permitted in several countries including Australia due to health concerns. According to AS , a roof tile is a moulded interlocking nonmetallic product used to form the field of the roof (AS , p. 5). Some early practice of roof tile, however, incorporates the use of non-interlocking roof tiles. The main function of roof tiles is to provide covering to a house or building. A roof tile is one type of roofing which covers broad range of materials, methods and design. Current practice of roof tiles manufacturing stretches from small-scale home industry of clay tiles to big-scale industrial productions of concrete and metal tiles. The dimension of roof tile profile is designed to accommodate wind and rain. In current practice, the dimension is designed by manufacturers. The building design and environmental conditions must also be considered. To comply with the standards, the roof tiles need to be designed to satisfy the requirements such as dimensional tolerances, freezing and thawing, durability, permeability, water absorption, transverse bending strength, resistance to salt attack, wind and weather resistance (ASTM C (Reapproved 2009), ASTM C (Reapproved 2009), ASTM C and AS ). It is Adriana N E Ticoalu C h a p t e r 6 141

160 therefore essential to understand how the roof tiles behave under the specified loading in order to design gomuti fibre composites roof tile. This study covers the evaluation of the transverse bending strength capacity of gomuti fibre composites roof tile. Many types of roof tiles are currently available on the market, such as timber shakes, terracotta and concrete tiles. Various roof tile profiles, materials, shapes and textures are also available. Figure 6.1 shows pictures of several types of roof tiles. Figure 6.1 Various types of roof tiles Structurally, roof tiles are laid uniformly on timber battens. The timber battens themselves rest on the rafters. Figures 6.2 and 6.3 show schematic pictures of a simple plain roof tile arrangement. Figure 6.2 Schematic of roof tile arrangement (front view) Adriana N E Ticoalu C h a p t e r 6 142

161 Figure 6.3 Schematic of roof tile arrangement (top view) Several publications have been found on natural fibre composite roof tiles, mostly incorporating cement as the matrix material. Savastano et al. (1999) reported the performance of roof tiles made of plant fibre and blast furnace slag cement. It was found that, for roof tiles containing coir fibre, the maximum load capacity was 454 N for the tile with 10.9 mm thickness (487 mm long x 263 mm wide). Silva et al. (2010) investigated the properties of corrugated and flat sheets sisal-cement composites having dimensions of 400 x 400 x 12 mm. It was found that the corrugated sheet provides increased inertia of the cross section and higher ultimate load compared to the flat sheet (Silva et al. 2010). With a three point bending test, the ultimate load for the flat sheet was found to be 3.91 kn and kn for the corrugated sheet (Silva et al. 2010). Adriana N E Ticoalu C h a p t e r 6 143

162 Table 6.2 presents the typical dimensions and weight of roof tiles. Based on the typical dimension of clay and concrete tiles, as outlined in Table 6.2, the targeted tile dimension chosen in this study is 330 mm (width) and 420 mm (length). Style Table 6.2 Typical dimensions and weight of roof tiles. Width, w (mm) Length, L (mm) Weight, m (kg) 1 Spanish S-tile (clay): 13 x 16 ½ in n/a 1 Spanish S-tile (clay): 9 x 14 in n/a 1 Flat shingle (clay): 6 8 x in n/a 1 Interlocking (clay): 9 13 x in n/a 1 Concrete tile: x 16 ½ 17 in n/a 2 CSR Monier Georgian Boral Linea Bristile Classic Source: 1) Bliss, 2006; 2) CSR Monier; 3)Boral; 4) Bristile 6.3 Performance requirements of roof tiles Roofing material is designed to be lightweight, fire resistant, water resistant and weather resistant (ultraviolet light and chemical debris). It provides coverage to the house or building. In the area where hail stones and snow are expected to occur, the roofing has to be designed accordingly. Therefore, specific strength and stiffness are required for roofing materials. The strength of roof tiles are measured with the transverse bending test. AS specifies that when roof tiles are tested for their average transverse Adriana N E Ticoalu C h a p t e r 6 144

163 breaking strength, according to AS , the transverse breaking load for each specimen shall not be less than N per millimetre of exposed width. Based on AS , for a 330 mm width of tile, the transverse breaking load for one tile shall not be less than N, which is the result of 330 mm x N/mm. Based on ASTM C (2009) and ASTM C (2009), one tile in dry condition must to sustain not less than 1112 N transverse loading with a testing span of 305 mm. Therefore, in this parametric study, the load of 1112 N will be the design load. Figure 6.4 presents the transverse bending strength test configuration and loading conditions that will be simulated in this study. As shown in the Figure, the test configuration depicts a three point bending test, with a specified span of 305 mm. The supports and loading element are of 50 mm wide timber battens. A detailed description of the roof tile profile and loading condition is provided in Table 6.3. Deflection limit of L/200 is chosen in this study since there are no specific requirements of the deflection limits for roof tile. Figure 6.4. Transverse bending strength test configuration Adriana N E Ticoalu C h a p t e r 6 145

164 Table 6.3 Roof tile profile under consideration. Description Value Width (w) 330 mm Total length of tile (L 0) 420 mm Test span (L) 305 mm Deflection limit (Span/200) 1.53 mm Design load per tile (P) 1112 N Therefore a parametric study has been devised to incorporate these specifications. Behaviour of the roof tile under transverse loading is investigated in this study. Effects of different properties variations on the behaviour of gomuti fibre composite roof tile were also evaluated. 6.4 Design of gomuti fibre composites roof tiles The dimensions of the roof tile under investigation in this study, which are 330 mm (width) and 420 mm (length), are chosen based on the typical dimensions of roof tile currently in the market. Figure 6.5 shows the plain roof tile profile that will be investigated. Figure 6.5. Flat roof tile profile Adriana N E Ticoalu C h a p t e r 6 146

165 Depending on the resin type and fibre treatment (as investigated in Chapter 4), gomuti fibre composites offer a range of tensile strength from MPa, flexural strength from MPa, compressive strength between MPa, and shear strength between MPa. Details of the ranges of mechanical properties of gomuti fibre composites are presented in Table 6.4. The values are the results from the experimental investigation. For application of gomuti fibre composites in construction industry, it is expected that the performance satisfy the criteria of the intended construction material. Table 6.4 shows the performance range of gomuti fibre composites for application in construction industry. With the performance range as shown in the table, particularly for applications targeted in construction industry, these properties are suitable such as for roof, decking, wall cladding, and lower load-bearing elements. Table 6.4 Range of mechanical properties of gomuti fibre composites for roof tile design. Properties Polyester Epoxy Vinylester Overall design range Tensile strength (MPa) Modulus of elasticity (GPa) Flexural strength (MPa) Compressive strength (MPa) Shear strength (MPa) Impact strength (kj/m 2 ) Based on the values of the mechanical properties of gomuti fibre composites as presented in Table 6.4, a range of modulus of elasticity of GPa was chosen as design parameters to simulate the behaviour of the roof tile. Furthermore, Adriana N E Ticoalu C h a p t e r 6 147

166 the range of tensile strength of t MPa justifies the chosen design variables of MPa. Furthermore, preliminary calculations were made to determine the minimum required thickness of the roof tile based on gomuti fibre composites properties and the design loading condition. Since the tile is designed to be subjected to transverse breaking strength, the minimum thickness based on strength is obtained by using tensile and compressive strength values. Meanwhile, the minimum thickness to satisfy the stiffness criteria is calculated based on the modulus and the deflection limit. Table 6.5 presents the tabulation of calculations of t-min based on the characteristics strength and stiffness capacity. The characteristics strength and stiffness were calculated using the following formulas: σσ cchaaaa = σσ mmmmmmmm 1.64ss (6.1) EE cchaaaa = EE mmmmmmmm 1.64ss (6.2) where, σσ cchaaaa = characteristics strength (MPa); σσ mmmmmmmm = mean strength (MPa); EE cchaaaa = characteristics modulus (GPa); EE mmmmmmmm = mean modulus (GPa). The minimum thicknesses based on shear and flexural strength were also calculated, however, the results show that the values are significantly lower. This shows that failures due to shear and flexure are not governing. Details of the calculation for t-min based on shear and flexural strengths are presented in Appendix E. The following basic mechanical formulas were used in the calculations: Adriana N E Ticoalu C h a p t e r 6 148

167 Bending strength σσ = MM ZZ (6.3) σσ = (PPPP 4 ) ( 1 6 bbh2 ) (6.4) Shear strength ττ = QQQQ bbbb (6.5) ττ = PP 2 (1 8 bbh2 ) bb( 1 12 bbh3 ) (6.6) Deflection δδ = PPLL3 48 EEEE (6.7) where, σ = bending strength (MPa); M = bending moment (N.mm); Z = Section modulus (mm 3 ); P = Load (N); b = section width (mm); h = height or thickness (mm); τ = shear strength (MPa); S = first moment of area (mm 3 ); I = moment of inertia (mm 4 ); L = span (mm); E = modulus of elasticity (MPa). Adriana N E Ticoalu C h a p t e r 6 149

168 Table 6.5 Determination of minimum required thickness based on the properties of material and loading condition. Composite samples UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Characteristics Tensile strength Characteristics Modulus Characteristics Compressive strength M/σ Z-max Calculated t - min (mm) (MPa) (GPa) (MPa) (Tensile) (Compressive) (mm3) based on strength based on deflection Adriana N E Ticoalu C h a p t e r 6 150

169 Table 6.5 shows that based on strength limits of gomuti fibre composites, the minimum thickness is between 5.85 mm and 8.64 mm. Further, it shows that there is small variations in the minimum thickness when strength is the design criteria. With the strength exhibited by gomuti fibre composites, a roof tile with a thickness of 6 9 mm can sustain the loading condition. Based on the shear and flexural strengths (Table E.1 in Appendix E), the calculated minimum thickness is between 0.06 and 0.09 mm and between 3.99 mm and 4.90 mm, respectively. Based on the serviceability limits, however, the minimum thickness (as in Table 6.5) is between mm and mm. Therefore, this range of deflection has become the controlling parameter. Based on these calculations, this study will cover the thickness variable from mm. Table 6.6 presents the design variables of thickness, modulus of elasticity and tensile strength that will be the focus in this parametric study. Table 6.6 Design variables for gomuti fibre composites roof tile. Thickness (mm) Modulus of Elasticity (GPa) Tensile strength (MPa) (t-tile) (E-tile) (Strength-tile) Behaviour of gomuti fibre composites roof tile The study of the behaviour of gomuti fibre composites roof tiles with variations of Young s modulus of the tile (E-tile), thickness of the tile (t-tile) and Adriana N E Ticoalu C h a p t e r 6 151

170 strength of the tile (Strength-tile), is presented in this section. Since there are substantial variations in the properties of natural fibre composites, numerical simulation can be complex. A Strand7 simulation of gomuti fibre composites roof tile under transverse loading is presented in Appendix E. The composite flat roof tile has been modelled using plate element (unidirectional laminate plate) with input parameters of modulus and thickness variation. Tile dimensions are fixed to 330 mm x 420 mm. The plate element has been meshed to 340 plates and 385 nodes. A static load modelled as plate face load with a magnitude of 1112 N was applied to the span of 305 mm. The simulation inputs are presented in Tables E.3(a) E.3(e). When subjected to the specified loading conditions, the roof tile shows the deflection modes as shown in Figure 6.6. In this Figure, the deflection along the length of the roof tile with variations of thickness (t-tile) and modulus (E-tile) is presented. The figure was constructed with deflection data obtained from a simulation. Details of the simulation is presented in Appendix E. As shown in Figure 6.6, significant difference in deflection is observed when t-tile is 10 mm, for the given value of modulus. As expected, increasing the thickness reduces the variations in the deflections. When t-tile is 20 mm, the difference in E-tile only causes marginal difference in the deflection. It can be inferred therefore that when thickness of the tile is designed to be 20 mm, there will be almost negligible effects from E-tile. It is evident from the deflection modes that differences in the stiffness affects the behaviour of the tile. Adriana N E Ticoalu C h a p t e r 6 152

171 (a) t tile: 10 mm (b) t tile: 12.5 mm (c) t tile: 15 mm (d) t tile: 17.5 mm (e) t tile: 20 mm Figure 6.6 Deflection along the length of tile for different modulus and different thicknesses. Adriana N E Ticoalu C h a p t e r 6 153

172 Under transverse bending load of 1112 N, by varying the elastic moduli of the tile (E-tile), the corresponding thickness of tile (t-tile) can be calculated. Figure 6.7 shows the plot of E-tile versus deflection. When lower modulus material is used, the t-tile increases. When higher modulus material is used, the t-tile decreases and the difference is smaller. The Figure shows the effects of variations of E-tile to the deflection, with different t-tile. As can be observed from the plots, a larger difference is evident for tile with 10 mm compared to the other thicknesses. The difference in modulus is almost negligible when the tile thickness increases to 20 mm. With deflection limit of 1.53 mm, tile with 17.5 mm and 20 mm satisfy the loading criteria. Meanwhile, for tile with 15 mm thickness, to satisfy the criteria, it needs to have a minimum modulus of 4.5 GPa. From Figures 6.7 and 6.8, it can be observed that with the deflection limit of 1.53 mm, any given E-tile over 4 GPa will satisfy the criteria when t-tile is over 15 mm. However, with thickness of 10 mm, the E-tile has to be designed to have a modulus of more than 15 GPa. Figure 6.7 E-tile vs deflection plot for the thickness variations Adriana N E Ticoalu C h a p t e r 6 154

173 Figure 6.8 E-tile vs deflection plot for the thickness variations Figure 6.9 shows the behaviour of the roof tile for different modulus and thicknesses. It is evident that with a lower E-tile, the thickness will have higher value, and when the E-tile is increased, the required thickness will also reduce. Figure 6.9 Behaviour of a roof tile with variations of elastic moduli Adriana N E Ticoalu C h a p t e r 6 155

174 Figure 6.10 shows the calculated P-max based on the strength of the tile and the variations of tile thickness. For the 10 mm thick tile, the capacity to sustain loading is lower compared to a thicker tile. Furthermore, even with a varying of the tile strength from 20 MPa to 50 MPa, there is only small difference in the loadbearing capacity. When the tile is 20 mm thick, a large difference is made by just changing the targeted tile strength. Figure 6.10 P-max calculated based on strength of the tile and thickness variations. Figure 6.11 shows that when the thickness of the tile is increased, the strength required will be lower. When the thickness is decreased, the required strength will be higher. Adriana N E Ticoalu C h a p t e r 6 156

175 Figure 6.11 Plate thickness vs plate stress 6.6 Gomuti fibre composites roof tile design guidelines Figure 6.12 shows curves of thickness versus modulus and thickness versus strength of the tile. The figure was constructed based on the determined design variables (see Table 6.6) of modulus of elasticity with a range of 4.0 to 6.5 GPa. The Figure shows that a tile with higher modulus requires less thickness while tile with lower modulus requires more thickness. Similarly, a tile with a higher strength requires less thickness while tile with lower strength requires more thickness. For a roof tile design, these following equations based on the plots in Figure 6.12 are obtained: σ = t -2 (strength) (6.8) E = t -3 (stiffness) (6.9) Adriana N E Ticoalu C h a p t e r 6 157

176 Figure 6.12 Roof tile design curves Adriana N E Ticoalu C h a p t e r 6 158

177 Figure 6.12 can be used to obtain the required modulus of elasticity and strength of a material for roof tile, based on a given thickness of the tile. Furthermore, the plots can also be used to obtained the required thickness and strength based on a given modulus. When the thickness is specified, for example 12.5 mm, the modulus of the tile has to be minimum of 8 GPa, with the strength of not less than 10 MPa. When a value of modulus is given, for example 6.25 GPa, the required minimum thickness is 13.6 mm. Further from Figure 6.12, when the values of modulus and design strength of a material are given, the design thickness could be determined. For example, a composite material has a modulus of 4.0 GPa and a strength of 10 MPa. By plotting the value of the modulus towards the curve of t-tile versus E-tile or by substituting the modulus in Equation 6.7, the minimum required thickness can be obtained. In this case, the thickness is 15.8 mm. By plotting the value of the strength towards the t-tile versus strength-tile curve, the minimum required thickness obtained is 12.4 mm. Therefore, the design thickness will be the 15.8 mm and stiffness is the governing criteria. In a special case, for example, the known modulus is 12 GPa and the strength is 10 MPa. From these values, the minimum required thickness are 10.9 mm and 12.4 mm, respectively for the stiffness and strength criteria. The design thickness in this case is the 12.4 mm, which is based on the strength criteria. In this case, the strength has become the governing criteria. By taking into account a safety factor (SF) of 2, the required characteristics strength will be twice the strength value. In the case of gomuti fibre composites, with a range of characteristics strength between MPa, which is based on the experimental data, although a safety factor of 2 is incorporated, the strength requirements are still satisfied. This clearly indicates that for gomuti fibre composites Adriana N E Ticoalu C h a p t e r 6 159

178 roof tile, stiffness is the governing criteria. In cases when a material possess a high modulus and low strength, the strength is more likely to become the governing criteria. Design parameters based on shear strength and impact strength may become important in several cases of roof tile, especially at the time when hail stones, falling debris, high-blast and heavy maintenance conditions are expected. These conditions will be investigated in future work. Table 6.7 presents examples of roof tile design based on Figure The values in the first three rows are actual values obtained from investigation for the 5 % treated gomuti fibre composites samples. It can be observed that for the particular composites samples, with a range of modulus from 4.8 to 6.3 GPa, the required thickness of a roof tile is between 13.5 to 14.8 mm and the design strength of 7 to 8.4 MPa. With SF = 2, the characteristics strength will be 16.8 MPa. Modulus of elasticity (E-tile) Table 6.7 Examples of roof tile design Thickness (t-tile) Design strength (strength-tile) (GPa) (mm) (MPa) Conclusions This chapter has outlined the design for a gomuti fibre composite roof tile. It was found that, under the designated transverse bending strength design load, the governing criteria was the stiffness of the composites. Adriana N E Ticoalu C h a p t e r 6 160

179 The strength and stiffness capacity of gomuti fibre composites for potential application as roof tile were evaluated by parametric study with the data from experimental results. Based on their strength, the composites can sustain the specified transverse bending load. It was observed in the simulation, however, that the composites of gomuti fibre and thermoset resins are most likely to exhibit tensile failure when the deflection reached 1.53 mm, with a minimum thickness of approximately mm. With E-tile between 4.0 MPa and 6.5 MPa, and strength of 20 MPa and 50 MPa, failure of the composites is determined by the stiffness. While strength criteria is not a governing factors in the design, when the design strength incorporate a safety factor, strength of the composites will become an important criteria. Based on the principles of mechanics of materials, graphs of t-tile versus E-tile and t-tile versus strength-tile was constructed. The graphs can be used to obtain the design thickness based on modulus, or to obtain modulus and strength based on the thickness. If a thickness is given, the modulus and the corresponding strength can be obtained from the graph. When a value of modulus is given, the required minimum thickness can be obtained. When the values of modulus and strength of a material are known, the corresponding thickness can be determined. As it is informed in literature, fibre composite design is more influenced by stiffness and/or strain-at-failure criteria. Increasing the stiffness by means of adjusting the tile profile into different type of tile such as pan-cover or S-style is an alternative option when thickness is to be retained at below 10 mm. Moreover, in some special cases, the shear strength or the impact strength may become a governing failure criteria. These design approaches will be investigated in a future work. Adriana N E Ticoalu C h a p t e r 6 161

180 This chapter has provided a numerical study on potential application of gomuti fibre composites as roof tile. It also has provided a fundamental design procedures for composites roof tiles in general as an alternative between the existing roof tile materials and thatch roofing. Adriana N E Ticoalu C h a p t e r 6 162

181 CHAPTER 7 CONCLUSIONS 7.1 General In this research work, the characteristics and behaviour of gomuti fibre and its composites with thermoset resins have been investigated. The strength and stiffness characteristics of the composites relates to the characteristics of the gomuti fibre and the type of matrix used. Furthermore, it was found that alkali treatment has improved the tensile properties of the fibre and also the adhesion between fibre and matrix. The use of epoxy and vinylester resins offer better bonding to the fibre compared to the use of polyester. Mechanical tests which included tensile, compressive, flexural and shear tests were performed to evaluate the characteristics and behaviour. Furthermore, impact tests were performed to evaluate the impact resistance capacity of the composites. Lastly, a parametric study was conducted to evaluate the feasibility of applications of gomuti fibre composites as a roof tile. 7.2 Characteristics of gomuti fibre and the effects of alkali treatment Compared to other natural fibres such as jute, flax and hemp, gomuti fibre exhibits a larger diameter, similar density, lower single fibre strength, lower single Adriana N E Ticoalu C h a p t e r 7 163

182 fibre modulus and higher elongation. Closer similarity of characteristics was found when gomuti fibre was compared to coir fibre. Although there were substantial variations in data, alkali treatment with 5 % and 10 % NaOH to gomuti fibre resulted in a smaller average diameter and higher average densities and tensile properties compared to untreated fibre. The 5 % treatment offers the best results. The untreated sample began to degrade at lower temperature, as shown by the TGA results, compared to the treated gomuti fibre samples. Results of TGA and DSC indicate that the alkali treatment resulted in higher degradation temperature, which shows improvements in the fibre s thermal stability. Based on its characteristics and behaviour, which is similar to coir fibre, gomuti can, therefore, be considered as a potential natural fibre for use in composite materials. 7.3 Strength and stiffness characteristics of gomuti fibre composites with thermoset polymer resins Based on the investigations, the use of different resins resulted in a significant difference in the composites properties which is related to fibre-matrix bonding. Fibre treatment with alkali also resulted in a modification of the properties of gomuti fibre composites. However, the difference in amount of treatment was found to be minor. It was inferred that treatment of gomuti fibre is necessary to improve fibre-matrix bonding where 5 % amount of NaOH gave the best properties and the use of vinylester resin was found to be the best option of matrix. Adriana N E Ticoalu C h a p t e r 7 164

183 Results of mechanical tests show that unidirectionally aligned gomuti fibre composites with approximately 20% fibre fraction by weight, exhibit tensile strengths and moduli approximately between 28.4 and 48.5 MPa and GPa, respectively; flexural strengths and moduli between 64.1 and 97.0 MPa and GPa, respectively; compressive strengths and moduli between MPa and GPa, respectively; and shear strengths between MPa. Furthermore, results indicated that the compressive strength of the composites are approximately 2 4 times the tensile strengths while shear strengths are approximately times the tensile strengths. The use of different resins has resulted in significant variations in the mechanical properties while fibre treatment provides only minor variations. It was found that gomuti fibre composites may offer more significant benefits when used for products with low to moderate strength requirements. The use of natural fibre as a composite ingredients with polymer thermoset resins may not necessarily increase the tensile strength of the resin but it reduces the amount of resin used. The study into properties and behaviour indicates that gomuti-fibre/thermoset composites may offer more significant benefits when used for products with low to moderate strength requirements. 7.4 Impact strength and behaviour characteristics of gomuti fibre composites with thermoset polymer resins. Impact strength characteristics of gomuti fibre composites have been evaluated using an instrumented drop-weight charpy impact test. It was found that, Adriana N E Ticoalu C h a p t e r 7 165

184 overall, samples of gomuti fibre composites in this study exhibit an impact strength of between 15 kj/m2 and 71.7 kj/m2. The use of epoxy resulted in lower impact strength which can be attributed to better fibre-matrix bonding. The use of polyester has the opposite effect. This confirms the findings from the tensile, flexural, compressive and shear tests. Treatment to the fibre resulted in lower impact strength which again confirm the better fibre-matrix bonding. It was observed that a different amount of fibre treatment resulted in small difference in the impact strength. It was also observed from the tests, that the incorporation of gomuti fibre into the resin matrices has resulted in the composites displaying a tough break failure mode, which is confirmed by the load-deflection plot. Pictures of the specimen after impact tests show that specimens with higher impact strength show excessive fibre pull-out and fibre-matrix debonding while specimens with lower impact strength show more flat cracks. 7.5 Gomuti fibre composite roof tile Under the designated transverse bending strength for roof tile design, results of the parametric study indicated that, with a tile dimension of 330 mm x 420 mm, the major factor governing the design of roof tile is the stiffness. The range of modulus of gomuti fibre composites is from 4.0 to 6.5 GPa, which gives larger design thicknesses compared to design thickness based on the strength criteria. A graph of two curves (strength and stiffness) was constructed, where by a given thickness, the values of modulus and design strength can be obtained. Furthermore, based on the modulus and strength, a design thickness or a range of Adriana N E Ticoalu C h a p t e r 7 166

185 design thickness can be obtained. It was found that for gomuti fibre composites with a range of modulus from 4.3 to 6.3 GPa, the required minimum thickness is approximately from 13.5 to 15.4 mm and design strength of 6.5 to 8.4 MPa. While strength criteria is not a governing factors in the design, when the design strength incorporate a safety factor, strength of the composites will become an important criteria. In special cases when a material exhibits high modulus and low strength, design based on the strength is likely to become major criteria. 7.6 Recommendation for future research Future research regarding general natural fibre composites and gomuti fibre composites are recommended for the following: - Single fibre tensile strength and modulus of gomuti fibre are clearly lower than that of flax, hemp and kenaf. The characteristics of gomuti fibre resembles closely to coir fibre particularly with the elongation and surface morphology. Substantial variations, however, exists within data of characteristics of natural fibres. Therefore, investigations with more than 25 specimens for each sample would increase the statistical confidence in the results. - Investigation on the longitudinal properties of gomuti fibre composites has given beneficial knowledge into the characteristics and behaviour of the composites. Experimental evaluations on the transverse properties of the composites will further confirm the characteristics and behaviour. Adriana N E Ticoalu C h a p t e r 7 167

186 - Controlling the volume fraction of natural fibre composites processed with hand lay-up is difficult. One proposed approach is evaluating the volume fraction by image analysis. - Evaluation of the impact strength of a roof tile with plate impact test is needed to propose a roof tile design specification. - Experimental testing for dynamic weather resistance of roof tiles is recommended for full scale evaluation of gomuti fibre composites roof tile. - Future study on gomuti fibre composites with regard to fibre-matrix debonding will be beneficial, as it will point out important information or findings for further full scale testing of gomuti fibre composites roof tile. - The long term effect of using natural fibre composites is an important topic of research which necessarily covers a longer period of observation plan to ensure the effectiveness of the research. Therefore, a continuous research program is recommended in observation of the long-term effects. The positive long term effects of using natural fibre composites is particularly in environment, economic and academic sectors. With increasing use of naturally degradable materials, it is expected that the production of synthetic products will be lessen, hence supporting a healthier environment. Economically, the production and development of materials from natural resources can promote better income and lifestyle to rural or remote communities. Furthermore, this topic has opened a wide potentiality of future research. In the case of long-term performance of natural fibre composites, while there is still a need of more studies, it is expected that the composites will add more capacity and service life to the items made of mere plant fibres. Adriana N E Ticoalu C h a p t e r 7 168

187 The long term challenges with natural fibre composites may relate to its natural degrade-ability, creep and fatigue behaviour. Adriana N E Ticoalu C h a p t e r 7 169

188 LIST OF REFERENCES Abernethy, RB 1996, The new Weibull handbook, 2nd edn, Gulf Publishing Company, Houston. Akesson, D, Skrifvars, M & Walkenstrom, P 2009, 'Preparation of thermoset composites from natural fibres and Acrylate Modified Soybean Oil Resins', Applied Polymer Science, vol. 114, pp Ali, A, Sanuddin, AB & Ezzeddin, S 2010, 'The effect of aging on Arenga pinnata fiber-reinforced epoxy omposite', Materials and Design. Alms, J, Yonko, P, McDowell, R & Advani, S 2009, 'Design and development of an I-beam from natural composites', Journal of Biobased Materials and Bioenergy, vol. 3, no. 2, pp Amico, S, Sydenstricker, T & da Silva, P 2005, 'Evaluation of the influence of chemical treatment on the tensile strength of sisal fibres by a weibull distribution analysis', Metals Materials and Processes, vol. 17, no. 3-4, pp Andersons, J, Sparnins, E, Nystrom, B & Joffe, R 2011, 'Scale effect of the tensile strength of flax-fabric-reinforced polymer composites', Journal of Reinforced Plastics and Composites, vol. 30, no. 23, pp Aranguren, MI & Reboredo, MM 2007, 'Plant-based reinforcements for thermosets: Matrices, Processing, and Properties', in S Fakirov & D Bhattacharyya (eds), Handbook of engineering biopolymers: homopolymers, blends and composites, Hanser. ASTM International 2003 (Reapproved 2009), ASTM C Standard Specification for Clay Roof Tiles. ASTM International 2003 (Reapproved 2009), ASTM C Standard Specification for Concrete Roof Tile. ASTM International 2005, ASTM D5379M 05, Standard Test Method for Shear Properties of Composite Materials by the V-Notched Beam Method. ASTM International 2007, ASTM D Standard Test Method for Tensile Properties of Single Textile Fibers, ASTM International, United States. ASTM International 2008, ASTM C Standard Test Method for Wind Resistance of Concrete and Clay Roof Tiles (Mechanical Uplift Resistance Method). Adriana N E Ticoalu List of references 170

189 ASTM International 2010, ASTM D Standard Test Method for Compressive Properties of Rigid Plastics, ASTM International, United States. ASTM-D , 'Standard Test Method for Compressive Properties of Rigid Plastics', ASTM-International. ASTM-D , 'Standard test method for tensile strength and Young's modulus for high-modulus single-filamanet materials'. ASTM-D5379/D5379M , 'Standard Test Method for Shear Properties of Composite Materials by the V-Notched Beam Method', ASTM-International. Åström, BT 1997, Manufacturing of polymer composites, Chapman & Hall, London. Australia, S 2006, Fibre reinforced polymer roof, AS , Australia, viewed 6 July Aziz, SH & Ansell, MP 2004, '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', Composites Science and Technology, vol. 64, pp Bachtiar, D, Sapuan, SM & Hamdan, MM 2008, 'The effect of alkaline treatment on tensile properties of sugar palm fibre reinforced epoxy composites', Materials and Design, vol. 29 pp Bachtiar, D, Sapuan, S & Hamdan, M 2009, 'The Influence of Alkaline Surface Fibre Treatment on the Impact Properties of Sugar Palm Fibre-Reinforced Epoxy Composites', Polymer-Plastics Technology and Engineering, vol. 48, no. 4, pp Bachtiar, D, Sapuan, SM, Zainudin, ES, Khalina, A & Dahlan, KZM 2010, 'The tensile properties of single sugar palm (Arenga pinnata) fibre', IOP Conference Series: Materials Science and Engineering, vol. 11, p Baley, C 2002, 'Analysis of the flax fibres tensile behaviour and analysis of the tensile stiffness increase', Composites: Part A, vol. 33, pp Beckermann, G & Pickering, K 2008, 'Engineering and evaluation of hemp fibre reinforced polypropylene composites: Fibre treatment and matrix modification', Composites: Part A, vol. 39, pp Benjamin Bax & Mussig, J 2008, 'Impact and tensile properties of PLA/Cordenka and PLA/flax composites', Composites Science and Technology, vol. 68, pp Adriana N E Ticoalu List of references 171

190 Bisanda, ETN 1991, 'Sisal Fibre Reinforced Composites, Ph.D. Thesis, University of Bath, United Kingdom', Cited in Mwaikambo, L.Y., Review of the history, properties and application of plant fibres, African Journal of Science and Technology (AJST), Science and Engineering Series, Vol. 7, No. 2, pp , Bisanda, ETN 1993, 'The Manufacture of Roofing Panels from Sisal Fiber Reinforced Composites', Journal of Materials Processing Technology, vol. 38, no. 1-2, pp , item: ISI:A1993KM Biswas, S, Kindo, S & Patnaik, A 2011, 'Effect of fiber length on mechanical behaviour of coir fiber reinforced epoxy composites', Fibers and Polymers, vol. 12, no. 1, pp Bliss, S 2006, Best Practices Guide to Residential Construction: Materials, Finishes, and Details John Wiley & Sons, Inc., Hoboken, New Jersey. Brahmakumar, M, Pavithran, C & Pillai, R 2005, 'Coconut fibre reinforced polyethylene composites: effect of natural waxy surface layer of the fibre on fibre/matrix interfacial bonding and strength of composites', Composites Science and Technology 65, vol. 65, pp Canche-Escamilla, G, Rodriguez-Laviada, J, Cauich-Cupul, J, Mendizabal, E, Puig, J & Herrera-Franco, P 2002, 'Flexural, impact and compressive properties of a rigid-thermoplastic matrix/cellulose fiber reinforced composites', Composites part A, vol. 33, pp Chawla, KK 1998, Fibrous materials, Cambridge University Press, United Kingdom. Chou, T-W 1992, Microstructural design of fiber composites, Cambridge University Press, New York, USA. Dandi, B, Sapuan, S & Hamdan, M 'Chemical composition of Arenga Pinnata fiber (Table 1)', in Cited in Suriani MJ, Hamdan MM, Sastra HY, and Sapuan SM. Study of interfacial adhesion of tensile specimens of Arenga Pinnata fiber reinforced composites. Multidiscipline Modeling in Materials and Structures 2007; 3: De Rosa, IM, Kenny, JM, Debora Puglia b & Carlo Santulli c, Fabrizio Sarasini a 2010, 'Morphological, thermal and mechanical characterization of okra (Abelmoschus esculentus) fibres as potential reinforcement in polymer composites'. Defoirdt, N, Biswas, S, Vriese, LD, Tran, LQN, Acker, JV, Ahsan, Q, Gorbatikh, L, Vuure, AV & Verpoest, I 2010, 'Assessment of the tensile properties of coir, bamboo and jute fibre', Composites Part A: Applied Science and Manufacturing, vol. 41, no. 5, pp Adriana N E Ticoalu List of references 172

191 Delft University of Technology, 2003, Composite Applications using Coir Fibres in Sri Lanka, Common Fund For Commodities Fast Track Project. Dweib, M, Hu, B, Shenton III, H & Wool, R 2006, 'Bio-based composite roof structure: Manufacturing and processing issues', Composite Structures, vol. 74, pp Dweib, MA, Hu, B, O'Donnell, A, Shenton, H & Wool, R 2004, 'All natural composite sandwich beams for structural applications', Composite Structures, vol. 63, pp El-Shekeil, Y, Sapuan, S, Khalina, A, Zainudin, E & Al-Shuja a, O 2012, 'Effect of alkali treatment on mechanical and thermal properties of Kenaf fiber-reinforced thermoplastic polyurethane composite', Journal of Thermal Analysis and Calorimetry, vol. 109, no. 3, pp El-Shekeil, Y, Sapuan, S, Khalina, A, Zainudin, E & Al-Shuja a, O 2012, 'Influence of chemical treatment on the tensile properties of kenaf fiber reinforced thermoplastic polyurethane composite', Express Polymer Letters, vol. 6, no. 12, pp Facca, AG, Kortschot, MT & Yan, N 2006, 'Predicting the elastic modulus of natural fibre reinforced thermoplastics'. Faruk, O, Bledzki, AK, Fink, H-P & Sain, M 2012, 'Biocomposites reinforced with natural fibers: ', Progress in Polymer Science, vol. 37, no. 11, pp Guades, E, Aravinthan, T, Manalo, A & Islam, M 2013, 'Damage modelling of repeatedly impacted square fibre-reinforced polymer composite tube', Materials & Design, vol. 47, pp Guades, E, Aravinthan, T, Manalo, A & Islam, M 2013, 'Experimental investigation on the behaviour of square FRP composite tubes under repeated axial impact', Composite Structures, vol. 97, pp Harish, S, Michael, DP, Bensely, A, Lal, DM & Rajadurai, A 2009, 'Mechanical property evaluation of natural fiber coir composite', Materials characterization, vol. 60, pp Hepworth, D, Bruce, D, Vincent, J & Jeronimidis, G 2000, 'The manufacture and mechanical testing of thermosetting natural fibre composites', Journal of material science, vol. 35 pp Hu, B, Dweib, M, Wool, RP & Shenton III, HW 2007, 'Bio-Based Composite Roof for Residential Construction', Journal of Architectural Engineering, vol. 13, pp Adriana N E Ticoalu List of references 173

192 Ishak, MR & Z. Leman, SMS, M.Y. Salleh and S. Misri 2009, 'The effect of sea water treatment on the impact and flexural strength of sugar palm fibre reinforced epoxy composites', International Journal of Mechanical and Materials Engineering (IJMME), vol. 4, no. 3, pp Ishak, MR, Sapuan, SM, Rahman, MZA, Leman, Z & Anwar, UMK 2011, 'Characterization of sugar palm (Arenga pinnata) fibres - Tensile and thermal properties', J Therm Anal Calorim. ISO , 'Fibre-reinforced plastic composites: Determination of flexural properties', International Standard. Joffe, R, Andersons, J & Wallstrom, L 2003, 'Strength and adhesion characteristics of elementary flax fibres with different surface treatments', Composites part A, vol. 34, pp Kabir, MM 2012, 'Effects of Chemical Treatments on Hemp Fibre Reinforced Polyester Composites', PhD Dissertation, University of Southern Queensland, Toowoomba, Queensland, Australia. Leman, Z, Sapuan, SM, Azwan, M, Ahmad, MMHM & Maleque, MA 2008, 'The Effect of Environmental Treatments on Fiber Surface Properties and Tensile Strength of Sugar Palm Fiber-Reinforced Epoxy Composites', Polymer-Plastics Technology and Engineering, vol. 47, pp Manfredi, LB, Rodriguez, ES, Wladyka-Przybylak, M & Vazquez, A 2006, 'Thermal degradation and fire resistance of unsaturated polyester, modified acrylic resins and their composites with natural fibre', Polymer Degradation and Stability, vol. 91, pp Manthey, NW, Cardona, F, Francucci, G & Aravinthan, T 2013, 'Thermomechanical properties of epoxidized hemp oil-based bioresins and biocomposites', Journal of Reinforced Plastics and Composites. DOI: / Misri, S, Leman, Z, Sapuan, SM & Ishak, MR 2009, 'Mechanical properties and fabrication of small boat using woven glass/sugar palm fibres reinforced unsaturated polyester hybrid composite', IOP Conference Series: Materials Science and Engineering, vol. 11, no. 1, pp Mogea, J, Seibert, B & Smits, W 1991, 'Multipurpose palms: the sugar palm (Arenga pinnata (Wurmb) Merr.)', Agroforestry Systems, vol. 13, pp Mussig, J 2008, 'Cotton Fibre-reinforced Thermosets Versus Ramie Composites: A Comparative Study using Petrochemical- and Agro-based Resins', J Polym Environ (2008) 16: Adriana N E Ticoalu List of references 174

193 Mwaikambo, LY & Ansell, MP 2002, 'Chemical modification of hemp, sisal, jute, and kapok fibers by alkalization', Journal of Applied Polymer Science, vol. 84, no. 12, pp Mwaikambo, L & Ansell, M 2003, 'Hemp fibre reinforced cashew nut shell liquid composites'. Nechwatal, A, Mieck, K-P & Reußmann, T 2003, 'Developments in the characterization of natural fibre properties and in the use of natural fibres for composites', Composites Science and Technology, vol. 63, no. 9, pp O Donnell, A, Dweib, M & Wool, R 2004, 'Natural fiber composites with plant oil-based resin', Composites Science and Technology, vol. 64, pp Ochi, S 2008, 'Mechanical properties of kenaf fibers and kenaf/pla composites', Mechanics of Materials, vol. 40, pp OKSMAN, K 2000, 'Mechanical Properties of Natural Fibre Mat Reinforced Thermoplastic', Applied Composite Materials 7: , Oksman, K, Wallstrom, L, Berglund, LA & Filho, RDT 2002, 'Morphology and mechanical properties of unidirectional sisal- epoxy composites', Journal of Applied Polymer Science, vol. 84, no. 13, pp Padmavathi, T, Naidu, SV & Rao, R 2012, 'Studies on Mechanical Behavior of Surface Modified Sisal Fibre - Epoxy Composites', Journal of Reinforced Plastics and Composites, vol. 31, no. 8, pp Pickering, KL, Institute of Materials, M & Mining 2008, Properties and Performance of Natural Fibre Composites, Woodhead Pub. and Maney Pub. Pickering, K, Beckermann, G, Alam, S & Foreman, N 2007, 'Optimising industrial hemp fibre for composites', Composites: Part A, vol. 38, pp Prasad, S, Pavithran, C & Rohatgi, P 1983, 'Alkali treatment of coir fibres for coir-polyester composites', Journal of Materials Science, vol. 18, no. 5, pp Rao, KMM, Rao, KM & Prasad, AVR 2010, 'Fabrication and testing of natural fibre composites: Vakka, sisal, bamboo and banana', Materials and Design, vol. 31, pp Rassmann, S, Paskaramoorthy, R & Reid, RG 2011, 'Effect of resin system on the mechanical properties and water absorption of kenaf fibre reinforced laminates', Materials & Design, vol. 32, no. 3, pp Ray, D, Sarkar, B & Bose, N 2002, 'Impact fatique behaviour of vinylester resin matrix composites reinforced with alkali treated jute fibres'. Adriana N E Ticoalu List of references 175

194 Ray, D, Sarkar, B, Basak, R & Rana, A 2002, 'Study of the Thermal Behavior of Alkali-Treated Jute Fibers', Journal of Applied Polymer Science, vol. 85, pp Razak, HA & Ferdiansyah, T 2005, 'Toughness Characteristics of Arenga pinnata Fibre Concrete', Journal of Natural Fibers, Vol. 2(2) 2005, vol. 2, no. 2, pp Romhany, G, Karger-Kocsis, J & Czigany, T 2003, 'Tensile Fracture and Failure Behavior of Technical Flax Fibers'. Rouison, D, Sain, M & Couturier, M 2004, 'Resin transfer molding of natural fiber reinforced composites: cure simulation'. Rouison, D, Sain, M & Couturier, M 2006, 'Resin transfer molding of hemp fiber composites: optimization of the process and mechanical properties of the materials', Composites Science and Technology, vol. 66, no. 7-8, pp Rout, J, Misra, M, Tripathy, SN, SK & Mohanty, A 2001, 'The influence of fibre treatment on the performance of coir-polyester composites', Composites Science and Technology 61 (2001) Sahari, J, Sapuan, SM, Ismarrubie, Z & Rahman, M 2011, 'Comparative study of physical properties based on different parts of sugar palm fibre reinforced unsaturated polyester composites', Key Engineering Materials, vol , pp Sastra, HY, Siregar, JP, Sapuan, SM, Leman, Z & Hamdan, MM 2005, 'Flexural Properties of Arenga pinnata Fibre Reinforced Epoxy Composites', American Journal of Applied Sciences, (Special Issue): 21-24, Sastra, HY, Siregar, JP, Sapuan, SM & Hamdan, MM 2006, 'Tensile Properties of Arenga pinnata Fiber-Reinforced Epoxy Composites', Polymer-Plastics Technology and Engineering, vol. 45, pp Satyanarayana, KG & Wypych, F 2007, 'Characterization of natural fibers', in S Fakirov & D Bhattacharyya (eds), Handbook of engineering biopolymers: homopolymers, blends and composites, Carl Hanser Verlag, Munich, pp Satyanarayana, KG, Guimarães, JL & Wypych, F 2007, 'Studies on lignocellulosic fibers of Brazil. Part I: Source, production, morphology, properties and applications', Composites Part A: Applied Science and Manufacturing, vol. 38, no. 7, pp Satyanarayana, KG, Sukumaran, K, Kulkarni, AG, Pillai, SGK & Rohatgi, PK 1986, 'Fabrication and properties of natural fibre-reinforced polyester composites', Composites, vol. 17, no. 4, pp Adriana N E Ticoalu List of references 176

195 Sawpan, MA 2009, 'Mechanical performance of industrial hemp fibre reinforced polylactide and unsaturated polyester composites', University of Waikato, New Zealand. Silva, GG, De Souza, D, Machado, J & Hourston, D 2000, 'Mechanical and Thermal Characterization of Native Brazilian Coir Fiber', Journal of Applied Polymer Science, vol. 76, pp Singh, S & Mohanty, A 2007, 'Wood fiber reinforced bacterial bioplastic composites: Fabrication and performance evaluation', Composites Science and Technology, vol. 67, pp Sinha, E & Rout, S 2009, 'Influence of fibre-surface treatment on structural, thermal and mechanical properties of jute fibre and its composite'. Spārniņš, E 2009, 'Mechanical Properties of Flax Fibers and Their Composites', Luleå University of Technology Standards Australia 1994, AS Plastic roof and wall cladding materials Methods of test, Part 0: General introduction and list of methods. Standards Australia 2001, AS Determination of tensile properties of plastics materials Part 4: Test conditions for isotropic and orthotropic fibrereinforced plastic composites, Standards Australia International Ltd, Australia. Standards Australia 2002, AS : Roof tiles. Standards Australia 2002, AS : installation of roof tile, Standards Australia, Australia. Standards Australia 2002, AS Methods of testing roof tiles, Method 3: Determination of transverse strength. Suriani, MJ, Hamdan, MM, Sastra, HY & Sapuan, SM 2007, 'Study of interfacial adhesion of tensile specimens of Arenga Pinnata fiber reinforced composites', Multidiscipline Modeling in Material and Structures, vol. 3, no. 2, pp Symington, MC, Banks, W, West, OD & Pethrick, R 2009, 'Tensile Testing of Cellulose Based Natural Fibers for Structural Composite Applications', Journal of Composite Materials, vol. 43, no. 9. Taha, I, Steuernagel, L & Ziegmann, G 2007, 'Optimization of the alkali treatment process of date palm fibres for polymeric composites', Composite Interfaces, vol. 14, no. 7-9, pp Adriana N E Ticoalu List of references 177

196 Thiruchitrambalam, M & Shanmugam, D 2012, 'Influence of pre-treatments on the mechanical properties of palmyra palm leaf stalk fiber-polyester composites', Journal of Reinforced Plastics and Composites. Thomason, J & Vlug, M 1996, 'Influence of fibre length and concentration on the properties of glass fibre-reinforced polypropylene: 4. Impact properties', Composites part A, vol. 28A, pp Ticoalu, A, Aravinthan, T & Cardona, F 2010, 'A review of current development in natural fiber composites for structural and infrastructure applications', in Southern Region Engineering Conference (SREC): proceedings of thesouthern Region Engineering Conference (SREC) November 2010, Toowoomba, Australia. Ticoalu, A, Aravinthan, T & Cardona, F 2011, 'Experimental investigation into gomuti fibres/polyester composites', in 21st Australasian Conference on the Mechanics of Structures and Materials, proceedings of the21st Australasian Conference on the Mechanics of Structures and Materials, Fragomeni, S, Venkatesan, S, Lam, NTK, Setunge, S (eds.), CRC Press/Balkema, Melbourne, Australia, pp Ticoalu, A, Aravinthan, T & Cardona, F 2012, 'A review on the characteristics of gomuti fibre and its composites with thermoset resins', Journal of Reinforced Plastics and Composites, vol. 32, no. 2, pp Ticoalu, A, Aravinthan, T & Cardona, F 2013, 'Properties and behaviour of gomuti fibre composites under tensile and compressive load'. Ticoalu, A, Aravinthan, T & Cardona, F 2013, 'A review on the characteristics of gomuti fibre and its composites with thermoset resins', Journal of Reinforced Plastics and Composites, vol. 32, no. 2, pp Tomczak, F, Sydenstricker, THD & Satyanarayana, KG 2007, 'Studies on lignocellulosic fibers of Brazil. Part II: Morphology and properties of Brazilian coconut fibers', Composites Part A: Applied Science and Manufacturing, vol. 38, no. 7, pp Trindade, WG, Hoareau, W, Megiatto, JD, Razera, IAT, Castellan, A & Frollini, E 2005, 'Thermoset Phenolic Matrices Reinforced with Unmodified and Surface- Grafted Furfuryl Alcohol Sugar Cane Bagasse and Curaua Fibers: Properties of Fibers and Composites', Biomacromolecules, vol. 6, no. 5, pp Valadez-Gonzalez, A, Cervantes-Uc, J, Olayo, R & Herrera-Franco, P 1999, 'Effect of fiber surface treatment on the fiber matrix bond strength of natural fiber reinforced composites', Composites: Part B, vol. 30, pp Adriana N E Ticoalu List of references 178

197 Van De Velde, K & Kiekens, P 2002, 'Thermal Degradation of Flax: The Determination of Kinetic Parameters with Thermogravimetric Analysis', Journal of Applied Polymer Science, vol. 83, pp Van Rijswijk, K, Brouwer, WD & Beukers, A 2001, 'Application of Natural Fibre Composites in the Development of Rural Societies'. Wambua, P, Ivens, J & Verpoest, I 2003, 'Natural fibres: can they replace glass in fibre reinforced plastics?', Composites Science and Technology, vol. 63, pp Wonderly, C, Grenestedt, J, Fernlund, G & Cěpus, E 2005, 'Comparison of mechanical properties of glass fiber/vinyl ester and carbon fiber/vinyl ester composites', Composites: Part B, vol. 36 pp Yan, L 2012, 'Effect of alkali treatment on vibration characteristics and mechanical properties of natural fabric reinforced composites', Journal of Reinforced Plastics and Composites, vol. 31, no. 13, pp Yan, L, Chouw, N & Yuan, X 2012, 'Improving the mechanical properties of natural fibre fabric reinforced epoxy composites by alkali treatment', Journal of Reinforced Plastics and Composites, vol. 31, no. 6, pp Yu, HN, Kim, SS, Hwang, IU & Lee, DG 2008, 'Application of natural fiber reinforced composites to trenchless rehabilitation of underground pipes', Composite Structures, vol. 86, pp Adriana N E Ticoalu List of references 179

198 APPENDIX A OVERVIEW OF NATURAL FIBRES AND THEIR COMPOSITES A.1 Type of natural fibres Natural fibre composites can be either a combination of natural fibre with synthetic resin, synthetic fibre with bio-resin or natural fibre with bio-resin. By definition, natural fibre is fibre that is not manmade, hence obtained from plants, animals or humans. Synthetic fibre is a classification for manmade fibre such as glass fibre, carbon fibre and aramid. Bio-resin means bio-degradable resin such as plant-based resin, while synthetic resin such as polyester, vinyl ester and epoxy are obtained from petrochemical products. Both synthetic and bio-resin can be either in the form of thermoset or thermoplastic resin. Many types of natural fibres are available around the world and are in abundance. Coconut (Cocos nucifera) fibre (coir) and jute (Corhorus capsularis) fibre are available in India and some countries in South-east Asia. Flax fibre is obtained from flax plants (Linum usitatissimum) that grow in temperate climates such as in European countries and China (Pickering 2008). Hemp (Cannabis sativa) for industrial purposes is cultivated in several European countries, China, Australia and New Zealand. Sisal (Agave sisalana), bamboo, banana, oil palm, and palm trees are some of the plants grown in tropical or sub-tropical climates from which fibres can obtained (Pickering 2008). Non-plant based fibres that are available and have been used in some countries are wool and chicken feathers. Figure A.1 illustrates the general grouping of fibres. Adriana N E Ticoalu A 1

199 Bast Plant fibre Core Fruit (Coir, cotton, kapok, palm) Other Naturally occuring fibre Animal fibre Wool Feather Cashmere Fibre Mineral fibre Asbestos Wollastonite Bio-based fibre Nylon Acetate Manmade fibre Glass Carbon Synthetic fibre Aramid Polypropylene Polyester Figure A.2. Schematic illustration of fibre grouping Adriana N E Ticoalu A 2

200 Traditionally, especially in rural developing countries, natural fibres have been cultivated and used extensively for textile, multipurpose durable rope, bags, brooms, fish nets and filters. Hemp fibre has been used for textiles and canvas (Pickering 2008). Several types of fibres such as coir and straw have also been used for applications in housing as roof material, wall insulation and in mudbricks or adobe. The most common fibre composite elements are made of synthetic fibre combined by synthetic resin such as glass fibre reinforced polymer (GFRP) and carbon fibre reinforced polymer (CFRP). These types of fibre composites have a wide range of applications including automotive, aircraft, transportation infrastructure, furniture, food container and housing. Notable characteristics of all synthetic fibre composites are a high-strength to weight ratio, durable, flexible shape and easily customisable according to the need. However, the all-synthetic fibre composites pose a long term cost and environmental concern related to the use of non-renewable petrochemical resources and the less bio-degradable nature of the materials. A.2 Existing usage of natural fibres Natural fibre composites use one or more natural ingredients in their production. They are a re-emerging material that had been used centuries ago for many applications but lost their attractiveness due to the introduction and mass production of GFRP and other synthetic composites. The first recorded usage of natural fibre composites was straw reinforced clay used as bricks in Ancient Egypt (Exodus 5:7 NKJV; Kozlowski & Machiewacz-Talarczyk 2006 cited in Rowell Adriana N E Ticoalu A 3

201 2008). With similar knowledge and technology, some countries have been producing mudbricks or adobe for housing as shown in Figure A.2. The re-introduction of natural fibre composites has channelled more research into the use of natural ingredients in fibre composites. Environmental awareness and long-term cost also contribute to increased research and development in NFC for various applications especially, those that require a low to moderate strength capacity. The use of natural fibre with synthetic resin, synthetic fibre with bio-resin and natural fibre with bio-resin have been reported in literature. (a) (b) Figure A.2 Bricks made of straw reinforced clay composites (adobe or mudbricks): (a) Brick making in Romania (Soare 2003); (b) Adobe brick house in Milyanfan village, Kyrgyzstan (Vmenkov 2007) Natural fibre composites can be the combinations of natural fibre synthetic polymer resin, synthetic fibre bio-resin or natural fibre bio-resin. Natural fibres such as sisal, hemp, kenaf and jute has been combined with polypropylene resin to develop natural fibre thermoplastic composites (Wambua et al. 2003). Brahmakumar et al. (2005) have studied the composite of coir and Low Density Polyethylene (LDPE). Van de Velde and Kiekens (2001) studied the thermoplastic Adriana N E Ticoalu A 4

202 pultrusion of flax reinforced polyethylene terephthalate and flax reinforced polypropylene. The composites of natural fibre and thermoplastic synthetic resin have found application in the automotive industry (Brouwer 2000), especially for interior parts including car seats and car rooves (Rijswijk et al. 2001). Unlike thermoplastic resin, thermoset resin hardens and does not liquefy after curing. This feature favours use in building and infrastructure industries (Composites for infrastructure 1998). Natural fibre-thermoset resin composites that have been studied by several researchers are hemp-polyester (Rouison et al. 2006), jute-epoxy (Gassan & Gutowski 2000) and flax-polyester (Rodríguez et al. 2005). Satyanarayana et al. (1986) reported the making of a coir-polyester roof. When significant strength is required, the combination of synthetic fibre and bioresin may be a feasible option. LOC Composites Pty Ltd developed and produced a fibre composite sandwich panel concept which uses plant based polymers for its resin and core material (Van Erp & Rogers 2008). The combination of natural fibre and bio resin is probably the perfect combination for an environmentally friendly material. However, research and development work to enhance the quality and strength of this type of composites is still in progress. Examples of developed biodegradable resins are cellulose acetate, copolyester, polycaprolactone (PCL), poly(ester amide), poly(ethylene terepthalate), (PET)-modified, polyglycolide (PGA), polyhydroxyalkanoates (PHA), poly(lactic acid) (PLA), poly(vinyl alcohol) (PVOH), starch and starch blends, and other blends (Mohanty et al., 2000). Examples of research and development on all-bio-composites are chopped hemp/cellulose acetate composites (Mohanty et al., 2004), acrylated epoxidized soybean oil (AESO) reinforced by Adriana N E Ticoalu A 5

203 flax, cellulose, pulp and hemp (O Donnell et al., 2004) and cashew nut shell liquid (CNSL) composites reinforced by untreated unidirectional hemp fibre (Mwaikambo & Ansell, 2003). Other developments of bio-composites that have been reported are unidirectional-kenaf-pla (Ochi, 2008) and hemp and flax-aeso (Akesson, 2009). The aim of this review is to provide an overview of the research into natural fibre composites, the properties of natural plant fibres, the properties of natural fibre composites and the existing and potential applications of the composites. A.3 Properties of natural fibre Being natural products, natural fibres from plants are generally biodegradable and are in abundant supply all around the world. Compared to synthetic fibres, natural fibres have lower density which leads to lighter products. Moreover, the fibres are cheaper, therefore offering significant cost advantages (O Donnell 2004). Further advantages of using natural fibres in composites are low specific weight, being a renewable resource, production requiring little energy, low cost, environmentally friendly processing, possibility of thermal recycling and acoustic insulating properties (Brouwer 2000). In terms of strength, natural fibres are still outweighed by synthetic fibres. Pictures of several natural fibres are shown in Figure A.3. Adriana N E Ticoalu A 6

204 Figure A.3. Pictures of various natural fibres A large variation is found in the properties of natural fibres. One type of natural fibre, when grown in different climates, may give different properties. In one particular plant, fibres that are obtained from different parts such as bast and core, may give different results. Cultivation conditions, growing period and the retting process have also been found to affect the properties of plant-based fibre (Ochi 2008; Pickering et al. 2007). Generally, flax fibre has the high tensile strength while coconut fibre (coir) and sugar-cane bagasse have the low tensile strength. Among many natural fibres, coir fibre exhibit the distinctive characteristics of high elongation. Adriana N E Ticoalu A 7

205 Natural fibres can be categorised in six types according to the part of the plants from which they are sourced (Rowell 2008 in Pickering 2008). The six types are bast fibres (e.g. jute, flax, hemp, ramie and kenaf), leaf fibres (e.g. banana, sisal, agave and pineapple), seed fibres (e.g. coir, cotton and kapok), core fibres (e.g. kenaf, hemp and jute), grass and reed (e.g. wheat, corn and rice) and all other types including wood and roots. Bast fibres are fibres located under a thin bark of the plant and they grow parallel to the length of the stem (Rowell 2008 in Pickering 2008). For kenaf, hemp and jute, the core part of the plants, inside the bast, is also a source of fibres. Natural fibre mainly consists of cellulose of approximately 30 % 90 % depending on the type of the fibre (Rowell 2008 in Pickering 2008). Lignin is the second major component while the remaining percentage may consists of moisture, pectin, ash and waxes. Before natural fibres can be processed for composites manufacturing, they must be pre-processed to remove them from the plants, and to remove dirt and unwanted substances. Several common processing methods are decortification, cotton gin to separate cotton from its seeds, the retting process (Rowell 2008 in Pickering 2008) and de-husking for coir. A few types of natural fibres can be found in the market in mat, rope, cloth or woven forms. However, many types of natural fibres are only available in the original form or chopped. To remove additional dirt and the waxy layer and to enhance the adhesion between natural fibres and the resin matrix, fibres are treated. The intention is to improve the interfacial adhesion of the composites which is expected to further improve the mechanical and thermal properties. Alkali treatment (Pickering et al. Adriana N E Ticoalu A 8

206 2007; Beckermann 2008, Kabir et al. 2010), heat or water treatment, acetylation treatment (Kabir et al. 2010), and seawater treatment (Leman et al. 2008) are some of the reported treatments. Improved mechanical properties of composites and stronger fibre can be achieved with the correct and optimum treatment. A.4 Strength and stiffness of natural fibre composites The properties and behaviour of natural fibre composites are affected by several factors. The type of fibre, form of fibre, type of resin, manufacturing process and volume fraction all influence the properties and behaviour of natural fibre composites, as they do for synthetic fibre composites. Additionally, in natural fibre composites, due to the character of natural fibre, the properties and behaviour of the composites can be affected by the type of plant from which the fibres are sourced, the cultivation process, interfacial behaviour between fibre and resin, and the treatment of the fibre before composite manufacturing. The strength property is probably most affected by type of fibre, fibre fraction and manufacturing process among all other factors. Tensile strength of natural fibre composites is achievable within the range of MPa. Coir composites show the lowest tensile strength whereas flax composites exhibit highest tensile strength which is closer to the strength of glass fibre composites. Similar patterns also can be observed for flexural strength. The flexural strength is within the range of MPa. Table A.1 presents the tensile and flexural properties of natural fibre composites. Adriana N E Ticoalu A 9

207 Table A.1. Natural fibre composites Fibre Resin Fibre fraction (wt. %) Process Tensile strength (MPa) Tensile modulus (GPa) Flexural strength (MPa) Flexural modulus (GPa) Flax Epoxy 40 (Vol.) N/A Hemp CNSL - Hand lay up Kenaf Polyester 20 Compression moulding Jute Polyester 50 Compression moulding Cotton Polyester - Hand lay up Coir Polyester 9 Hand lay up Banana Polyester 14 Hand lay up Sisal Polyester Glass fibre Polyester Hand lay-up (Gowda et al., 1999; Ishak et al., 2009; Mwaikambo & Ansell, 2003; Satyanarayana et al., 1986; Singh et al., 1996; Van de Weyenberg et al., 2003) A.5 Impact strength of natural fibre composites In structural design, the most commonly used primary design criteria are bending strength, bending stiffness, tensile strength and tensile stiffness (Marsh, 2000). However, for certain applications such as roofing, impact strength is also important. In an Australian standard, for glass fibre reinforced polyester to be used as roof materials, impact resistance must be shown to be 1.96 J at 23 0 C and has the tensile strength not less than 50 MPa with no visible cracking or crazing or hole after testing. Coir and cotton composites are known to have better impact resistance than other natural fibre composites. Study on impact properties will give an idea of how the material behave when subjected to impact loading such as a sudden heavy load. Adriana N E Ticoalu A 10

208 A.6 Potential applications of natural fibre composites Natural fibre composites generally have the feature of flexibility in design which allow the designer to tailor the properties of the composites to achieve the desired end product. To date, natural fibre composites have been well accepted in the automotive, packaging and furniture industries. Further applications have been in biomedical applications for bone and tissue repair and reconstruction (Cheung et al. 2009). For application in the construction and building industry, the development of natural fibre composites is still in its infancy. This is due to strength requirement issues. In structural applications, natural fibre composites have been used to develop load-bearing elements such as sandwich beam (Dweib et al. 2003), I-shaped beam (Alms et al. 2009), rooves (Bisanda 1993; Dweib et al. 2006; Satyanarayana et al. 1986) and multipurpose panels (O Donnell et al. 2004; Van Erp & Rogers 2008). Meanwhile, in infrastructure applications, natural fibre composites have been used for water pipe rehabilitation (Yu et al. 2008) and water tanks (Rijswijk 2001). Other use of NFPC includes fibre boards, panel and flush doors for low-cost housing needs, panel and roofing sheets, car rooves, catamaran hulls, tabletops, shelves, fluid containers, pedestrian bridges (Rijswijk et al. 2003), flooring, hand rails, window frame, door panel and furniture. Table A.2 lists some of the existing studies on applications of natural fibre composites for building and construction. Adriana N E Ticoalu A 11

209 Table A.2. Existing research into the applications of natural fibre composites Application Fibre Resin Sandwich beam 1 Cellulose (recycled paper) AESO I beam 2 Jute (burlap) ESO Cellular beam 3 Hemp UPE Cellular plate 3 Flax UPE Roof 4 Coir Polyester Roof 5 Sisal (woven mat) CNSL Roof 6 Cellulose (recycled paper) AESO Multipurpose panel 7 N/A Plant based polymers Multipurpose panel 8 Cellulose (recycled paper) AESO Underground pipes (rehabilitation) 9 Jute (mat) UPE Source: 1 Dweib et al. 2004; 2 Alms et al. 2009; 3 Burgueno et al. 2004; 4 Satyanarayana et al 1986; 5 Bisanda 1993; 6 Dweib et al. 2006; 7 Van Erp and Rogers 2008; 8 O Donnell 2004; 9 Yu et al. 2008). Natural fibre composites offer a broad range of opportunity for research and development. Major drawbacks that limits their usage and applications are the large variations in their properties. Therefore, research into natural fibre composites are highly beneficial. Adriana N E Ticoalu A 12

210 APPENDIX B FIBRE CHARACTERISATION DETAILS B.1. Fibre diameter Table B.1 Diameter of untreated and treated gomuti fibre. No. Diameter (µm) UN 5N 10N MEAN S.D Adriana N E Ticoalu B 1

211 B.2. Fibre density Table B.2 Density of untreated and treated gomuti fibre. No. Density (g/cm 3 ) UN 5N 10N Mean S.D B.3. Fibre tensile strength Table B.3 Strength of untreated and treated gomuti fibre. No. Fibre strength (MPa) UN 5N 10N MEAN S.D Adriana N E Ticoalu B 2

212 B.4. Fibre Young s modulus Table B.4 Young s modulus of untreated and treated gomuti fibre. No. Young's modulus (GPa) UN 5N 10N MEAN S.D Adriana N E Ticoalu B 3

213 B.5. Fibre strain-at-failure Table B.5 Strain-at-failure of untreated and treated gomuti fibre. No. Strain-at-failure (%) UN 5N 10N MEAN S.D Adriana N E Ticoalu B 4

214 B.6. One-way ANOVA with MATLAB Statistics Toolbox Table B.6 ANOVA table for fibre diameter. Source SS Df MS F Prob>F Columns Error Total Table B.7 ANOVA table for fibre strength. Source SS Df MS F Prob>F Columns e-005 Error Total Table B.8 ANOVA table for fibre Young s modulus. Source SS Df MS F Prob>F Columns Error Total Adriana N E Ticoalu B 5

215 B.7. Pictures of untreated and treated gomuti fibre under microscope Adriana N E Ticoalu B 6

216 B.8. Representative figures of obtaining E value from stress-strain curves (as for Figure 3.6) Adriana N E Ticoalu B 7

217 Adriana N E Ticoalu B 8

218 B.9. Tensile capacity and sectional stiffness details of gomuti fibre Table B.9 Tensile capacity and sectional stiffness of untreated (UN) gomuti fibre. Sample d area Peak load Stress MoE EA (mm) (mm^2) (N) (MPa) (GPa) UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN UN Mean SD Adriana N E Ticoalu B 9

219 Table B.10 Tensile capacity and sectional stiffness of treated (5N) gomuti fibre. Sample d area Peak load Stress MoE EA (mm) (mm^2) (N) (MPa) (GPa) 5N N N N N N N N N N N N N N N N N N N N N N N N N Mean SD Adriana N E Ticoalu B 10

220 Table B.11 Tensile capacity and sectional stiffness of treated (10N) gomuti fibre. Sample d area Peak load Stress MoE EA (mm) (mm^2) (N) (MPa) (GPa) 10N N N N N N N N N N N N N N N N N N N N N N N N N Mean SD Adriana N E Ticoalu B 11

221 APPENDIX C MECHANICAL PROPERTIES TESTING DETAILS C.1. Tensile test data No. Ave.thick. Ave. width Table C.1.1 Tensile test results of UN.PE Length Peak load Peak stress MoE v (mm) (mm) (mm) (N) (MPa) (GPa) (mm/mm ) Ave SD Ave.thick. Table C.1.2 Tensile test results of 5N.PE Ave. width Length Peak load Peak stress MoE v No. mm/m (mm) (mm) (mm) N MPa GPa m Ave SD Adriana N E Ticoalu C 1

222 No. Ave.thick. Table C.1.3 Tensile test results of 10N.PE Ave. width Length Peak load Peak stress MoE v (mm) (mm) (mm) N MPa GPa mm/m m Ave SD Ave.thick. Ave. width Table C.1.4 Tensile test results of UN.E Length Peak load Peak stress MoE v No. mm/m (mm) (mm) (mm) N MPa GPa m Ave SD Ave.thick. Ave. width Table C.1.5 Tensile test results of 5N.E Length Peak load Peak stress MoE v No. mm/m (mm) (mm) (mm) N MPa GPa m Ave SD Adriana N E Ticoalu C 2

223 No. Ave.thick. Table C.1.6 Tensile test results of 10N.E Ave. width Length Peak load Peak stress MoE v (mm) (mm) (mm) N MPa GPa mm/m m Ave SD No. Ave.thick. Table C.1.7 Tensile test results of UN.VE Ave. width Length Peak load Peak stress MoE v (mm) (mm) (mm) N MPa GPa mm/m m Ave SD Table C.1.8 Tensile test results of 5N.VE No. Ave.thick. Ave. width Length Peak load Peak stress MoE v (mm) (mm) (mm) (N) (MPa) (GPa) (mm/mm) Ave SD Adriana N E Ticoalu C 3

224 Table C.1.9 Tensile test results of 10N.VE No. Ave.thick. Ave. width Length Peak load Peak stress MoE v (mm) (mm) (mm) (N) (MPa) (GPa) (mm/mm) Ave SD Table C.1.10 Summary of tensile strength of gomuti fibre composites No. UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Table C.1.11 Summary of tensile modulus of gomuti fibre composites No. UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E F.10N.VE Adriana N E Ticoalu C 4

225 C.2. Flexural test data Table C.2.1 Flexural test results of UN.PE No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Table C.2.2 Flexural test results of UN.E No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Table C.2.3 Flexural test results of UN.VE No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Adriana N E Ticoalu C 5

226 Table C.2.4 Flexural test results of 5N.PE No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Table C.2.5 Flexural test results of 5N.E No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Table C.2.6 Flexural test results of 5N.VE No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Adriana N E Ticoalu C 6

227 Table C.2.7 Flexural test results of 10N.PE No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Table C.2.8 Flexural test results of 10N.E No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Table C.2.9 Flexural test results of 10N.VE No. h b F L Flexural strength (mm) (mm) (N) (mm) (MPa) Ave SD Adriana N E Ticoalu C 7

228 Table C.2.10 Summary of flexural strength of gomuti fibre composites No. UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Table C.2.11 Summary of flexural modulus of gomuti fibre composites No. UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE C.3. Compressive test data Compressive test results of UN.PE No. thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Adriana N E Ticoalu C 8

229 No. Compressive test results of 5N.PE thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Compressive test results of 10N.PE No. thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Compressive test results of UN.E No. thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Adriana N E Ticoalu C 9

230 No. Compressive test results of 5N.E thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Compressive test results of 10N.E No. thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Compressive test results of UN.VE No. thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Adriana N E Ticoalu C 10

231 No. Compressive test results of 5N.VE thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD No. Compressive test results of 10N.VE thickness height width Peak load Peak stress Compr. Mod. mm mm mm N MPa GPa Ave SD Summary of compressive strength of gomuti fibre composites No. UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Summary of compressive modulus of gomuti fibre composites No. UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Adriana N E Ticoalu C 11

232 C.4. Shear test data Shear test results of UN.PE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Shear test results of UN.E No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Shear test results of UN.VE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Adriana N E Ticoalu C 12

233 Shear test results of UN.VE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Shear test results of 5N.PE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Shear test results of 5N.E No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Adriana N E Ticoalu C 13

234 Shear test results of 5N.VE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Shear test results of 10N.PE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Shear test results of 10N.PE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Adriana N E Ticoalu C 14

235 Shear test results of 10N.VE No Notch Height mm Notch Width mm Deflection at Peak mm Peak Load N Peak Shear Stress MPa Mean S.D Summary of shear strength data No. UN.PE UN.E UN.VE 5N.PE 5N.E 5N.VE 10N.PE 10N.E 10N.VE Adriana N E Ticoalu C 15

236 C.5. Two-way ANOVA with MATLAB Two-way ANOVA is a statistical inference technique to analyse whether the data has any correlation. For 2 factors of analysis, a two-way ANOVA is used. ANOVA table for tensile strength. Source SS df MS F Prob>F Columns Rows Interaction Error Total ANOVA table for compressive strength. Source SS df MS F Prob>F Columns Rows Interaction Error Total ANOVA table for flexural strength. Source SS df MS F Prob>F Columns Rows Interaction Error Total ANOVA table for shear strength. Source SS df MS F Prob>F Columns Rows Interaction Error Total Adriana N E Ticoalu C 16

237 C.6. Representative figure of elastic-linear region for Figure 4.10 (Tensile modulus was measured by an extensometer up to 0.3% strain). C.7. Representative figures of elastic-linear region for Figure 4.14 (Compressive modulus). Adriana N E Ticoalu C 17

238 C.8. Representative figures of elastic-linear region for Figure 4.18 (Flexural modulus). Adriana N E Ticoalu C 18

239 C.9. Comparison of composite panel with different fibre content: (a) 10 %, (b) 20%, (c) >20% Adriana N E Ticoalu C 19

240 APPENDIX D IMPACT TEST DATA Test no Width (mm) Table D.1 Impact test result of UN.PE Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total absorbed energy (J) Mean S.D Test no Width (mm) Table D.2 Impact test result of UN.E Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total absorbed energy (J) Mean S.D Test no Width (mm) Table D.3 Impact test result of UN.VE Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total absorbed energy (J) Mean S.D Adriana N E Ticoalu D 1

241 Test no Width (mm) Table D.4 Impact test result of 5N.PE Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total energy (J) Mean S.D Test no Width (mm) Table D.5 Impact test result of 5N.E Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total absorbed energy (J) Mean S.D Specim en Thickness (mm) Table D.6 Impact test result of 5N.VE Width (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Absorbed Energy (J) Mean S.D Adriana N E Ticoalu D 2

242 Test no Width (mm) Table D.7 Impact test result of 10N.PE Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total energy (J) Mean S.D Test no Width (mm) Table D.8 Impact test result of 10N.E Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total energy (J) Mean S.D Test no Width (mm) Table D.9 Impact test result of 10N.VE Thickness (mm) Energy to failure (J) Maximum load (kn) Impact velocity (m/s) Total energy (J) Mean S.D Adriana N E Ticoalu D 3

243 APPENDIX E SIMULATION OF GOMUTI FIBRE COMPOSITES ROOF TILES E.1 Calculation of t-min based on shear strength and flexural strength Known data: P = 1112 N L = 305 mm M = N.mm b = 330 mm d-max = 1.53 mm Q = N Table E.1 Calculation of t-min based on shear strength and flexural strength Composite samples Shear strength calculated t-min Flexural strength calculated t-min (MPa) (mm) (MPa) (mm) UN.PE UN.E UN.VE N.PE N.E N.VE N.PE N.E N.VE Adriana N E Ticoalu E 1

244 E.2 Detail of Strand7 simulation Figure E.1 shows the roof tile plate model under modelled loading conditions in a transverse bending strength test. The tile was modelled as a plate element (unidirectional laminate plate). The condition of interest is the strength and deflection in the longitudinal direction for a transverse bending strength. The supports were modelled as node restrain along the width of tile. The tile was subjected to a 1112 kn plate face load distributed on an area of (330 x 25) mm. Figure E.1 Roof tile model: Plate element The typical deflection mode of the tile subjected to loading is presented in Figure E.2 where it can be observed that there is mid-span deflection in the direction of y(-) and deflection at the two ends of the tile in the direction of y(+). Figure E.3 shows the stress contour of the modelled tile, in the direction of xx. Adriana N E Ticoalu E 2

245 Figure E.2 Deflection shape and stress contour of the model E.4 Simulation inputs and results Table E.3(a) Calculation of deflection based on E-tile variations and t-tile = 10 mm E-tile t Defl. (+) Defl. Midspan v E2 G (GPa) (mm) (mm) (mm) Table E.3(b) Calculation of deflection based on E-tile variations and t-tile = 12.5 mm E-tile (GPa) v E2 G t (mm) Defl. (+) (mm) Defl. Midspan (mm) Adriana N E Ticoalu E 3