Static and Dynamic Mechanical Properties of

6/3/2014

https://www.lap-publishing.com//system/covergenerator/build/85751

Static and Dynamic Mechanical Behaviour

Strain rate effect is a widely recognized crucial factor that influences the mechanical properties of a material. Despite its acknowledged importance, the understanding of how such a factor interacts with the sensitivity of polymeric materials (in terms of its mechanical properties) is still less reported and remains unclear. Therefore, in this study, an experimental technique, based on the compression Split Hopkinson Pressure Bar (SHPB), was introduced to perform dynamic compression testing; whereas a conventional universal testing machine was used to perform static compression testing, to experimentally investigate the interactive effect of strain rates towards the compressive properties of various thermoplastic-based materials

Mohd Firdaus Omar

Static and Dynamic Mechanical Properties of Thermoplastic Materials

Mohd Firdaus Omar Dr. Mohd Firdaus Omar is a senior lecturer at Universiti Malaysia Perlis (UniMAP). He received his PhD and BEng(honors) from the Universiti Sains Malaysia (USM). His research interests include polymer composites, dynamic response, and impact behavior of engineering materials.

https://www.lap-publishing.com//system/covergenerator/build/85751

Omar

978-3-659-40638-6

1/2

ACKNOWLEDGEMENTS

Alhamdulillah, all praises to Allah for the strength and His blessing in completing this thesis. A special appreciation goes to the Universiti Malaysia Perlis (UniMAP) and Ministry of Higher Education (MOHE) for sponsoring and providing the financial scholarship for this research work.

It is with immense gratitude that I acknowledge the support and help of my supervisor, Professor Hazizan Md Akil, for his guidance and constant support in my academic work and daily life. His invaluable help, constructive comments and suggestions throughout the experimental stage of this thesis have contributed to the success of this research. Not only that, his humble attitude in life has inspired me to be a better person like him. Special appreciation is expressed to Professor Takashi Yokoyama from the Okayama University of Science, Department of Mechanical Engineering, for his valuable suggestions, during several discussions, regarding this research effort. I also wish to express my deepest gratitude to Professor Zainal Arifin Bin Ahmad for his encouragement and advice, which he offered from time to time.

Correspondingly, my special thanks to Professor Hanafi Bin Ismail (Dean) and all the technical staff of the School of Materials and Mineral Resources, Universiti Sains Malaysia for their continual support and technical advice, especially in handling the research facilities. I owe my deepest appreciation to Mr Shahrul, Mr Segar, Mr Helmi, Mr Kemuridan, Mr Shahid, Mr Khairi, Mr Shahril, Mr Syafiq, Mr Shahrizol, Mr Faizal, Mr Mokhtar, Mr Halim and Mr Farid for their kindness and help during this i

research effort. Without them, I believe that the research will not be completed on time. I would also like to express my utmost gratitude to Universiti Sains Malaysia for providing me with a personal short term grant (USM-RU-PGRS) which was very helpful throughout my three years of research.

I am indebted to my many colleagues especially Firdaus Nazeri, Rashid, Bisyrul, Dr. Nik Nuriman, Hafiz, Fadli, Chang Boon Peng, Helmi, Helfy, Kahar, Nabil and Amin for their moral support and motivation whenever I needed it. In addition, my acknowledgement goes to the assistance and inspiration of all those who contributed directly or indirectly to this research effort.

Last but not least, I would like to dedicate this thesis to my parents, Haji Omar bin Zanion and Hjh Saliah Binti Lazem, and my wife, Annurin ’Ainulhidayah, who have always stood by me especially during my hard time. Their love, understanding and encouragement are always appreciated.

ii

TABLE OF CONTENTS

Acknowledgements……………………………………………………

i

Table of Contents……………………………………………………...

iii

List of Tables…………………………………………………………..

x

List of Figures……………………………………………………........ xii List of Symbols……………………………………………………….. xx List of Abbreviations…………………………………………………. xxii Abstrak……………………………………………………………….. xxiv Abstract……………………………………………………………….. xxvi

CHAPTER ONE: INTRODUCTION 1.1 Static and dynamic mechanical properties of materials

1

1.2 Development of dynamic testing

1

1.3 Static/dynamic mechanical behaviour of thermoplastic polymers

3

1.4 Static/dynamic mechanical behaviour of polypropylene based

4

composites 1.5 Problem statements

6

1.6 Objectives of study

7

1.7 Organisation of thesis

8

CHAPTER TWO: LITERATURE REVIEWS 2.1 Introduction

10

2.2 Polymers

10

2.2.1 Thermoplastic polymers

10

2.2.1.1 Characteristics of thermoplastics iii

10

2.2.1.2 Semi-crystalline versus amorphous thermoplastics

11

2.3 Composite material

13

2.4 Polymer matrix reinforced composites (PMCs)

14

2.4.1 Benefits and drawbacks of PMCs

17

2.5 Particulate filled polymer composites

17

2.5.1 Cost

18

2.5.2 Particle size

18

2.5.3 Particle shape

21

2.6 Particulate filled thermoplastic composites (PFTCs)

22

2.7 Static and dynamic mechanical testing

26

2.7.1 Dynamic mechanical testing

27

2.7.2 Drop weight impact

27

2.7.3 Taylor impact

30

2.7.4 Expanding ring

31

2.7.5 Split Hopkinson pressure bar apparatus (SHPBA)

33

2.7.5.1 History and development of SHPBA

37

2.7.5.2 Theory behind the conventional SHPBA

40

2.7.5.3 Specimen’s geometry consideration

44

2.7.5.4 SHPB testing on soft materials

45

2.8 Static/dynamic mechanical behaviour of thermoplastic polymers

46

and their composites 2.9 Strain rate sensitivity of thermoplastic polymers and their

51

composites 2.10 Summary

53

CHAPTER THREE: MATERIALS AND METHODS 3.1 Introduction

55 iv

3.2 Materials

55

3.2.1 Thermoplastic polymers

55

3.2.1.1 Polypropylene homopolymer (PP)

55

3.2.1.2 Polyethylene (PE)

56

3.2.1.3 Polycarbonate (PC)

57

3.2.2 Particulate fillers

58

3.2.2.1 Zinc oxide (ZnO)

58

3.2.2.2 Mica

58

3.2.2.3 Silica (SiO2)

59

3.3 Fabrication of polymer specimens 3.3.1 Compression moulding

59 59

3.4 Fabrication of polymer composite specimens

60

3.4.1 Compounding process

60

3.4.2 Compression moulding

61

3.4.3 Specimen cutting

61

3.5 Material characterisations

62

3.5.1 Particle size analyser

62

3.5.2 High resolution transmission electron microscopy (HRTEM)

62

3.5.3 Density measurement

62

3.5.4 Differential scanning calometry (DSC)

63

3.5.5 Dynamic mechanical analysis (DMA)

64

3.6 Mechanical tests

65

3.6.1 Static compression testing

65

3.6.2 Dynamic compression testing

66

3.6.2.1 Split Hopkinson Pressure Bar Apparatus (SHPBA) 3.6.3 Static tensile testing

66 67

3.7 Post damage analysis

67

3.7.1 Field emission scanning electron microscopy (FESEM) v

67

3.7.2 Energy dispersive X-ray spectroscopy (EDXS)

68

3.8 Experimental chart

69

CHAPTER FOUR: CALIBRATIONS AND VERIFICATIONS OF THE SHPB RESULTS 4.1 Introduction

70

4.2 Mechanical impedance

72

4.3 Dynamic stress equilibrium

75

4.4 Specimen’s slenderness ratio

77

4.5 Calibration of the SHPB set-up

79

4.6 Verification of the average strain rate

85

4.7 Summary

87

CHAPTER FIVE: MEASUREMENT AND PREDICITION ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF THERMOPLASTIC POLYMERS 5.1 Introduction

88


90

5.2.1 Dynamic Mechanical Analysis (DMA)

90

5.3 Stress/strain characteristic

96

5.4 Stiffness and strength properties

101

5.5 Strain rate sensitivity, thermal activation volume and strain

103

energy 5.6 The solution in the numerical equations

107

5.7 Summary

110

vi

CHAPTER SIX : EFFECT OF MOLECULAR STRUCTURE ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF THERMOPLASTIC POLYMER 6.1 Introduction

112


113

6.2.1 Density analysis

113

6.2.2 Crystallinity measurement

114


116

6.4 Yield behaviour

120

6.5 Stiffness and strength properties

123

6.6 Strain rate sensitivity, thermal activation volume and strain

125

energy 6.7 The solution in the numerical equations

128

6.8 Summary

130

CHAPTER SEVEN: MEASUREMENT ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF POLYPROPYLENE BASED COMPOSITES USING NANO AND MICRO FILLERS 7.1 Introduction

132

POLYPROPYLENE/NANO-ZINC OXIDE

134

NANOCOMPOSITES 7.2 Material characterisations

134


134


141

7.4 Stiffness properties

146 vii

7.5 Strength properties

147

7.6 Rate sensitivity, thermal activation volume and strain energy

150

7.7 Fracture surface analysis

152

POLYPROPYLENE/MICRO-MICA MICROCOMPOSITES

155


155


155


160


164


165


168

7.13 Fracture surface analysis

172

7.14 Summary

174

CHAPTER EIGHT: EFFECT OF PARTICLE SIZE ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF POLYPROPYLENE BASED COMPOSITES 8.1 Introduction

176


178

8.2.1 Particle size confirmation

178


182


185


187


190

8.7 Post damage analysis

192

8.7.1 Physical analysis

192 viii

8.7.2 Fracture surface analysis

193

8.8 Summary

196

CHAPTER NINE: CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK 9.1 Conclusions

198

9.2 Suggestions for further work

200

REFERENCES

202

APPENDICES

224

Appendix A (List of International and National Journal)

224

ix

LIST OF TABLES Pages Table 2.1

Example of both semi-crystalline thermoplastics and

13

amorphous thermoplastics and their characteristics Table 2.2

The current development of the SHPB apparatus

40

Table 3.1

The properties of polypropylene homopolymer

55

Table 3.2

The properties of three different types of

56

polyethylene Table 3.3

The properties of polycarbonate polymer used in this

57

study Table 3.4

The typical properties of ZnO nanoparticles

58

Table 3.5

The typical properties of Mica microparticles

58

Table 3.6

The typical properties of SiO2 particles

59

Table 3.7

The recipes used and identifications for PP/nano-

60

ZnO composites and PP/micro-Mica composites Table 4.1

The mechanical impedance characteristic of all tested

74

specimens Table 5.1

The overall properties of various thermoplastic

100

polymers under both static and dynamic loadings Table 5.2

The strain rate sensitivity and thermal activation

106

volume of various polymeric materials under different region of strain rates Table 5.3

The parameters used for the numerical equations

108

Table 6.1

Measured density values of various types of

114

polyethylene specimens

x

Table 6.2

DSC characterisation of polyethylene specimens

116

Table 6.3

Typical properties of PE specimens under a wide

119

range of strain rate investigated Table 6.4

The strain rate sensitivities and thermal activation

126

volumes of various types of PE under different region of strain rates investigated Table 6.5

The parameters used for the numerical equations

128

Table 7.1

The rate sensitivity and the thermal activation

150

volume of the neat PP and the PP/nano-ZnO composites under static and dynamic regions Table 7.2

The rate sensitivity and the thermal activation

170

volume of the neat PP and the PP/micro-Mica composites under static and dynamic regions Table 7.3

The comparison of energy absorbing capacity

171

between PP/micro-Mica composites and PP/nanoZnO composites under a wide range of strain rate investigated Table 8.1

The rate sensitivities and thermal activation volume of the PP/SiO2 (as function of particle sizes) measured under static and dynamic regions

xi

191

LIST OF FIGURES Pages Figure 2.1

Molecular arrangement of amorphous and semi-

11

crystalline thermoplastics Figure 2.2

The classification of PMCs

16

Figure 2.3

Idealised view of the way filler particles disperse

19

and of the different forms of particle types that might be encountered Figure 2.4

Complex particle dispersion behaviour, as often

20

encountered with fine, precipitated fillers Figure 2.5

Some types of common particle shapes in

21

particulate-filled composites Figure 2.6

Schematic diagram of strain rate regimes (in

27

reciprocal seconds) and the techniques that are suitable for obtaining them Figure 2.7

Schematic diagram of drop weight impact machine

29

Figure 2.8

Sequence of deformation after impact of cylindrical

31

projectile against rigid wall Figure 2.9

Principle of expanding ring technique ; (a) steel

32

block with explosive in core ; (b) steel block with explosive in core ; (c) section of ring Figure 2.10 Schematic diagram of conventional SHPBA

34

Figure 2.11 The Lagrangian x-t diagram illustrating wave

34

movement in the Hopkinson Bars Figure 2.12 Schematic diagram of SHPB output signal

36

Figure 2.13 Schematic diagram of traditional SHPB device in

38

1914 xii

Figure 2.14 The usage and development of SHPB apparatus

39

from 1940 to 1998 Figure 2.15 Typical diagram of the Hopkinson Bars

44

Figure 2.16 The compressive stress/strain curves of (A) HDPE,

48

(B) PP and (PC) under a wide range of strain rate investigated Figure 2.17 Typical stress–strain curves for epoxy–SiO2

50

nanocomposites at room temperature Figure 3.1

Schematic molecular structures representation of

57

different types of polyethylene Figure 4.1

The variation of stress uniformity with the number

77

of wave cycles for different bar–specimen relative impedances Figure 4.2

The oscilloscope readings of the (A) 0.33

80

slenderness ratio, (B) 0.50 slenderness ratio, and (C) 1.33 slenderness ratios Figure 4.3

Time histories of applied stress of the (A) 0.33

81

slenderness ratio, (B) 0.5 slenderness ratio, and (C) 1.33 slenderness ratios Figure 4.4

Strain gauge signals on the oscilloscope during

83

calibration Figure 4.5

Comparison of stress vs. time characteristic, derived

83

from strain gauge signals during calibration Figure 4.6

Oscilloscope traces from split Hopkinson pressure

84

bar test on polyethylene (Vs = 12.57 ms-1) Figure 4.7

Time histories of compressive strength on each face of polypropylene specimen

xiii

85

Figure 4.8

Dynamic true stress–strain and strain rate–strain

86

curve in compression on polypropylene (PP) with 16.4 ms-1 striking velocity Figure 5.1

The DMA curves of (A) Storage modulus, (B) Tan

93

delta, (C) Loss modulus, and (D) Superposition curve of polyethylene Figure 5.2


94

delta, (C) Loss modulus, and (D) Superposition curve of polypropylene Figure 5.3


95

delta, (C) Loss modulus, and (D) Superposition curve of polycarbonate Figure 5.4

Typical true stress/strain curves of several

98

polymeric materials at different level of strain rates: (A) PE, (B) PP and (C) PC Figure 5.5

The yield stresses of tested polymers under a wide

99

range of strain rate investigate Figure 5.6

The yield strains of tested polymers under a wide

100

range of strain rate investigated Figure 5.7

The compression modulus of tested polymers under

102

various loading rates Figure 5.8

The ultimate compressive strength of tested

102

polymers under various loading rates Figure 5.9

The strain energy of various polymers under a wide

106

range of strain rate (strain, ε = 0.025) Figure 5.10 The relationship between experimental and numerical values of yield stress for various xiv

109

polymeric materials. The experimental values are from the curves in Figure 5.5 Figure 5.11 The relationship between experimental and

110

numerical values of unstable strain for various polymeric materials Figure 6.1

DSC thermograms represent melting region of

115

various polyethylene specimens Figure 6.2

DSC thermograms represent crystalline region of

116

various polyethylene specimens Figure 6.3

The typical stress/strain curves for (A) LDPE, (B)

119

LLDPE and (C) HDPE under various loading rates Figure 6.4

The yield stress values of LDPE, LLDPE and

121

HDPE specimens under various strain rates Figure 6.5

The yield strain values of LDPE, LLDPE and

122

HDPE specimens under various strain rates Figure 6.6

A schematic diagram illustrating a polymer

122

crystalline spherulite Figure 6.7

The compression modulus of PE specimens of

124

LDPE, LLDPE and HDPE over a wide range of strain rates Figure 6.8

The ultimate compressive strength (UCS) of LDPE,

125

LLDPE and HDPE over a wide range of strain rates Figure 6.9

The strain energy of LDPE, LLDPE and HDPE

127

specimens under static and dynamic loadings Figure 6.10 The

relationship

between

experimental

and

numerical values of yield stress for PE specimens. The experimental values (yield stress) are from the curves in Figure 6.4 xv

129

Figure 6.11 The

relationship

between

experimental

and

130

numerical values of unstable strain for PE specimens Figure 7.1

The DMA curves of (A) Storage modulus and tan

138

delta, (B) Loss modulus, and (C) Superposition curve of PP/1% ZnO composite Figure 7.2


139



140


True compression stress/strain curves of the (A) PP

144

neat, (B) PP/1% ZnO, (C) PP/3% ZnO and (D) PP/5% ZnO composites under a wide range of strain rate investigated Figure 7.5

The typical features of stress/strain curves of

145

PP/nano-ZnO composites under dynamic loading Figure 7.6

Toughening mechanism with rigid particle under

145

compression loading Figure 7.7

The compression modulus of the pure PP and the

147

PP/nano-ZnO composites under various levels of strain rates investigated Figure 7.8

The yield strength and the ultimate strength values of the pure PP and the PP/nano-ZnO composites under a wide range of strain rate investigated

xvi

149

Figure 7.9

The strain energy of the pure PP and the PP/nano-

152

ZnO composites under various loading rates Figure 7.10 The FESEM micrographs of the fractured surface of

154

(A) PP/1% ZnO, (B) PP/3% ZnO, and (C) PP/5% ZnO composites at 16.47 s-1 of striking velocity Figure 7.11 The DMA curves of (A) Storage modulus and tan

157

delta, (B) Loss modulus, and (C) Superposition curve of PP/10% Mica composite Figure 7.12 The DMA curves of (A) Storage modulus and tan

158

delta, (B) Loss modulus, and (C) Superposition curve of PP/15% Mica composite Figure 7.13 The DMA curves of (A) Storage modulus and tan

159

delta, (B) Loss modulus, and (C) Superposition curve of PP/20% Mica composite Figure 7.14 True compression stress/strain curves of the (A) PP

162

neat, (B) PP/10% Mica, (C) PP/15% Mica and (D) PP/20% Mica composites under a wide range of strain rate investigated Figure 7.15 The determination of aspect ratio for both ZnO and

163

Mica (Chiu et al., 2008) particles Figure 7.16 The compression modulus of the neat PP and the

165

PP/micro-Mica composite under various levels of strain rates investigated Figure 7.17 The yield strength of the neat PP and the PP/micro-

167

Mica composite under various levels of strain rates investigated Figure 7.18 The compressive strength of the neat PP and the PP/micro-Mica composite under various levels of xvii

168

strain rates investigated Figure 7.19 The strain energy of the neat PP and the PP/micro-

170

Mica composites under various loading rates Figure 7.20 The FESEM micrographs of the fractured surface of

173

(A) PP/10% Mica, (B) PP/15% Mica, and (C) PP/20% Mica composites at 16.47 ms-1 of striking velocity Figure 8.1

(A-B) The particle sizes distribution results and

181

SEM images of SiO2 microparticles. (C-D) The TEM images of SiO2 nanoparticles used in this study Figure 8.2

True stress/strain curves of PP/SiO2

185

nanocomposites with various particles sizes under different levels of loading rates Figure 8.3

The compression modulus values of PP/SiO2

186

nanocomposites under various loading rates as a function of particle sizes Figure 8.4

The yield strength values of PP/SiO2

189

nanocomposites under various loading rates as a function of particle sizes Figure 8.5

The ultimate compressive strength values of

189

PP/SiO2 nanocomposites under various loading rates as a function of particle sizes Figure 8.6

The strain energy of PP/SiO2 composites as a function of particle sizes under a wide range of strain rate investigated (ε: 0.025) xviii

192

Figure 8.7

The photographs of PP/SiO2 composites specimen

193

under (A) static loading (0.1 s-1) and (B) dynamic loading (650 s-1) as a function of particle sizes Figure 8.8

The fracture surface of PP/SiO2 composites with various particle sizes (A) 3 µm, (B) 1 µm, (C) 20 nm (D) 11 nm under 1100s-1 of strain rate

xix

196

LIST OF SYMBOLS

·

e

Strain rate

DH m

Melting Heat of The Sample

DH 0

Melting Heat of 100% Crystalline Thermoplastic

w

Test Frequency

V*

Thermal Activation Volume

r

Density

d

Tan Delta

q

Bragg Angle

“k”

Number of Wave Cycles

Ab

Cross-Sectional Area of Bar

As

Cross-Sectional Area of Specimen

Co

Wave Velocity

dhkl

Distance Between Adjacent Planes

do

Displacement Amplitude

E

Bridge Voltage

E

Young Modulus

E’

Storage Modulus

E’’

Loss Modulus

eo

Voltage Change In The Bridge

G

Amplifier Gain Factor

k

Boltzmann Constant

lg

Specimen Gauge Length

lo

Initial Length of The Specimen

n

Interger Number

sg

Strain Gage Factor xx

T

Absolute Temperature

V

Mechanical Impedance Value

v

Velocity of Deformation

β

Strain Rate Sensitivity Parameter

ε

Strain Pulse

εi

Incident Strain Pulse

εr

Reflected Strain Pulse

εt

Transmitted Strain Pulse

λ

Wavelength of X-ray Beam

σ

Stress

σ average

Average Stress

σ back

Back Stress

σ front

Front Stress

σi

Internal Stress

σy

Yield Stress

xxi

LIST OF ABBREVIATIONS

ABS

Acrylonitrile Butadiene Styrene

ASTM

American Society For Testing and Materials

CaCO3

Calcium Carbonate

CMCs

Ceramic Matrix Composites

DMA

Dynamic Mechanical Analysis

DSC

Differential Scanning Calometry

EDX

Energy Dispersive X-ray

FESEM

Field Emission Scanning Electron Microscopy

FT-IR

Fourier Transform Infrared Spectroscopy

GFRC

Glass Fiber Reinforced Composite

HDPE

High Density Polyethylene

HRTEM

High Resolution Transmission Electron Microscopy

KBr

Potassium Bromide

LDPE

Low Density Polyethylene

LLDPE

Linear Low-Density Polyethylene

MMCs

Metal Matrix Composites

PA-6

Nylon 6

PA-66

Nylon 66

PC

Polycarbonate

PE

Polyethylene

PEEK

Polyether Ether Ketone

PET

Polyethylene Terephthalate

PFTCs

Particulate Filled Thermoplastic Composites

PMCs

Polymer Matrix Reinforced Composites

PMMA

Polymethyl Methacrylate

POM

Polyoxymethylene xxii

PP

Polypropylene

PS

Polystyrene

PTFE

Polytetrafluoroethylene

PVC

Polyvinyl Chloride

SHPBA

Split Hopkinson Pressure Bar Apparatus

SiO2

Silica

TGA

Thermogravimetric Analysis

UCS

Ultimate Compressive Strength

UHMWPE Ultra High Molecular Weight Polyethylene UTM

Universal Testing Machine

XRD

X-ray Diffraction

ZnO

Zinc Oxide

xxiii

SIFAT-SIFAT MEKANIKAL STATIK DAN DINAMIK BAGI BAHAN-BAHAN TERMOPLASTIK ABSTRAK

Dalam kajian ini, teknik eksperimen, berasaskan pemampatan pecahan Hopkinson tekanan bar (SHPB), telah diperkenalkan untuk menjalankan ujian mampatan dinamik manakala mesin ujian konvensional sejagat telah digunakan untuk menjalankan ujian mampatan statik. Keduadua teknik digunakan untuk secara eksperimennya menyiasat kesan interaktif kadar terikan terhadap sifat-sifat mampatan pelbagai bahan berasaskan termoplastik. Semua bahan-bahan berasaskan termoplastik yang digunakan dalam kajian ini telah dihasilkan menggunakan proses penekanan panas. Keputusan SHPB pada awalnya telah disahkan dan ditentukur.

Hasil

kajian

menunjukkan

bahawa

semua

spesimen

termoplastik yang diuji (iaitu PP, PE, dan PC) mempamerkan pergantungan besar pada kadar terikan yang dikenakan; dimana tegasan alah, modulus mampatan dan kekuatan mampatan, semuanya telah meningkat dengan peningkatan kadar terikan. Menariknya, kedua-dua persamaan Eyring dan persamaan hukum kuasa asas hampir selari dengan keputusan uji kaji bagi keseluruhan kadar terikan yang disiasat. Kesan struktur molekul, terhadap sifat-sifat mekanikal statik dan dinamik bagi termoplastik polimer, juga telah ditentukan menggunakan spesimen polietilena dengan struktur molekul yang berbeza (iaitu LDPE, LLDPE and HDPE). Keputusan menunjukkan bahawa struktur molekul polietilena telah memberi kesan kepada sifat-sifat mekanikal dari segi takat alah, kekakuan, kekuatan, kadar kepekaan, isipadu pengaktifan, dan tenaga yang diserap. Bagi komposit berasaskan termoplastik, dua jenis partikel pengisi telah ditambah ke dalam matrik polipropilena; iaitu zink oksida dan mika. Ia boleh dilihat secara xxiv

jelas bahawa pengenalan pengisi meningkatkan sifat-sifat mampatan komposit, termasuk modulus mampatan, serta kekuatan alahnya. Kajian juga telah mendapati bahawa kandungan partikel mempamerkan hubungan yang tidak ketara dengan sensitiviti kadar tekanan dan isipadu haba pengaktifan, bagi kedua-dua polipropilena diperkuat sistem komposit. Bagi kesan ciri-ciri partikel-matrik, serbuk partikel silika telah dimanipulasi untuk menyiasat secara eksperimen mengenai hubungan antara saiz partikel dan sifat-sifat mekanikal komposit di bawah pelbagai kadar terikan yang di kenakan. Menariknya, saiz partikel-partikel silika memberikan kesan yang jelas ke atas sifat-sifat mampatan komposit berasaskan polipropilena. Secara kuantitatifnya, komposit dengan silika bersaiz nano mencatatkan sifat-sifat mampatan yang lebih tinggi, dari segi kekuatan alah, kekuatan muktamad dan kekakuan berbanding komposit dengan silika besaiz mikro, untuk semua kadar terikan yang di siasat.

xxv

STATIC AND DYNAMIC MECHANICAL PROPERTIES OF THERMOPLASTIC MATERIALS

ABSTRACT

In this study, an experimental technique, based on the compression Split Hopkinson Pressure Bar (SHPB), was introduced to perform dynamic compression testing whereas a conventional universal testing machine was used to perform static compression testing. These two techniques were used to experimentally investigate the interactive effect of strain rates towards the compressive properties of various thermoplastic-based materials. All of the thermoplastic-based materials used in this study were fabricated using a hot press process. The SHPB results were initially verified and calibrated. The results indicated that all tested thermoplastic specimens (i.e. PP, PE, and PC) showed a great dependency on the strain rate applied; where the yield stress, compression modulus, and compressive strength, were all proportionally increased as the strain rate was increased. Interestingly, both Eyring and basic power law equations were almost agreed with the experimental results over a wide range of strain rates investigated. The effect of molecular structure, on the static and dynamic mechanical properties of thermoplastic polymer, was also determined using polyethylene specimens with different molecular structures (i.e. LDPE, LLDPE, and HDPE). The results indicated that the molecular structure of polyethylene did affect its mechanical properties in terms of yield behaviour, stiffness, strength, rate sensitivity, activation volume, and absorbed energy. For thermoplastic based reinforced composites, two types of particulate fillers were added into the polypropylene matrix namely zinc oxide and mica. It can be clearly seen that the introduction of filler xxvi

increased the composites’ compressive properties, including their compression modulus, as well as their yield strength. It was also found that the particle content showed an insignificant relationship with strain rate sensitivity and thermal activation volume, for both polypropylene reinforced composite systems. As for the effect of particle-matrix characteristics, silica particles were manipulated to experimentally investigate the correlation between particle size and the mechanical properties of composites under a wide range of strain rates investigated. Interestingly, the size of the silica particles gave significant effects on the compressive

properties

of

the

polypropylene-based

composites.

Quantitatively, composites with nano-sized silica recorded higher compressive properties, in terms of yield strength, ultimate strength and stiffness as compared to composites with micro-sized, for all strain rates investigated.

xxvii

CHAPTER 1 INTRODUCTION

1.1Static and dynamic mechanical properties of materials The mechanical performance of materials is determined by their static and dynamic mechanical characteristics. Unfortunately, the majority of scientists only focus on the static rather than on the dynamic mechanical behaviour of materials. This is attributed to the limited number of dynamic facilities, as well as the difficulty in performing dynamic testing. Intensive efforts should be carried out in the future in order to gain a better understanding of the dynamic perspective of the behaviour of materials. 1.2Development of dynamic testing The knowledge of the characteristics of a material at dynamic loading is becoming ever more essential with the desire to produce products or structures that are capable of withstanding high velocity impacts. Based on this, several conventional mechanical tests have been developed over the years to obtain the mechanical characteristics of materials at high strain rates, using screw or hydraulic loading systems (Hamouda and Hashmi, 1998; Field et al., 2004). For example, a pendulum impact machine, such as the Charpy or Izod, can yield a strain rate of up to 100 s-1, but only provide the absorbed energy information up to fracture. Meanwhile, other common high strain rate facilities are the drop-weight impact and the servo-hydraulic tester. Although a drop-weight impact test can give both impressive and convenient results, but the test is still limited by several factors, such as the sensitivity towards contact conditions 1

between the impactor and the specimen (Hsiao et al., 1999). In addition, the drop-weight test is also restricted to lower strain rate conditions (i.e. between 1 to 10 ms-1) (Richardson and Wisheart, 1996). On the other hand, the servo-hydraulic test also has a similar restriction with the drop-weight impact test, where it is only credible for intermediate strain rates (Othman et al., 2009). A desire to scrutinize the characteristics of materials at very high strain rates revealed a most promising technique, namely the Split Hopkinson Pressure Bar (SHPB) technique. The SHPB technique was initiated by Kolsky (1949) and developed by Hauser (1966), where a stress pulse travelled through an elastic input bar, through a short sample, and finally into an elastic output bar. The important characterisation of the SHPB technique is that it is highly dependent on the capability of the technique to obtain a stress–strain curve as the output, which holds useful information as to the characteristics of materials. Even though Kolsky introduced this technique almost five decades ago, it was only intensively used by researchers during the early 1970s. More recently, the SHPB technique has become the standard method for measuring the dynamic mechanical properties of materials in the range of 102 s-1 to 104 s-1 strain rates. (Evora and Shukla, 2003; Field et al., 2004). In the SHPB set-up, a semiconductor strain gauge is mounted on each Hopkinson bar. Meanwhile, the stress and strain within the specimen are obtained from an analysis of the signals from these two gauges. One of the basic and fundamental assumptions of the SHPB technique is the stress homogeneity within the sample. The technique assumes that the stress field is homogenous within the sample and that the propagating waves in the bar have a negligible attenuation and dispersion. However, a conventional SHPB is not suitable for low impedance materials, such as polymers, 2

polymeric foams and rubbers, because the transmitted signal is too small to be captured by the strain gauge mounted on the transmitter bar (Song and Chen, 2005; Van Sligtenhorst et al., 2006). Besides that, the equilibrium state is reached slowly when testing soft materials. Based on this limitation, the conventional theory of the SHPB technique will be invalid, and other solutions must be found. Lately, two common approaches have emerged to overcome this dilemma. In the first method, the application of a pulse shaper was used to induce a faster dynamic equilibrium achievement (Frew et al., 2005; Vecchio and Jiang, 2007). On the other hand, the second method was to use a low-impedance pressure bar, e.g. a polymer bar, which has an impedance value closer to that of the materials being tested (Johnson et al., 2010). It is believed that a closer impedance mismatch will significantly enhance the propagation of the transmitted pulse. Based on this concern, it is convenient to say that the SHPB test is still reliable for the performance of dynamic testing on soft specimens, especially polymeric based materials. 1.3Static and dynamic mechanical behaviour of thermoplastic polymers Thermoplastic polymers have been extensively used as engineering components that are purposefully designed to resist impact, ranging from bottles and pipes to helmets and body armours. Among the many types of thermoplastic polymers, polyethylene (PE), polypropylene (PP) and polycarbonate (PC) have attracted much interest from scientists as well as industry sectors. Based on this, many studies have been conducted to investigate their overall characteristics, especially their mechanical performances (Li et al., 1995; Karian, 2003; Nitta and Maeda, 2010). Unfortunately, most of the previous researches were only focused on their static mechanical behaviour. Since the applications of these thermoplastic 3

polymers have been extended from conservative to various engineering applications, the strain rate factor should not be neglected and requires extra precaution from the researcher. Based on this consideration, some of the researchers have taken the initiative to experimentally investigate the mechanical properties of PE, PP, as well as PC at various levels of strain rates (Walley and Field, 1994; Mulliken and Boyce, 2006; Cao and Wang, 2012). Walley and Field (1994) reported that PE and PP show different patterns in terms of the rate sensitivity (i.e. maximum stress) as a function of the applied strain rate. It was experimentally proven that the PP specimen exhibits a bilinear relationship, where the rate sensitivity of stress increases sharply at a strain rate of about 103 s-1. Conversely, the PE specimen does not unambiguously show a change of slope over a wide range of strain rates. Apart from that, Mulliken and Boyce (2006) found that the PC specimen shows a different magnitude of increment in terms of yield stress under both static and dynamic loading. They also reported that the slope of the yield stress increment is much greater under dynamic loading than that of static loading. Based on the recorded results, it is believed that the knowledge of the dynamic mechanical characteristic of these thermoplastic polymers (i.e. PE, PP, and PC) is still presently unclear. Therefore, it is believed that a more detailed and systematic study should be carried out in the future in order to achieve a conclusive explanation on the highlighted issues. 1.4 Static and dynamic mechanical behaviour of polypropylene based composites In general, polypropylene (PP) is renowned as a high volume commodity plastic, with a remarkable cost/performance balance, which has 4

contributed to its commercial success. Regrettably, this thermoplastic polymer is still referred to as a low cost engineering plastic and is inappropriate for crucial engineering applications. It is widely accepted that the incorporation of fillers into a PP matrix has shown great potential in increasing the longevity and durability of PP, fulfilling various requirements of engineering applications. For the past few years, many studies have been carried out on polypropylene-based composites using micro and nano-sized particulate reinforcement (Balasuriya et al., 2001; Alcock et al., 2007). They found that PP composites that are reinforced with nano-sized particles exhibit greater properties compared to their micro-scale reinforced counterparts (Jeong et al., 2005; Thostenson et al., 2005). Furthermore, it was also found that nanocomposites, with a good dispersion of fillers, show significant improvements in terms of their mechanical, thermal, electrical, optical, and physical-chemical properties, even at relatively low filler contents (Javni et al., 2002; Friedrich et al., 2005; Cho et al., 2006). Zinc oxide (ZnO) and mica particles are promising fillers for reinforcing the PP matrix due to their outstanding properties as compared to other conventional fillers of a similar nature (Chiang et al., 2005; Cheng et al., 2007; Rashid et al., 2008; Rashid et al., 2011). Apart from that, it is believed that both composite systems have great potential as engineering products due to their capability to provide a good balance between impact resistance, production cost, and weight. Typically, the majority of engineering products are mainly subjected to dynamic loading and therefore it is critical to investigate the dynamic behaviours of these two composite systems in order to avoid any mishaps during service. As with virgin polymers, existing works are highly focused on their static mechanical behaviour (i.e. PP reinforced with ZnO and mica particles). 5

Unfortunately, the influence of fillers on strain rate sensitivity and the dynamic behaviour of both composite systems have often not been considered. This phenomenon might be attributed to the nature of the composite, which can complicate the specimen’s geometrical design for dynamic testing (Hamouda and Hashmi, 1998). Nevertheless, several researchers have come out with an optimised specimen’s geometry to overcome this drawback and claim that the dynamic facilities are also suitable and reliable for composite materials, especially polymer matrix composites (PMCs) (Hao et al., 2005; Guo and Li, 2007). Therefore, this is a great opportunity to discover the capabilities and possibilities of these composites to replace conventional materials, especially in dynamic loading applications. 1.5Problem statements It is generally acknowledged that the applications of thermoplasticbased products have been extended from conservative to more challenging applications like engineering components, constructions, load-bearing applications, etc. Hence, the strain rate effect should be the first priority factor to be investigated, since almost all of the highlighted applications are mainly involved with both static and dynamic conditions. In addition, the knowledge of rate sensitivity is also important during material selection in order to estimate the magnitude of changes in material’s properties. Without this knowledge, it is almost impossible to predict and prevent the unexpected failure during service. Recently, there is a very limited number of works that are concerned with the dynamic behaviour as well as the rate sensitivity of thermoplastic-based materials. In addition, numerical studies on the dynamic mechanical properties of these composites are also infrequently reported and need an additional effort to further clarify the 6

relationship between the experimental and numerical results, which is important for engineering design and simulation purposes. Based on the highlighted issues, we believe that a systematic study is necessary to fulfil the lack of information in this area. Apart from external factors like the strain rate effect, it was also believed that the internal structures of polymer (Liu and Baker, 1992; Wood-Adams et al., 2000; Wood-Adams, 2001) and the filler-matrix related characteristics of the polymer composites (i.e. as particle size, particle–matrix interface adhesion, particle shape and geometry) may also influence the mechanical properties of the polymeric specimens. However, we recognize that a similar kind of study under a dynamic range of strain rates has never been reported in the past and remains a major challenge in the development of a better understanding on the mechanical behaviour of thermoplastic-based products under various loading conditions. 1.6 Objectives of study The objectives of this study are: 1) To compare the static and dynamic mechanical properties of several thermoplastic polymers 2) To examine the effect of molecular structures on the static and dynamic compressive properties of thermoplastic polymers. 3) To measure the static and dynamic mechanical properties of polypropylene based composites using nano and micro fillers. 4) To investigate the effect of particle size on the static and dynamic compressive properties of polypropylene-based composites.

7

1.7 Organisation of thesis This thesis has been divided into altogether nine chapters. Each chapter gives the information about the research interest as mentioned in the objectives earlier.

· Chapter 1 covers the introduction of the thesis. It contains a general overview on the development of static and dynamic testing and a brief introduction about dynamic studies on polymeric materials, a problem statement, objectives of the project and organisation of the thesis. · Chapter 2 contains some fundamental concepts of the split Hopkinson pressure bar technique, together with some reviews of related works reported in previous literature. · Chapter

3

explains

the

material

specifications,

research

methodology, and experimental procedures which are carried out in this study. · Chapter 4 discusses the calibrations and verifications of the SHPB results. · Chapter 5 discusses the effect of the strain rate on several thermoplastic polymers (i.e. PE, PP and PC). In addition, in this chapter, the experimental results were also compared with two established equations namely the Eyring and power basic equations. 8

· Chapter 6 discusses the effect of molecular structure on the static and dynamic compressive properties of the thermoplastic polymer, PE. As with Chapter 5, both the experimental and numerical results were validated using two established equations namely the Eyring and power basic equations. · Chapter 7 discusses the effect of the strain rate and particle content on the static and dynamic compressive properties of polypropylenebased composites (i.e. PP/nano-ZnO composites and PP/micro-Mica composites).

· Chapter 8 discusses the effect of particle size on the static and dynamic compressive properties of polypropylene-based composites (i.e. PP/SiO2 composites).

· Chapter 9 concludes the findings of the project and the evaluation that has been made in order to assess the achievements of the objectives. Some of the suggestions for further study have been explained.

9

CHAPTER 2 LITERATURE REVIEWS

2.1 Introduction This chapter summarises the principle of thermoplastic polymers followed by a brief overview of thermoplastic-based composites, explaining their increasing use in a wide range of engineering applications. In addition, a literature survey was carried out on the development of the dynamic facilities, especially a Split-Hopkinson pressure bar apparatus (SHPBA). Works on the static and dynamic behaviours of thermoplastic polymers and their composites were also extensively reviewed. 2.2 Polymers Basically, the word polymer is derived from two different Greek roots which are ‘poly-‘, meaning many, and ‘mer’, meaning part or segment. Therefore, a polymer can be defined as the repetition of many similar segments (i.e. mer) that are connected together to form a long chain. In general, polymers are classified into three different classes which are thermoplastics, thermosets and elastomers (Harper, 2002). Among those classes, thermoplastic polymers have been widely used in both conservative and extreme applications.

2.2.1 Thermoplastic polymers 2.2.1.1 Characteristics of thermoplastics 10

A thermoplastic polymer usually begins in the form of a pellet, and then becomes softer (i.e. pliable and plastic) with increasing heat. As it cools, it will reversely transform back to the solid state without any crosslinked formation. This process (i.e. heating and cooling) can be repeated over and over, though continual recycling will ultimately degrade the polymer. In general, a thermoplastic polymer is subdivided into two distinct classes of molecular arrangement, which are semi-crystalline and amorphous as can be seen in Figure 2.1. These differences will significantly affect the behaviour of the thermoplastic material, especially during processing.

Figure 2.1: Molecular arrangement of amorphous and semi-crystalline thermoplastics (http://www.azom.com/article.aspx?ArticleID=83)

2.2.1.2 Semi-crystalline versus amorphous thermoplastics

Thermoplastic polymers like polypropylene (PP), polyethylene (PE), nylons (PA), polyacetal (POM), and thermoplastic polyesters (PET) are referred to as semi-crystalline thermoplastics where in the solid state, a 11

great proportion of their molecular chains are structurally ordered and closely packed in certain specific alignments. Meanwhile, polycarbonates (PC), polystyrene (PS), polyvinyl chloride (PVC) and acrylics (PMMA) are examples of amorphous thermoplastics. This indicates that in the solid state, their molecular chains are randomly arranged and this is attributed to the complex entanglement. It should be noted that at very high temperatures (i.e. melting state), both types of thermoplastic polymers will portray similar features of amorphous molecular structures. The key characteristics of semi-crystalline thermoplastics are translucent or opaque white colour, sharp melting point, good resistance to stress cracking and good fatigue resistance. Meanwhile, most amorphous thermoplastics tend to be naturally transparent, soften over a broad range of temperatures, prone to stress cracking and poor fatigue resistance. Table 2.1 shows the common examples for both types of thermoplastic polymers and their specific characteristics.

12

Table 2.1: Example of both semi-crystalline thermoplastics and amorphous thermoplastics and their characteristics. (http://www.slideshare.net/Annie05/amorphous-and-semi-crystallinepresentation)

2.3 Composite material A composite material is a material system that consists of two or more constituent materials with significant differences in terms of their physical or chemical properties, and which remain separate at the macroscopic or microscopic scale within the finished structure. The history of composite materials started in the early 20th century. During that time, fibreglass was first introduced to reinforce several high performance thermoset resins, such as polyester and epoxy resins. After a few years, the composites industry evolved from thermosets to plastic resins in order to fulfil a wide range of light weight applications. It was reported that the earliest applications of glass fibre reinforced composite (GFRC) products 13

were in the marine industry (Strong, 2002). In 1943, the first plane with a GFRC fuse ledge was flown at the Wright-Patterson Air Force base. Although this composite system was introduced almost seven decades ago but the GFRCs still dominate the recent composites market (i.e. covering approximately 90% of the composites market). The rapid development of composite systems has revealed numerous types of newer and stronger reinforcements. Not only that, the alternative materials from metals and ceramics have also been manipulated as competitive matrix materials. To avoid confusion, composite families are divided into three distinct classes depending on the nature of their matrix. The most promising composites in recent industries is polymer matrix reinforced composites (PMCs). This composites system is based on the polymer matrix in either thermosets or thermoplastics. Although most of the PMCs are reinforced with fibre, but recently particulate fillers have started to gain much attention from researchers as well as industries. Other types of composites systems are called metal matrix composites (MMCs) and ceramic matrix composites (CMCs) (Callister Jr, 1994). 2.4 Polymer matrix reinforced composites (PMCs) According to Othman (2007), PMCs are classified into two major groups which are thermoplastics and thermosets. The thermoplastic group is subdivided into four other groups, i.e. glass mat, fibre reinforced, natural fibre thermoplastic composites and mineral reinforced thermoplastics. Meanwhile, the thermoset group is subdivided into two groups, i.e. carbon reinforced

and

natural/synthetic

fibre

reinforced.

The

detailed

classifications of the polymer matrix composites are depicted in Figure 2.2. 14

A different classification of PMCs was made by Alger (1997). He classified PMCs into three major groups which are: · Polymer-polymer combinations (i.e. polymer blends), · Polymer-gas combination (i.e. expended, cellular or foamed polymers), · Polymer-stiff filler combinations (i.e. polymer-fibre or polymer particulate composites). Among

these

highlighted

PMCs,

the

polymer-stiff

filler

combinations have attracted much attention from both researchers and industries due to their outstanding end properties. However, it is important to realise that these highlighted benefits can be only achieved with the proper selection of the constituent materials and manufacturing techniques. Otherwise, the end results might be different or lesser. It is easy to be impressed with the potential benefits offered by PMCs but their drawbacks must also be considered. Normally, PMCs that contain natural fillers tend to absorb moisture, thus decreasing their overall performance, especially their mechanical properties (Dhakal et al., 2007; Mazuki et al., 2010). Apart from that, PMCs may also encounter obscurity with a high coefficient of thermal expansion characteristic which results in dimensional instability. The anisotropic nature of PMCs may also contribute to difficulties in the design process (Mallick, 1993). In terms of their thermal resistance ability, PMCs show inferior performance at elevated temperatures. These disadvantages might restrict their utilization in some fields of application.

15

16

Figure 2.2: The classification of PMCs (Othman, 2007)

2.4.1 Benefits and drawbacks of PMCs Polymer matrix composites (PMCs) offer a range of potential benefits over conventional neat polymers especially for high-end applications. The most common reason why PMCs are implemented in many critical applications like structural components is related with their capability to provide outstanding stiffness at lower weight than that of neat polymers (Callister Jr, 1994). Apart from that, other additional benefits offered by PMCs can be summarised as follows: · Mechanical properties are outstanding · Processing of PMCs does not involve high pressure and high temperature · High abrasion resistance can be achieved · Corrosion resistance is remarkable · Low thermal expansion can be achieved · Impact and damage tolerance characteristics are excellent · Low production cost 2.5 Particulate-filled polymer composites

As defined by the word “particulate”, the reinforcing phase for this kind of composites are normally spherical or at least has dimensions of similar order in all directions (Lin, 2010). The introduction of particulate fillers into polymer matrices will significantly improve nearly every property of the virgin polymers, including their processing ability, dimensional stability, chemical resistance, strength, stiffness, etc. However, to achieve optimum properties, several factors should be considered carefully at the beginning stage of the filler selection. According to Rothon 17

(2003), important filler characteristics such as cost, particle size, particle shape and geometry are the main elements that require extra consideration during the implementation of particulate fillers into a polymer matrix. Therefore, brief descriptions of each factor will be discussed in the following subtopics. 2.5.1 Cost Initially, the main reason for using filled composites is to reduce the cost of the raw materials. It is generally known that the polymer is more expensive than the particulate filler. However, it is inappropriate to directly compare filled composites with their raw materials in terms of cost saving criteria due to several principal reasons. Firstly, filled composites undergo more complex fabrication stages like the compounding process, which requires a bigger capital investment, more manpower as well as energy. Secondly, the prices of the raw materials (i.e. the filler and matrix) are normally quoted according to their weight, whilst those of their composites are based on their volume. Therefore, it will be more suitable if the comparison is made based on their cost-property performance, where the composite will definitely beat the original material (i.e. the filler and matrix). 2.5.2 Particle size Particulate size is another factor that affects the end properties of filled composites. For synthetic fillers, the particulate size is highly dependent on the conditions during the synthesis process (i.e. precipitation) and possibly by any additional coating process. Meanwhile, for natural fillers, the particulate size is determined by the extraction process from the 18

raw deposits, including the mining and separation stages. Until recently, numerous methods have been implemented to measure the particle size including optical scattering, diffraction from particulate suspensions and sieving. Theoretically, particle size can be divided into three distinct categories namely primary particles, agglomerates and aggregates. The term “agglomerates” refers to a collection of weakly bonded particles, whereas the term “aggregates” refers to a collection of strongly bonded particles. For a better understanding of the highlighted issues, Rothon (2003) has suggested an idealised view of particle types and their breakdown during the composite formation as illustrated in Figure 2.3.

Figure 2.3: Idealised view of the way filler particles disperse and of the different forms of particle types that might be encountered (Rothon, 2003) 19

In most cases, filler systems do not follow the trend shown in Figure 2.3 where the steps shown are often less sharp and overlap. The majority of them exhibit more complicated profiles as depicted in Figure 2.4. From the recorded profile in Figure 2.4, it can be seen that those agglomerates, sometimes referred to as flocs, can arise due to the loss of colloidal stability in the polymerising systems, or to reticulation (filler network formation) above the glass transition, especially in cured elastomers, an effect often observed with carbon blacks. This phenomenon becomes more difficult and serious for synthetic products, especially those formed by precipitation. For this kind of filler system, strong and complex aggregates are present. Normally, these aggregates will break down slowly and thus alter the ideal particle dispersion profile in Figure 2.3

Figure 2.4: Complex particle dispersion behaviour, as often encountered with fine, precipitated fillers (Rothon, 2003)

20

2.5.3 Particle shape As with particle size, the particle shape is an additional factor that influences the properties of the final product. The shape of the particle is often determined by the genesis of the filler, the crystal structure and processes it has undergone. Previous researchers have proposed several terms that describe the shape of a particle including spherical, flaky, platy, blocky, irregular, acicular, needle, etc. Therefore, some typical particle shapes that are likely to be found in most particulate-filled composites are illustrated in Figure 2.5.

Figure 2.5: Some types of common particle shapes in particulate-filled composites (Rothon, 2003) For synthetic fillers, their shape depends on both the production conditions and the chemical composition. For example, precipitated calcium carbonate (CaCO3) can be produced in various forms including 21

aragonite, calcite or vaterite by merely changing the precipitation conditions. These precipitation conditions can be manipulated (i.e. during drying and milling) to either produce single crystals or complicated aggregates. Meanwhile, the external shape of the mineral fillers is determined by their crystal structures as well as by the environmental conditions in which the mineral was formed. If permitted to grow without restraint, then the particle will be bounded by crystal faces in a regular way which is derived from regular atomic arrangement. Nevertheless, under certain critical circumstances like under pressure, temperature or the effects of impurities, the crystal may adopt different shapes or habits such as cubic, fibrous (i.e. fine, long, needles), acicular (i.e. needle-like), lamellar (i.e. plate-like) and prismatic. Although, it is almost impossible to form perfect crystals, but even poorly formed ones will always show evidence of their intrinsic symmetry. 2.6 Particulate-filled thermoplastic composites (PFTCs) As discussed in the previous section, thermoplastic polymers tend to soften appreciably as they are heated, thus decreasing their mechanical performance. Even worse, they start to lose their shape at elevated temperatures. Therefore, for the past few years, rapid and progressive efforts have been developed to overcome this dilemma. This can be seen in the early work on reinforced thermoplastic matrices by Leong and his coworkers (Leong et al., 2004b). Surprisingly, they found that the addition of mineral fillers (i.e. talc and CaCO3) increases the modulus and crystallization temperature of unfilled thermoplastics. The incorporation of mineral fillers including kaolin, talc, calcium carbonate and mica into thermoplastic polymers has become a common 22

practice in the plastic industry. The main purpose for this kind of action is related to the cost reduction of moulded products. Apart from the cost reduction, fillers are also used to improve the mechanical properties of thermoplastics, such as strength, rigidity, hardness and durability (Katz and Milewski, 1987; Rusu et al., 2001; Chan et al., 2002). However, the optimum filler loading should be determined carefully since excessive fillers may adversely affect the ductility, processability and strength properties of composites (Premalal et al., 2002; Fu et al., 2008). Lately, PE and PP are the most popular semi-crystalline thermoplastic polymers to be used as matrices in compounding with fillers. Meanwhile, polystyrene (PS) and polycarbonates (PC) have recorded a similar popularity trend for amorphous thermoplastic polymers. Many studies have been demonstrated on particulate-filled polyethylene composites in order to fully characterise their overall performance. As pointed out by Zhao et al. (2005), modulus

of

PE/clay

composites

have

been

the strength and

found

to

increase

perpendicularly with increasing clay loading, whereas the notched impact strength shows a contrary trend. In addition, the thermal stability of PE/clay composites is far better than that of unfilled PE up to certain clay loadings. It is believed that organoclay can play two conflicting functions in the thermal stability of polymer/clay nanocomposites. At low clay loading, the clay layers become effective barriers, thus significantly increasing the thermal stability of the PE/clay composites. However, with increasing clay loading, the catalysing effect rapidly rises and becomes dominant, so that the thermal stability is decreased. The addition of metallic filler into the PE matrix is believed to increase the thermal conductivity properties of the neat PE as previously reported by Kumlutas et al. (2003). They proved that the addition of conductive particles (i.e. aluminium) into the HDPE matrix 23

gradually increased the thermal conductivity of the composites as compared to unfilled HDPE. Another common commodity plastic is PP. Progressive attention has been specifically made in order to extend its applications from conservative to more challenging applications. For this reason, several researchers have been intensively involved in works that are related with reinforced PP. For example, Svoboda et al. (2001), reported that the presence of clay filler increases the tensile modulus but decreases the elongation up to a certain extent. Typically, the pure PP and the PP/clay composites with a low clay content exhibit yielding behaviour in the stress-strain characteristic with a maximum elongation of up to 200%. Meanwhile, PP/clay composites that had been reinforced with a clay content higher than 7% did not show any yielding behaviour, where samples were immediately broken after reaching the maximum loading (i.e. stress). More recently, a study by Manchado et al. (2005) investigated the effect of different fillers (i.e. single-walled carbon nanotubes and carbon black) on both the thermal and mechanical properties of reinforced PP composites. Initially, it was found that the introduction of both reinforcements significantly increased the Young’s modulus of the composites up to a certain extent. However, the increment trend was somehow different between these two fillers. They mentioned that a further increase in single-wall carbon nanotubes proportion in the composites (i.e. 1 wt %) provided a marked decrease in the tensile modulus, whereas carbon black fillers recorded an increment pattern with increasing filler content. The difference is mainly attributed to the morphology of both fillers. For a similar interface area, carbon blacks with more isometric particles may induce a significant difference in their aspect ratio, meaning that the former are able to entangle and interconnect more easily and more often than that of the latter. Meanwhile, increasing the 24

single-walled carbon nanotubes concentration in the composites may encourage the formation of aggregates. It is assumed that the aggregates of nanotube ropes will significantly reduce the aspect ratio (length/diameter) of the reinforcement, hence reducing the rigidity of the composites. For amorphous thermoplastics, PC has received remarkable attention from scientists due to its several advantages. Lately, several researchers have proposed a newer technique to fabricate reinforced PC composites instead of the conventional melting process. One of the new approaches was introduced by Pham et al. (2008), where they impregnated the carbon nanotube sheets with PC by means of a vacuum-aided solution infiltration. The benefits of this technique are closely related with the capability in producing large sized end products which are readily scalable for mass production (Wang et al., 2007). Besides that, it can also produce thermoplastic composites with a highly loaded and well-structured carbon nanotube network (Gou et al., 2004; Wang et al., 2007). Choi et al. (2005) have successfully studied the mechanical and thermal properties of PC/carbon nanofibre composites. They reported that both the hardness and Young’s modulus of the PC/carbon nanofibre composites increased with increasing carbon nanofibre loading up to 25 wt %. This increment was contributed by the state of the filler dispersion within the PC matrix. The well dispersed carbon nanofibres within the PC matrix relatively enhanced the interfacial bonding between the filler and the matrix and therefore improved the mechanical properties. On the other hand, the thermal properties of pure PC and PC/carbon nanofibre composites were successfully determined using thermogravimetric analysis (TGA). It was reported that all the specimens started to decompose above the melting point of the PC. These decomposition features were similar at 25

around 350 oC for all the specimens and then started to shift to higher temperatures with increasing carbon nanofibre content when the decomposition temperature surpassed 430 oC. This retarding effect was attributed to the interaction between the carbon nanofibres and was likely to be a result of absorption by the carbon surface of the free radicals that existed during the polymer decomposition (Troitskii et al., 1997) 2.7 Static and dynamic mechanical testing The perspective of static, quasi-static and dynamic mechanical testing is always referred to as the strain rate applied to the specimen, as can be found in Figure 2.6. For a strain rate that is lower than 1 s -1, the mechanical properties of materials can be easily characterised using conventional mechanical testing machines, such as the universal testing machine (UTM) (Guo and Li, 2007; Fu and Wang, 2009). In contrast with static testing, the dynamic mechanical facilities require a more complicated principal in order to precisely characterise the dynamic mechanical properties of the material. Not only that, there are many additional requirements and assumptions that need to be satisfied before proceeding to the actual test. Both factors have become the main reasons why recently scientists are more focussed on the static mechanical behaviour of materials rather than on their dynamic mechanical behaviour.

26

Figure 2.6: Schematic diagram of strain rate regimes (in reciprocal seconds) and the techniques that are suitable for obtaining them (Field et al., 2004) 2.7.1 Dynamic mechanical testing For the past years, there have been various dynamic facilities available such as the drop weight impact, Taylor impact, expanding ring, plate impact, split Hopkinson pressure bar, etc. A brief explanation of their basic principles, advantages and disadvantages will be further discussed in the following section. 2.7.2 Drop weight impact Initially, the drop weight impact machine was purposefully designed to demonstrate the low to intermediate speed impact test. The fundamental of this machine was based on the dropped weight concept where the impact velocity is dependent on the earth’s gravity. For the machine design, a specimen was fixed on top of the steel base. After that, an impactor was 27

elevated at a certain specific height before being released on the specimen surface. During the collision between the specimen’s surface and the impactor, the kinetic energy of the impactor was then absorbed by the progressive failure of the specimen. The kinetic energy absorption continued until the impactor was totally stopped. The load cell was held at the space between the specimen and the steel base in order to record the generated crushing force during the collision. The crushing force results were then captured and recorded using a data acquisition system. The impact velocity of the impactor before hitting the specimen was sensed using a speed sensor. All important parameters like the crushing force data, the dropped weight mass and the impact velocity were then manipulated and used in the validation of the numerical analysis. Gunawan et al. (2011) divided their drop weight impact system into 4 subsystems, namely (1) the frame that consists of the guide columns, base plate and concrete block; (2) the impactor assembly, which consists of the impactor frame, projectile, roller, and weighting mass; (3) the clamp and hoist mechanism; and finally (4) the instrumentation. Their drop weight set-up is illustrated in Figure 2.7.

28

No 1 2 3 4

Name Guide column Steel Plate Concrete base Impactor assembly A. Impactor head B. Weighing mass C. Wheels D. Frame Clamp Hoist Speed sensor Load cell Specimen

5 6 7 8 9

Qty 2 1 1 1 4 1 set 1 1 1 1 1

Figure 2.7: Schematic diagram of the drop weight impact machine (Gunawan et al., 2011) The main advantage offered by this machine is related to a cost effective solution as compared to other dynamic facilities that use a gas gun. Lin et al. (2006) investigated the mechanical behaviour of epoxy reinforced modified montmorillonite (cloisite 30B) and titanium dioxide nanocomposites at dynamic loadings using the falling mass impact tester, and found prominent improvements in terms of impact strength, by the filler weight contents. Although a falling mass impact test can give both impressive and convenient results, but the tests are still limited by several factors, such as sensitivity towards the contact conditions between the 29

impactor and the specimen (Hsiao et al., 1999). In addition, the system is also restricted to lower strain rate loading conditions (between 1 ms -1 and 10 ms-1) since the striking velocity of the impactor totally relies on the height of the machine (Richardson and Wisheart, 1996). 2.7.3 Taylor impact The Taylor impact experiment was first developed by Taylor during late 1940s (Taylor, 1946). His main purpose in demonstrating the technique was to investigate the dynamic strength of ductile materials under compression. For the experimental design, this method propels a cylindrical projectile towards the specimen, which is normally rigid and symmetrical in shape. From the measurement of the initial velocity of the projectile, the velocity of the target and the change of shape, the dynamic behaviour of the material can be obtained using the respective equations. As an example, Taylor (1946) recorded the progression of the incident that arises during the crash between the projectile and the specimen, as can be seen in Figure 2.8.

30

Initial shape

Final shape

Figure 2.8: Sequence of deformation after impact of cylindrical projectile against rigid wall (Taylor, 1946) Although the Taylor impact can provide a higher strain rate condition than that of the drop weight impact test but somehow it is only credible to certain specific specimen geometries. As revealed by Field et al. (2004), the technique will face a difficulty in determining the dynamic deformation of a specimen in disc form. For this reason, the usage of the Taylor impact is rarely implemented and is slowly diminishing, especially for current dynamic characterisations. 2.7.4 Expanding ring Another dynamic facility that can be used to characterise the dynamic mechanical behaviour of materials is based on the expanding ring, as suggested by Hoggat and Recht (1969). The basic principle of the technique involves a hollow cylinder with an explosive core as a method of 31

initiating a shock wave as depicted in Figure 2.9 (Meyers, 1994). The velocity histories are then manipulated using a set of simple equations in order to calculate the stress/strain characteristic of the material. A variety of detonation products can be used to provide various strain rate conditions.

Figure 2.9: Principle of expanding ring technique ; (a) steel block with explosive in core ; (b) steel block with explosive in core ; (c) section of ring (Meyers, 1994) The major advantage offered by this technique is the ability to produce an extremely high strain rate test, whilst the main drawback for this technique is the difficulty in the data reduction measurement that leads to inaccuracy in the stress measurement (Hamouda and Hashmi, 1998). Daniel and LaBedz (1983) performed the expanding ring for the unidirectional 00 to 900 graphite/epoxy specimens of up to 500 s-1. They found that the technique can give reliable results where composites with 0 0 fibres showed some significant increment in modulus but no dramatic 32

changes in strength. For composites with 900, a much higher modulus and strength than the static value was exhibited. 2.7.5 Split Hopkinson pressure bar apparatus (SHPBA) Among those dynamic facilities mentioned in the above discussion, the SHPBA is the most promising and widely used for material characterisation between strain rates of 102 s-1 to 104 s-1. Typically, the SHPBA consists of three separate bars, namely a striker bar, an incident bar and a transmitter bar as shown in Figure 2.10. At the beginning stage of the SHPBA test, the specimens are clamped between the incident and the transmitter bar (refer to Figure 2.10). The striker bar is accelerated by the pressure from the air gun and then launched through the gun barrel before colliding with the incident bar. During the collision, a compressive strain pulse (ε i) is generated in the incident bar and travels towards the specimen. Due to the impedance mismatch between the bar and the specimen surface, some of the generated pulse is reflected back (ε r) to the incident bar and the remaining strain pulse (ε t) will be transmitted through the specimen into the transmitter bar. The propagation of the strain pulse along the Hopkinson bars can be well understood by the Lagrangian x-t diagram which is shown in Figure 2.11. The generated incident and the transmitted and reflected pulses are then captured by the piezoelectric strain gauges mounted on the incident and transmitter bar using a special adhesive. The output voltage of the Wheatstone circuit due to the change of resistance in the strain gauge when deformation occurred in the Hopkinson bars is then transferred to the transducer amplifier to amplify the voltages produced by the strain gauges. The signal is then captured using an oscilloscope and is saved in the computer for data processing. 33

Figure 2.10: Schematic diagram of a conventional SHPBA

Figure 2.11: The Lagrangian x-t diagram illustrating wave movement in the Hopkinson Bars A typical diagram of the SHPB output signal is illustrated in Figure 2.12. The Y-axis signifies the voltage in mV while the X-axis represents the sampling rate (in µs). The signal with the green colour represents the 34

voltage measured by the strain gauge mounted on the incident bar, where it shows the incident and reflected pulses. Conversely, the signal with the blue colour represents the voltage measured by the strain gauge attached on the transmitter bar, where it shows the transmitted pulse. The recorded voltage is then converted to a strain pulse using several specific equations as follows:

e0 = G=

Esg e

(2.1)

4 V e0

(2.2)

Substituting the Equation 2.1 into the Equation 2.2 yields:

e =

4V GEsg

(2.3)

where eo, E, sg, G, V and ε are the voltage changes in the bridge, bridge voltage, strain gauge factor, amplifier gain factor, recorded voltage (i.e. incident/reflected/transmitted)

and

strain

pulse

(i.e.

incident/reflected/transmitted), respectively. For our SHPB set-up, the E, sg, and G were fixed to 10 V, 2.08 and 1000, respectively. The conversion of the strain pulse enables the calculation of the stress, strain as well as the strain rate which will be discussed in detail under the SHPB theory subtopic.

35

Figure 2.12: Schematic diagram of SHPB output signal The main element that is attributed to the high popularity of the SHPBA is related to its capability to easily obtain the stress/strain curve as the output result, which holds useful information to characterise materials. In addition, the versatility of the test configuration has also contributed to the high usage of the SHPB, where it is available in compression (Hosur et al., 2001), tensile (Rong and Sun, 2012) and torsion (Hokka et al., 2012). However, there are various factors affecting the accuracy of the SHPB result such as specimen–bar interface friction, extension–shear coupling and the rise time of the loading pulse that need to be addressed during the SHPBA test (Ninan et al., 2001) . Based on previous studies, it can be concluded that the SHPB is suitable with almost all materials, including metals (Chen et al., 2003), ceramics (Ravichandran and Subhash, 1994), polymers (Walley and Field, 1994) as well as composites (Chan et al., 2002). Nevertheless, additional modifications and precautions should be taken into consideration during the SHPB test for soft specimens, like polymers, foam and elastomers.

36

2.7.5.1 History and development of SHPBA The initial idea in manipulating a pressure bar to obtain the characteristics of a material characteristic was first introduced by Hopkinson (1872). He initiated a procedure that allowed the transaction of dropping weight energy to a wire and started to measure the deformation of the wire before failure. In 1914, his son Bertram Hopkinson, continued his work by using a bar in order to obtain the pressures developed by the blast on impact from a bullet (Hopkinson, 1914). A schematic diagram of an orthodox SHPB device fabricated by Bertram Hopkinson is illustrated in Figure 2.13. As referred to in Figure 2.13, he used a bar (B) that was being suspended by two sets of wires. This bar was parallel to a box (D) which was also suspended. A secondary rod (C) was placed at the end of the main rod and held in place by a small magnetic force. A bullet was then shot at the end of the main rod (A). The collision induced a pressure pulse that was imparted into the main rod. The generated pulse travelled down the main rod into the secondary rod causing it to fly off and to be caught by the box. The displacement of the box and the secondary rod were measured with a basic measurement device, enabling the calculation of the momentum. The very simple measurement devices available at that time prohibited the accuracy of the result.

37

Figure 2.13: Schematic diagram of traditional SHPB device in 1914 (Hopkinson, 1914) A very limited number of works were involved with this kind of research until 1948 when Davis (1948) demonstrated a crucial study on the Hopkinson pressure bar. At that time, more accurate devices were used to precisely measure the displacement of the secondary bar (i.e. the end bar). This research was then continued by Kolsky (1949) using a three-bar system which contained the striker bar, the incident bar and the transmitted bar. He mounted the condenser units on both the incident and the transmitter bars to gain knowledge on characterising the mechanical properties of the tested specimen. It is also believed that a new promising era of the SHPBA was begun from this kind of research. Figure 2.14 shows the usage and the development of the SHPB from 1940 to 1998.

38

The histogram data was based on the published paper in any given year where an SHPB was used to determine the dynamic mechanical properties of various materials. From Figure 2.14, it is clearly seen that the usage of the SHPB apparatus for determining the dynamic mechanical properties of various materials started to become well-known in the late 1970s. Additionally, the recent development and modification of the SHPB apparatus by various experts are combined together and summarised in Table 2.2. The rapid development of the SHPB is driven by the desire to scrutinise the characteristics of various materials at dynamic conditions for crucial applications.

Figure 2.14: The usage and development of the SHPB apparatus from 1940 to 1998 (Field et al., 2004)

39

Table 2.2: The current development of the SHPB apparatus (Field et al., 2004) Year 1980 1985 1991

19911993 19922003 19972002 1998 1998 19982002 1999 2003 20032008

Development Gorham and Field develop the miniaturised direct impact Hopkinson bar Albertini develops large SHPB for testing structures and concrete Nemat-Nasser develops one pulse loading SHPBs (compression, tension and torsion) and soft recovery techniques Use of torsional SHPB for measurement of dynamic sliding friction and shearing properties of lubricants Development of polymer SHPB for testing foams Use of wave separation techniques to extend the effective length of a Hopkinson bar system Development of magnesium SHPB for soft materials Development of radiant methods for heating metallic SHPB specimens quickly Analysis of wave propagation in nonuniform viscoelastic rods performed Development of one pulse torsion SHPB Extension of Hopkinson bar capability to intermediate strain rates Application of speckle metrology to specimen deformation

References (Gorham, 1980) (Albertini et al., 1985) (Nemat-Nasser et al., 1991) (Feng and Ramesh, 1991) (Zhao, 1997);(Wang et al., 1994) (Zhao and Gary, 1997); (Bacon, 1999) (Gray III et al., 1998) (Macdougall, 1998); (Lennon and Ramesh, 1998) (Bacon, 1999) (Chichili and Ramesh, 1999) (Othman et al., 2003) (Grantham et al., 2003)

2.7.5.2 Theory behind the conventional SHPBA In the conventional SHPBA, the behaviour of materials is obtained from the difference in interface velocities (V1, V2 in Figure 2.15). As the 40

elastic pulse deforms the sample length, the distance between the incident and transmitter bars decrease since V1>V2. This deformation occurs over a period of time which enables the calculation of the strain rate using the following equation: d e s V1 - V2 = dt Ls

(2.4)

Unfortunately, it is very difficult and almost impossible to measure the velocity at the end of each bar. Thus, an alternative approach using elastic wave propagation in the incident and transmitter bars is often adopted. Theoretically, the wave velocity in the material is defined as:

Co =

E

(2.5)

r

Where, Co, E and r are the wave speed, Young’s modulus and density respectively. It is believed that the longitudinal wave propagates through the elastic media at this speed (Salisbury, 2001). In order to determine the stress, strain and strain rate history of the specimen, the incident strain εi (t), the reflected strain εr (t) and the transmitted strain εt (t) can be manipulated. The relationship between the velocities at the interface and the strain can be expressed by the following equations: V1 = C0e i At (t=0)

(2.6)

V2 = C0e t

At t>0, the incident and reflected waves are overlaid and therefore reduce the velocity of V1. As a result, V1 becomes: 41

(2.7)

V1 = Co (e i - e r )

Equations 2.6 and 2.7 will be inserted into Equation 2.4 in order to calculate the strain rate. Thus, the strain rate equation can be summarised as follows: ·

es =

Co (e i - e r - e t ) L

(2.8)

In our case, the Hopkinson bars are made from the same material and therefore the stress in the sample is obtained by using the following equation:

ss =

F1 (t ) + F2 (t ) 2 As

(2.9)

Where, F1 and F2 are the applied force (i.e. by the bar) at the specimen surface. Meanwhile, As represents the cross-sectional area of the sample. As the E=σ/ε (stress/strain), the forces in the bar can be related to the strains in the bar by: F1 = Eb Ab (e i + e r )

(2.10)

F2 = Eb Ab (e t )

For the specimen’s stress, Equation 2.10 is substituted into Equation 2.9 and then summarised as follows: 42

ss =

Eb Ab (e i + e r + e t ) 2 As

(2.11)

Where, Eb and Ab are the Young’s modulus and the cross-sectional area of the bar respectively. Generally, to achieve an equilibrium state, F1 = F2 and εi+εr=εt. Hence, the stress, strain and strain rates can be simplified and summarised into the following equations:

s s (t ) = Eb

Ab e t (t ) As

(2.12)

t

e s = -2 ·

e = s

Co e r (t )dt Ls ò0

(2.13)

C d e (t ) = -2 o e r (t ) dt Ls

(2.14)

The derivations of Equations 2.12 to 2.14 are closely related with the following assumptions and ideas (Li and Lambros, 1999): 1. The bars must remain elastic throughout the SHPB test. 2. The propagation of the wave in the Hopkinson bars was approximated by a one-dimensional theory, where the wave dispersion and attenuation were totally negligible. 3. The pulse is uniform and homogeneous over the cross section of the bar. 4. The friction and radial inertia effects were negligible and the specimen remains in equilibrium throughout the test. It is necessary to enforce the first assumption since the Hopkinson equations are mainly adopted from elastic wave equations. For the second 43

assumption, it is clearly stated that the wave dispersion and attenuation are totally negligible so that the strain measured by the strain gauges is assumed to be the same with the strain experienced at the interface. Dispersion is a result of a bar’s phase velocity dependence on the frequency, which in effect distorts the wave as it propagates (Kaiser, 1998). The third assumption indicates that the generated pulse must be uniform and homogenous over the cross section of the Hopkinson bars. This property must exist to prevent any needless pulse, such as a distortion pulse. Therefore, it is suggested that the pulse is fully developed in four (Davies, 1948) to ten (Follansbee, 1985) bar diameters from the interface. Ultimately, for the fourth assumption, it needs a detailed explanation on the specimen’s geometry consideration and invites serious discussion among the experts. Hence, a typical discussion on the highlighted issue is made in the following section.

Figure 2.15: Typical diagram of the Hopkinson Bars

2.7.5.3 Specimen’s geometry consideration The majority of SHPB experts agree that the geometry of the specimens will significantly influence the stress equilibrium within the specimen’s body (Kolsky, 1949; Davies and Hunter, 1963; Dioh et al., 44

1993). As pointed out by Davies and Hunter (1963), the equilibrium can be achieved if the back and forth pulses within the specimen is more than p times. Therefore, to achieve that condition, they suggested that the optimum slenderness ratio (length/diameter) for metals as well as polymers is 0.5. More recently, a study on the effect of the specimen’s thickness on stress equilibrium was performed by Wu and Gorham (1997). They found that a thinner specimen will be able to achieve faster uniformity of deformation but will increase the effect of friction. Their finding almost agrees with the earlier result reported by Dioh et al. (1993). Dioh and his co-workers suggest that it is crucial to choose a suitable specimen’s geometry in order to correctly determine the properties of materials, especially at dynamic loading. An optimum specimen’s geometry will reduce the effect of dispersive distortion and therefore improve the uniformity of deformation. 2.7.5.4 SHPB testing on soft materials As highlighted in the above argument, soft specimens might face difficulty in dealing with the conventional SHPB since the transmitted pulse is significantly decreased causing the signal to noise ratio to decrease. Thus, it is necessary to boost the transmitted pulse in order to meet the SHPB requirements. Many studies have been performed to overcome this dilemma. As revealed by Chen et al. (1999b), the intensity of the transmitted pulse can be increased by reducing the cross-sectional area of the transmitter bar. This action will significantly reduce the impedance value of the transmitter bar and therefore increase the sensitivity of the transmitted pulse measurement. Besides that, the method also prevents any needless effects like dispersion and attenuation problems. A year later in 2000, Chen et al. (2000) proposed another solution for increasing the signal 45

in the transmitter bar. This time, they embedded a small quartz crystal into the transmitter bar. Initially, the quartz crystal will replace the conventional strain gauges with 3 times more sensitivity than that of conventional strain gauges. The nature of quartz crystal with a similar shape and impedance also eliminates the undesired effects of placing this alternative gauge at the middle of the transmitter bar. However, problems like reflections and refractions will occur when the generated pulse reaches the quartz crystal due to the discontinuities between the crystal-bar interfaces. More recent studies by Frew et al. (2005) and Vecchio and Jiang (2007) suggested another alternative method that can be easily implemented during the SHPB test on soft materials. They believed that the usage of a pulse shaper will increase the transmitter pulse, thus inducing faster stress equilibrium within the specimen body. On the other hand, Johnson et al. (2010), proposed the idea of implementing a lower impedance bar, e.g. a polymer bar, which has an almost similar impedance value to that of the tested specimen. By doing this, it will significantly boost the propagation of the transmitted pulse. As a pre-conclusion, with minor modifications, it is convenient to say that the SHPBA is both reliable and consistent in determining the dynamic mechanical behaviour of soft specimens, especially for polymeric materials. 2.8 Static and dynamic mechanical behaviour of thermoplastic polymers and their composites Comparative studies between the static and dynamic mechanical properties of thermoplastic polymers were rarely reported in the past. However, a crucial preliminary investigation was initiated by several groups of researchers (Rietsch and Bouette, 1990; Walley and Field, 1994; 46

Nakai and Yokoyama, 2008). The most detailed and prominent finding was reported by Walley and Field (1994). They exposed the thermoplastic specimens to a wide range of investigated strain rates from 0.001 s -1 (static) to 10000 s-1 (dynamic). The stress/strain results for HDPE, PP and PC are shown in Figure 2.16. From the stress/strain curves in Figure 2.16, it is clearly indicated that PP and PC experienced pronounced load drops at all the investigated strain rates. Conversely, HDPE exhibits some strain hardening before flowing at constant stress above a strain of about 0.3. Although the results might be useful to interpret the behaviour of selected thermoplastic polymers under various loading conditions, but somehow there is a limitation in terms of their experimental set-up and parameters. For more conclusive findings, the average strain rate parameter should be constant from static to dynamic loading for all the tested specimens. By implementing this, it is easier to compare materials from one to another. Based on this limitation, it is believed that systematic works should be carried out in the future in order to gain reliable results especially for comparison purposes.

47

Figure 2.16: The compressive stress/strain curves of (A) HDPE, (B) PP and (PC) under a wide range of investigated strain rates (Walley and Field, 1994) For polymer composites, attention was highly focused on the thermoset based composites (Kusaka et al., 1998; Hosur et al., 2001; Guo and Li, 2007; Naik et al., 2010a; Naik et al., 2010b). However, the knowledge can also be manipulated to understand the mechanical 48

behaviour of thermoplastic based composites. As reported by Guo and Li (2007), they found that the yield stress as well the flow stress of the Epoxy/SiO2 nanocomposites increased dramatically with an increasing applied strain rate as depicted in Figure 2.17. A similar finding was reported by Chen et al. (1999a), where they believed that the stress increment was closely related with the secondary molecular processes. Increasing the strain rate will significantly decrease the molecular structure of the polymer chains, thus making the material stiffer. Interestingly, they also found that in some cases, the strain hardening effect has a secondary effect on the strain rate effect. This phenomenon can be clearly observed at curves (b) (at strain-rate 0.01 s−1) and overrun curve (c) (at strain-rate 0.2 s−1) at a strain of about 0.5 (refer to Figure 2.17). Based on this observation, the common sense deduction that the positive rate sensitivity makes the material stronger is inappropriate in some cases. In this case, the strain hardening effect plays a primary role where a decrease in the strainhardening effect may be related to the micro damage accumulated during loading within the composite body. At high loading rates, the mobility of the polymer chains decreases, and thus more ruptures have to take place to adapt to the deformation of the specimen. In previous studies, only a few works were discussed on the effect of particle loading on the static and dynamic mechanical behaviour of the polymer composites and therefore, the related knowledge remains unclear.

49

Figure 2.17: Typical stress–strain curves for Epoxy–SiO2 nanocomposites at room temperature (Guo and Li, 2007) 50

2.9 Strain rate sensitivity of thermoplastic polymers and their composites The strain rate sensitivity of a material is referred to as the magnitude of the changes of the properties of a material towards applied strain rates. This knowledge is essential during material selection especially for extreme applications where major change in material’s properties is prohibited. Usually, this parameter is mostly presented in terms of the proportion of the material, like Young’s modulus, yield stress (Dasari and Misra, 2003; Zebarjad and Sajjadi, 2008), flow stress (Malatynski and Klepaczko, 1980; Nakai and Yokoyama, 2008), etc.

In current practice, it is always

beneficial if these characteristics are presented as a value in order to easily compare the magnitude of the strain rate sensitivity between one material and another.

Recently, there have been several types of strain rate

sensitivity parameters available which were previously proposed (Dasari and Misra, 2003; Nakai and Yokoyama, 2008; Picu et al., 2005). However, in this case, the parameters suggested by Dasari and Misra (2003) and Nakai and Yokoyama (2008) are the most suitable parameters to be used for polymeric materials. Zebarjad and Sajjadi (2008) determined the strain rate sensitivity of HDPE/CaCO3 nanocomposites based on their yield stress value under various levels of strain rate investigated. Since the yield stress of polymer composites is thermally activated, their strain rate sensitivity can be described by using the well-known Eyring rate expression as follows: æ · 2e s y = s + KLn ç · çe è 0 0 y

ö ÷ ÷ ø

(2.15)

51

·

·

where σy, σ0y, e 0 , e and K are the yield stress at a specific temperature, the ·

yield stress at a specified temperature when the strain rate is equal to e 0 , the reference strain rate, and a measure of the strain rate dependency, respectively. A similar approach has been implemented by Sarang and Misra (2004) where they also manipulated the Eyring equations to determine the strain rate sensitivity of the Wollastonite-reinforced ethylene–propylene copolymer composites. However this strain rate sensitivity parameter is only suitable for measuring the magnitude of changes in the yield stress. A more versatile strain rate sensitivity parameter was proposed by Malatynski and Klepaczko, in the early 80’s (Malatynski and Klepaczko, 1980).

The parameter was intensively used by Nakai and Yokoyama

(2008) to measure the strain rate sensitivity of polymeric specimens (i.e. ABS, PA-6, PA-66 and PC) up to 2.5% of strain. The proposed parameter is as follows:

b=

s2 - s1 ·

·

ln(e 2 / e1 )

·

·

e 2 > e1

(2.16)

ε = 0.025

where σ1 and σ2 is the flow stress at the fixed strain (in this case, the strain can be used up to fracture point) under different strain rates. Instead of versatility, the parameter also offers the simplest way to measure the strain rate sensitivity of materials from certain specific strains up to fracture point. Based on this concern, the parameter has been comprehensively used in this study to obtain the strain rate sensitivity of tested thermoplastic polymers and their composite specimens. 52

2.10 Summary Generally, thermoplastic is subdivided into two distinct classes of molecular arrangement which are semicrystalline and amorphous. These two types of thermoplastics have been used worldwide in human daily routines. Amongst

many types

of plastics, polypropylene

(PP),

polyethylene (PE) and polycarbonate (PC) are the most widely produced and everyone comes into contact with them daily. Now days, the introduction of filler with various forms including particulate, short and long fibres into thermoplastic matrix have been intensively studied. This phenomenon was due to the evolution of applications for thermoplasticbased product, where it was extended from conservative to more challenging applications especially under extreme conditions. Due to the excellent balance between impact resistance and production cost, particulate filled thermoplastic composites have gain extra attention from both scientist and industry sectors. In order to simulate actual service environments, static and dynamic mechanical testing should be performed on thermoplastic-based products to fully characterise their overall mechanical performance. This knowledge is essential to prevent any calamity and unexpected failure during service. For static testing, the mechanical properties of materials can be easily characterised using conventional mechanical testing machine such as universal testing machine (UTM). However, it becomes more complicated with dynamic testing since it requires many additional requirements and assumptions that need to be satisfied before proceed to the actual test. This was the main reason why recently scientists are more interested on the static behaviour of material rather than their dynamic behaviour. In current 53

practise, there are several dynamic facilities available such as drop weight impact, Taylor impact, expending ring, split Hopkinson pressure bar and etc. Among those dynamic facilities, the split Hopkinson pressure bar is the most promising and widely used for material characterisation between strain rates of 102 s-1 to 104 s-1. The main factor that attributed to the high reputation of split Hopkinson pressure bar is related with its capability to easily obtain the stress/strain curve as output results which holds useful information to characterised materials. In addition, the versatility of test configuration has also contributed to the highly usage of split Hopkinson pressure bar where it’s available in compression, tensile and torsion modes. It has been reported that soft specimens like polymeric material may encounter obscurity during split Hopkinson pressure bar test. Not only that, the polymer composites also experience difficulty with split Hopkinson pressure bar test due to their anisotropic nature. However, several modifications on the conventional split Hopkinson pressure bar apparatus have been proposed by the previous researchers to overcome this dilemma. There are several previous studies were carried out on the static and dynamic mechanical properties of thermoplastic polymers and their composites. However, the information and knowledge is still remains unclear since there is limitation in terms of their experimental set-up and parameter. Based on this limitation, it is believed that systematic works should be carried out in the future in order to gain reliable results especially for comparison purposes. For thermoplastic-based polymer composites, only few works were discussed on the effect of particle loading on the static and dynamic mechanical behaviour of the polymer composites and therefore, the related knowledge remains indistinct.

54

CHAPTER 3 MATERIALS AND METHODS

3.1 Introduction In this chapter, detail explanations on the materials and methods used in this study were disclosed. All technical specifications such as the data of materials were obtained from the respective material suppliers. In addition, an overview of the overall methodology is provided at the end of the chapter. 3.2 Materials 3.2.1 Thermoplastic polymers 3.2.1.1 Polypropylene homopolymer (PP) Polypropylene homopolymer (TITANPRO PX-617) was supplied by Titan Petchem (M) Sdn. Bhd. Table 3.1 shows the properties of the polypropylene used in this study according to the manufacturer’s data. Table 3.1: The properties of polypropylene homopolymer Properties

Values

Melt flow rate, at 230oC Density Tensile strength at yield Elongation at yield Flexural modulus Notched Izod Impact strength at 23oC Heat Deflection Temperature at 4.6 kg/cm2 Rockwell Hardness Water absorption after 24 hours

1.7 g/10 min 0.9 g/cm3 34.3 MPa 9.0% 1.7 GPa 6 J/cm 102oC

ASTM Method D 1238 D 1505 D 638 D 638 D 790 B D 256 A D 648

88 R Scale 0.02%

D 785 A D 570

55

Test

3.2.1.2 Polyethylene (PE) Three types of polyethylene (PE), namely low density polyethylene (TITANLENE LDF200GG), linear low density polyethylene (TITANEX 2021) and high density polyethylene (TITANEX HI-2081) were used in this study. All types of polyethylene (PE) were supplied by Titan Petchem (M) Sdn. Bhd. Table 3.2 shows the properties of all polyethylene polymers as quoted by the respective supplier. Table 3.2: The properties of three different types of polyethylene Properties

LDPE

Melt flow rate, at 230oC (g/10 2 min) Density (g/cm3) 0.922 o 95 Vicat softening point, C o 160 – Melt temperature, C 180°C

LLDPE HDPE ASTM Test Method 2 20 D 1238 0.922 95 160 – 200°C

0.957 D 1505 124 D 1525 180 – 250°C

Molecular structures of LDPE, LLDPE and HDPE are illustrated in Figure 3.1. From the schematic diagram in Figure 3.1, LDPE appears in high branches of the molecular structures, whereas LLDPE appears in linear and typical short chain branches. In addition, HDPE shows an almost linear pattern of molecular structure with a little amount of branch points.

56

Figure 3.1: Schematic molecular structures representation of different types of polyethylene

3.2.1.3 Polycarbonate (PC) Polycarbonate (PTS Tristar PC-088) was obtained from Polymer Technology and Services, LLC. Details properties of polycarbonate used in this study are given in Table 3.3. Table 3.3: The properties of polycarbonate polymer used in this study Properties

Values

Melt flow rate, at 300oC Density Tensile strength at yield Flexural modulus Notched Izod Impact strength at 23oC Heat Deflection Temperature at 4.6 kg/cm2 Water absorption after 24 hours at 73oC

8 g/10 min 1.2 g/cm3 62.1 MPa 2.2 GPa 7.5 J/cm 102oC

ASTM Method D 1238 D 792 D 638 D 790 B D 256 A D 648

0.15%

D 570

57

Test

3.2.2 Particulate fillers 3.2.2.1 Zinc oxide (ZnO) Zinc oxide (ZnO) nanoparticles used as filler was obtained from Sigma Aldrich.

This nanoparticle was produced by using the French

process. The average particle size of the ZnO nanoparticles is 35 nm. The specifications of the ZnO nanoparticles are shown in Table 3.4. Table 3.4: The typical properties of ZnO nanoparticles Properties Form Formula Concentration Average particle sizes pH Density Purity

Description/ Value Nanoparticles ZnO 50 wt % in H2O 35 nm 7±0.1 (for aqueous systems) 1.6 - 1.8 g/cm3 99.8%

3.2.2.2 Mica Mica microparticles were purchased from Bidor Mineral Sdn. Bhd. and the properties of the particles is illustrated in Table 3.5. Table 3.5: The typical properties of Mica microparticles Properties Form Empirical formula Average particle sizes Surface area Density Molecular weight

Description/ Value Microparticles KAl3Si3O10(OH)1.8F0.2 12 µm 11.61 m2/g 2.77 - 2.88 g/cm3 398.71 g/mol 58

3.2.2.3 Silica (SiO2) Initially, the silica particles were obtained from Sigma Aldrich where the initial particle size is approximately 3 µm. In this study, the planetary ball mill (Pulverisette 6) was used to reduce the particles size into three different sizes (i.e. 1 μm, 20 nm, and 11 nm) with milling time of 30, 240 and 480 minutes, respectively. After that, the SiO2 particles were sieved using an Endecott’s siever (32 µm) before sent to the particle size analyser (Malvern 2000S Mastersizer) and transmission electron microscopy (TEM) for particle size confirmation. The properties of the SiO2 particles used in this study are shown in Table 3.6. Table 3.6: The typical properties of SiO2 particles Properties Form Formula Surface area Average particle sizes pH Bulk density

Description/ Value Microparticles SiO2 325 m2/g ± 25 m2/g ~ 3 µm 3.6 – 4.3 at 40 g/l 2.3 lb/cu.ft at 25°C

3.3 Fabrication of polymer specimens 3.3.1 Compression moulding Cylindrical specimens were fabricated using a hot compression technique. The polymer pellets (i.e. LDPE, LLDPE, HDPE, PP and PC) were compression moulded in an electrically heated hydraulic press (TS-30 Ton hydraulic hot press), using the button mould, into sizes of 12 mm in diameter and 27 mm in thickness. After that, the moulded specimens were then cut using a bench saw (CB-75F), into size of 12 mm (diameter) x 18 mm (length) and 12 mm (diameter) x 6 mm (length) for both static and 59

dynamic mechanical testings, respectively. The optimum processing temperatures used for PP, LDPE, LLDPE, HDPE and PC were 180oC, 110oC, 115oC, 130oC and 275oC, respectively. 3.4 Fabrication of polymer composite specimens 3.4.1 Compounding process The particulate fillers (i.e. ZnO and Mica) and PP granules were prepared to produce composites with different composition as shown in Table 3.7. Meanwhile, for the effect of particles sizes, 3% by weight of silica particle was concisely added into the polypropylene matrix for each type of PP/SiO2 composite. Table 3.7: The recipes used and identifications for PP/nano-ZnO composites and PP/micro-Mica composites Identifications PP (wt %) PP PP/1% ZnO PP/3% ZnO PP/5% ZnO

100 99 97 95

Zinc oxide (wt %) 0 1 3 5

Identifications PP (wt %)

Mica (wt %)

PP PP/10% Mica PP/15% Mica PP/20% Mica

0 10 15 20

100 90 85 80

During the first stage of sample preparation, each particle (i.e. ZnO, Mica and SiO2) were dried at 80oC in a vacuum oven for 24 hours, in order to remove moisture. After that, particles (i.e. ZnO, Mica and SiO2) and PP were then mixed using a heated 2-roll mill, with a mixing temperature of 180oC, for a period of approximately 20 minutes. The processing condition involved melting the PP in the heated roll-mill for 5 minutes. Well-dried 60

particles (i.e. ZnO, Mica and SiO2) were then gradually added, until all of the particles were embedded into the PP matrix. After 15 minutes, the compounded specimen was discharged from the mixing instrument, and then immediately sheeted through a laboratory mill, at a 1 mm nip setting. This was for the purpose of specimen handling during compression moulding. 3.4.2 Compression moulding The compounded composite sheets (i.e. PP/ZnO, PP/Mica and PP/SiO2 composites) were then cut into small pieces before being hot pressed. The sheeted specimens were compression moulded in an electrically heated hydraulic press (TS-30 Ton hydraulic hot press), using a button mould, into 12 mm diameter and 27 mm thick shapes. A thin film was placed in between the mould during compression process in order to get a smooth surface of the specimen. Firstly, the button mould was heated to 180oC before loading the sheeted specimens. Prior to the 5 minutes compression, pre heating was carried out for 15 minutes. After that, the mould was immediatly transfered to a cold press with a hydraulic pressure of 100 kg/cm2 for 5 minutes. The flashers at the edge of the specimens were carefully trimmed out. Finally, the moulded composite specimens were stored at ambient temperature for 24 hours. 3.4.3 Specimen cutting The moulded composite specimens were then cut using a bench saw (CB-75F), into size of 12 mm (diameter) x 18 mm (length) and 12 mm (diameter) x 6 mm (length) for both static and dynamic mechanical testings, respectively. 61

3.5 Material characterisations 3.5.1 Particle size analyser Various characterisation methods can be used to determine particle size distribution. In this study, the particle size distribution of silica microparticles was measured by wet sieving method, using a laser Granulometer (Malvern 2000S Mastersizer). In this process, 0.3 wt % of sodium hexametaphosphate [(NaPO3)6] was implemented as dispersant. The results of particle size distribution were computed by the machine, where the data was measured based on a mean diameter, D (v, 0.5).

3.5.2 High resolution transmission electron microscopy (HRTEM) A conventional particle size analyser will experience difficulty in obtaining the nanoscale of particles, due to the agglomeration problem. Therefore, morphology analysis was performed to obtain the average size of silica nanoparticles, using High Resolution Transmission Electron Microscopy (HRTEM). HRTEM was performed using a Philips TECHNAI 20 (200 kV) where sample preparation involved sonicating the nanofillers (i.e. ZnO and SiO2) in a 2% surfactant solution for at least half an hour, and then placing a few drops of the resulting suspension onto holey carbon grids. 3.5.3 Density measurement The densities of the samples were measured using a density balance (Precisa XT 220 A) according to ASTM: D792-08. Five readings were taken for each set of specimen and the equation is as follows: 62

Density ( g/cm³ ) =

r C 0.9975(Room Temperature) , r+w

(3.1)

Where r and w are apparent weight of samples in air and water, respectively. 3.5.4 Differential scanning calometry (DSC) Differential scanning calorimetry (DSC) is an established method to determine the crystallinity of the polymers (Kong and Hay, 2002; Kong and Hay, 2003). Therefore, in this experiment, the percentage of crystallinity for each polyethylene specimens (i.e. LDPE, LLDPE and HDPE) was obtained using a Mettler-Toledo DSC 1 STAR System with nitrogen gas cooling assembly. The heat flow and temperature calibrations were performed according to the instrument manual. The experiments were carried out in an inert dynamic atmosphere of high purity nitrogen set at a flow rate of 50 ml min−1. Approximately 10 mg of the polyethylene specimens were placed in a standard 40 µl aluminium crucible pan. The DSC test was run using two scans where the first scan was performed to eliminate the residual stress from the polyethylene specimens. In the first scan, the sample was heated from 30oC to 150oC at a heating rate of 10oC min−1, kept isothermally at 150oC for 3 min and cooled to 30oC at a cooling rate of 10oC min−1. Meanwhile, the second scan was carried out to determine the crystalline temperature (Tc) and melting temperature (Tm), which ranged from 30oC to 150oC at a similar heating rate of 10oC min−1. The degree of crystallinity for each polyethylene specimens were calculated using the following equation:

63

Crystallinity =

DH m C 100% DH 0

(3.2)

Where DH m and DH 0 is the melting heat of the sample and the melting heat of 100% crystalline, respectively. In this case, DH 0 for polyethylene is 270.03 J/g. The term DH 0 is a reference value and represents the heat of melting if the polymer were 100% crystalline. This reference heat of melting has been established for each of the commonly used polymers and melting heat of a 100% crystalline polyethylene, whose value is estimated at 270.03 J/g (Rusu et al., 2001; Lopes and Felisberti, 2006). 3.5.5 Dynamic mechanical analysis (DMA) Dynamic Mechanical Analysis (DMA) testing was performed using a Mettler Toledo (Model 861) under a compression configuration. This configuration was purposefully chosen to simulate a similar environment to that of the static and the dynamic mechanical tests. Cylindrical specimens of approximately 1 mm thickness were cut from the sheet stock, so that the final specimens had an approximate dimension of 12 mm (diameter) x 1 mm (thickness). The specimens were tested over the entire temperature range of the DMA instrument, from -140oC to 200oC (i.e. depends on the tested polymeric specimens), at three different frequencies of 1 Hz, 10 Hz, and 100 Hz, respectively. According to Mulliken and Boyce (2006), the test frequency is converted to a strain rate by examining one-quarter of a cycle in the sinusoidal load program. The time duration of this quarter cycle is obtained from the test frequency. Meanwhile, the strain amplitude achieved during the DMA test, can be calculated from the prescribed displacement and the known specimen gauge length. It is believed that the increment of strain along the DMA test is approximated to be linear, and 64

therefore, an average strain rate can be calculated using the following equation: æ do ö ç ÷ · strain çè lg ÷ø = e= time æ 1 1 ö ç ÷ è4wø

(3.3)

Where do, lg and w are the displacement amplitude, specimen gauge length, and test frequency, respectively. The displacement amplitude and the specimen gauge length were fixed to certain specific values of 0.2 µm and 8 mm, respectively. Based on the calculated values, the range of particular frequencies used in this study (i.e. 1 Hz, 10 Hz, and 100 Hz) were corresponded to the strain rates of 0.0001 s-1, 0.001 s-1, and 0.01 s-1, respectively. 3.6 Mechanical tests 3.6.1 Static compression testing The frictional effect is essential factor that needs to be avoided during static testing. Based on this, it was crucial to select a suitable slenderness ratio (l/d) of compression specimen, in order to minimize this undesired effect, where ASTM: E9-09 was highly recommended (Nakai and Yokoyama, 2008). Specifically, the compression specimens were cut into a slenderness ratio of approximately 1.5 (i.e. length: 18 mm/diameter: 12 mm). In current practice, the compression specimen was compressed under a constant crosshead speed of 10.8 mm/min and 108 mm/min; which corresponds to strain rates of 0.01 s-1 and 0.1 s-1, respectively; using an Axial-Torsion Universal Testing Machine (INSTRON-5982). As a 65

precaution, a thin film of lubricant was pasted onto both ends of the compression specimens to eliminate the needless effect (i.e. Frictional effect) during the test. Five measurements were taken for each loading rate, in order to quantify the average behaviour of the tested specimens. 3.6.2 Dynamic compression testing 3.6.2.1 Split Hopkinson Pressure Bar Apparatus (SHPBA) In SHPB testing, the specimen geometry is one of the vital requirements which need to be considered. Thus, the issue of specimen geometry, especially thickness, was actively debated by the previous researchers (Davies and Hunter, 1963; Dioh et al., 1993; Guo and Li, 2007). However, the majority of experts agreed that the thickness effect was not due to wave propagation effects in the specimen, but more on the radial inertia and friction considerations (Woldesenbet and Vinson, 1997; Zhao, 1998; Nakai and Yokoyama, 2008). Therefore, the tests reliability is highly related to the slenderness ratio of the specimen, rather than the specimen's thickness. In this experiment, the slenderness ratio used was based on previous studies, which was 0.5 (Davies and Hunter, 1963; Nakai and Yokoyama, 2008). Therefore, the specimen length was created, so that the l/d (d=12 mm) value was equal to 0.5. For convenient comparative study, all specimens were exposed to three dynamic average strain rates which are 650 s-1, 900 s-1 and 1100 s-1, respectively. In real practice, there is no specific international standard test method for SHPB testing since every single SHPB apparatus available in the world is slightly different from one to another. However, in this study, several critical aspects such as apparatus set-up, specimen geometry and data acquisition produce were carefully performed based on the guidelines provided by the American society for metals (ASM) under the topic of high strain rate testing (Nemat66

Nasser, 2000). Similar to the static test, five measurements were performed for each loading rates in order to determine the average dynamic mechanical behaviour of all tested specimens. 3.6.3 Static tensile testing The value of tensile modulus for all tested specimens is needed in order to measure their mechanical impedance. Therefore, static tensile testing was performed using an Axial-Torsion Universal Testing Machine (INSTRON-5982) at room temperature according to ASTM: D638. In this study, the tested specimens were cut into dumbbell shape using a dumbbell cutter. All tensile specimens (i.e. dumbbell shape) were pulled under constant crosshead speed of 50 mm/min and 50 mm gauge length. Prior to the test, the thickness of the specimens were measured by using thickness gauge. Similar to the compression testing, five measurements were conducted for each specimen and the tensile modulus values were averaged to obtain a mean value. 3.7 Post damage analysis 3.7.1 Field emission scanning electron microscopy (FESEM) Fractographic analysis of all tested specimens was performed using a Field Emission Scanning Electron Microscope (FESEM) (ZEISS SUPRA 35 VP). The specimens were prepared from cross sections of the fracture surface. The test specimens were attached to an aluminium mount with a carbon double-side and a sputter, with a platinum layer, by using a Palaron SC 515 sputter coater to eliminate the electron charging effect.

67

3.7.2 Energy dispersive X-ray spectroscopy (EDXS) Energy Dispersive X-ray Spectroscopy (EDXS) (EDAX ZEISS SUPRA-35) was used to analyse the composition of particle elements on the fracture surface.

68

B

SiO2 with various particle sizes + PP

Post damage analysis · SEM · EDX

B Mechanical tests · Static compression testing · SHPB testing

Characterisations (Particle size analyser, HRTEM, SEM)

69

Stage 1: Calibration and Verifications (Stress equilibrium, mechanical impedance, verification and calibration of SHPB results and etc...)

A

Mechanical tests · Static compression testing · SHPB testing

Characterisations (DMA)

A

Numerical solutions · Eyring model · Basic power law equation

Polyethylene (LDPE, LLDPE and HDPE)

Mechanical tests · Static compression testing · SHPB testing

Characterisations · Density · Crystallinity measurements

Stage 3: Effect of molecular structures on static and dynamic compressive properties of thermoplastic polymer

Numerical solutions · Eyring model · Basic power law equation

Thermoplastic polymers (PP, PE and PC)

Stage 2: Measurement and prediction predi on static and dynamic mechanical properties of several thermoplastic polymers

Compounding and compression moulding processes

Stage 5: Effect of particle size on static and dynamic compressive properties of polypropylene based composites

Compression moulding process

Identifications

Stage 4: Measurement on static and dynamic mechanical properties of polypropylene based composites using nano and micro fillers PP/nano-ZnO ZnO @ Mica Characterisations composites + B (DMA) PP PP/micro-Mica composites Mechanical tests Post damage analysis · Static compression · SEM testing · EDX · SHPB testing

A

3.8 Experimental chart

CHAPTER 4 CALIBRATIONS AND VERIFICATIONS OF THE SHPB RESULTS

4.1 Introduction The main challenge in the SHPB operation and research was mainly related with the difficulty in obtaining a reliable stress wave measurement. The problem has become more complicated for soft specimens where several considerations should be taken into account, especially the design of the specimen’s geometry (Field et al., 2004; Chen et al., 2002; Salisbury, 2001). In addition, it was also reported that the conventional Hopkinson bars should be modified to avoid any drastic impedance mismatch between the specimen-bar interfaces (Salisbury, 2001). Moreover, the majority of the experts agreed that the specimen dimension is critical to achieving stress equilibrium during the high strain rate testing (Dioh et al., 1993; Davies and Hunter, 1963). Due to the lack of SHPB standard procedures, most of the previous researchers proposed their own calibration styles so as to provide some guidance to people who will work with this innovative apparatus. For example, Woldesenbet and Peter (2009) calibrated their SHPB set-up by validating the results obtained from the strain gauges mounted on the Hopkinson Bars. For this reason, a specimen was fitted with a strain gauge and then tested. For an ideal SHPB set-up, the recorded strain from both measurements (i.e. the specimen and the bar) should be parallel and similar. On the other hand, Naik et al. (2008) demonstrated another calibration method that was related with the stress measurement within the Hopkinson bars. During this calibration, the two elastic bars were wrung together without a specimen sandwiched between them. With this, the incident and transmitter bars can be treated as a single bar. For an 70

excellent SHPB set-up, the stress histories of both the incident and transmitter bars should be perfectly matched. This indicates that the stress states within the incident bar and the transmitter bar are exactly the same, thus ensuring that the SHPB apparatus is perfectly aligned and friction free. Many of the previous researchers were interested in proving the homogeneity of the stress state within their tested specimens (Wu and Gorham, 1997; Nakai and Yokoyama, 2008). Therefore, a new calibration method was purposefully designed and developed. In this calibration practice, the stress history at the front and back surfaces of the specimen was measured. By showing a similar stress characteristic at both specimen ends (i.e. front and back), it was experimentally proven that the stress homogeneity within the specimen was successfully achieved during the SHPB testing. This chapter presents the calibrations and verifications of the SHPB results. This preliminarily step is essential to ensure that the SHPB used in this study has fully satisfied the requirements of the high strain rate testing before being run on the actual specimens. For that reason, several critical aspects like the mechanical impedance and the stress equilibrium were obtained. In addition, the effect of the slenderness ratio towards the stress homogeneity within the specimen was also carried out. On the other hand, the SHPB set-up was calibrated using two established calibration methods which are with and without specimens. Ultimately, the strain rate verification was also demonstrated in order to determine the average strain rate involved during the dynamic loading.

71

4.2 Mechanical impedance Generally, the mechanical impedance of a material can be defined as the ratio of force to particle velocity (Graff, 1991). In the SHPB testing, it is necessary to obtain the mechanical impedance of both the specimen and the Hopkinson bar since it determines the ratio between the magnitudes of the reflected and transmitted waves at the specimen-bar interface. Moreover, the issue has become crucial for low impedance materials (i.e. polymers, foam, soft materials, etc.) as the majority of the incident waves would be reflected at the interface. Based on this concern, we have calculated the impedance values of the Hopkinson bars and the specimens, using the following equation (Kaiser, 1998;Wehmer, 2008) : (4.1)

V = s rC0

Where s, r and Co is the cross-sectional area, density and wave velocity respectively of all the measured specimens. In addition, the wave velocity can be measured using the following equation (Wehmer, 2008)

C0 =

E

(4.2)

r

Where, E and r is the Young’s modulus and density respectively of the measured specimens. All the calculated data have been grouped together and shown in Table 4.1. For the thermoplastic specimens, PC recorded the highest mechanical impedance values followed by PP and PE. Interestingly, all the PP-based composite specimens showed a higher mechanical impedance value than that of pure PP. Moreover, the impedance values 72

were also increased steadily with increasing particle loading for both composite systems (i.e. PP/nano-ZnO and PP/micro-Mica composites). This phenomenon was mainly attributed to the introduction of rigid particles into the PP matrix where it linearly

increased the density

(Moresco et al., 2010), the Young’s modulus (Leong et al., 2004a; Fu et al., 2008) as well as the wave velocity, thus enhancing the mechanical impedance of all the composite specimens. On the other hand, the ratio of impedance (β) between the silver steel bars (Vb) and the specimen (Vs) being tested is given by:

b=

Vs x100 Vb

(4.3)

All the calculated β values are also included in Table. 4.1. From the calculated results in Table 4.1, it can be summarised that the ratio of impedance for all the tested thermoplastic polymers and their composites is in between 2.3% to 3.9%. This preliminary approach is almost similar with the one that was implemented by the previous work of (Wehmer, 2008). However, Wehmer (2008) reported that the rubber-modified syntactic foams recorded higher β values which ranged from 3.2% to 5.3%. This indicates that the specimen-bar impedance mismatch in the current study is slightly lower and better compared to the one that was reported by Wehmer (2008). Alternatively, the β values can also be manipulated in order to determine the stress equilibrium within the specimen which will be discussed in the following subtopic.

73

Note:

952 ± 1.000 911 ± 2.000 1071 ± 24.000 919 ± 1.000 926 ± 2.000 932 ± 2.000 956 ± 5.000 1018 ± 3.000 1049 ± 1.000

Density ( kg/m3) 7830 0.911 ± 0.051 1.015 ± 0.094 2.406 ± 0.088 1.025 ± 0.088 1.043 ± 0.063 1.127 ± 0.114 1.216 ± 0.092 1.365 ± 0.060 1.491 ± 0.022

978.229 ± 54.773 1055.538 ± 97.782 1498.832 ± 64.291 1056.098 ± 90.677 1061.296 ± 64.146 1099.649 ±111.258 1127.815 ± 85.532 1157.957 ± 51.013 1192.205 ± 17.628

Wave velocity (ms-1) 5227.885 105.325 ± 5.898 108.754 ± 10.077 181.549 ± 8.786 109.767 ± 9.425 111.148 ± 6.722 115.910 ±11.730 121.941 ± 9.270 133.319 ± 5.886 141.442 ± 2.096

4629. 565

Impedance (kg s-1)

2.275 2.349 3.922 2.371 2.401 2.504 2.634 2.880 3.055

Reference

Impedance ratio, β (%)

machine, respectively

74

(2) Density and Young’s modulus values of each specimen were experimentally obtained using a density analyser and UTM

values without concerning the standard deviations

(1) For the purpose of simplicity and clarity, the percentage ratios of all the tested specimens were calculated based on the mean

Composites

Composites

Polymers

Hopkinson Bars

Silver steel

PE PP PC PP/ 1 wt% ZnO PP/ 3 wt% ZnO PP/ 5 wt% ZnO PP/ 10 wt% Mica PP/ 15 wt% Mica PP/ 20 wt% Mica

Specifications

Specimens

Young’s modulus (GPa) 214

Table 4.1: The mechanical impedance characteristic of all tested specimens

4.3 Dynamic stress equilibrium For an ideal SHPB bar experiment, the sample should be in dynamic stress equilibrium and deform at a nearly constant strain rate over most of the test duration (Hamouda and Hashmi, 1998; Frew et al., 2002; Nakai and Yokoyama, 2008). Generally, by striking the end of the bar (i.e. the input bar), a compressive stress pulse is generated that immediately begins travelling towards the specimen. Upon arrival at the specimen, a certain amount of the generated stress pulse is reflected back towards the impact end, and the remainder of the stress pulse is transmitted through the specimen to the second bar (i.e. the transmitter bar). At this moment, the interface between the specimen and the transmitter bar is still unloaded. Again, due to the impedance mismatch, some of the stress pulse is reflected back to the specimen and a portion of it is transmitted to the transmitter bar. During this time of loading, the specimen is not in a state of equilibrium due to the difference in terms of the front-end stress (i.e. the specimenincident bar interface) and the back-end stress (i.e. the specimen-transmitter bar interface). However, it is believed that the stress in the specimen initially increases from zero and reaches a homogenous state, up to a certain number of back and forth loading waves within the specimen. Thus, it is always useful to calculate the amount of the wave cycles in the specimen. This critical information can be manipulated in order to precisely measure the accuracy of the recorded data. Based on this concern, we have estimated the wave cycles for the stress equilibrium state of various polymeric specimens using several specific equations as follows:

The stress increment Ds k = s k - s k -1 describes the incremental change of stress in the sample after the incident wave travels “k” times from the 75

front end to the back end of the sample. Theoretically, it would be expressed by (Yang and Shim, 2005) :

æ 1- b ö Ds k = ç ÷ è 1+ b ø

k -1

r cvo 1+ b

(4.4)

In addition, the sample stress s k can be obtained by summation while considering the impedances from (Yang and Shim, 2005) :

ìï æ 1 - b ök üï 1 s k = å Ds i = - ( ro co vo ) í1 - ç ÷ ý 2 i =1 ïî è 1 + b ø ïþ k

(4.5)

With the square wave transversing the sample “k” number of times, the stress equilibrium of the sample can be quantified by the parameter and can be defined by (Yang and Shim, 2005) :

ak =

Ds k

sk

=

2b (1 - b )k -1 1 + b k - (1 - b )k

(4.6)

By substituting the β values into Equation 4.6, we calculated the wave cycles to reach equilibrium for PP, PP/5% ZnO and PP/20% Mica, respectively, as depicted in Figure 4.1. Those specimens were chosen as the representative of each thermoplastic and their composite specimens. As can be seen in Figure 4.1, all the calculated specimens show almost similar characteristics in terms of their stress difference between both ends. Quantitatively, the PP specimen takes 29 wave cycles for the stress equilibrium difference to become less than 5% when the input pulse is a perfect square wave. Interestingly, all the calculated composite specimens recorded lower wave cycles (i.e. 28 cycles: PP/5% ZnO and 27 cycles: 76

PP/20% Mica) for a stress equilibrium difference of less than 5%. Based on this finding, it can be observed that the addition of a rigid particle into a polymer matrix slightly improves the ability of the composites specimen to achieve stress equilibrium during the SHPB testing. On the other hand, a similar characterisation method has been employed by the previous study on the polymeric foam specimens (Wehmer, 2008). Interestingly, they found that their foam specimen exhibits lower wave cycles than that of our calculated specimens where it only takes between 11 and 13 cycles for the stress equilibrium difference to become less than 5%.

Figure 4.1: The variation of stress uniformity with the number of wave cycles for different bar–specimen relative impedances

4.4 Specimen’s slenderness ratio In the SHPB test, the issue of specimen geometry, especially thickness, was actively debated by previous researchers (Davies and 77

Hunter, 1963; Dioh et al., 1993; Guo and Li, 2007). However, most of the experts agreed that the thickness of the specimen did not greatly affect the wave propagation in the specimen as much as the radial inertia and friction considerations (Woldesenbet and Vinson, 1997; Zhao, 1998; Nakai and Yokoyama, 2008). For verification purposes, we performed a series of SHPB tests on the PP/1% ZnO specimen with three different slenderness ratios (i.e. length/diameter) of 0.33, 0.50 and 1.33 respectively. The length of the PP/1% ZnO composite specimen was purposely cut into three different sizes (i.e. 16 mm, 6 mm and 4 mm), while the diameter of the specimen was fixed at 12 mm in order to achieve the required slenderness ratios. Basically, the strain gauges mounted on the incident and transmitter bars enable the stress waves to be measured. A one dimensional study of wave propagations has revealed that the following quantities can be derived from measurements of the incident pulse εi , the transmitted pulse εt , and the reflected pulse εr (Lindholm, 1964; Wu and Gorham, 1997). Therefore, the stress on the front (loaded) face of the specimen is calculated based on the following equation (Wu and Gorham, 1997):

s front = E

A (e i + e r ) As

(4.7)

Where E and A are the Young's modulus and area of the pressure bar respectively, and As is the area of the specimen. Meanwhile, the stress on the rear face of the specimen is given by (Wu and Gorham, 1997):

s back = E

A (e t ) As

(4.8) 78

Hence, the average stress is:

s average =

EA (e i + e r + e t ) 2 As

(4.9)

The oscilloscope readings and the calculated stresses (i.e. front and back) for each slenderness ratio are illustrated in Figures 4.2 and 4.3 respectively. Based on the graphs in Figures 4.2 and 4.3, it was observed that a composite specimen with a slenderness ratio of 0.5 gives a good description in terms of the stress homogeneity between the specimen’s back and front stresses than that of other slenderness ratios. Moreover, a similar finding has been reported by Nakai and Yokoyama, (2008) on several polymeric materials. Based on this concern, in our actual SHPB test, all the tested specimens were cut based on a slenderness ratio of 0.5, meaning that the diameter and the thickness of the specimens were 12 mm and 6 mm respectively. 4.5 Calibration of the SHPB set-up In the real world, there are two common calibration methods (i.e. with and without specimens) that can be manipulated to verify the accuracy and the reliability of the SHPB set-up (Naik et al., 2008;Yokoyama, 2003). In the current practice, we have implemented both calibration set-ups before the actual tests were run on the specimens. For the first calibration set-up, the two Hopkinson bars (elastic bars) were wrung together without a specimen between them. Both ends of the elastic bars were lubricated with a thin (wax) lubricant in order to prevent any frictional effects

79

C

B

80

slenderness ratios

Figure 4.2: The oscilloscope readings of the (A) 0.33 slenderness ratio, (B) 0.50 slenderness ratio, and (C) 1.33

A

C

B

σ Back 2 1

σ Front

81

slenderness ratios

Figure 4.3: Time histories of the applied stress of the (A) 0.33 slenderness ratio, (B) 0.5 slenderness ratio, and (C) 1.33

A

Therefore, both elastic bars can be treated together as a single bar, which was then impacted by the striker bar with 16.47 ms-1 striking velocity. The oscilloscope readings, captured during calibrations, are presented in Figure 4.4. Gauge 1 represents the voltage measured by the strain gauge mounted on the incident bar. On the other hand, Gauge 2 represents the voltage measured by the strain gauge mounted on the transmitter bar. It can be observed that the amplitude and the duration of both the incident and the transmitted pulse are nearly the same. Moreover, no reflected pulses were recorded during this calibration. Apart from that, the stress histories from the strain gauge signals are presented in Figure 4.5. Typically, the stress history was based on the signals captured from both gauges mounted on the Hopkinson bars, where σ

1

and σ

2

indicated the

stresses within the incident bar and the transmitter bar respectively. The stress history profile, shown in Figure 4.5, is in good agreement with the profile presented by Naik et al. (2008). Based on the recorded stress history in Figure 4.5, it is clearly seen that the stress obtained as in σ 1 and σ2 matched very well. This indicates that the stress states within the incident bar and the transmitter bar are the same.

82

Figure 4.4: Strain gauge signals on the oscilloscope during calibration

Figure 4.5: Comparison of stress vs. time characteristic, derived from strain gauge signals during calibration

83

Another way of verifying and calibrating the SHPB set-up was closely related with the stress equilibrium within the specimen. Although there have been numerous discussions on this issue but here, we have been encouraged to prove it experimentally. Hence, for this reason, the stress equilibrium of the polyethylene specimen was obtained experimentally at 12.5 ms-1 of the striking velocity, Vs. Typical oscilloscope records from the SHPB test on the polyethylene specimen is shown in Figure 4.6. In Figure 4.6, the upper trace shows the transmitted strain pulse (ε t ), whereas the lower trace shows the incident (εi) and reflected (εr) strain pulse.

Figure 4.6: Oscilloscope traces from split Hopkinson pressure bar test on polyethylene (Vs = 12.57 ms-1) Meanwhile, the resulting axial stress history at both the front and the back ends of the specimen is displayed in Figure 4.7. These specimen stress histories were calculated using both Equations 4.7 and 4.8. From Figure 84

4.7, it is clearly seen that both axial stress histories agree well with each other; specifically the assumption and requirement of the dynamic stress equilibrium across the specimen that is verified experimentally.

Figure 4.7: Time histories of compressive strength on each face of polypropylene specimen 4.6 Verification of the average strain rate Theoretically, the strain rate can be calculated by dividing the velocity of the deformation with the initial length of the specimen as depicted in the Equation 4.9.

·

e=

d e d æ l - l0 ö 1 dl v = ç = ÷= dt dt è l0 ø l0 dt l0

(4.10)

85

where v and lo is the velocity of deformation and the initial length of the specimen respectively. However, in the case of the SHPB, the dynamic strain rate measurement is not a straightforward calculation as compared to any conventional static testing. Generally, the strain rate can usually be obtained from the initial slope of the strain rate–time relation. However, in this study, the strain rate significantly varies with time and is not constant during the impact test. For example, the strain rate calculation for the polypropylene specimen at 16.4 ms-1 of striking velocity (Vs) is illustrated in Figure 4.8. From Figure 4.8, it was experimentally proven that the strain rate of the polyethylene specimen is not constant with time during the test. In this case, Nakai and Yokoyama (2008) suggested that the strain rate is calculated by dividing the area under the strain rate–strain curve, up to the maximum strain under loading. By implementing this method, the strain rate recorded during the SHPB test for polypropylene (PP) was found to be approximately 1100 s-1.

Figure 4.8: Dynamic true stress–strain and strain rate–strain curve in compression on polypropylene (PP) with 16.4 ms-1 striking velocity 86

4.7 Summary In this chapter, critical aspects like mechanical impedance, dynamic stress equilibrium and strain rate verification were successfully clarified. Moreover, the SHPB set-up was also carefully calibrated using two established calibration procedures. Based on the calibration results, it is convenient to say that the SHPB apparatus used in this study has fully satisfied the requirements of the high strain rate testing. Apart from that, it was experimentally proven that our SHPB set-up is in perfect working condition (i.e. accurately aligned and friction free) and therefore, ready for further investigations.

87

CHAPTER 5 MEASUREMENT AND PREDICITION ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF THERMOPLASTIC POLYMERS

5.1 Introduction

The demands of light material have continuously increased for the past few years. A lot of funds and works have been invested to create lighter material without scarifying its initial performances. To date, thermoplastic polymer is one of the promising materials that completely satisfied the lenient requirements with added excellent balance between impact resistance and weight. For these reasons, it has received remarkable attention from both industrial and educational sectors. Although thermoplastic polymer consists a lot of excellent abilities, their mechanical characteristics have become the primary criteria which determine the overall performances (Kuo and Jeng, 2010; Benkhenafou et al., 2011). Recently, many sophisticated techniques exist to characterise the mechanical properties of thermoplastic polymers. Nevertheless, the techniques are totally different between static and dynamic assessments. As Hamouda and Hashmi (1998) pointed out; the universal testing machine (UTM) used in static properties measurement, would not be relevance for dynamic measurement, due to its difficulties to provide high strain rate condition to the specimen. Therefore, unique dynamic facility was first introduced by Kolsky (1949), in 1949, to fulfil and satisfy the high strain rate testing requirements which is the split Hopkinson pressure bar (SHPB) apparatus. Year by year, the apparatus has experienced magnificent 88

evolution by the previous researchers (Haugou et al., 2006; Marais et al., 2004; Sasso et al., 2008), where now, it has become the standard method of measuring material dynamic mechanical properties, in the range of 102 s-1 to 104 s-1 strain rates (Field et al., 2004; Naik et al., 2010a).

In thermoplastic polymer, things like chain structures, type of branching and molecular weight might be key drivers that affect their mechanical characteristic (Tang et al., 2001; Antic et al., 2003; Magniez et al., 2010). Apart from internal issues, it is believed that external factor like strain rate effect may also cause huge impacts to the mechanical behaviour of thermoplastic polymers. Since the applications of thermoplastic polymers have been extended from conservative to various engineering applications, the strain rate factor should not be neglected and need extra attention from the researchers. For the past few years, several experimental approaches have been performed to study the strain rate effect towards the mechanical properties of polymers (Sarva et al., 2007; Gómez-del Río, 2010; Mulliken and Boyce, 2006). However, understanding on how loading rates manipulate the sensitivity of thermoplastic polymer is presently unclear and remains a major challenge in dynamic perspective. In addition, numerical studies on dynamic mechanical properties of thermoplastic polymers are also infrequently reported and need an additional effort to further clarify the relationship between experimental and numerical results, which is important for engineering design and simulation purpose. Based on the highlighted issues, the present chapter was conducted to obtain quantitative estimation of the strain rates effect on static and dynamic mechanical properties of several thermoplastic polymers. To achieve the objectives, an experimental technique, based on the 89

compression split Hopkinson pressure bar (SHPB), was introduced to perform high strain rate testing whereas, a conventional universal testing machine was used to perform static compression testing on various thermoplastic polymers. Three common thermoplastic polymers were used in this study which is polyethylene (PE), polypropylene (PP) and polycarbonate (PC). Previously, there is no specific comparison was made between PE, PP and PC polymers in terms of their mechanical characteristics under dynamic loading. Although, several studies have been reported on some of the thermoplastic polymers, the collected data are not comparable due to the different set-up of SHPB apparatus. Therefore, for comparison purpose, the stress/strain characteristics of all tested thermoplastic polymers were characterised under static, as well as dynamic loading. Two specific numerical equations were proposed and implemented to predict the yield stress and unstable strain of tested polymers. Additionally, the correlation between strain rates, with strain rate sensitivity, thermal activation volume and strain energy were also obtained to fully characterise the mechanical behaviour of the tested polymers. 5.2 Material characterisations 5.2.1 Dynamic Mechanical Analysis (DMA) The

Dynamic

Mechanical

Analysis

(DMA)

is

a

unique

characterisation technique that can give essential information regarding the thermomechanical characteristic of the material (Akay, 1993; Mazuki et al., 2011). In this study, the DMA apparatus was used in order to clarify and understand the effect of the strain rate (i.e. frequency) towards material transition over a range of selected temperature. Therefore, representative storage modulus (E’), tan delta (d) and loss modulus (E’’) taken at three different frequencies (i.e. 1 Hz, 10 Hz and 100 Hz) for PE, PP and PC are 90

plotted in Figure 5.1, Figure 5.2 and Figure 5.3, respectively. Those selected frequencies (i.e. 1 Hz, 10 Hz and 100 Hz) were corresponded to the strain rates of 0.0001 s-1, 0.001 s-1 and 0.01 s-1, respectively. Generally, it was observed that, the E’ values of all tested polymers (Figure 5.1 A, 5.2 A and 5.3 A) decrease with increasing temperature due to the increased of segmental mobility (Joseph et al., 2003). In addition, it was also found that the E’ has increase significantly with increasing strain rates over a range of tested temperature. Similar finding was also reported by the Joseph et al. (2003) where they found that E’ for PP increases steadily with increasing strain rates. This phenomenon was closely related to the segmental motion of the polymer chains during test. At higher strain rates, the polymer chains is restricted due to the insufficient time to re-oriented themselves thus enhance the rigidity of the material. On the other hand, the tan delta (d) of all tested specimens were also highly affected by the applied strain rates. Theoretically, the maximum value of d is referred to the glass transition temperature (Tg) of the specimen (Hsu and Langer, 1985). Based on the d curves in Figures 5.1 (B), 5.2 (B) and 5.3 (B), it is interesting to note that an obvious shifting of the Tg value was recorded for all tested thermoplastic specimens as a function of the applied strain rates. Statistically, the Tg of the PE, PP and PC specimens were shifted approximately 0.6 oC, 5 oC and 9 oC higher in between 0.0001 s-1 to 0.001 s-1 of strain rates, respectively. Meanwhile, for 0.001 s-1 to 0.01 s-1 of strain rates, the PE, PP and PC specimens had shifted approximately 0.4 oC, 5 oC and 9 oC to higher temperature, respectively. The dependency of glass transition temperature towards applied strain rates (i.e. frequencies) has another implication that sometimes goes unnoticed. By doing the frequency’s superposition up to the dynamic region, it will definitely revealed the different in terms of the material transition under 91

various loading conditions up to certain specific temperature. For example, at room temperature (~30oC), the viscoelastic characteristic of the PP specimen has been changed from a rubbery-like behaviour at static loading (0.0001 s-1) to a glassy-like behaviour at dynamic loading (1000 s -1) as illustrated in Figure 5.2 (D). According to Deschanel et al. (2009), the transition (i.e. from static to dynamic condition) has altered the intermolecular interaction in the amorphous domains of the polymeric material and therefore, it started to pose significant resistant towards deformation.

Apart from that, the loss modulus (E’’) also called the viscous or imaginary modulus (Menard, 1998) of all tested thermoplastic specimens is depicted in Figures 5.1 (C), 5.2 (C) and 5.3 (C), respectively. As can be seen in Figures 5.1 (C), 5.2 (C) and 5.3 (C), the E’’ values of all tested polymers increase significantly with increasing strain rate. Quantitatively, PE, PP and PC had recorded 3 MPa, 5 MPa and 310 MPa of E’’ increment for the strain rate range of 0.0001 s-1 to 0.001 s-1, respectively. Meanwhile, for the strain rate range of 0.001 s-1 to 0.01 s-1, PE, PP and PC had recorded 2 MPa, 8 MPa and 2 MPa of E’’ increment, respectively. This phenomenon was attributed to the mobility of the polymer chains since the loss factors are most sensitive to the molecular motions (Mazuki et al., 2011). Increasing strain rate will directly restricted the segmental motion of the polymer chains thus, increase the loss modulus (E’’) of all tested thermoplastic polymer specimens.

92

~ 0.4 0C shifted ~ 0.6 0C shifted

D

B

polyethylene 93

Figure 5.1: The DMA curves of (A) Storage modulus, (B) Tan delta, (C) Loss modulus, and (D) Superposition curve of

C

A

~5oC shifted ~5oC shifted

D

B

polypropylene 94


C

A

D

C

~ 9oC shifted ~ 9 C shifted o

polycarbonate

95


B

A

5.3 Stress/strain characteristic The stress-strain characteristic is one of the most important considerations to evaluate the mechanical characteristic of the materials. Therefore,

the

true

compressive

stress-strain

curves

of

various

thermoplastic polymers measured under static and dynamic loading rates (i.e. 0.01 s-1, 0.1 s-1, 650 s-1, 900 s-1 and 1100 s-1) is summarised in Figure 5.4 (A) , Figure 5.4 (B) and Figure 5.4 (C), respectively. From the graph in Figure

5.4

(A-C),

each

thermoplastic

polymer

shows

different

characteristic in terms of stress/strain behaviour where PE and PP are classified into yield behaviour classes. On the other hand, PC displays ductile-brittle behaviour of stress/strain characteristic. Apart from that, the stress/strain characteristics of all the tested specimens have been dominated by the strain rate effect where the flow stresses were increased as the strain rates increased during specific amount of strain. For example, at a fixed strain of 0.025 for PP, the difference between strain rate of 0.01 s-1 and 1100 s-1 was approximately 100 MPa. Similar finding has been reported by the previous study where the authors believed that the increment was attributed to the strengthening effect of the material towards strain rate applied (Guo and Li, 2007).

96

97

Figure 5.4: Typical true stress/strain curves of several polymeric materials at different level of strain rates: (A) PE, (B) PP and (C) PC

To further clarify the strengthening effects of all tested polymers under a wide range of strain rates, the yield stress and strain values were pointed out and illustrated in Figure 5.5 and Figure 5.6, respectively. Obviously, all tested specimens showed a positive trend in yield stress, with an increasing strain rate. In opposition, the yield strain decreased significantly with an increasing strain rate. This decrease was attributed to an increase in adiabatic temperature within the sample as an increasing strain rates (Guo and Li, 2007) where, the accumulated heat, made the material stiffer (i.e. more brittle). Another explanation of this phenomenon was reported by (Nakai and Yokoyama, 2008), where it was directly related to the accumulated micro damage during deformation. The authors 98

believed that, the higher the strain rates applied, the severe would be the damage experienced by the specimens where more ruptures have to take place in order to adapt larger deformations at higher strain rates. As a result, the combination of both effects had made all thermoplastic polymers recorded lower yield strain, with an increasing strain rate. PC had recorded the highest yield stress values followed by PP and PE for all tested strain rates. This was due to the molecular structure (i.e. side group) of the tested thermoplastic specimens. PC with stiffer molecular structure (additional bulky structure) requires higher stress/force to perform deformation. Meanwhile, PE with uncomplicated side group may easily deformable and therefore recorded the highest yield strain values as compared to other thermoplastic specimens for all tested strain rates. All the information gathered from the stress strain curves are grouped and shown in Table 5.1.

Figure 5.5: The yield stresses of tested polymers under a wide range of strain rate investigated 99

Figure 5.6: The yield strains of tested polymers under a wide range of strain rate investigated Table 5.1: The overall properties of various thermoplastic polymers under both static and dynamic loadings Polymers Strain rates (s-1) 0.01 0.1 PE 650 900 1100 0.01 0.1 PP 650 900 1100 0.01 0.1 PC 650 900 1100

Yield stress (MPa)

Yield strain

UCS (MPa)

12.73±1.63 14.55±1.73 24.67±2.05 27.33±2.17 30.00±2.38 30.00±1.65 33.64±1.82 72.00±3.01 73.81±3.87 75.60±4.03 64.55±1.92 113.64±2.05 172.80±3.35 180.00±4.05 224.00±4.15

0.055±0.0008 0.043±0.0010 0.021±0.0012 0.018±0.0015 0.017±0.0021 0.039±0.0007 0.030±0.0009 0.013±0.0011 0.011±0.0015 0.010±0.0018 0.034±0.0012 0.031±0.0016 0.020±0.0019 0.014±0.0021 0.012±0.0024

16.36±2.24 21.82±2.76 34.67±3.36 48.00±3.54 50.67±3.92 42.73±1.95 68.18±2.53 135.24±3.03 146.43±3.15 153.52±3.65 72.73±1.76 124.54±2.35 196.00±2.98 218.00±2.75 262.00±2.45

100

Compression modulus (GPa) 0.23±0.11 0.34±0.15 1.17±0.22 1.52±0.24 1.76±0.26 0.77±0.12 1.12±0.16 5.67±0.23 6.77±0.25 6.80±0.27 1.89±0.18 2.76±0.18 8.60±0.26 12.86±0.29 18.67±0.32

5.4 Stiffness and strength properties Stiffness and strength properties are the best statistical indicator to evaluate the performance of the materials. Therefore, the compression modulus and ultimate compressive strength (UCS) of all tested thermoplastic polymers were carried out and demonstrated in Figure 5.7 and Figure 5.8, respectively. From the graphs in Figure 5.7 and Figure 5.8, it can be clearly seen that the compression modulus as well as compressive strength was significantly increased with increasing strain rates. The increment in compression modulus was closely related to the secondary molecular processes where an increasing strain rates decreased the mobility of polymer chains, thus making the polymeric material stiffen (Chen et al., 1999a; Guo and Li, 2007; Gómez-del Río, 2010). On the other hand, the enhancement of the compressive strength was attributed to the shifting of the molecular relaxation for polymeric materials during impact (i.e. from static to dynamic loading), as previously discussed in DMA results. Accumulated heat during dynamic loading had also altered the shifting of the glass transition for polymeric materials. This shifting had influenced the intermolecular interaction in the amorphous region thus, begins to pose significant resistant to deformation (Deschanel et al., 2009). As a result, all thermoplastic polymers become stronger as the strain rates increase (i.e. from static to dynamic loading). Statistically, PC recorded the highest strength and stiffness properties followed by PP and PE for all tested strain rates. Theoretically, PC with stiffer molecular structure (additional bulky structure) requires higher stress/force to perform deformation and therefore recorded the highest strength and stiffness properties than that of other thermoplastic specimens.

101

Figure 5.7: The compression modulus of tested polymers under various loading rates

Figure 5.8: The ultimate compressive strength of tested polymers under various loading rates 102

5.5 Strain rate sensitivity, thermal activation volume and strain energy

Each material reacts differently under various loading condition. The magnitude of reaction can be statistically measured using several methods. For example, the sensitivity of flow stresses can be measured by using strain rate sensitivity parameter. However, strain rate sensitivity is always presented in terms of stress-strain curves, which show the effect of strain rates on the proportion of materials, such as Young’s modulus, strength, flow stress etc. In real world, it is always useful if the strain rate sensitivity can be carried out in terms of values in order to easily compare the materials. Therefore, in this study the strain rate sensitivity of the thermoplastic specimens was based on the definition of strain rate sensitivity, which can be expressed as follows (Chiou et al., 2005; Nakai and Yokoyama, 2008):

b=

s2 - s1 ·

·

ln(e 2 / e1 )

·

·

e 2 > e1

(5.1)

ε = 0.025

Where σ1 and σ2 is the flow stress at the fixed strain (in this case, the strain used is 0.025) under different strain rates. Ideally, the selection of the 0.025 as the reference strain was based on several factors. Firstly, at 0.025 of strain, the flow stresses of each specimen under various strain rates remained below the ultimate compressive stress point thus, giving useful information in terms of their stress/strain behaviour. Moreover, most of the previous literatures also employed exactly the same point of strain when calculating the strain rate sensitivity of their polymeric specimens (Nakai and Yokoyama, 2008). Therefore, for comparison purposes, we had 103

decided to calculate the strain rate sensitivity of thermoplastic specimens at a similar specific strain which is 0.025. Alternatively, there is a way to relate the strain rate sensitivity with thermal activation volume of the material as reported by the previous work (Chiou et al., 2005). Therefore, the derivation of the thermal activation volume, taking into account the strain rate sensitivity, can be expressed as follows (Chiou et al., 2005):

· é ù e ê ln( · 2 ) ú ê e1 ú = kT V * = kT ê ú ê s2 - s1 ú b ê ú êë úû

(5.2)

Where k is the Boltzmann Constant and T is the absolute temperature. All calculated strain rate sensitivities and thermal activation volumes of each sample, have been grouped and illustrated in Table 5.2. From Table 5.2, the strain rate sensitivity of all polymeric specimens increased as the range of strain rate increased. Statistically, PP had recorded the highest level of strain rate sensitivity in static and quasi-static region (i.e. static to dynamic) as compared to other thermoplastic specimens. However, at dynamic region, PC displayed the highest strain rate sensitivity compared to that of PP and PE. On the other hand, the thermal activation volume of each specimen decreased significantly with an increasing strain rate. Overall, PE has recorded the highest amount of thermal activation volume for all tested loading rates as compared to PP and PC. Theoretically, in polymeric material, thermal activation volume is referred to the free volume between polymer chains structures which influenced by the localised motions of segment or possibly side group of 104

polymer chains (White, 1981). Hence, at high strain rate where the mobility of the molecular chain is restricted, it will indirectly contribute to the lower thermal activation volume. Meanwhile, at low strain rate, the entanglement of the polymer chains has increase thus, increasing the thermal activation volume.

The impact response can be identified as ability of the material to absorb and dissipate energy during loading. Theoretically, under shock loading, impact can be classified as high energy or low energy (O'Brien et al., 1991). It is believed that, high energy impact, typically results in immediate failure. However, damage from low energy impact is hardly detected and may lead to sudden and catastrophic failure. Based on this concern, the capacity of the absorbed energy (i.e. strain energy) has become our primary considerations in the present study. Usually, energy absorption can be determined by measuring the area under the stress-strain curve up to certain specific strain. In this experiment, we calculated the strain energy of the tested thermoplastic specimens up to 0.025 of strain. The relationship between strain energy and strain rates of three different thermoplastic polymers is illustrated in Figure 5.9. From the graph in Figure 5.9, it can be seen that the strain energy increased steadily with increasing strain rates. This observation is in consistent with the work reported by the previous study (Yi et al., 2001) where the authors believed that the increment in strain energy capacity was attributed to the increase of flow stresses, initiation strain, and propagation strain as an increasing strain rates. Moreover, at dynamic loading, tested specimens absorbed more energy in order to deform since their molecular mobility was restricted. 105

Table 5.2: The strain rate sensitivity and thermal activation volume of various polymeric materials under different region of strain rates

Polymers

Range of strain rates (s-1)

b=

PP PC

0.01 to 0.1 0.1 to 650 650 to 1100 0.01 to 0.1 0.1 to 650 650 to 1100 0.01 to 0.1 0.1 to 650 650 to 1100

·

·

ln(e2 / e1 ) (MPa)

Classification ·

PE

s2 - s1

·

e2 > e1 ε = 0.025 0.621 1.653 4.087 3.105 7.647 48.869 1.889 7.180 119.846

Static Static to dynamic Dynamic Static Static to dynamic Dynamic Static Static to dynamic Dynamic

· é ù ê ln( e· 2 ) ú ê ú kT V * = kT ê e1 ú = ê s2 - s1 ú b ê 3 ú ëê (m ) ûú

·

·

e2 > e1 ε = 0.025 6.069 x 10-27 2.280 x 10-27 9.222 x 10-28 1.214 x 10-27 4.929 x 10-28 7.713 x 10-29 1.995 x 10-27 5.249 x 10-28 3.145 x 10-29

Figure 5.9: The strain energy of various polymers under a wide range of strain rate (strain, ε = 0.025)

106

5.6 The solution in the numerical equations The yield behaviour of polymers is normally considered as a thermally activated process incorporating strain-rate effect (Fu and Wang, 2009). Therefore, many numerical approaches have been performed to predict this particular behaviour. Lately, Eyring theory was widely employed to clarify the yield behaviour and to predict the yield stress of polymeric materials (Richeton et al., 2003; Richeton et al., 2005; Richeton et al., 2006; Fu and Wang, 2009). Fu and Wang (2009) proposed the simplest way of cooperative models to modify the original Eyring equation. Hence, the yield stress, σy could be defined as: 1

æ · ön 2kT -1 ç e ÷ s y = si + sinh ·* ç ÷ V èe ø

(5.3)

·*

Where σi, k, T, V, n, e are internal stress, Boltzmann constant, absolute

temperature,

activation

volume,

material

parameter

and

characteristic strain, respectively. Using the experimental data, the yield stress was modelled according to Equation 5.3. Apart from yield stress prediction, we were also interested to predict the axial strain at yield point since it has close relationship with yield behaviour of polymeric materials. According to Fu and Wang (2009), the axial strain at yield point could also be defined as unstable strain, εy. Here, using a basic power law equation, we proposed the relationship between unstable strains and strain rate of three different thermoplastic polymers as follows:

·*

m

(5.4)

ey = A - Be

107

·

·*

·

Where A, B and m are material parameters. e = ( e / e 0 ) is the ·

dimensionless strain rate and e 0

is the reference strain rate. For the

purpose of simplicity, the reference strain rate is assumed to be 1. All parameters and constants used in Equations 5.3 and 5.4 are grouped together and shown in Table 5.3. Table 5.3: The parameters used for the numerical equations

Polymers

σi, (MPa)

2kT V

(MPa)

·*

e

(s-1)

n

A

B

m

PE

12

60

1.7 x 106

3.1

0.057

0.0058

0.275

PP

30

140

1.7 x 106

3.3

0.041

0.0055

0.248

PC

63

187

1.7 x 106

8.18

0.036

0.0054

0.210

The numerical and experimental values of yield stress and unstable strain is illustrated in Figure 5.10 and Figure 5.11, respectively. As shown by the dot line in Figure 5.10, the model gives a good description of the yield stress within the range of the strain rate investigated. Quantitatively, we found that, the error of the Eyring model is approximately 3%, 2% and 11% for PE, PP and PC, respectively. The error was calculated based on Rsquared values of both numerical and experimental lines for each of tested thermoplastic polymers using Equation 5.5. Consequently, based on the calculated error, we believe that Eyring constitutive equation is both acceptable and relevance to predict the yield stress of tested polymers over a wide range of strain rates investigated.

108

R 2 (experimental) -R 2(model) R 2(experimental)

(5.5)

C 100

Figure 5.10: The relationship between experimental and numerical values of yield stress for various polymeric materials. The experimental values are from the curves in Figure 5.5

Similar to yield stress prediction, the R-squared values has been manipulated to measure the accuracy of the proposed model for unstable strain of tested thermoplastic polymers. As illustrated in Figure 5.11, the proposed model shows a good correlation with the experimental results for unstable strain under both static and dynamic loading conditions. Overall, the model exhibits error of about 3%, 3% and 2% for PE, PP and PC, respectively. Therefore, it proved that the model results almost agree well with the experimental results and can be utilised to engineering designs and for simulation purpose. 109

Figure 5.11: The relationship between experimental and numerical values of unstable strain for various polymeric materials 5.7 Summary In the present chapter, series of static and dynamic testing on the three thermoplastic polymers (i.e. PE, PP, and PC) have been successfully performed at different levels of strain rate up to nearly 1100 s -1 using the conventional universal testing machine and the Split Hopkinson Pressure Bar apparatus, respectively. From the results, it was observed that, tested thermoplastic polymers show different characteristic of stress/strain curves. Typically, PP and PE display yield behaviour classes of stress/strain curve characteristic. Meanwhile, PC shows more ductile-brittle behaviour of stress/strain curve characteristic. On the other hand, the mechanical properties of all tested thermoplastic polymers show great dependency on 110

the strain rate applied. The yield stress, compression modulus and compressive strength were proportionally increased as the strain rate increased. However, the yield strain shows a contradictory pattern where it was gradually decreased with applied strain rate. Apart from that, the strain rate sensitivity, the thermal activation volume and strain energy of the PE, PP and PC were also measured. It was found that, the strain rate sensitivity and strain energy of all thermoplastic specimens were significantly increased with increasing strain rate, whereas the thermal activation values show contrary trend. Interestingly, in this study, the yield stress and unstable strain were successfully modelled for a wide range of strain rate and the numerical results almost agree well with the experimental results.

111

CHAPTER 6 EFFECT OF MOLECULAR STRUCTURE ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF THERMOPLASTIC POLYMER

6.1 Introduction

As discussed in Chapter 5, all tested thermoplastic polymers (i.e. PE, PP and PC) show significant dependency towards applied strain rates up to certain extent. Apart from that, it was believed that, the internal structure of polymer has also play as a crucial factor that determines its overall properties. Therefore, as far as polymeric material is concerned, the relationship between structures of polymer (i.e. molecular branching, isomerism, molecular weight and polarity) and its mechanical behaviour should not be neglected and needs critical attention. For the past few years, several studies have been performed to investigate the effect of molecular structures on the overall characteristics of polymers, especially its mechanical properties (Liu and Baker, 1992; Wood-Adams et al., 2000; Wood-Adams, 2001). However, we realise that a similar kind of study under a dynamic range of strain rates has never been reported in the past and remains a major challenge in the development of better understanding on the mechanical behaviour of polymeric materials under high strain rate conditions. Based on this concern, the present chapter was carried out to investigate the effect of molecular structures on static and dynamic compression properties of thermoplastic polymer. To achieve the goals, an 112

experimental technique based on the compression split Hopkinson pressure bar (SHPB) was introduced to perform high strain rate testing whereas a conventional universal testing machine was used to perform static compression testing on different types of polyethylene. Three different types of polyethylene were used in these studies which are low density polyethylene (LDPE), low linear density polyethylene (LLDPE) and high density polyethylene (HDPE). Sample responses were then characterised and discussed in terms of their stress/strain curve, yield behaviour, stiffness, strength, rate sensitivity, thermal activation volume and strain energy, respectively. Additionally, two specific mathematical models were employed to predict the yield behaviour of all tested samplings and also to validate the accuracy of the experimental results. 6.2 Material characterisations 6.2.1 Density analysis

Density is, in fact, a crucial factor that influences the polymer abilities. Moreover, the density also has a close relationship with the degree of crystallinity where the connection of both parameters can be described by the following equation (Rosen, 1982): 1

r

=

wc

rc

+

wa

ra

=

wc

rc

+

(1 - wc )

(6.1)

ra

Where w and r are the weight fraction and the density, respectively. Meanwhile, the subscripts c and a refers to the crystalline and amorphous phases, respectively. Based on Equation 6.1, it can be summarised that the density of polymers shows a linear relationship with the degree of crystallinity. Therefore, here, we measured the density of all polyethylene 113

specimens as our initial assessment. All recorded density values were grouped together and shown in Table 6.1. From Table 6.1, HDPE recorded the highest density values followed by LLDPE and LDPE. It is believed that the differences in density values arose largely from branching that occurred during polymerisation. The branch points sterically hinder packing into crystal lattice in their immediate vicinity, and thus lowered down the density values (Rosen, 1982). As HDPE appeared almost linear with little

branch points, it would easily allow the uniformity of the

polymer chain, enhanced the crystalline formation and hence, permit the polymer to achieve high density values (Harper, 2002). Table 6.1: Measured density values of various types of polyethylene specimens Readings (g/cm3)

Types of polyethylene

1

2

3

4

5

Mean

Std.

Values

Deviations

LDPE

0.915

0.917

0.915

0.916

0.915 0.916

±0.000894

LLDPE

0.926

0.930

0.929

0.930

0.929 0.929

±0.001643

HDPE

0.954

0.953

0.952

0.950

0.952 0.952

±0.001483

6.2.2 Crystallinity measurement DSC thermograms in Figure 6.1 and Figure 6.2 show the thermal history of various polyethylene specimens. Instead of crystallinity determination, the melting temperatures (Tm) as well as crystalline temperature (Tc) also become our secondary consideration. Universal analysis software was used to calculate the crystallinity of the polyethylene specimens based on 270.03 J/g for the 100% crystalline polyethylene, in this study. All DSC results are summarised in Table 6.2. From the DSC 114

results in Table 6.2, it can be clearly seen that HDPE recorded the highest percentage of crystallinity followed by LLDPE and LDPE, respectively. In a semicrystalline polymer, the percentage of crystallinity represents the amount of the crystalline area and also the indication of their molecular structure, whether it is closely or loosely packed. With higher percentages of crystallinity recorded by HDPE, it indicated the existence of a higher amount of crystalline formation, which indirectly confirmed their closely packed structure. Based on this observation, it is interesting to note that the DSC results agreed well with the density measurement, where the higher degree of crystalline formation would linearly increase the density of the polyethylene specimens.

Figure 6.1: DSC thermograms represent melting region of various polyethylene specimens

115

Figure 6.2: DSC thermograms represent crystalline region of various polyethylene specimens

Table 6.2: DSC characterisation of polyethylene specimens

Polyethylene

Tc (oC)

Tm (oC)

DH m (J/g)

DH 0 (J/g)

Crystallinity (%)

LDPE

88.35

110.45

72.76

270.03

26.94

LLDPE

107.69

130.33

110.07

270.03

40.76

HDPE

111.55

132.82

164.69

270.03

60.99

6.3 Stress/strain characteristic Stress-strain curves are an extremely important graphical measure of a material’s mechanical properties. Interestingly, the curves will provide a preliminary overview of several aspects of the material’s mechanical 116

proportions including yield behaviour, stiffness, strength, mode of failure, etc. Thus, the true compressive stress-strain curves of polyethylene, with a different range of densities (LDPE, LLDPE and HDPE), measured under static and dynamic loadings rate (0.01 s-1, 0.1 s-1, 600 s-1, 950 s-1, and 1100 s-1) are summarised in Figure 6.3 (A-C). The graphs in Figures 6.3 (A-C) show that all types of polyethylene display similar characteristics of stress/strain curves, where they can be classified into ductile behaviour classes. On the other hand, it is clearly seen that the strain rates play a primary role in determining the stress/strain characteristic of all tested polyethylene specimens where the flow stresses increased significantly with increasing strain rates. For example, at a specific strain of 0.05, the LDPE, LLDPE and HDPE specimens recorded the difference in flow stresses approximately 37.73 MPa, 38.18 MPa and 39.38 MPa between 0.01 s-1 and 1100 s-1 of strain rates, respectively. From the same recorded data, HDPE showed the highest values of flow stresses increment than that of LLDPE and LDPE. This phenomenon was attributed to the enhancement of crystalline formation in the HDPE molecular structures (Rosen, 1982). Since the polymer chains of HDPE are more closely packed in the crystalline areas than amorphous, most of them are available per unit area to support the applied stress at a certain deformation (i.e. strain). For that reason, the easier the polymer chains forming uniformity in the crystalline area, the stronger and less deformable the polymer would be. All the information gathered from the stress/strain curves are grouped together and shown in Table 6.3.

117

118

Figure 6.3: The typical stress/strain curves for (A) LDPE, (B) LLDPE and (C) HDPE under various loading rates Table 6.3: Typical properties of PE specimens under a wide range of strain rate investigated Types Strain of PE rates (s-1)

LDPE

LLDPE

HDPE

0.01 0.1 650 900 1100 0.01 0.1 650 900 1100 0.01 0.1 650 900 1100

Yield stress (MPa) 6.40±1.84 7.73±1.92 17.72±2.14 19.54±2.26 21.36±2.42 7.60± 1.82 12.27±1.94 18.18±2.04 21.36±2.12 22.73±2.16 12.73±1.63 14.55±1.73 24.67±2.05 27.33±2.17 30.00±2.38

Yield strain

UCS (MPa)

0.061±0.0012 0.059±0.0014 0.020±0.0016 0.015±0.0018 0.013±0.0024 0.057±0.0014 0.047±0.0014 0.019±0.0016 0.015±0.0018 0.013±0.0020 0.055±0.0008 0.043±0.0010 0.021±0.0012 0.018±0.0015 0.017±0.0021

8.80±3.24 10.45±3.68 34.55±3.82 40.45±4.12 44.55±4.14 10.80±2.32 18.18±2.42 36.82±2.84 46.36±3.12 48.18±3.16 16.36±2.24 21.82±2.76 41.76±3.36 54.12±3.54 59.41±3.92

119

Compressio n modulus (GPa) 0.10±0.04 0.13±0.02 0.89±0.04 1.30±0.06 1.70±0.08 0.13±0.02 0.26±0.04 0.95±0.06 1.45±0.08 1.72±0.10 0.23±0.08 0.34±0.12 1.17±0.14 1.52±0.14 1.76±0.16

6.4 Yield behaviour In a global view, the yielding behaviour of polyethylene may be addressed as the deformation of the network set-up by the entanglements under conditions of a high internal viscosity (Hobeika et al., 2000). Normally, the yield point reacts as a statistical boundary that separates the elastic and plastic regions. It would be of interest to determine the stress and strain values at the yield point, so that we can distinguish the effect of the molecular structure towards the elastic deformation of polyethylene specimens at both static and dynamic loadings. Based on this concern, the yield stress and strain values are pointed out and illustrated in Figure 6.4 and Figure 6.5, respectively. Of the three tested specimens, HDPE and LDPE recorded the highest yield stress and yield strain values, respectively over a wide range of strain rate investigated. Theoretically, the yielding behaviour is driven by the internal friction which comes from inter-lamellar and intra-lamellar shear process. Therefore, the lamella fibrils structure of polymer is illustrated in Figure 6.6 in order to have clearer justification on the yielding phenomenon of polyethylene specimens. The shear process is induced by the stretching change of the amorphous-crystalline texture in stepwise fashion: first locally, then by rearrangement of the whole lamellar structure, followed by breaking up the lamellar and forming fibrils, and lastly by disentanglement which leads to fracture of the samples (Hobeika et al., 2000). During the compression loading, the applied force tended to squeeze and rearrange the lamellar structure. However, due to the linear structure with a little amount of branch points (HDPE), the tendency to form an aligned, bundle and dense lamellar (i.e. during lamellar rearrangement) was higher than that of LLDPE and LDPE. Therefore, a higher external force (i.e. stress) was 120

required to break up the dense lamellar structure in HDPE. For this reason, HDPE had recorded the highest yield stress values compared to LDPE and LLDPE. Meanwhile, higher yield strain values indicate that the polymer is more ductile and easily deformable. High branches of the side groups that exist in the LDPE molecular structure becomes a significant barrier towards the creation of align and dense lamellar structures. The spaces and gaps that exist in-between the LDPE inter-lamellar and intra-lamellar will turn into a liability and weaker points which will indirectly make the shear process become easier. Therefore, LDPE will experience a large deformation with only a small amount of applied force.

Figure 6.4: The yield stress values of LDPE, LLDPE and HDPE specimens under various strain rates

121

Figure 6.5: The yield strain values of LDPE, LLDPE and HDPE specimens under various strain rates

Figure 6.6: A schematic diagram illustrating a polymer crystalline spherulite 122

6.5 Stiffness and strength properties Stiffness is a term used to describe the force needed to achieve a certain deformation of a structure (Zabashta, 1974) whereas, strength is referred to the material ability to withstand load (Galeski, 2003). In many cases, both properties have been rapidly used as a basic criterion during evaluation of material performance. A similar basis was taken into consideration to evaluate the performance of the polyethylene specimens as a function of molecular structures. Based on this concern, the compression modulus and ultimate compressive strength (UCS) were carried out and demonstrated in Figure 6.7 and Figure 6.8, respectively. From the graphs in Figure 6.7 and Figure 6.8, it can be seen that the compression modulus, as well as compressive strength of PE specimens, was significantly influenced by its molecular structure. As Gupta et al. (2005) pointed out; the effect of the molecular branching on the mechanical properties is coupled to its influence on the deformation behaviour of the crystalline phase. This observation is in line with several previous studies (Liu and Baker, 1992; Kim and Park, 1996). They speculated that a branch points prevents the formation of crystalline regions in its immediate vicinity leading to formation of two different lamellae, thereby establishing that portion of chains as a bridge or a tie molecule. The existing of the bridge or a tie molecule has further worsens their strengthening ability where it encouraged slippage and disentanglement during loading (Escudero Acevedo et al., 2008). As a result, HDPE with minimal branch point exhibits higher strength and stiffness properties than that of other PE specimens under a wide range of strain rate investigated. Another explanation on strengthening mechanism of tested PE specimens was attributed to their chain’s mobility during loading. It is 123

believed that at a high strain rate, the polymer chains is restricted due to the insufficient time to re-oriented themselves thus increase the rigidity (Shergold et al., 2006). Additionally, restriction of the polymer chains mobility at a high strain rate loading may also enhance the formation of additional intermolecular force (i.e. Van Der Waals) between structures. These accumulated intermolecular forces would give a strengthening effect to the polymer as the strain rate increased. As a result, all polyethylene specimens (LDPE, LLDPE and HDPE) became stronger as the strain rates increased from static to dynamic loading. Statistically, HDPE recorded the highest stiffness and strength values than that of LLDPE and LDPE over a wide range of strain rates investigated

Figure 6.7: The compression modulus of PE specimens of LDPE, LLDPE and HDPE over a wide range of strain rates

124

Figure 6.8: The ultimate compressive strength (UCS) of LDPE, LLDPE and HDPE over a wide range of strain rates 6.6 Strain rate sensitivity, thermal activation volume and strain energy Both Equations 5.1 and 5.2 have been implemented in order to calculate the strain rate sensitivity and the thermal activation volume of all polyethylene specimens. In summary, all calculated strain rate sensitivities and thermal activation volumes of each PE specimen have been grouped together and illustrated in Table 6.4. From Table 6.4, the magnitude of the strain rate sensitivity of the polyethylene specimens (i.e. LDPE, LLDPE and HDPE) increased significantly with an increase in strain rates. Quantitatively, LLDPE shows the highest dependency towards applied strain rates under both static and dynamic regions. However, in quasi-static region (static to dynamic), LDPE recorded the highest sensitivity values compared to that of LLDPE and HDPE. For the thermal activation volume analysis, the results had shown a decreasing pattern with an increasing 125

strain rate where HDPE recorded the highest amount of the thermal activation volume under both dynamic and quasi-static regions. As discussed previously, a rapid transfer of mass or specifically the localised motions of segment or possible side group of polymer chains would influence the thermal activation volume of the polymers (Zabashta, 1974;White, 1981). At a low strain rate, the entanglement of the polymer chains were mobile enough (Karger-Kocsis and Czigány, 2000) since they have sufficient time to re-orientate themselves (Shergold et al., 2006) and therefore, resulted in higher thermal activation volume. Meanwhile, at a high strain rate, the motion of polymer chains was strongly hampered where it had indirectly contributed to the lower thermal activation volume. On the other hand, based on the results reported in Table 6.4, it can be observed that the molecular structures of PE do not have any significant relationship with the rate sensitivity as well as the thermal activation volume under a wide range of strain rates investigated. Table 6.4: The strain rate sensitivities and thermal activation volumes of various types of PE under different region of strain rates investigated

b=

Types of PE

(MPa) ·

LDPE LLDPE

HDPE

Static Static to dynamic Dynamic Static Static to dynamic Dynamic Static Static to dynamic Dynamic 126

·

·

ln(e 2 / e1 )

Range of strain Classification rates (s-1)

0.01 to 0.1 0.1 to 650 650 to 1100 0.01 to 0.1 0.1 to 650 650 to 1100 0.01 to 0.1 0.1 to 650 650 to 1100

s2 - s1

·

· é ù e ê ln( · 2 ) ú ê e1 ú = kT V * = kT ê ú ê s2 - s1 ú b ê ú ëê (m3) ûú

e 2 > e1 ε = 0.025 0.592 2.071 20.74 1.777 1.916 23.327 0.621 1.653 4.087

·

·

e 2 > e1 ε = 0.025

6.367 x 10-27 1.820 x 10-27 1.817 x 10-28 2.121 x 10-27 1.967 x 10-27 1.6158 x 10-28 6.069 x 10-27 2.280 x 10-27 9.222 x 10-28

The relationship between strain energy and strain rates for LDPE, LLDPE and HDPE specimens is illustrated in Figure 6.9. From the graph in Figure 6.9, it can be seen that the polyethylene specimens show a similar pattern of strain energy behaviour. Typically, the capacity of strain energy for all PE specimens had increased greatly with the increasing strain rates. It indicated that, at a higher strain rate, more energy was required by the polyethylene specimens in order to deform, since the mobility of the polymer chains was restricted. For molecular structures point of view, HDPE with dense structure will require more energy to deform as compared to LLDPE and LDPE with many weaker points (i.e. bridge or tie molecules). As a result, HDPE recorded the highest capacity of energy absorption, followed by LLDPE and LDPE for all tested strain rates.

Figure 6.9: The strain energy of LDPE, LLDPE and HDPE specimens under static and dynamic loadings

127

6.7 The solution in the numerical equations Numerical solutions are other basic principles that need to be implemented in order to validate the experimental results. In this study, we were interested in predicting the yield behaviour and the unstable strain of the polyethylene specimens under a wide range of strain rate investigated. Using both Equations 5.3 and 5.4, the numerical and experimental values of yield stress and unstable strain for all PE specimens are illustrated in Figure 6.10 and Figure 6.11, respectively. All parameters and constants used in Equations 5.3 and 5.4 are grouped together and shown in Table 6.5. Table 6.5: The parameters used for the numerical equations σi, Polymers

(MPa)

2kT V

(MPa)

·*

e

(s-1)

n

A

B

m

LDPE

7

40

1.7 x 106

3.1

0.057

0.0058

0.275

LLDPE

8

50

1.7 x 106

3.1

0.057

0.0058

0.275

HDPE

12

60

1.7 x 106

3.1

0.057

0.0058

0.275

Interestingly, the line graph in Figure 6.10 proves that the Eyring model gives a fine description of the yield stresses within the investigated range. Quantitatively, the calculated errors of the Eyring equation were approximately 0.9%, 9.6% and 2% for LDPE, LLDPE and HDPE, respectively. Basically, the errors were calculated based on R-squared values of both numerical and experimental lines using Equation 5.5. Based on the calculated errors, we believe that the Eyring constitutive model is acceptable to predict the yield behaviour of the polyethylene specimens under a wide range of strain rates investigated. Meanwhile, for prediction of the unstable strain, a similar approach had been implemented to justify 128

the accuracy of the proposed model as illustrated in Figure 6.11. From the graph in Figure 6.11, it can be clearly seen that the proposed model also exhibited a good correlation between numerical and experimental results where the errors are approximately 1.2%, 0.2% and 3.3% for LDPE, LLDPE and HDPE, respectively. For comparison purposes, a previous constitutive model proposed by Fu et al. (2009) has been included in the graph. However, due to the different configuration of the testing (i.e. tension), it seemed like the model did not give a good interpretation of the results between numerical and experimental approaches.

Figure 6.10: The relationship between experimental and numerical values of yield stress for PE specimens. The experimental values (yield stress) are from the curves in Figure 6.4

129

Figure 6.11: The relationship between experimental and numerical values of unstable strain for PE specimens

6.8Summary In this chapter, the effect of molecular structure on the static and dynamic mechanical properties of PE specimens was successfully carried out using a conventional universal testing machine and a split Hopkinson pressure bar apparatus, up to nearly 1100 s-1 of strain rates, respectively. Results show that, the density and crystallinity of the polyethylene specimens (i.e. LDPE, LLDPE and HDPE) were significantly influenced by their molecular structures. In addition, all polyethylene specimens (i.e. LDPE, LLDPE and HDPE) showed almost similar characteristics of stress/strain curves which can be categorised into ductile behaviour with HDPE exhibits greater compressive properties in terms of yield stress, 130

stiffness and strength properties than that of a complex molecular structure (i.e. LLDPE and LDPE). On the other hand, yield stress, compression modulus, compressive strength, rate sensitivity as well as strain energy of all polyethylene specimens were proportionately increased as the strain rate increased. However, yield strain and thermal activation volume showed a contradictory pattern where it gradually decreased with applied strain rate. Of the three PE specimens, LLDPE recorded the highest strain rate sensitivity values under both static and dynamic regions of strain rates. Meanwhile, for quasi-static region (i.e. static to dynamic), LDPE had shown the highest value of strain rate sensitivity compared to LLDPE and HDPE. Interestingly, agreement between experimental and prediction values offered by both Eyring and power basic equations were found to be good over wide range of strain rate investigated.

131

CHAPTER 7 MEASUREMENT ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF POLYPROPYLENE BASED COMPOSITES USING NANO AND MICRO FILLERS

7.1 Introduction

For the past few decades, organic-inorganic composites have attracted remarkable interest from scientist, as well as industries, due to its superior properties, which include morphology, thermal, and mechanical properties. Furthermore, the use of mineral fillers, such as silica, muscovite, and zinc oxide, in a polymer matrix, significantly reduces production costs and weight; which are the two main elements of concern, in engineering applications. Polymer micro and nano-composites, based on polypropylene (PP), have initiated many alternatives and advanced composites, particularly for engineering applications, due their excellent balance between impact resistance, production costs, and weight. With their rapid development, there is a growing interest in increasing the effectiveness of compatibility between the organic mineral filler and the inorganic PP matrix; thus giving more prominent properties. For example, Garcıa-López et al. (2003) suggested that the organic compound can be used as modified filler, whilst functional compounds can be manipulated as compatibilizers. Amongst the many types of available nano fillers, zinc oxide (ZnO) particles have become important filler for reinforcing the PP matrix. This is largely due to its remarkable electrical properties, compared to that of other conventional fillers of a similar nature (Cheng et al., 2007). The introduction of ZnO particles into the PP matrix was also reported to 132

give a significant improvement in terms of the mechanical properties of the PP based composites; including their wear resistance, tensile strength, impact strength, and crystallization, along with added antimicrobial characteristics (Sanpo, 2008). On the other hand, another important silicate group which is Mica has attracted much less attention. Generally, Mica is an abundant, naturally occurring mineral with high strength, good thermal stability, good corrosion and electrical resistance. Therefore, Mica reinforced PP composites have a number of excellent characteristics, including improved strength, high stiffness, good dimensional stability, enhanced heat resistance and electrical resistivity and reduced cost (Rashid et al., 2008; Rashid et al., 2011).

For many years, a lot of works have been conducted on the mechanical behaviour of the polypropylene reinforced Zinc oxide (PP/ZnO) (Zaman et al., 2012; Altan and Yildirim, 2012) composites as well as the polypropylene reinforced Mica (PP/Mica) composites (Trotignon et al., 1982; Rashid et al., 2008). Unfortunately, none of them focused on their dynamic mechanical behaviour due to the several difficulties and limitations. As pointed out by Hamouda and Hashmi (1998), composite materials may experience difficulty with the dynamic testing, due to their anisotropic nature, which can complicate the design of the specimen’s geometry. In addition, the limited number of dynamic facilities also leads to the lack of information in the area of the dynamic properties for composite material. Therefore, until recently, the response of the particulate filled polymer composites towards applied strain rates up to the dynamic region, were unclear, and consequently, extra attention is required.

133

Based on these limitations, this chapter was purposefully designed to measure the static and dynamic compressive properties of two different systems of polypropylene based composites which are PP/nano-ZnO composites and PP/micro-Mica composites. An experimental technique, based on the compression Split Hopkinson Pressure Bar (SHPB), was introduced to perform high strain rate testing; where a conventional universal testing machine was used to perform static compression testing on all PP/nano-ZnO composites and PP/micro-Mica composites. In this study, both PP/nano-ZnO and PP/micro-Mica composites were prepared by a hot compression technique, with nano-ZnO and micro-Mica content of 1, 3, 5% and 10, 15 20% by weight, respectively. Both composite responses were then characterised in terms of their stress/strain curves, yield strength, ultimate strength, and stiffness. Furthermore, a correlation between the applied strain rate and filler content with rate sensitivity, thermal activation volume, and strain energy up to a certain deformation (i.e. 0.025 of strain), was also made. Ultimately, post damage analysis was performed on fracture surface of both composite systems to further identify the failure mechanism experienced by the specimens under loading, especially under dynamic loading.

POLYPROPYLENE/NANO-ZINC OXIDE COMPOSITES 7.2 Material characterisations 7.2.1 Dynamic Mechanical Analysis (DMA)

Similar with the neat thermoplastic polymer, the DMA apparatus was used in order to clarify and understand the effect of strain rate (i.e. frequency) towards composite material transition over a range of selected 134

temperatures (i.e.

-40oC to 140oC). Therefore, representative storage

modulus (E’), tan delta (d) and loss modulus (E’’) was taken at three different strain rates (i.e. 0.0001 s-1, 0.001 s-1, and 0.01 s-1) for PP/1% ZnO, PP/3% ZnO and PP/5% ZnO composites which are plotted in Figure 7.1 (A-B), Figure 7.2 (A-B) and Figure 7.3 (A-B) , respectively . The figures clearly show that the E’ of all PP/nano-ZnO specimens (i.e. Figures 7.1 A, 7.2 A and 7.3 A) decreased upon increasing temperature, due to the increased segmental mobility. Theoretically, in a semi-crystalline material, like polypropylene based composites, only the amorphous part undergoes segmental motion, whilst the crystalline region remains as a solid, until reaching its melting temperature (Tm) (Joseph et al., 2003). Furthermore, it was found that the E’ had increased significantly with the increasing strain rate. For example, at room temperature, the PP/5% ZnO composite statistically recorded approximately 10% and 15% of E’ increment for strain rates in the range of 0.0001 s-1- 0.001 s-1and 0.001 s-1- 0.01 s-1, respectively. Similar findings were also reported by Joseph et al. (2003), where they found that E’ for pure PP increases steadily with an increasing strain rate (i.e. frequency). This phenomenon is closely related to the segmental motion of polymer chains during testing.

At higher strain rates, the polymer chains is

restricted due to the insufficient time to re-oriented themselves thus enhance the rigidity of the composite material. Apart from that, the tan delta (d) value was also highly affected by the applied strain rate. In general, the maximum value of d refers to the glass transition temperature (Tg) of the specimen (Hsu and Langer, 1985). Based on the d curves shown in Figure 7.1 (A), 7.2 (A) and 7.3 (A), it is 135

interesting to note that an obvious shifting of the Tg value was recorded for all PP/nano- ZnO composites as a function of the applied strain rates. For example, the Tg of the PP/5% ZnO composite shifted approximately 3oC higher between 0.0001 s-1- 0.001 s-1 strain rates. Meanwhile, for 0.001 s-10.01 s-1 strain rates, the specimen shifted approximately 6 oC higher than the initial Tg value. Performing the frequency’s superposition up to the dynamic region, reveals a difference in terms of material transition under various loading conditions, up to a specific temperature. Typically, at room temperature (~30oC), the viscoelastic characteristic of all PP/5% ZnO composite changed from a rubbery-like behaviour at a static loading (0.0001 s-1) to a glassy-like behaviour at a dynamic loading (1000 s-1) as illustrated in Figures 7.1 (C), 7.2 (C) and 7.3 (C), respectively. According to Deschanel et al. (2009), this transition (i.e. from static to dynamic condition) alters the intermolecular interaction in the amorphous domains of the composite, and therefore, it starts to pose a significant resistance towards deformation. On the other hand, the loss modulus (E’’) of all tested PP/5% ZnO composite is illustrated in Figures 7.1 (B), 7.2 (B) and 7.3 (B), respectively. The figures reveal that the E’’ values of all tested polymers increase significantly with increasing strain rate. Quantitatively, PP/1% ZnO, PP/3% ZnO and PP/5% ZnO composites had recorded 7 MPa, 6 MPa and 17 MPa of E’’ increment for the strain rate range of 0.0001 s-1 to 0.001 s-1, respectively. Meanwhile, for the strain rate range of 0.001 s -1 to 0.01 s-1, PP/1% ZnO, PP/3% ZnO and PP/5% ZnO composites had recorded 14 MPa, 13 MPa and 18 MPa of E’’ increment, respectively. This phenomenon was closely related to the mobility of the polymer chains since the loss factors are most sensitive to the molecular motions (Mazuki 136

et al., 2011). Increasing strain rate will directly restricted the segmental motion of the polymer chains thus, increase the loss modulus (E’’) of all tested PP/nano- ZnO composites.

137

~ 5oC shifted ~ 5oC shifted

C

B

PP/1% ZnO composite 138

Figure 7.1: The DMA curves of (A) Storage modulus and tan delta, (B) Loss modulus, and (C) Superposition curve of

A

~3oC Shifted ~4oC Shifted

C

B

PP/3% ZnO composite

139


A

PP/5% ZnO composite 140


C

~6oC shifted ~3oC shifted

7.3 Stress/strain characteristic The true compressive stress/strain curves of the pure PP and the PP/nano-ZnO composite, as a function of filler contents measured under various loading rates (i.e. 0.01 s-1, 0.1 s-1, 650 s-1, 900 s-1 and 1100 s-1), are summarised in Figures 7.4 (A-D). The initial observation clearly shows that the pure PP reacts differently to the PP/nano-ZnO composite, in terms of their stress/strain curve characteristics. Visually, the pure PP shows a basic compression stress/strain characteristic for a semi-crystalline polymer, without showing any peculiar descriptions, like constant flow stress, strain hardening, and strain softening behaviours (Haward, 1993; Kontou and Farasoglou, 1998). Meanwhile, all PP/nano-ZnO composite exhibited more complicated stress/strain curve characteristics, particularly under dynamic loadings (i.e. 650 s-1, 900 s-1, and 1100 s-1). This trend is similar to work reported by many glassy polymers (Stachurski, 1997; Mulliken and Boyce, 2006). Specifically, the stress/strain curves of these composites can be divided into five stages, which are initial linear elasticity (A), non-linear transition to global yield (B), strain softening (C), flow at constant stress (D), and strain hardening (E); as illustrated in Figure 7.5 (Stachurski, 1997; Mulliken and Boyce, 2006). This phenomenon was presumably due to the toughening mechanism introduced by the ZnO particles, when reinforcing the PP matrix. Theoretically, the toughening mechanism of semi-crystalline polymers, with the addition of mineral fillers, can be described using a micro-mechanistic model (Zuiderduin et al., 2003; Eiras and Pessan, 2009). This is very similar to the cavitations mechanism in rubber toughened systems. Until recently, the toughening mechanism of polymers reinforced with rigid particles under compression loading, has not been reported. 141

Therefore, assuming that a rigid particle behaves in a similar manner with tension loading, we have proposed a similar mechanism with different loading modes, as shown in Figure 7.6. The micro-mechanistic model consists of the following three stages, which are: 1. Stress Concentration - The modifier particles act as stress concentrators, due to the different elastic properties between filler and the polymer matrix. 2. Debonding - Stress concentration gives rise to a build-up of triaxial stress around the filler particles, which leads to debonding at the particle/polymer interface. 3. Shear Yielding - The voids caused by debonding alter the stress state of the polymer matrix. This also reduces the sensitivity towards crazing, as the volume strain is released. The shear yielding mechanism becomes operative and the material is able to absorb large quantities of energy upon fracture. Based on these concerns, it can be speculated that the introduction of ZnO particles into the PP matrix slightly altered the toughening and hardening mechanisms of the pure PP where it was successfully proven by the specimen’s stress-strain curves, shown in Figures 7.4 (A-D).

142

143

Figure 7.4: True compression stress/strain curves of the (A) PP neat, (B) PP/1% ZnO, (C) PP/3% ZnO and (D) PP/5% ZnO composites under a wide range of strain rate investigated

144

Figure 7.5: The typical features of stress/strain curves of PP/nano-ZnO composites under dynamic loading

Figure 7.6: Toughening mechanism with rigid particle under compression loading 145

7.4 Stiffness properties The compression modulus of all PP/nano-ZnO composites was obtained under various levels of strain rates, as shown in Figure 7.7. For comparison purposes, the stiffness information of pure PP, under similar strain rates, was taken from our previous results in Figure 5.7. The bar chart in Figure 7.7 shows that the modulus values for all tested specimens was raised gradually with increasing strain. For example, the PP/5% Zn0 composite recorded a relative increment of about 5.5 GPa and 3.9 GPa between 0.001 s-1 to 650 s-1 and 650 s-1 to 1100 s-1 strain rates, respectively. Several previous works have reported on this issue (Chen et al., 1999; Guo and Li, 2007). According to Chen et al. (1999a) and Guo and Li (2007), the increment in compression modulus is directly related to the secondary molecular process; where an increasing strain rate will restrict the molecular mobility of the polymer chains, and thus, make the material stiffer.

However, particle content has become an additional criterion that influences the compression properties of PP/nano-ZnO composites, which can be observed from the trend shown in Figure 7.7. Statistically, at a strain rate of 0.01 s−1 and 1100 s−1, the compression modulus of pure PP is 0.77 GPa and 6.80 GPa. Meanwhile, at 5% ZnO content, the compression modulus was found to increase relatively to 0.85 GPa and 10.31 GPa. Several previous works have reported this issue (Leong et al., 2004a; Fu et al., 2008). Based on the work reported by Leong et al. (2004a), the increment in compression modulus was attributed to the inherent rigidity provided by the rigid particles (i.e. ZnO particle), and not that of the PP matrix. As a result, the higher the amount of ZnO (i.e. rigid particle) the 146

greater will be the modulus. Furthermore, it was also reported that the increase in rigidity is associated with a lack of chain mobility and deformability of the matrix. This is due to the presence of additional rigid particles (ZnO), which disturb the deformation of the crystalline region in the PP matrix (Fu et al., 2008). Both reasons can be used to explain the increment of the compression modulus under various loading rates.

Figure 7.7: The compression modulus of the pure PP and the PP/nano-ZnO composites under various levels of strain rates investigated


Further analysis was carried out on the strength properties of the tested composites. The yield strength and Ultimate Compressive Strength (UCS) of the PP/nano-ZnO composites were obtained under a wide range 147

of strain rates, as shown in Figure 7.8. Similar to the stiffness properties evaluation, the strength value under similar strain rates for pure PP was taken from our previous results in both Figure 5.5 and Figure 5.8, for comparison purposes. Figure 7.8 results clearly show that all specimens exhibited a positive increment, in terms of yield strength and ultimate strength, under all tested strain rates. For example, the PP/3% ZnO composite recorded approximately 13 MPa and 25 MPa increments of yield strength and the ultimate strength, under the static region (0.01 s-1 - 0.1 s-1), and 13 MPa and 33 MPa increments, under the dynamic region (650 s -1 1100 s-1). These increments were closely related to the viscoelastic properties (i.e. material transition) of the composite during loading, as previously discussed in the DMA results (refer to Figures 7.1 - 7.3). An increasing strain rate will significantly alter the intermolecular interaction within the amorphous domains of the PP matrix, and will therefore, enhance the strength properties of all PP/nano-ZnO composites (Deschanel et al., 2009).

In addition to strain rate, particle content also plays a secondary role that determines the compression properties of the tested specimens. Generally, all PP/nano-ZnO composites showed greater strength properties than that of pure PP, as illustrated in Figure 7.8. In polymer matrix composites, it is believed that the increment of strength values occurs up to a certain optimal filler content, before it starts to drop; due to a weaker particle dispersion and the agglomeration problem (Fu et al., 2008). In this study, the strength values of PP/nano-ZnO composites were increased gradually, up to 5% of the filler content, for each applied strain rate. A similar trend was also reported by Zhao et al. (2006) and Zaman et al. (2012), where they found an increment in yield strength up to 5% of the 148

filler content, with a similar composite system. Typically, the strength properties of the polymer composites were highly related to the state of the particle’s dispersion within the polymer matrix (Fu et al., 2008; Rashid et al., 2008). Based on this, we performed a morphology analysis on each of the PP/nano-ZnO composites, using the SEM equipment. All SEM micrographs, representing each of the PP/nano-ZnO composites, were collected together and are illustrated in Figure 7.10. These micrographs show that the white spots representing ZnO particles covered almost everywhere on the PP surface. This indicates that the state of ZnO nanoparticles distribution within PP matrix for all PP/nano-ZnO composites was in a good condition, thus resulting in remarkable increments of both yield and ultimate strength properties (i.e. up to 5% particle loading).

Figure 7.8: The yield strength and the ultimate strength values of the pure PP and the PP/nano-ZnO composites under a wide range of strain rate investigated 149

7.6 Rate sensitivity, thermal activation volume and strain energy Both Equations 5.1 and 5.2 have been implemented in order to calculate the strain rate sensitivity and the thermal activation volume of all PP/nano-ZnO composites specimens. In summary, all calculated strain rate sensitivities and thermal activation volumes of each PP/nano-ZnO composites specimen have been grouped together and illustrated in Table 7.1. Again, for comparison purpose, the rate sensitivities and the thermal activation volume of the pure PP were taken from previous results in Table 5.2. Table 7.1: The rate sensitivity and the thermal activation volume of the neat PP and the PP/nano-ZnO composites under static and dynamic regions

Range of strain Specimens rates (s-1)

b=

PP/1% Zn0 PP/3% Zn0 PP/5% Zn0

0.01 to 0.1 650 to 1100 0.01 to 0.1 650 to 1100 0.01 to 0.1 650 to 1100 0.01 to 0.1 650 to 1100

Regions

Static Dynamic Static Dynamic Static Dynamic Static Dynamic

·

·

ln(e 2 / e1 )

(MPa) ·

PP neat

s2 - s1

·

e 2 > e1 ε = 0.025 3.1050 48.8690 2.8952 45.9364 2.7360 46.6564 2.6725 48.2514

· é ù e ê ln( · 2 ) ú ê e1 ú = kT V * = kT ê ú ê s2 - s1 ú b ê ú (m3) ûú ëê

·

·

e 2 > e1 ε = 0.025

1.2140 x 10-27 7.7130 x 10-29 1.3019 x 10-27 8.2052 x 10-29 1.3776 x 10-27 8.0786 x 10-29 1.4104 x 10-27 7.8115 x 10-29

From the results shown in Table 7.1, it can be seen that all tested specimens showed positive increments, in terms of their rate sensitivity of flow stress, with increasing strain rates (i.e. from static to dynamic regions). However, thermal activation volume shows a contrary pattern; 150

where it linearly decreases with increasing strain rate. Even though there has been little discussion on this issue, we believe that this phenomenon attributes to the mobility of polymer chains during applied strain rates. At a high strain rate, a higher flow stress is typically required to perform deformation, as the mobility of the polymer chain is restricted. Therefore, increasing flow stress will directly increase rate sensitivity. However, the rapid transfer of mass along the polymer chain was also affected by the restriction of the chain’s mobility at a high strain rate, thus decreasing the thermal activation volume. Furthermore, it can be seen that the particle loading does not give any significant trend on the rate sensitivity and thermal activation volume of the PP/nano-ZnO composites. This typically shows that PP/nano-ZnO composites recorded lower rate sensitivities than that of pure PP. However, the variation is not consistent with particle content and the applied strain rate.

In this experiment, the strain energy of pure PP and PP/nano-ZnO composites was calculated up to 0.025 of strain, as shown in Figure 7.9. The line graph in Figure 7.9 shows that the strain energy rose gradually with increasing strain rate. Similar findings were reported by Yi et al. (2001). The authors agreed that this phenomenon attributed to the increase of flow stresses, initiation strain, and propagation strain, as increasing strain rates. Furthermore, it is noticeable that filler content also gave a positive improvement, in terms of the strain energy value for each applied strain rate. Generally, the filler will act as an important energy absorbance during loading, and to certain extent, the higher the filler contents, the better will be the energy absorbing efficiency.

151

Figure 7.9: The strain energy of the pure PP and the PP/nano-ZnO composites under various loading rates

7.7 Fracture surface analysis Fracture surface analysis was carried out using SEM equipment. EDX analysis was also performed, in order to identify the presence of ZnO particles on the fracture surface of all PP/nano-ZnO composites. Figure 7.10 shows the SEM micrograph of the dynamic compression fracture at 16.47 ms-1 striking velocity. From fractographic analysis, it was observed that a coarser appearance, with a massive plastic deformation and catastrophic fracture, occurred on the specimen’s surface, due to the enhancement of the applied stress during a high strain rate loading. Meanwhile, the white spots that appeared on the fracture surface of all PP/nano-ZnO composites were confirmed by EDX analysis to be ZnO nanoparticles. The elements of Zinc (Zn) and Oxygen (O) represent the 152

ZnO particles. Meanwhile, Platinum (Pt) element was contributed to by the coating material. Furthermore, it can be confirmed that the state of nanoparticles distribution, on the matrix of all PP/ZnO nanocomposites, is in a good condition and well dispersed. Even though the agglomeration issue occurred (i.e. especially for higher filler contents), it was still under control and restrained. This observation is in good agreement with the explanation made by Figures 7.7 and 7.8; where composites with higher filler contents recorded higher strength and stiffness properties.

153

Figure 7.10: The FESEM micrographs of the fractured surface of (A) PP/ 1% ZnO, (B) PP/ 3% ZnO, and (C) PP/ 5% ZnO composites at 16.47 s-1 of striking velocity

154

POLYPROPYLENE/MICRO-MICA COMPOSITES 7.8 Material characterisations 7.8.1 Dynamic Mechanical Analysis (DMA) Representative storage modulus (E’), tan delta (d) and loss modulus (E’’) was taken at three different strain rates (i.e. 0.0001 s-1, 0.001 s-1, and 0.01 s-1) for all PP/micro-Mica composite as depicted in Figure 7.11 (A-B) to 7.13 (A-B), respectively. As expected, the PP/micro-Mica composite show similar DMA trend with the results reported for PP/nano-ZnO composites. It can be seen that the E’ and the E’’ of all PP/micro-Mica composite composites increased steadily with increasing strain rate (i.e. frequency). This again confirm that both E’ and E’’ values were determined by the characteristic of the PP matrix, where higher strain rate will slightly disturb the segmental motion of the PP molecular chains, thus enhance the rigidity of the composites. In addition, the tan delta (d) characteristic of all PP/micro-Mica composite also react similarly with the PP/nano-ZnO composites where it has recorded significant transition (i.e. Tg has move to higher temperature) as an increasing strain rates. However, the magnitude of tan delta (d) transition is somehow different with the results reported for PP/nano-ZnO nanocomposites. Cumulatively, the Tg of PP/10% Mica, PP/15% Mica and PP/20% Mica composites shifted of about 3oC, 3oC and 5oC higher between 0.0001 s-1- 0.001 s-1 strain rates, respectively. On the other hand, for 0.001 s-1- 0.01 s-1 strain rates, the Tg of PP/10% Mica, PP/15% Mica and PP/20% Mica composites shifted approximately 3oC, 4oC and 5oC higher than their initial Tg value. 155

From the superposition curves of all PP/micro-Mica composite, at room temperature (~30oC), it can be seen that the viscoelastic properties of the composites has slightly changed from a rubbery-like behaviour at a static loading (0.0001 s-1) to a glassy-like behaviour at a dynamic loading (1000 s-1) as illustrated in Figures

7.11 (C), 7.12 (C) and 7.13 (C),

respectively. According to Deschanel et al. (2009), this transition (i.e. from static to dynamic condition) alters the intermolecular interaction in the amorphous domains of the composite, and therefore, it starts to pose a significant resistance towards deformation. This again proved that that the viscoelastic properties of both composite systems (i.e. PP/micro-Mica composite and PP/nano-ZnO composites) were largely contributed by the PP characteristic rather than ZnO or Mica particles.

156


C

B

PP/10% Mica composite

157


A


C

B

PP/15% Mica composite 158


A


C

B

PP/20% Mica composite

159


A

7.9 Stress/strain characteristic The pure PP and PP/micro-Mica composites, with different filler contents (i.e. 10, 15 and 20%) by weight, were exposed to static and dynamic compressions under various levels of strain rates. For the evaluation purpose, the average strain rates were kept at 0.01, 0.1, 650, 900 and 1100 s-1 as shown in Figures 7.14 (A-D), respectively. From Figures 7.14 (A-D), it is noticeable that the stress/strain characteristic of these PP/micro-Mica composites is somehow different with the one recorded for PP/nano-ZnO composites especially under dynamic loading (i.e. 650, 900 and 1100 s-1). Typically, based on Figures 7.14 (A-D), it can be seen that the stress/strain characteristics of PP/micro-Mica composites do not show obvious description of hardening and strengthening characteristic as compared to the PP/ZnO nanocomposites,

under similar strain rate

investigated. This different was mainly attributed to the aspect ratio of the embedded particles. Theoretically, aspect ratio could be defined as the length of the longest Feret diameter over the length of the shortest Feret Flongest

diameter Fshortest which illustrated in Figure 7.15 (Moschakis et al., 2005). According to Adams (1993), high aspect ratio particles (i.e. Mica) tend to cause high stress in the polymer matrix near the particle edge, thus facilitate a lot of weakest failure points which indirectly bring down the properties of the composite. Meanwhile, low aspect ratio particles (i.e. ZnO) will lower down the maximum stress on coalescence. This action will significantly retard and block the propagation of the crack and therefore enhance the strengthening and hardening characteristics of the composite (Cook et al., 1964).

160

161

Figure 7.14: True compression stress/strain curves of the (A) PP neat, (B) PP/10% Mica, (C) PP/15% Mica and (D) PP/20% Mica composites under a wide range of strain rate investigated 162

163

Figure 7.15: The determination of aspect ratio for both ZnO and Mica (Chiu et al., 2008) particles


Further analysis was carried out on the stiffness properties of the PP/micro-Mica composite, as depicted in Figure 7.16. From the bar chart in Figure 7.16, it is clearly seen that the compression modulus rises up gradually, with an increasing strain rate. This finding is in line with the trend recorded for PP/nano-ZnO composites and previously reported work (Li and Lambros, 2001). It was believed that, the increment in compression modulus was attributed to the decrease in molecular mobility of the polymer chains, and thus made the composites become stiffer. On the other hand, from the particle content point of view, an almost linear improvement, in terms of compression modulus for the PP/microMica composite, was observed with increasing particle contents over the range of the strain rates investigated. For example, at a strain rate of 650 s-1 and 1100 s-1, the compression modulus for the pure PP is 5.67 GPa and 6.80 GPa. Meanwhile, at 20% of Mica content, the compression modulus was found to increase relatively to 9.2 GPa and 14.6 GPa. Again, the trend of the results is parallel with the results reported for PP/nano-ZnO composites. This indicates that the stiffness properties of polymer matrix composites are mainly influenced by the concentration of rigid particles within the polymer matrix without concerning the compatibility between matrix and filler.

164

Figure 7.16: The compression modulus of the neat PP and the PP/microMica composite under various levels of strain rates investigated

7.11 Strength properties In this study, the relationship between the strength properties of PP/micro-Mica composite and the applied strain rates (i.e. 0.01, 0.1, 650, 900 and 1100 s-1) is portrayed in both Figures 7.17 and 7.18. From the bar graphs in Figures 7.17 and 7.18, it is interesting to note, that these composites react similarly with pure PP as well as PP/nano-ZnO composites where both yield and ultimate compressive strength increased significantly with increasing strain rate. This again confirmed that the enhancement of the strength properties, under dynamic loading compared to static loading, is directly related to the shifting of the glass transition for viscoelastic material during the compression test, as previously discussed 165

in the DMA results (i.e. Figures 7.11, 7.12 and 7.13). This transition reveals that, all PP/micro-Mica composite has slightly transformed from a rubbery-like behaviour at static loading to a glassy-like behaviour at dynamic loading, where intermolecular interactions in the amorphous domains of PP matrix begin to pose a significant resistance to deformation (Deschanel et al., 2009). As a result, it can be observed that, the higher the strain rate applied, the greater would be the strength properties (i.e. yield and ultimate compressive strength) of the PP/micro-Mica composite, up to certain extent. In addition to strain rate effect, we had also determined the relationship between the filler content of Mica microparticles with static and dynamic compressive properties of PP/micro-Mica composite. From the bar graph in Figure 7.17, it was observed that the yield strength of the PP/micro-Mica composite increase gradually with increasing mica contents up to 20 wt %. Similar finding were also reported by previous works (Suwanprateeb, 2000; Nielsen and Landel, 1994). Both authors agreed that, in compression, filler generally produces an increase in the strength of a polymer matrix, where compression tends to close cracks and flaws that are perpendicular to the applied stress, and thus increase the yield strength. However, the ultimate compressive strength is slightly different as compared to the yield strength, since it was referred as the maximum stress before failure. Therefore, this property was mainly influenced by the fillermatrix characteristic rather than the test configuration. From the bar graph in Figure 7.18, it is interesting to note that, the increment of ultimate compressive strength is somehow different between static and dynamic loadings. Increments in ultimate compressive strength, at static loading conditions, are observed up to 15% of filler content, before the value started to reduce at 20% of filler content. In comparison, at dynamic 166

loading, the increase can only be seen at 10% of filler content and the strength started to reduce at 15% of filler content. Based on the previous finding, it was reported that, at high amounts of mica particles, there is an overlap between one flake and another, as several misaligned flakes, or flakes stacked on top of each other, act as stress concentrators (Nielsen and Landel, 1994). This stress concentrators will drastically decreases the ultimate compressive strength value of the PP/micro-Mica composite especially at higher amounts of mica contents.

Figure 7.17: The yield strength of the neat PP and the PP/micro-Mica composite under various levels of strain rates investigated

167

Figure 7.18: The compressive strength of the neat PP and the PP/microMica composite under various levels of strain rates investigated

7.12 Rate sensitivity, thermal activation volume and strain energy As can be seen in Table 7.2, the PP/micro-Mica composites react similarly with the PP/nano-ZnO nanocomposites in terms of the rate sensitivity and the thermal activation volume. This can be seen from the pattern recorded in the Table 7.2, where all PP/micro-Mica composite specimens showed positive increment in terms of the rate sensitivity with increasing strain rate, whilst the thermal activation volume shows reversed trend. It was believed that, increasing strain rates will significantly reduce the PP chains mobility thus; higher flow stress is required to perform the deformation up to 0.025 of strain. Apart from that, the chains mobility restriction during higher strain rate has also indirectly reduced the thermal activation volume of the PP/micro-Mica composites. Similar with PP/nano168

ZnO composites, there is insignificant relationship was recorded between Mica particle contents and strain rate sensitivity as well as thermal activation volume of the PP/micro-Mica composite under a wide range of strain rate investigated.

The line graph in Figure 7.19 shows the relationship between the strain energy (i.e. absorbed energy) and the applied strain rate for PP/micro-Mica composites. The figure indicates that the strain energy shows positive increment with the applied strain rate, as well as the particle contents. This trend is in line with the results reported for PP/nano-ZnO composites. Interestingly, it was found that the intensity of strain energy is somehow different between these two composite systems. Typically, PP/micro-Mica composite absorbed more energy than that of PP/nano-ZnO nanocomposites up to 0.025 of strain (i.e. deformation). For convenient comparison, the different between PP/micro-Mica composites and PP/nano-ZnO composites were group together and shown in Table 7.3. Based on the comparative result from Table 7.3, it is statistically proven that the PP/micro-Mica composites absorbed higher energy than that of PP/nano-ZnO composites for each of applied strain rate up to 0.025 deformations. This phenomenon was mainly attributed to the geometrical structure of the particles used in this study (Dou et al., 2010). According to Duo et al. (2010), the soft PP surrounds the rigid filler will act as a buffer and effectively absorbs the energy during loading. Moreover, an optimization is reached between the PP and preferential orientation of the Mica flakes and therefore, achieves maximum value of the energy absorbing capacity.

169

Table 7.2: The rate sensitivity and the thermal activation volume of the neat PP and the PP/micro-Mica composites under static and dynamic regions

b=

s2 - s1 ·

·

ln(e2 / e1 )

Specimens

Range of

(MPa)

Regions

·

strain

PP/10% Mica PP/15% Mica PP/20% Mica


·

e2 > e1 ε = 0.025

rates (s-1) PP neat

· é ù e ê ln( · 2 ) ú ê ú kT V * = kT ê e1 ú = ê s2 - s1 ú b ê ú êë (m3) úû

·

·

e2 > e1 ε = 0.025


3.105 42.768 5.791 42.451 2.895 42.152 3.341 35.09

1.214 x 10-27 8.813 x 10-29 6.509 x 10-28 8.879 x 10-29 1.302 x 10-27 8.942 x 10-29 1.128 x 10-27 1.074 x 10-28

Figure 7.19: The strain energy of the neat PP and the PP/micro-Mica composites under various loading rates 170

0.01 0.1 650 900 1100

PP/ 15% Mica 0.4333 0.4958 2.0886 2.1834 2.2583

PP/ 20% Mica 0.4769 0.5692 2.0923 2.3038 2.3385

PP/1% ZnO 0.2000 0.2542 1.3376 1.6417 1.7375

171

PP/3% ZnO 0.2046 0.2636 1.3591 1.6773 1.7809

( J/m3)

rates, s-1

PP/ 10% Mica 0.2250 0.4834 1.7792 1.8792 2.0250

Polypropylene Based Composites

Strain

composites under a wide range of strain rate investigated

PP/5% ZnO 0.2288 0.2846 1.4925 1.7923 1.8951

10% Mica- 1% ZnO 11.1111 47.4141 24.8201 12.6384 14.1975

15% Mica- 3% ZnO 52.7810 46.8334 34.9277 23.1794 21.1398

20% Mica- 5% ZnO 52.0235 50.0000 28.6670 22.2024 18.9609

Cumulative different, %

Table 7.3: The comparison of energy absorbing capacity between PP/micro-Mica composites and PP/nano-ZnO

7.13 Fracture surface analysis Figure 7.20 shows the SEM micrograph of the PP/micro-Mica composites surface fracture at 16.47 ms-1 striking velocity. From fractographic analysis, it was observed that, a lot cavitations and holes appear on the specimen’s surface especially at high amount Mica contents (i.e. 15% and 20%). Generally, at high amount of Mica content, the Mica filler tends to stacked on top of each other, thus act as stress concentrator (Nielsen and Landel, 1994). Therefore, composite with higher amount of Mica contents tended to be pulled-out (i.e. debonded) under a dynamic loading due to the higher stress concentrator; and thus, enhanced the formation of holes and cavitations. This holes and cavitations will slightly disturb the effectiveness of the stress transfer between the matrix and the Mica particle, and vice versa (Rashid et al., 2011). This observation is in good agreement with the explanation made in Figure 7.18, where composites with high amount of Mica contents (i.e. 15% and 20%) recorded lower ultimate compressive properties.

172

Figure 7.20: The FESEM micrographs of the fractured surface of (A) PP/10% Mica, (B) PP/15% Mica, and (C) PP/20% Mica composites at 16.47 ms-1 of striking velocity 173

7.14 Summary

In this chapter, static and dynamic compression testing has been successfully performed on two different types of polypropylene based composites which are PP/nano-ZnO nanocomposites and PP/micro-Mica composites, respectively. Both composite systems were tested under different levels of strain rate up to nearly 1100 s -1, using the conventional universal testing and the split Hopkinson pressure bar apparatus, respectively. For the characterisation purpose, the DMA analysis was demonstrated in order identify their thermomechanical characteristic under various strain rates (i.e. frequencies) investigated. The DMA results for both composites systems shows similar characteristic in term of the Tg transition as increasing strain rates, where at room temperature, their viscoelastic properties has slightly changed from a rubbery-like behaviour at a static loading (0.0001 s-1) to a glassy-like behaviour at a dynamic loading (1000 s-1). From the stress/strain characteristic observation, it was found that PP/nano-ZnO nanocomposites show obvious description of hardening and strengthening mechanism. Meanwhile, those hardening and strengthening description was unclear for PP/micro-Mica composites, especially at dynamic loading. Generally, the compression modulus and the yield strength of both composite systems increase significantly with increasing strain rate as well particle content. However, for ultimate compressive strength characteristic, these two composite systems show significant variation from one to another. Apart from that, the strain rate sensitivity and the thermal activation volume of both composite systems were also calculated. Based on the calculated values, it was found that, the rate 174

sensitivity of both composite systems increases significantly with increasing strain rate, whilst the thermal activation volume shows contrary trend. In contrast to strain rate, the particle content shows insignificant relationship with strain rate sensitivity and thermal activation volume for both PP/nano-ZnO composites and PP/micro-Mica composites In this study, the strain energy of both composites systems was measured up to 0.025 deformations, as a function of strain rate as well as particle content. In general, both composites systems recorded positive increment in terms of the strain energy with increasing strain rate and particle content. However, based on the statistical measurement, it was found that, PP/micro-Mica composites absorbed more energy than that of PP/nano-ZnO composites for each of applied strain rates up to certain extent. Ultimately, the fractographic analysis was successfully determined using the SEM and EDX analysis in order to understand the failure mechanism experienced by these two composites systems, especially under dynamic loading.

175

CHAPTER 8 EFFECT OF PARTICLE SIZE ON STATIC AND DYNAMIC COMPRESSIVE PROPERTIES OF POLYPROPYLENE BASED COMPOSITES

8.1 Introduction

It was proven that the geometrical structures of filler and the particle contents play as significant factor that influenced the static and the dynamic compressive properties of the particulate filled polymer composites, as previously discussed in Chapter 7. Apart from that, the interaction between particle and matrix has also plays as a secondary role that determines their mechanical properties. Thus, the effect of particle-matrix related characteristics, such as particle size, particle–matrix interface adhesion, particle shape and geometry, require extra attention during their mechanical properties evaluation. For the past few decades, many studies have been conducted on the effect of particle size; especially under static loading (Sumita et al., 1983; Lazzeri et al., 2004; Zhu et al., 1999; Ji et al., 2002). Previous findings have experimentally proved that particle size has an obvious effect on the mechanical properties of particulate filled polymer composites. During the early 1980s, Sumita et al. (1983) emphasised their interest of replacing micro-scale silica with its nanoscale counterpart; since nanosilica particles confer remarkable mechanical properties. They also found that these nanoparticles gave higher rigidity to virgin polymers. More recent studies by Lazzeri et al. (2004) noted that the modulus of aluminium

hydroxide

filled

polypropylene

composites,

decreased

marginally with mean particle size (i.e. where the average particle diameter 176

varied from 1 to 25 µm). In contrast, Zhu et al. (1999) found that particle size effect was insignificant with the modulus of organo-soluble polyimide (PI)/Silica composites. However, the latest findings by Ji et al. (2002) and Mishra et al. (2005) slightly disagreed with the results reported by Zhu et al. (1999). They found that when particle size is decreased to a certain critical size, such as 30 nm, there will be an obvious effect of particle size on the modulus of particulate filled polymer composites. Indeed, it was experimentally observed that PP/CaCO3 composites, containing smaller nanoparticles, have a higher Young’s modulus than that of composites with larger nanoparticles (Mishra et al., 2005). As a pre-conclusion, it seems that there is less effect on composite modulus when the particle size is above its critical size. However, the effect will be remarkable when the particle size is reduced below its critical value.

The strength properties of particulate filled polymer composites are also greatly affected by particle size. Theoretically, for these composites, their strengths were strongly dependent on the stress transfer between the particles and the matrix. For a given volume fraction, smaller particles may have a higher surface area than that of larger particles, and therefore induce excellent bonding between particle and matrix. This remarkable bonding will effectively transfer the applied stress to the particles from the matrix and clearly improve strength (Cho et al., 2006; Zhang et al., 2004). However, for poorly bonded, strength reduction occurs by adding particles (Tjong and Xu, 2001). As far as particle size is concerned, the majority of previous studies only focused on the static behaviour of the particulate filled polymer composites. Unfortunately, there is a very limited number of works that were involved with the characteristics of these composites under dynamic loading. This indicates that the effect of particle size under 177

dynamic perspective issue, still remains unclear and further investigation is still required.

Based on this, the main objective of the present chapter is to study the particle size dependant on the static and dynamic compression properties of PP/SiO2 composites under various levels of strain rates. To achieve the goals, an experimental technique, based on the compression Split Hopkinson Pressure Bar (SHPB), was introduced to perform high strain rate testing; where a conventional universal testing machine was used to perform static compression testing on PP/SiO2 composites. In this study, PP/SiO2 composites were prepared using four different sizes of SiO2 particles, ranging from micro to nano sizes (i.e. 3 µm, 1 µm, 20 nm, and 11 nm) with appropriate shapes, respectively. For characterisation purposes, the size and shape of the SiO2 particles were confirmed using three established methods, namely a particle size analyser, Scanning Electron Microscopy (SEM),

and

High

Resolution

Transmission

Electron

Microscopy (HRTEM). Composite responses were then characterised in terms of stress/strain curves, yield strength, ultimate strength, and stiffness. Additionally, a correlation between the particle size of SiO2, and strain rate sensitivity and thermal activation volume up to a certain deformation (i.e. 0.025 of strain), was also made. Ultimately, post damage analysis was performed to further identify the failure mechanism experienced by the composites under both static and dynamic loadings. 8.2 Material characterisations 8.2.1 Particle size confirmation It is important to ensure that the SiO2 particle used in this study has a systematic difference in terms of size, but not in shape. We minimised the 178

possible difference in terms of shape by choosing an appropriate silica grade. Meanwhile, four different sizes of SiO2 particles, ranging from micro- scales to nano-scales, were randomly chosen as filler to achieve the main objective of the study. The confirmation of particle size was done by using a particle size analyser and Transmission Electron Microscopy (TEM). Practically, the particle size analyser is the most commonly used to obtain the particle size distribution. However, it typically shows a better accuracy in the micro-scale rather than the nano-scale of particles; due to the agglomeration problem. Meanwhile, TEM images are widely accepted as a practical standard for evaluating particle size (especially for nanoparticles), size distribution, and morphology (Isley and Penn, 2006). Therefore, Figure 8.1 (A-D) show the cumulative results and the TEM images of four different sizes of SiO2 particle used in this study. Based on the cumulative results shown in Figures 8.1 (A) and (B), the mean particle size of both SiO2 microparticles are approximately 3 µm and 1 µm, respectively. Meanwhile, the TEM images of another two SiO2 particles are clearly displayed in Figures 8.1 (C) and (D). Quantitatively, the mean size of the SiO2 nanoparticles in Figures 8.1 (C) and (D) are approximately 20 nm and 11 nm, respectively. The shape of the silica particle used in this experiment was identified using SEM and TEM images, as illustrated in Figures 8.1 (A-D). Although, the TEM images is not really clear to identify the particle shape, but based on the SEM images in Figure 8.1 (A-B), it can be confirmed that the silica particle used in this experiment was consisted of various shapes, such as spherical, cubical, and hexagonal, which are referred to as irregular shapes (Ahmad et al., 2008). In summary, it was proven that the SiO2 particles used in this experiment were adequately sized (i.e. 3 µm, 1 µm, 20 nm, and 11 nm) and of an appropriate particle shape. 179

180

Figure 8.1: (A-B) The particle sizes distribution results and SEM images of SiO2 microparticles. (C-D) The TEM images of SiO2 nanoparticles used in this study 181


The true compressive stress/strain curves of the PP/SiO2 composites, as a function of filler sizes measured under various loading rates (i.e. 0.01 s-1, 0.1 s-1, 650 s-1, 900 s-1 and 1100 s-1), are summarised in Figure 8.2 (AD). From an initial observation, it is clearly seen that all PP/SiO2 composites behave differently under various loadings, where obvious failure points can only be scrutinised under a dynamic loading rather than a static loading. These results are in agreement with previously reported works on several polymers (Jordan et al., 2007;Siviour et al., 2004). This phenomenon was presumably due to the rate of deformation and the toughening mechanism experienced by the specimens during loading. At dynamic loading, the PP/SiO2 specimens collided with higher striking velocities, which were directly attributed to the higher deformation rates. This induces a severe catastrophic shear yielding mechanism, and thus, leads to the accumulation of micro-damage within the composites. Furthermore, at dynamic loading, the mobility of the polymer chains decreases and therefore, more ruptures have to take place in order to adapt to a larger deformation of the composites (Guo and Li, 2007). Based on this, the catastrophic failure is severe and even clearer under a dynamic loading than that of a static loading.

On the other hand, the effect of particle size also significantly influences the stress/strain characteristic of all tested specimens; especially under dynamic loading. For example, at 1100 s -1 of strain rate, composites with smaller particles (i.e. nano- sized) showed continual strain-hardening throughout the loading. Meanwhile, composites with larger particles (i.e. 182

micro-sized) exhibited a more distinct yield point and only showed some strain-hardening at larger deformation (i.e. strain). This result indicates that the toughness behaviour of tested composites show a linear relationship with particle size. This trend is similar to the work reported by Lange and Radford (1971) where they found that the epoxy reinforced aluminium trihydrate, recorded a decrease of fracture toughness with a decreasing particle diameter. According to Dubnikova et al. (2004), the decrease of the toughening effect with a decreasing particle size can be explained by the role of particle size as the adhesive factor of a fracture. This adhesive factor is due to the increase of debonding stress with the particle size decrease, and the incomplete particle debonding. The difficulty of voiding (i.e. smaller d size: refer to Figure 7.6) leads to a composite brittle fracture as a result of the high yield stress and restriction of plastic flow (Dubnikova et al., 2004).

183

184

Figure 8.2: True stress/strain curves of PP/SiO2 nanocomposites with various particles sizes under different levels of loading rates

8.4 Stiffness properties The effect of particle size on the compression modulus of PP/SiO2 composite is illustrated in Figure 8.3. The results indicate that the relationship between particle size and compression modulus are almost identical under both static and dynamic loadings, where the compression modulus increases steadily with decreasing particle sizes. Basically, all PP/SiO2 composites recorded higher compression modulus than that of neat polypropylene, due to the introduction of a rigid particle to the polymer matrix. Nevertheless, the intensity of increment was slightly different between micro- and nano-sized particles. Quantitatively, at 0.01 s-1 and 1100 s-1 of strain rate, PP/SiO2 with a particle size of 11 nm recorded 185

approximately 32% and 26% compression modulus increment, compared to that of neat polypropylene, respectively. Meanwhile, PP/SiO2 with a particle size of 3 µm recorded a much lower increment in terms of compression modulus, which was approximately 6% and 3%, respectively, when compared to the neat polypropylene under similar loading rates. This trend indicates that at certain critical sizes, such as 30 nm, there will be an obvious effect of particle size on the modulus, which was predicted theoretically by Ji et al. (2002). Indeed, Mishra et al. (2005) experimentally proved that polymer composites (PP/CaCO3) containing smaller particles have a higher modulus than composites with larger particles.

Figure 8.3: The compression modulus values of PP/SiO2 nanocomposites under various loading rates as a function of particle sizes

186

8.5 Strength properties The yield and ultimate compressive strength (UCS), as a function of filler sizes were determined and are illustrated in Figures 8.4 and 8.5, respectively. From the bar graph shown in Figure 8.4, it is interesting to note that for a given particulate weight fraction, the yield stress of the PP/SiO2 composite increases with a decreasing particle size over the wide range of strain rates investigated. These results are consistent with those obtained by Leidner and Woodhams (1974), Buggy et al. (2005) and Fu et al. (2008), respectively. Even though most of the previous works (Buggy et al., 2005; Leidner and Woodhams, 1974; Fu et al., 2008) only focused on the static behaviour of Polymer Matrix Composites (PMCs), but somehow their knowledge is still relevant in describing the dynamic behaviour of the composites tested in this experiment. Basically, the yield strength of particle filled composites is defined as the maximum stress that the composite can sustain under uniaxial loading (i.e. compression loading) before plastic deformation occurs. Therefore, it is widely accepted that yield strength relies highly on the effectiveness of stress transfer between matrix and filler (Buggy et al., 2005; Fu et al., 2008; Leidner and Woodhams, 1974). Smaller particles have a higher total surface area for a given particle loading. This indicates that the yield strength increases with the increasing surface area of the filled particles; through a more efficient stress transfer mechanism (Fu et al., 2008).

Meanwhile, the bar graph in Figure 8.5 clearly shows that the PP/SiO2 composites react differently in terms of the ultimate compressive strength values, under both static and dynamic loadings. For a static loading, the introduction of a rigid particle steadily increased the ultimate 187

compressive strength, regardless of particle size. Meanwhile, for dynamic loading, the composites that were incorporated with larger particle sizes (i.e. 3 µm and 1 µm) recorded lower UCS values than that of the neat polypropylene. Conversely, for composites with smaller particle sizes (i.e. 20 nm and 11 nm), the trend was reversed. This trend was similar to many Polymer Matrix Composites (PMCs) that were exposed under a tensile loading (Pukanszky and Voros, 1993; Nakamura et al., 1992). Generally, in compression, filler may produce an increase in the ultimate compression strength of the polymer composite, where compression tends to close cracks and flaws that are perpendicular to the applied stress; and therefore, increase the compressive strength (Suwanprateeb, 2000; Nielsen and Landel, 1994). However, the strength of two-phase composites is still highly dependent on the effectiveness of the stress transfer between filler and matrix. For poorly bonded particles (i.e. micro-sized), the stress transfer at the particle/matrix interface is inefficient. This will tend to get worse at dynamic loading, where a higher deformation rate will enhance the discontinuity in the form of debonding, due to the non-adherence of the particle to the polymer. In contrast, for composites that contain well-bonded particles, the addition of particles into a polymer will lead to an increase in compressive strength; especially for nanoparticles with a high surface area (Fu et al., 2008).

188

Figure 8.4: The yield strength values of PP/SiO2 nanocomposites under various loading rates as a function of particle sizes

Figure 8.5: The ultimate compressive strength values of PP/SiO2 nanocomposites under various loading rates as a function of particle sizes 189

8.6 Strain rate sensitivity, thermal activation volume and strain energy From the results calculated in Table 8.1, it is interesting to note that the SiO2 particle size gives a significant effect on both the strain rate sensitivity and the thermal activation volume (refer β and V* values) under both static and dynamic loadings. It can be seen, that smaller particles steadily decreased the strain rate sensitivity of the PP/SiO2 composites. Quantitatively, the percentage of strain rate sensitivity reduction between 3 µm and 11 nm was approximately 78% and 73%, under both static and dynamic loadings, respectively. The results presented here, will be useful during the material selection; especially for extreme applications, where less sensitivity properties are required. Conversely, smaller particle sizes increased the thermal activation volume of the PP/SiO2 specimens. In current practice, there is no scientific explanation that can be used to describe this trend. However, we believed that the particle’s surface area play an important role in determining the rate sensitivity of tested specimens. Theoretically, the flow stress could be defined as force over area, (σ: F/A). Meanwhile, during loading, the applied stress will be transfer from matrix to particle and vice versa. Therefore, composites that contained smaller particles (i.e. larger surface area for a given weight fraction) would attributed to the lower flow stresses distribution within the composites body and therefore induced lower strain rate sensitivity of flow stresses. On the other hand, the thermal activation volume is corresponded to the localised motion of segment or possibly side group of polymer chains (White, 1981). Hence, for smaller particles, we believed that the localised motion of polymer chains is less restricted than that of bigger particles thus, recorded higher thermal activation volume.

190

On the other hand, the relationship between the particle size and the strain energy of the PP/SiO2 composites is illustrated in Figure 8.6. Results show that there is insignificant correlation between these two parameters. This results are inconsistent with several previously reported works (Javni et al., 2002; Ng et al., 1999), where they found positive increment in terms of the energy absorption (i.e. strain energy) capacity as decreasing particle sizes. The inconsistency was attributed to the selection of deformation point. Those previous work measured their specimen’s strain energy up to failure point. Meanwhile, in this study, we had standardised the deformation point up to 0.025 of strain in order to easily compare the material from one to another. This might be one of the main reasons that attributed to the different pattern between current and previous results.

Table 8.1: The rate sensitivities and thermal activation volume of the PP/SiO2 (as function of particle sizes) measured under static and dynamic regions

PP/SiO2 with Range of different strain particle rates (s-1) sizes (µm) 3 1 0.020 0.011


b=

s2 - s1 ·

·

ln(e 2 / e1 )

Classification

(MPa) ·

·

· é ù e ê ln( · 2 ) ú ê e1 ú = kT V * = kT ê ú ê s2 - s1 ú b ê ú ëê ûú

e 2 > e1 ε = 0.025


2.789 30.413 1.451 20.282 1.160 10.131 0.619 8.154

191

·

·

e 2 > e1 ε = 0.025

1.351 x 10-27 1.250 x 10-28 2.598 x 10-27 1.858 x 10-28 3.249 x 10-27 3.720 x 10-28 6.089 x 10-27 4.622 x 10-28

Figure 8.6: The strain energy of PP/SiO2 composites as a function of particle sizes under a wide range of strain rate investigated (ε: 0.025)

8.7 Post damage analysis 8.7.1 Physical analysis Figures 8.7 (A) and (B) show photographs of the PP/SiO2 composite specimens, under static (0.1 s-1) and dynamic loadings (650 s-1), as a function of particle size. The figures clearly show that the failure characteristics of the composites were completely different under both static and dynamic loadings. It was observed that all PP/SiO2 specimens experienced severe catastrophic fractures under a dynamic loading. Meanwhile, no fracture was observed in the case of a static loading. Similar observations were reported by Blackman et al. (2003). Interestingly, the photographs in Figures 8.7 (A) and (B) showed a good correlation with the stress/strain characteristic discussed in Figure 8.2. It was physically proved 192

that the catastrophic fractures can only be seen under a dynamic loading; rather than a static loading.

Figure 8.7: The photographs of PP/SiO2 composites specimen under (A) static loading (0.1 s-1) and (B) dynamic loading (650 s-1) as a function of particle sizes

8.7.2 Fracture Surface Analysis Fracture surface analysis was performed using SEM equipment, as illustrated in Figure 8.8 (A-D). In addition, the EDX analysis was also carried out to distinguish between the presences of SiO2 particles on the fracture surface, of all composite specimens. Figure 8.8 (A-D) shows the SEM micrograph of the dynamic compression fracture at a dynamic loading of 1100 s-1 of strain. From fractographic analysis, it was observed that larger particles (i.e. 3 µm and 1 µm) tended to be pulled-out (i.e. debonded) under a dynamic loading, due to them being poorly bonded; and thus, enhanced the formation of holes and voids. Those holes and voids did 193

not appear on the fracture surface of the composite with smaller particles sizes. Theoretically, the formation of holes and cavitation’s disturbs the effectiveness of the stress transfer between the matrix and the SiO2 particle, and vice versa (Rashid et al., 2011). This observation is in good agreement with the explanation made in Figures 8.4 and 8.5, where composites with larger particles (i.e. 3 µm and 1 µm) recorded lower compression properties, in terms of yield strength and Ultimate Compressive Strength (UCS); compared to composites with smaller particles (i.e. 20 nm and 11 nm). Furthermore, the EDX analysis confirmed the presence of SiO2 particles by showing the elements that exist at the pointed area ‘x’ for all PP/SiO2 composites. The elements of Silicon (Si) and Oxygen (O) represent the SiO2 particles. Meanwhile Platinum (Pt) was contributed by the coating material.

194

195

Figure 8.8: The fracture surface of PP/SiO2 composites with various particle sizes (A) 3 µm, (B) 1 µm, (C) 20 nm (D) 11 nm under 1100s-1 of strain rate

8.8 Summary

The effect of particle size on the static and dynamic compression properties of PP/SiO2 composites was successfully investigated using a conventional universal testing machine and a split Hopkinson pressure bar apparatus, up to nearly 1100 s-1 of strain rates, respectively. Both the size and the shape of the SiO2 particles used in this study were successfully obtained using a conventional particle size analyser, Scanning Electron Microscopy (SEM),

and

High

Resolution

Transmission

Electron

Microscopy (HRTEM). Results show that the stress/strain characteristic of 196

the PP/SiO2 composites was influenced by particle size, where the bigger the particle the greater would be the toughening effect. Apart from that, we also found that, the size of the SiO2 particles gave significant effects on the compressive properties of the PP/SiO2 composites. Quantitatively, PP/SiO2 composites with smaller particle sizes (i.e. 20 nm and 11 nm) recorded higher compression properties, in terms of yield strength, ultimate strength, and stiffness, as compared to PP/SiO2 composites with larger sizes (i.e. 3 µm and 1 µm), for all strain rates tested. In addition, the reduction in SiO2 particle size also significantly decreased the rate sensitivity of the PP/SiO2 composites, whereas the thermal activation volume showed a contrary trend. From physical observations, it was proved that the fractures can only be seen under a dynamic loading; rather than a static loading. Furthermore, from the fracture surface analysis, it was observed that the PP/SiO2 composite with larger particle sizes (i.e. 3 µm and 1 µm), tended to be pulled-out (i.e. debonded) under a dynamic loading, due to poor bonding, and therefore, enhanced the formation of holes and voids. These holes and voids did not appear on the fracture surface of the composite with smaller particles sizes.

197

CHAPTER 9 CONCLUSIONS AND SUGGESTIONS FOR FURTHER WORK

9.1 Conclusions

Static and dynamic mechanical tests have been successfully performed on several thermoplastic polymers (i.e. PE, PP and PC) and polypropylenebased composites (i.e. PP/nano-ZnO composites and PP/micro-Mica composites) up to nearly 1100 s-1 of strain rate using the split Hopkinson pressure bar apparatus. From the overall results, the following conclusions can be drawn:

· The mechanical properties of all the tested thermoplastic polymers (i.e. PE, PP, and PC) showed a great dependency on the strain rate applied. The yield stress, compression modulus and compressive strength were proportionally increased as the strain rate increased. However, the yield strain showed a contradictory pattern where it was gradually decreased with applied strain rates. It was found that the strain rate sensitivity and the strain energy of all the thermoplastic specimens were significantly increased with increasing strain rates, whereas the thermal activation values showed a contrary trend. Interestingly, it was found that the Eyring and power basic equations gave a good interpretation of the experimental results over a wide range of strain rates investigated.

198

· For the effect of molecular structures, it was found that the HDPE exhibits greater compressive properties in terms of yield stress, stiffness and strength properties than that of a complex molecular structure (i.e. LLDPE and LDPE). On the other hand, the yield stress, compression modulus, compressive strength, rate sensitivity as well as the strain energy of all the polyethylene specimens were proportionately increased as the strain rate increased. Conversely, the yield strain and thermal activation volume showed a contradictory pattern where it gradually decreased with the applied strain rate. Of the three PE specimens, LLDPE recorded the highest strain rate sensitivity values under both the static and dynamic regions of strain rates. Meanwhile, for the quasi-static region (i.e. static to dynamic), LDPE had shown the highest value of strain rate sensitivity compared to LLDPE and HDPE. Interestingly, agreement between the experimental and prediction values offered by both the Eyring and power basic equations were also found to be good over a wide range of strain rates investigated. · For polypropylene-based composites using micro and nano fillers, it was found that PP/nano-ZnO nanocomposites show an obvious description of hardening and strengthening mechanism. Meanwhile, that hardening and strengthening description was unclear for PP/micro-Mica

composites,

especially

at

dynamic

loading.

Generally, the compression modulus and the yield strength of both composite systems increase significantly with increasing strain rates as well as particle content. However, for the ultimate compressive strength characteristic, these two composite systems showed significant variations from one another. In addition, the magnitude of rate sensitivity is somehow different between these two composite 199

systems. Generally, both composite systems recorded positive increments in terms of the strain energy with increasing strain rates as well as the particle content. However, based on the statistical measurement, it was found that PP/micro-Mica composites absorbed more energy than that of PP/nano-ZnO composites for each of the applied strain rates up to 0.025 deformations. · Ultimately, for the effect of particle size, the results showed that the stress/strain

characteristic

of

the

PP/SiO2

composites

was

significantly influenced by particle size, where the bigger the particle the greater would be the toughening effect. Apart from that, it was also found that the size of the SiO2 particles had a significant effect on the compressive properties of the PP/SiO2 composites. Quantitatively, PP/SiO2 composites with nano-sized particles (i.e. 20 nm and 11 nm) recorded higher compression properties in terms of yield strength, ultimate strength and stiffness, as compared to PP/SiO2 composites with nano-sized particles (i.e. 3 µm and 1 µm), for all strain rates tested. In addition, the reduction in SiO2 particle size also significantly decreased the rate sensitivity of the PP/SiO2 composites, whereas the thermal activation volume showed a contrary trend 9.2 Suggestions for further work Although a lot of time and work have been sacrificed to accomplish the various objectives of the study within a given time frame, but somehow there is much more that can be done in the future to gain a better understanding of the static and dynamic behaviour of polymeric materials, especially thermoplastic-based materials. Based on this concern, the 200

following works are suggested and further research could be carried out according to these suggestions or other more creative research ideas: · Further improvement on the gas gun would allow higher striking velocities, thus resulting in greater strain rates. This improvement could be achieved by using a longer cylinder or the development of a fast acting spool-type valve. · As far as polymers are concerned, other internal factors like molecular weight, tacticity and isomerism should also be considered during the static and dynamic mechanical properties evaluation. This might be useful to have a more conclusive finding on the effect of molecular structures on both the static and dynamic mechanical properties of polymeric materials. · For polymer matrix composites (PMCs), it is suggested to perform the filler surface treatment in order to have better adhesion between the filler and the matrix. This might significantly increase the properties of PMCs, especially their dynamic mechanical behaviour. · Apart from particle size, other filler-matrix characteristics such as particle geometry, particle-matrix adhesion, particle surface area, particle aspect ratio, etc., should also be considered in the future. This might be attributed to the more conclusive finding on the effect of the filler-matrix characteristic on the static and dynamic mechanical properties of PMCs.

201

REFERENCES ADAMS, J. 1993. Particle size and shape effects in materials science: examples from polymer and paper systems. Clay Minerals, 28 (4), p.509-530. AHMAD, F. N., JAAFAR, M., PALANIANDY, S. & AZIZLI, K. A. M. 2008. Effect of particle shape of silica mineral on the properties of epoxy composites. Composites Science and Technology, 68 (2), p.346-353. AKAY, M. 1993. Aspects of dynamic mechanical analysis in polymeric composites. Composites Science and Technology, 47 (4), p.419-423. ALBERTINI, C., BOONE, P. & MONTAGNANI, M. 1985. Development of the Hopkinson bar for testing large specimens in tension. J Phys France, 46(C5), p.499–504. ALCOCK, B., CABRERA, N. O., BARKOULA, N. M., SPOELSTRA, A. B., LOOS, J. & PEIJS, T. 2007. The mechanical properties of woven tape all-polypropylene composites. Composites Part A: Applied Science and Manufacturing, 38 (1), p.147-161. ALGER, M. S. M. 1997. Polymer science dictionary.United Kingdom: Kluwer Academic Publishers. ALTAN, M. & YILDIRIM, H. 2012. Effects of compatibilizers on mechanical and antibacterial properties of injection molded nanoZnO filled polypropylene. Journal of Composite Materials, DOI: 10.1177/0021998312436999. ANTIC, V., GOVEDARICA, M. & DJONLAGIC, J. 2003. The effect of segment length on some properties of thermoplastic poly (ester– siloxane) s. Polymer International, 52 (7), p.1188-1197. BACON, C. 1999. Separation of waves propagating in an elastic or viscoelastic Hopkinson pressure bar with three-dimensional effects. International Journal of Impact Engineering, 22 (1), p.55-69. 202

BALASURIYA, P. W., YE, L. & MAI, Y. W. 2001. Mechanical properties of wood flake-polyethylene composites. Part I: effects of processing methods and matrix melt flow behaviour. Composites Part A: Applied Science and Manufacturing, 32 (5), p.619-629. BENKHENAFOU, F., CHEMINGUI, M., FAYOLLE, B., VERDU, J., FEROUANI, A. K. & LEFEBVRE, J. M. 2011. Fracture behaviour of a polypropylene film. Materials & Design, 32 (3), p.1515-1519. BLACKMAN, B. R. K., WILLIAMS J.G, PAVAN A. 2003. Fracture of Polymers, Composites and Adhesives II.United Kingdom: Elsevier ltd & ESIS, 231. BUGGY, M., BRADLEY, G. & SULLIVAN, A. 2005. Polymer-filler interactions in kaolin/nylon 6, 6 composites containing a silane coupling agent. Composites Part A: Applied Science and Manufacturing, 36 (4), p.437-442. CALLISTER JR, W. D. 1994. Materials Science and Engineering–an Introduction, .USA: John Wiley& Sons Inc. CAO, K. & WANG, Y. 2012. Effects of strain rate and temperature on the tension behavior of polycarbonate. Materials & Design, 38, p.53-58. CHAN, C. M., WU, J., LI, J. X. & CHEUNG, Y. K. 2002. Polypropylene/calcium carbonate nanocomposites. Polymer, 43 (10), p.2981-2992. CHEN, L. P., YEE, A. F. & MOSKALA, E. J. 1999a. The molecular basis for the relationship between the secondary relaxation and mechanical properties of a series of polyester copolymer glasses. Macromolecules, 32 (18), p.5944-5955. CHEN, W., LU, F., FREW, D. & FORRESTAL, M. 2002. Dynamic compression testing of soft materials. Journal of applied mechanics, 69 (3), p.214-223.

203

CHEN, W., LU, F. & ZHOU, B. 2000. A quartz-crystal-embedded split Hopkinson pressure bar for soft materials. Experimental Mechanics, 40 (1), p.1-6. CHEN, W., SONG, B., FREW, D. & FORRESTAL, M. 2003. Dynamic small strain measurements of a metal specimen with a split Hopkinson pressure bar. Experimental Mechanics, 43 (1), p.20-23. CHEN, W., ZHANG, B. & FORRESTAL, M. 1999b. A split Hopkinson bar technique for low-impedance materials. Experimental Mechanics, 39 (2), p.81-85. CHENG, H., CHEN, C. & TSAY, C. 2007. Transparent ZnO thin film transistor fabricated by sol-gel and chemical bath deposition combination method. Applied Physics Letters, 90, p.012113. CHIANG, H., WAGER, J., HOFFMAN, R., JEONG, J. & KESZLER, D. 2005. High mobility transparent thin-film transistors with amorphous zinc tin oxide channel layer. Applied Physics Letters, 86, p.013503. CHICHILI, D. & RAMESH, K. 1999. Recovery experiments for adiabatic shear localization: a novel experimental technique. Journal of applied mechanics, 66, p.10. CHIOU, S., CHENG, W. & LEE, W. 2005. Strain rate effects on the mechanical properties of a Fe-Mn-Al alloy under dynamic impact deformations. Materials Science and Engineering: A, 392 (1-2), p.156-162. CHIU, C.-W., CHU, C.-C., CHENG, W.-T. & LIN, J.-J. 2008. Exfoliation of smectite clays by branched polyamines consisting of multiple ionic sites. European Polymer Journal, 44 (3), p.628-636. CHO, J., JOSHI, M. & SUN, C. 2006. Effect of inclusion size on mechanical properties of polymeric composites with micro and nano particles. Composites Science and Technology, 66 (13), p.1941-1952.

204

CHOI, Y. K., SUGIMOTO, K. I., SONG, S. M. & ENDO, M. 2005. Mechanical and thermal properties of vapor-grown carbon nanofiber and polycarbonate composite sheets. Materials Letters, 59 (27), p.3514-3520. COOK, J., GORDON, J., EVANS, C. & MARSH, D. 1964. A mechanism for the control of crack propagation in all-brittle systems. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 282 (1391), p.508-520. DANIEL, I. M. & LABEDZ, R. H. 1983. Method for compression testing of composite materials at high strain rates. Compression testing of homogeneous materials and composites, ASTM STP, 808, p.121-139. DASARI, A. & MISRA, R. 2003. On the strain rate sensitivity of high density polyethylene and polypropylenes. Materials Science and Engineering: A, 358 (1), p.356-371. DAVIES, E. & HUNTER, S. 1963. The dynamic compression testing of solids by the method of the split Hopkinson pressure bar. Journal of the Mechanics and Physics of Solids, 11 (3), p.155-179. DAVIES, R. M. 1948. A critical study of the Hopkinson Pressure Bar.London: Philosophical Transaction Royal Society London, 375475. DESCHANEL, S., GREVISKES, B., BERTOLDI, K., SARVA, S., CHEN, W., SAMUELS, S., COHEN, R. & BOYCE, M. 2009. Rate dependent finite deformation stress-strain behavior of an ethylene methacrylic acid copolymer and an ethylene methacrylic acid butyl acrylate copolymer. Polymer, 50 (1), p.227-235. DHAKAL, H., ZHANG, Z. & RICHARDSON, M. 2007. Effect of water absorption on the mechanical properties of hemp fibre reinforced unsaturated polyester composites. Composites Science and Technology, 67 (7), p.1674-1683. DIOH, N., LEEVERS, P. & WILLIAMS, J. 1993. Thickness effects in split Hopkinson pressure bar tests. Polymer, 34 (20), p.4230-4234. 205

DOU, Q., LI, L. & MENG, M. R. 2010. Improving the toughness of mica composites [Online]. Plastics Research Online. Available: http://www.4spepro.org/view.php?article=002647-2010-02-09 [Accessed 20 January 2011]. DUBNIKOVA, I., BEREZINA, S. & ANTONOV, A. 2004. Effect of rigid particle size on the toughness of filled polypropylene. Journal of Applied Polymer Science, 94 (5), p.1917-1926. EIRAS, D. & PESSAN, L. A. 2009. Mechanical properties of polypropylene/calcium carbonate nanocomposites. Materials Research, 12, p.517-522. ESCUDERO ACEVEDO, M., QUIJADA, R. & CAMPOS VALLETTE, M. 2008. Study of the effect of branching in degradation of polyethylenes obtained via metallocene catalyst Journal of the Chilean Chemical Society, 53, p.1490-1493. EVORA, V. & SHUKLA, A. 2003. Fabrication, characterization, and dynamic behavior of polyester/TiO2 nanocomposites. Materials Science and Engineering A, 361 (1-2), p.358-366. FENG, R. & RAMESH, K. 1991. Dynamic behavior of elastohydrodynamic lubricants in shearing and compression. Journal de Physique IV, 01 (C3), p.69-76. FIELD, J., WALLEY, S., PROUD, W., GOLDREIN, H. & SIVIOUR, C. 2004. Review of experimental techniques for high rate deformation and shock studies. International Journal of Impact Engineering, 30 (7), p.725-775. FOLLANSBEE 1985. “The Hopkinson Bar,”Mechanical Testing, Metals Handbook.Metals Park, Ohio: American Society For Metal, 198– 217. FREW, D., FORRESTAL, M. & CHEN, W. 2002. Pulse shaping techniques for testing brittle materials with a split hopkinson pressure bar. Experimental Mechanics, 42 (1), p.93-106. 206

FREW, D., FORRESTAL, M. & CHEN, W. 2005. Pulse shaping techniques for testing elastic-plastic materials with a split Hopkinson pressure bar. Experimental Mechanics, 45 (2), p.186-195. FRIEDRICH, K., ZHANG, Z. & SCHLARB, A. 2005. Effects of various fillers on the sliding wear of polymer composites. Composites Science and Technology, 65 (15-16), p.2329-2343. FU, S. & WANG, Y. 2009. Tension testing of polycarbonate at high strain rates. Polymer testing, 28 (7), p.724-729. FU, S. Y., FENG, X. Q., LAUKE, B. & MAI, Y. W. 2008. Effects of particle size, particle/matrix interface adhesion and particle loading on mechanical properties of particulate-polymer composites. Composites Part B: Engineering, 39 (6), p.933-961. GALESKI, A. 2003. Strength and toughness of crystalline polymer systems. Progress in Polymer Science, 28 (12), p.1643-1699. GARCıA-LÓPEZ, D., PICAZO, O., MERINO, J. & PASTOR, J. 2003. Polypropylene–clay nanocomposites: effect of compatibilizing agents on clay dispersion. European Polymer Journal, 39 (5), p.945950. GÓMEZ-DEL RÍO, T., RODRÍGUEZ, J. 2010. Compression yielding of polypropylenes above glass transition temperature. European Polymer Journal, 46 (6), p.1244-1250. GORHAM, D. 1980. Measurement of stress-strain properties of strong metals at very high rates of strain. Mechanical properties at high rates of strain, 1979, p.16-24. GOU, J., MINAIE, B., WANG, B., LIANG, Z. & ZHANG, C. 2004. Computational and experimental study of interfacial bonding of single-walled nanotube reinforced composites. Computational Materials Science, 31 (3), p.225-236.

207

GRAFF, K. F. 1991. Wave motion in elastic solids.London: Dover Publication Inc. , 431-482. GRANTHAM, S., SIVIOUR, C., PROUD, W., WALLEY, S. & FIELD, J. 2003. Speckle measurements of sample deformation in the split Hopkinsons pressure bar. J. Phys. IV France 110, p.405-410. GRAY III, G., IDAR, D., BLUMENTHAL, W., CADY, C. & PETERSON, P. Year. High-and low-strain rate compression properties of several energetic material composites as a function of strain rate and temperature. In: 11th International Detonation Symposium Aug. 31-Sep 4, 1998 1998 Snowmass Village, CO. Los Alamos National Lab., NM (United States). GUNAWAN, L., DIRGANTARA, T. & PUTRA, I. S. 2011. Development of a Dropped Weight Impact Testing Machine. International Journal of Engineering & Technology, 11 (6), p.120-126. GUO, Y. & LI, Y. 2007. Quasi-static/dynamic response of SiO2-epoxy nanocomposites. Materials Science and Engineering: A, 458 (1-2), p.330-335. GUPTA, P., WILKES, G. L., SUKHADIA, A. M., KRISHNASWAMY, R. K., LAMBORN, M. J., WHARRY, S. M., TSO, C. C., DESLAURIERS, P. J., MANSFIELD, T. & BEYER, F. L. 2005. Does the length of the short chain branch affect the mechanical properties of linear low density polyethylenes? An investigation based on films of copolymers of ethylene/1-butene, ethylene/1hexene and ethylene/1-octene synthesized by a single site metallocene catalyst. Polymer, 46 (20), p.8819-8837. HAMOUDA, A. & HASHMI, M. 1998. Testing of composite materials at high rates of strain: advances and challenges. Journal of Materials Processing Technology, 77 (1-3), p.327-336. HAO, X., GAI, G., LU, F., ZHAO, X., ZHANG, Y., LIU, J., YANG, Y., GUI, D. & NAN, C.-W. 2005. Dynamic mechanical properties of whisker/PA66 composites at high strain rates. Polymer, 46 (10), p.3528-3534. 208

HARPER, C. A. 2002. Handbook of plastics, elastomers, and composites.New York: McGraw-Hill Professional, Inc. HAUGOU, G., MARKIEWICZ, E. & FABIS, J. 2006. On the use of the non direct tensile loading on a classical split Hopkinson bar apparatus dedicated to sheet metal specimen characterisation. International Journal of Impact Engineering, 32 (5), p.778-798. HAUSER, F. 1966. Techniques for measuring stress-strain relations at high strain rates. Experimental Mechanics, 6 (8), p.395-402. HAWARD, R. N. 1993. Strain hardening Macromolecules, 26 (22), p.5860-5869.

of

thermoplastics.

HOBEIKA, S., MEN, Y. & STROBL, G. 2000. Temperature and strain rate independence of critical strains in polyethylene and poly (ethyleneco-vinyl acetate). Macromolecules, 33 (5), p.1827-1833. HOGGATT, C. & RECHT, R. 1969. Stress-strain data obtained at high rates using an expanding ring. Experimental Mechanics, 9 (10), p.441-448. HOKKA, M., KOKKONEN, J., SEIDT, J., MATRKA, T., GILAT, A. & KUOKKALA, V. T. 2012. High Strain Rate Torsion Properties of Ultrafine-Grained Aluminum. Experimental Mechanics, p.1-9. HOPKINSON, B. 1914. A method of measuring the pressure produced in the detonation of high explosive or by the impact of bullets. Proc. R. Soc. Lond. A, 213 (89), p.437-456. HOPKINSON, J. 1872. On the rupture of iron wire by a blow. Proceeding of the Manchester Literary and Philosophical Society (2nd ed.), 11, p.40-45. HOSUR, M., ALEXANDER, J., VAIDYA, U. & JEELANI, S. 2001. High strain rate compression response of carbon/epoxy laminate composites. Composite structures, 52 (3), p.405-417. 209

HSIAO, H., DANIEL, I. & CORDES, R. 1999. Strain rate effects on the transverse compressive and shear behavior of unidirectional composites. Journal of Composite Materials, 33 (17), p.1620-1642. HSU, T. T. P. & LANGER, R. 1985. Polymers for the controlled release of macromolecules: Effect of molecular weight of ethylene- vinyl acetate copolymer. Journal of biomedical materials research, 19 (4), p.445-460. ISLEY, S. L. & PENN, R. L. 2006. Relative brookite and anatase content in sol-gel-synthesized titanium dioxide nanoparticles. The Journal of Physical Chemistry B, 110 (31), p.15134-15139. JAVNI, I., ZHANG, W., KARAJKOV, V., PETROVIC, Z. & DIVJAKOVIC, V. 2002. Effect of nano-and micro-silica fillers on polyurethane foam properties. Journal of cellular plastics, 38 (3), p.229. JEONG, S., YEO, S. & YI, S. 2005. The effect of filler particle size on the antibacterial properties of compounded polymer/silver fibers. Journal of Materials Science, 40 (20), p.5407-5411. JI, X. L., JING, J. K., JIANG, W. & JIANG, B. Z. 2002. Tensile modulus of polymer nanocomposites. Polymer Engineering & Science, 42 (5), p.983-993. JOHNSON, T., SARVA, S. & SOCRATE, S. 2010. Comparison of Low Impedance Split-Hopkinson Pressure Bar Techniques in the Characterization of Polyurea. Experimental Mechanics, 50 (7), p.931-940. JORDAN, J. L., SIVIOUR, C. R., FOLEY, J. R. & BROWN, E. N. 2007. Compressive properties of extruded polytetrafluoroethylene. Polymer, 48 (14), p.4184-4195. JOSEPH, P. V., MATHEW, G., JOSEPH, K., GROENINCKX, G. & THOMAS, S. 2003. Dynamic mechanical properties of short sisal fibre reinforced polypropylene composites. Composites Part A: Applied Science and Manufacturing, 34 (3), p.275-290. 210

KAISER, M. A. 1998. Advancements in the split Hopkinson bar test. MSc thesis, Virginia Polytechnic Institute and State University. KARGER-KOCSIS, J. & CZIGÁNY, T. 2000. Strain rate dependence of the work of fracture response of an amorphous poly(ethylenenaphthalate) (PEN) film. Polymer Engineering & Science, 40 (8), p.1809-1815. KARIAN, H. G. 2003. Handbook of polypropylene and polypropylene composites.New York: Marcel Dekker Inc., 1-10. KATZ, H. S. & MILEWSKI, J. V. 1987. Handbook of fillers for plastics.New York: Van Nostrand Reinhold, 1-40. KIM, Y.-M. & PARK, J.-K. 1996. Effect of short chain branching on the blown film properties of linear low density polyethylene. Journal of Applied Polymer Science, 61 (13), p.2315-2324. KOLSKY, H. 1949. An investigation of the mechanical properties of materials at very high rates of loading. Proceedings of the Physical Society. Section B, 62, p.676. KONG, Y. & HAY, J. 2002. The measurement of the crystallinity of polymers by DSC. Polymer, 43 (14), p.3873-3878. KONG, Y. & HAY, J. 2003. The enthalpy of fusion and degree of crystallinity of polymers as measured by DSC. European Polymer Journal, 39 (8), p.1721-1727. KONTOU, E. & FARASOGLOU, P. 1998. Determination of the true stress–strain behaviour of polypropylene. Journal of materials science, 33 (1), p.147-153. KUMLUTAS, D., TAVMAN, I. H. & TURHAN ÇOBAN, M. 2003. Thermal conductivity of particle filled polyethylene composite materials. Composites Science and Technology, 63 (1), p.113-117.

211

KUO, H.-C. & JENG, M.-C. 2010. Effects of part geometry and injection molding conditions on the tensile properties of ultra-high molecular weight polyethylene polymer. Materials & Design, 31 (2), p.884893. KUSAKA, T., HOJO, M., MAI, Y.-W., KUROKAWA, T., NOJIMA, T. & OCHIAI, S. 1998. Rate dependence of mode I fracture behaviour in carbon-fibre/epoxy composite laminates. Composites Science and Technology, 58 (3–4), p.591-602. LANGE, F. F. & RADFORD, K. C. 1971. Fracture energy of an epoxy composite system. Journal of materials science, 6 (9), p.1197-1203. LAZZERI, A., THIO, Y. & COHEN, R. 2004. Volume strain measurements on CaCO3/polypropylene particulate composites: the effect of particle size. Journal of Applied Polymer Science, 91 (2), p.925-935.

LEIDNER, J. & WOODHAMS, R. 1974. The strength of polymeric composites containing spherical fillers. Journal of Applied Polymer Science, 18 (6), p.1639-1654. LENNON, A. & RAMESH, K. 1998. A technique for measuring the dynamic behavior of materials at high temperatures. International Journal of Plasticity, 14 (12), p.1279-1292. LEONG, Y. W., ABU BAKAR, M. B., ISHAK, Z. A. M., ARIFFIN, A. & PUKANSZKY, B. 2004a. Comparison of the mechanical properties and interfacial interactions between talc, kaolin, and calcium carbonate filled polypropylene composites. Journal of Applied Polymer Science, 91 (5), p.3315-3326. LEONG, Y. W., ISHAK, Z. A. M. & ARIFFIN, A. 2004b. Mechanical and thermal properties of talc and calcium carbonate filled polypropylene hybrid composites. Journal of Applied Polymer Science, 91 (5), p.3327-3336.

212

LI, X., HRISTOV, H. A., YEE, A. F. & GIDLEY, D. W. 1995. Influence of cyclic fatigue on the mechanical properties of amorphous polycarbonate. Polymer, 36 (4), p.759-765. LI, Z. & LAMBROS, J. 1999. Determination of the dynamic response of brittle composites by the use of the split Hopkinson pressure bar. Composites Science and Technology, 59 (7), p.1097-1107. LI, Z. & LAMBROS, J. 2001. Strain rate effects on the thermomechanical behavior of polymers. International journal of solids and structures, 38 (20), p.3549-3562. LIN, J., CHANG, L., NIEN, M. & HO, H. 2006. Mechanical behavior of various nanoparticle filled composites at low-velocity impact. Composite Structures, 74 (1), p.30-36. LIN, O. H. 2010. Preparation and Characterization of Modified Nanosilica/ Polypropylene Composites. Degree of Doctor of Philosophy, Universiti Sains Malaysia. LINDHOLM, U. 1964. Some experiments with the split hopkinson pressure bar*. Journal of the Mechanics and Physics of Solids, 12 (5), p.317-335. LIU, T. & BAKER, W. 1992. The effect of the length of the short chain branch on the impact properties of linear low density polyethylene. Polymer Engineering & Science, 32 (14), p.944-955. LOPES, C. M. A. & FELISBERTI, M. I. 2006. Composite of low-density polyethylene and aluminum obtained from the recycling of postconsumer aseptic packaging. Journal of Applied Polymer Science, 101 (5), p.3183-3191. MACDOUGALL, D. 1998. A radiant heating method for performing hightemperature high-strain-rate tests. Measurement Science and Technology, 9, p.1657.

213

MAGNIEZ, K., DE LAVIGNE, C. & FOX, B. L. 2010. The effects of molecular weight and polymorphism on the fracture and thermomechanical properties of a carbon-fibre composite modified by electrospun poly (vinylidene fluoride) membranes. Polymer, 51 (12), p.2585-2596. MALATYNSKI, M. & KLEPACZKO, J. 1980. Experimental investigation of plastic properties of lead over a wide range of strain rates. International Journal of Mechanical Sciences, 22 (3), p.173-183. MALLICK, P. 1993. Fiber-reinforced composites: materials, manufacturing, and design.New York: Marcel Dekker Inc., 40-120. MANCHADO, M., VALENTINI, L., BIAGIOTTI, J. & KENNY, J. 2005. Thermal and mechanical properties of single-walled carbon nanotubes–polypropylene composites prepared by melt processing. Carbon, 43 (7), p.1499-1505. MARAIS, S., TAIT, R., CLOETE, T. & NURICK, G. 2004. Material testing at high strain rate using the split Hopkinson pressure bar. Latin American Journal of Solids and Structures, 1 (3), p.219-339. MAZUKI, A. A. M., AKIL, H. M., SAFIEE, S., ISHAK, Z. A. M. & BAKAR, A. A. 2011. Degradation of dynamic mechanical properties of pultruded kenaf fiber reinforced composites after immersion in various solutions. Composites Part B: Engineering, 42 (1), p.71-76. MENARD, K. P. 1998. Dynamic Mechanical Analysis: A Practical Introduction.Boca Raton: CRC Press, 40-120. MEYERS, M. 1994. Dynamic Behaviour of Materials.New York: John Wiley & Son Inc., 23-151. MISHRA, S., SONAWANE, S. & SINGH, R. 2005. Studies on characterization of nano CaCO3 prepared by the in situ deposition technique and its application in PP- nano CaCO3 composites. Journal of Polymer Science Part B: Polymer Physics, 43 (1), p.107113. 214

MORESCO, M., ROSA, S. M. L., SANTOS, E. F. & NACHTIGALL, S. M. B. 2010. Agrofillers in polypropylene composites: A relationship between the density and the mechanical properties. Journal of Applied Polymer Science, 117 (1), p.400-408. MOSCHAKIS, T., MURRAY, B. S. & DICKINSON, E. 2005. Microstructural evolution of viscoelastic emulsions stabilised by sodium caseinate and xanthan gum. Journal of Colloid and Interface Science, 284 (2), p.714-728. MULLIKEN, A. & BOYCE, M. 2006. Mechanics of the rate-dependent elastic-plastic deformation of glassy polymers from low to high strain rates. International journal of solids and structures, 43 (5), p.1331-1356. NAIK, N., CH, V. & KAVALA, V. 2008. Hybrid composites under high strain rate compressive loading. Materials Science and Engineering: A, 498 (1-2), p.87-99. NAIK, N. K., VENKATESWARA RAO, K., VEERRAJU, C. & RAVIKUMAR, G. 2010a. Stress-strain behavior of composites under high strain rate compression along thickness direction: Effect of loading condition. Materials & Design, 31 (1), p.396-401. NAIK, N. K., YERNAMMA, P., THORAM, N. M., GADIPATRI, R. & KAVALA, V. R. 2010b. High strain rate tensile behavior of woven fabric E-glass/epoxy composite. Polymer Testing, 29 (1), p.14-22. NAKAI, K. & YOKOYAMA, T. 2008. Strain Rate Dependence of Compressive Stress-Strain Loops of Several Polymers. Journal of Solid Mechanics and Materials Engineering, 2 (4), p.557-566. NAKAMURA, Y., YAMAGUCHI, M., OKUBO, M. & MATSUMOTO, T. 1992. Effects of particle size on mechanical and impact properties of epoxy resin filled with spherical silica. Journal of Applied Polymer Science, 45 (7), p.1281-1289.

215

NEMAT-NASSER, S. 2000. In: KUHN, H. & MEDLIN, D. (eds.) ASM Handbook: Mechanical Testing and Evaluation. Ohio, USA: ASM International. NEMAT-NASSER, S., ISAACS, J. B. & STARRETT, J. E. 1991. Hopkinson techniques for dynamic recovery experiments. Proceedings of the Royal Society of London. Series A: Mathematical and Physical Sciences, 435 (1894), p.371-391. NG, C., SCHADLER, L. & SIEGEL, R. 1999. Synthesis and mechanical properties of TiO2-epoxy nanocomposites. Nanostructured Materials, 12 (1-4), p.507-510. NIELSEN, L. E. & LANDEL, R. F. 1994. Mechanical properties of polymers and composites.New York Marcel Dekker. NINAN, L., TSAI, J. & SUN, C. 2001. Use of split Hopkinson pressure bar for testing off-axis composites. International Journal of Impact Engineering, 25 (3), p.291-313. NITTA, K.-H. & MAEDA, H. 2010. Creep behavior of high density polyethylene under a constant true stress. Polymer Testing, 29 (1), p.60-65. O'BRIEN, T., HAHN, H. & LAGACE, P. 1991. Composite materials: fatigue and fracture. Astm Intl. OTHMAN, N. 2007. Characterisation and Properties of Bentonite/Polypropylene Composite. Ph.D Thesis, Universiti Sains Malaysia. OTHMAN, R., BISSAC, M., COLLET, P. & GARY, G. 2003. Testing with SHPB from quasi-static to dynamic strain rates. J. Phys. IV France 110, p.397-404. OTHMAN, R., GUÉGAN, P., CHALLITA, G., PASCO, F. & LEBRETON, D. 2009. A modified servo-hydraulic machine for 216

testing at intermediate strain rates. International Journal of Impact Engineering, 36 (3), p.460-467. PHAM, G. T., PARK, Y. B., WANG, S., LIANG, Z., WANG, B., ZHANG, C., FUNCHESS, P. & KRAMER, L. 2008. Mechanical and electrical properties of polycarbonate nanotube buckypaper composite sheets. Nanotechnology, 19, p.325705. PICU, R., VINCZE, G., OZTURK, F., GRACIO, J., BARLAT, F. & MANIATTY, A. 2005. Strain rate sensitivity of the commercial aluminum alloy AA5182-O. Materials Science and Engineering A, 390 (1-2), p.334-343. PREMALAL, H. G. B., ISMAIL, H. & BAHARIN, A. 2002. Comparison of the mechanical properties of rice husk powder filled polypropylene composites with talc filled polypropylene composites. Polymer Testing, 21 (7), p.833-839. PUKANSZKY, B. & VOROS, G. 1993. Mechanism of interfacial interactions in particulate filled composites. Composite Interfaces(The Netherlands), 1 (5), p.411-427. RASHID, E. S. A., ARIFFIN, K., AKIL, H. M. & KOOI, C. C. 2008. Mechanical and thermal properties of polymer composites for electronic packaging application. Journal of Reinforced Plastics and Composites, 27 (15), p.1573-1584. RASHID, E. S. A., RASYID, M. F. A., AKIL, H. M., ARIFFIN, K. & KOOI, C. C. 2011. Effect of ion exchange treatment on the properties of muscovite filled epoxy composite. Applied Clay Science, 52 (3), p.295-300. RAVICHANDRAN, G. & SUBHASH, G. 1994. Critical appraisal of limiting strain rates for compression testing of ceramics in a split Hopkinson pressure bar. Journal of the American Ceramic Society, 77 (1), p.263-267.

217

RICHARDSON, M. O. W. & WISHEART, M. J. 1996. Review of lowvelocity impact properties of composite materials. Composites Part A: Applied Science and Manufacturing, 27 (12), p.1123-1131. RICHETON, J., AHZI, S., DARIDON, L. & RÉMOND, Y. 2003. Modeling of strain rates and temperature effects on the yield behavior of amorphous polymers. J. Phys. IV France, 110, p.39-44. RICHETON, J., AHZI, S., DARIDON, L. & RÉMOND, Y. 2005. A formulation of the cooperative model for the yield stress of amorphous polymers for a wide range of strain rates and temperatures. Polymer, 46 (16), p.6035-6043. RICHETON, J., AHZI, S., VECCHIO, K., JIANG, F. & ADHARAPURAPU, R. 2006. Influence of temperature and strain rate on the mechanical behavior of three amorphous polymers: Characterization and modeling of the compressive yield stress. International journal of solids and structures, 43 (7-8), p.2318-2335. RIETSCH, F. & BOUETTE, B. 1990. The compression yield behaviour of polycarbonate over a wide range of strain rates and temperatures. European Polymer Journal, 26 (10), p.1071-1075. RONG, Z. & SUN, W. 2012. Experimental and numerical investigation on the dynamic tensile behavior of ultra-high performance cement based composites. Construction and Building Materials, 31, p.168-173. ROSEN, S. L. 1982. Fundamental principles of polymeric materials.New York: Wiley Interscience 48-59. ROTHON, R. N. 2003. Particulate-filled polymer composites.United Kingdom: Rapra Technology Limited, 53-148. RUSU, M., SOFIAN, N. & RUSU, D. 2001. Mechanical and thermal properties of zinc powder filled high density polyethylene composites. Polymer Testing, 20 (4), p.409-417.

218

SALISBURY, C. 2001. Spectral analysis of wave propagation through a polymeric hopkinson bar. MSc Thesis, University of Waterloo. SANPO, N. Year. Anti-Bacterial Property of Cold Sprayed ZnO-Al Coating. In: BioMedical Engineering and Informatics, BMEI 27-30 May 2008 2008. IEEE, 488-491. SARANG, S. & MISRA, R. 2004. Strain rate sensitive behavior of wollastonite-reinforced ethylene–propylene copolymer composites. Materials Science and Engineering: A, 381 (1), p.259-272. SARVA, S., DESCHANEL, S., BOYCE, M. & CHEN, W. 2007. Stressstrain behavior of a polyurea and a polyurethane from low to high strain rates. Polymer, 48 (8), p.2208-2213. SASSO, M., NEWAZ, G. & AMODIO, D. 2008. Material characterization at high strain rate by Hopkinson bar tests and finite element optimization. Materials Science and Engineering: A, 487 (1-2), p.289-300. SHERGOLD, O. A., FLECK, N. A. & RADFORD, D. 2006. The uniaxial stress versus strain response of pig skin and silicone rubber at low and high strain rates. International Journal of Impact Engineering, 32 (9), p.1384-1402. SIVIOUR, C., GIFFORD, M., WALLEY, S., PROUD, W. & FIELD, J. 2004. Particle size effects on the mechanical properties of a polymer bonded explosive. Journal of materials science, 39 (4), p.1255-1258. SONG, B. & CHEN, W. 2005. Split Hopkinson pressure bar techniques for characterizing soft materials. Latin American Journal of Solids and Structures, 2 (2), p.113-152. STACHURSKI, Z. 1997. Deformation mechanisms and yield strength in amorphous polymers. Progress in Polymer Science, 22 (3), p.407474.

219

STRONG, A. B. 2002. History of Composite Materials-opportunities and necessities. [Accessed 9 May 2012]. SUMITA, M., SHIZUMA, T., MIYASAKA, K. & ISHIKAWA, K. 1983. Effect of reducible properties of temperature, rate of strain and filler content on the tensile yield stress of nylon 6 composites filled with ultrafine particles. Journal of macromolecular science. Physics, 22 (4), p.601-618. SUWANPRATEEB, J. 2000. Calcium carbonate filled polyethylene: correlation of hardness and yield stress. Composites Part A: Applied Science and Manufacturing, 31 (4), p.353-359. SVOBODA, P., ZENG, C., WANG, H., LEE, L. J. & TOMASKO, D. L. 2002. Morphology and mechanical properties of polypropylene/organoclay nanocomposites. Journal of Applied Polymer Science, 85 (7), p.1562-1570. TANG, L., TAM, K., YUE, C., HU, X., LAM, Y. & LI, L. 2001. Influence of the molecular weight of ethylene vinyl acetate copolymers on the flow and mechanical properties of uncompatibilized polystyrene/ethylene–vinyl acetate copolymer blends. Polymer International, 50 (1), p.95-106. TAYLOR, G. 1946. The testing of materials at high rates of loading. J. Inst. Civil Engrs, 26 (8), p.486-519. THOSTENSON, E., LI, C. & CHOU, T. 2005. Nanocomposites in context. Composites Science and Technology, 65 (3-4), p.491-516. TJONG, S. & XU, S. 2001. Ternary polymer composites: PA6, 6/maleated SEBS/glass beads. Journal of Applied Polymer Science, 81 (13), p.3231-3237. TROITSKII, B., TROITSKAYA, L., YAKHNOV, A., LOPATIN, M. & NOVIKOVA, M. 1997. Retardation of thermal degradation of PMMA and PVC by C60. European Polymer Journal, 33 (10), p.1587-1590. 220

TROTIGNON, J. P., SANSCHAGRIN, B., PIPERAUD, M. & VERDU, J. 1982. Mechanical properties of mica-reinforced polypropylene composites. Polymer Composites, 3 (4), p.230-238. VAN SLIGTENHORST, C., CRONIN, D. S. & WAYNE BRODLAND, G. 2006. High strain rate compressive properties of bovine muscle tissue determined using a split Hopkinson bar apparatus. Journal of Biomechanics, 39 (10), p.1852-1858. VECCHIO, K. S. & JIANG, F. 2007. Improved Pulse Shaping to Achieve Constant Strain Rate and Stress Equilibrium in Split-Hopkinson Pressure Bar Testing. Metallurgical and Materials Transactions A, 38 (11), p.2655-2665. WALLEY, S. M. & FIELD, J. E. 1994. Strain Rate Sensitivity of Polymers in Compression from Low to High Rates. DYMAT Journal, 1 (3), p.211-277. WANG, L., LABIBES, K., AZARI, Z. & PLUVINAGE, G. 1994. Generalization of split Hopkinson bar technique to use viscoelastic bars. International Journal of Impact Engineering, 15 (5), p.669686. WANG, S., LIANG, Z., PHAM, G., PARK, Y. B., WANG, B., ZHANG, C., KRAMER, L. & FUNCHESS, P. 2007. Controlled nanostructure and high loading of single-walled carbon nanotubes reinforced polycarbonate composite. Nanotechnology, 18, p.095708. WEHMER, P. 2008. High Strain Rate Characteristics of Rubber Modified Syntactic Foams. MSc thesis, University of Arkansas. WHITE, J. 1981. On internal stress and activation volume in polymers. Journal of Materials Science, 16 (12), p.3249-3262. WOLDESENBET, E. & PETER, S. 2009. Volume fraction effect on high strain rate properties of syntactic foam composites. Journal of materials science, 44 (6), p.1528-1539.

221

WOLDESENBET, E. & VINSON, J. R. Year. Effect of specimen geometry in high-strain-rate testing of graphite/epoxy composites. In: 38th AIAA/ASME/ASCE/AHS/ASC. Struc. Dyn. and Matl. Conf., 1997. 927-934. WOOD-ADAMS, P. M. 2001. The effect of long chain branches on the shear flow behavior of polyethylene. Journal of Rheology, 45, p.203. WOOD-ADAMS, P. M., DEALY, J. M., DEGROOT, A. W. & REDWINE, O. D. 2000. Effect of Molecular Structure on the Linear Viscoelastic Behavior of Polyethylene. Macromolecules, 33 (20), p.7489-7499. WU, X. & GORHAM, D. 1997. Stress Equilibrium in the Split Hopkinson Pressure Bar Test. Le Journal de Physique IV, 7 (C3), p.3-3. YANG, L. M. & SHIM, V. P. W. 2005. An analysis of stress uniformity in split Hopkinson bar test specimens. International Journal of Impact Engineering, 31 (2), p.129-150. YI, F., ZHU, Z., ZU, F., HU, S. & YI, P. 2001. Strain rate effects on the compressive property and the energy-absorbing capacity of aluminum alloy foams. Materials Characterization, 47 (5), p.417422. YOKOYAMA, T. 2003. Impact Tensile Stress–strain Characteristics of Wrought Magnesium Alloys. Strain, 39 (4), p.167-175. ZABASHTA, Y. F. 1974. On the dislocation mechanism of polymer deformation: Defects of spiral molecules. Mechanics of Composite Materials, 10 (4), p.493-496. ZAMAN, H. U., HUN, P. D., KHAN, R. A. & YOON, K. B. 2012. Morphology, mechanical, and crystallization behaviors of micro-and nano-ZnO filled polypropylene composites. Journal of Reinforced Plastics and Composites, 31 (5), p.323-329.

222

ZEBARJAD, S. M. & SAJJADI, S. A. 2008. On the strain rate sensitivity of HDPE/CaCO3 nanocomposites. Materials Science and Engineering: A, 475 (1-2), p.365-367. ZHANG, Q., TIAN, M., WU, Y., LIN, G. & ZHANG, L. 2004. Effect of particle size on the properties of Mg(OH)2-filled rubber composites. Journal of Applied Polymer Science, 94 (6), p.2341-2346. ZHAO, C., QIN, H., GONG, F., FENG, M., ZHANG, S. & YANG, M. 2005. Mechanical, thermal and flammability properties of polyethylene/clay nanocomposites. Polymer Degradation and Stability, 87 (1), p.183-189. ZHAO, H. 1997. Testing of polymeric foams at high and medium strain rates. Polymer Testing, 16 (5), p.507-516. ZHAO, H. 1998. A study of specimen thickness effects in the impact tests on polymers by numeric simulations. Polymer, 39 (5), p.1103-1106. ZHAO, H. & GARY, G. 1997. A new method for the separation of waves. Application to the SHPB technique for an unlimited duration of measurement. Journal of the Mechanics and Physics of Solids, 45 (7), p.1185-1202. ZHAO, H. & LI, R. K. Y. 2006. A study on the photo-degradation of zinc oxide (ZnO) filled polypropylene nanocomposites. Polymer, 47 (9), p.3207-3217. ZHU, Z. K., YANG, Y., YIN, J. & QI, Z. N. 1999. Preparation and properties of organosoluble polyimide/silica hybrid materials by sol– gel process. Journal of Applied Polymer Science, 73 (14), p.29772984. ZUIDERDUIN, W. C. J., WESTZAAN, C., HUÉTINK, J. & GAYMANS, R. J. 2003. Toughening of polypropylene with calcium carbonate particles. Polymer, 44 (1), p.261-275.

223

APPENDICES Appendix A (List of journal publication) International Journal Publications:

1) Omar MF, Akil HM, Ahmad ZA. Measurement and prediction of compressive properties of polymers at high strain rate loading. Materials & Design. 2011;32(8-9):4207-4215. 2) Omar MF, Akil HM, Ahmad ZA. Effect of molecular structures on dynamic compression properties of polyethylene. Materials Science and Engineering: A. 2012;538(0):125-134. 3) Omar MF, Akil HM, Ahmad ZA. Static and dynamic compressive properties of mica/polypropylene composites. Materials Science and Engineering: A. 2011;528(3):1567-1576. 4) Omar MF, Akil HM, Ahmad ZA. Mechanical properties of nanosilica/polypropylene composites under dynamic compression loading. Polymer Composites. 2011;32(4):565-575. 5) Omar MF, Akil HM, Ahmad ZA. Particle Size - Dependent on the Static and Dynamic Compression Properties of Polypropylene/Silica Composites, Material & Design, Accepted, In Press. DOI: http://dx.doi.org/10.1016/j.matdes.2012.09.026 6) Mohd FO, Akil H, Ahmad ZA. Effect of Particle Sizes on Rate Sensitivity and Dynamic Mechanical Properties of Polypropylene/Silica (PP/SiO2) Nanocomposites. Advanced Materials Research. 2012;364:181-185. 7) Omar MF, Akil HM, Ahmad ZA, Mahmud S. The effect of loading rates and particle geometry on compressive properties of polypropylene/zinc oxide nanocomposites: Experimental and numerical prediction. Polymer Composites. 2012;33(1):99-108.

224

8) Omar MF, Akil HM, Ahmad ZA. Measurement on Strain Rate Sensitivity and Dynamic Mechanical Properties of Various Polymeric Materials. Key Engineering Materials. 2011;471:385-390. 9) Akil HM, Omar MF, Ahmad ZA. High-strain-rate behavior of polymer matrix nanocomposites. Plastic Research Online: DOI: 10.1002/spepro.003612. 10) Omar MF, Akil HM, Ahmad ZA. Static and Dynamic Compressive Properties of Polypropylene/Zinc Oxide nanocomposites, Mechanics of materials, Under review.

National Journal Publications: 1) Akil HM, Omar MF, Ahmad ZA. Measurement of dynamic mechanical properties of nanosilica/polypropylene composites using split Hopkinson pressure bar apparatus (SHPBA), Journal of Industrial Technology (SIRIM), 2009;19(1):31-40.

225

Static and Dynamic Mechanical Properties of

Static and Dynamic Mechanical Properties of

Suggest Documents

Static and Dynamic Mechanical Properties of Long-Span Cable

Dynamic and Static Mechanical Properties of Crosslinked ... - MDPI

static and dynamic mechanical analysis of chemically

Mechanical properties of rock under coupled static-dynamic loads

DYNAMIC AND STATIC PROPERTIES OF ... - Semantic Scholar

Static and Dynamic Properties of Synaptic ... - BioMedSearch

Dynamic Mechanical Properties of Resilin

Dynamic Mechanical Properties of Resilin

Dynamic mechanical properties of hydroxyapatite

Static Mechanical Properties of Materials - CEIB

Studies on mechanical, thermal and dynamic mechanical properties of

Quasi-static and dynamic mechanical performance of glass microsphere

Static and Dynamic Mechanical Behaviors of Electrostatic ... - MDPI

Art. 36 Mechanical, dynamic mechanical and morphological properties ...

Dynamic Mechanical Properties of Aramid Fabrics ...

Dynamic Mechanical Properties of Resilin - Biomedical Engineering ...

Dynamic Mechanical Properties of ABS Material ... - CiteSeerX

Dynamic mechanical properties of human lenses

dynamic mechanical properties of crosslinked natural rubber ...

Dynamic mechanical properties of decorative papers impregnated

Static and dynamic properties of laterally coupled DFB ... - EE, Technion

Static and dynamic properties of vortex pairs in asymmetric ...

Static and dynamic properties of the interface between a polymer ...

New insights in static and dynamic properties of Gibbs ...