WI’~."4%ag

4

“ '-‘ mains-"ms”

I“!

‘ "v «31‘ .4) j.

‘__¢—.

 

 

fa“
.

 

 

.3.
1'

 

h? -

#44,,- dg " .
’:,ig!§{v' :tb'fgi t ‘ ’ I’m
H,‘ ' 1

W1
:9
1132". . ‘ $5
:5“ “ Li I ,
'1. <

. "I
if! *1
(g: .-
gi‘éa
{FD <

‘91
y:

P.
a"
"1‘?" ' 51‘ .1?" . t ‘
5 L " - l ', . 1 -
#5315 2‘51 53:4; .1 , 13" 1" it V; ‘ 4:53. ‘f
0 1- f
"' ‘qf‘k‘ pf ' J}?

 

 

 

 

 

ice 3
57016523
LIB RARY
Michigan State
U niversity
This is to certify that the
dissertation entitled

EFFECT OF SURFACE CHEMISTRY ON THE INTERFACIAL
ADHESION AND MECHANICAL PROPERTIES OF
NATURAL FIBER REINFORCED POLYMER COMPOSITES

presented by

Guangda Shi

has been accepted towards fulfillment
of the requirements for

Ph . D . degree in Chemical Engineering

 

i//fl 'f’ng) ‘£?
/ ‘ / ~ .
CyéibnétAzUl/'/, Cfflflg, ‘

Major professor U

 

Date January 16, 2003

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

PLACE IN RETURN Box to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDue.p65—p.15

 

EFFECT OF SURFACE CHEMISTRY ON THE INTERFACIAL ADHESION AND
MECHANICAL PROPERTIES OF NATURAL FIBER REINFORCED POLYMER
COMPOSITES
By

Guangda Shi

A DISSERTATION

Submitted to
Michigan State University
in partial fiIlfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemical Engineering and Materials Science

2003

ABSTRACT
EFFECT OF SURFACE CHEMISTRY 0N THE INTERFACIAL ADHESION AND
MECHANICAL PROPERTIES OF NATURAL FIBER REINFORCED POLYMER
COMPOSITES
By

Guangda Shi

The fiber-matrix adhesion between cellulose fiber and polymer matrix was
studied to understand the interfacial bonding mechanism in a natural fiber reinforced
composite system. The effects of fiber surface modification, both physical and chemical,
on the interfacial adhesion were observed and quantified using AFM and XPS analysis.
Both natural and regenerated cellulose fiber were utilized in this research to improve the
applicability of the experimental results.

Various fiber and matrix modifications were used to alter the interfacial chemistry
between cellulose fibers and polymer matrix, hence changing the bonding characteristic
at the composite interface. It’s concluded that the matrix properties, such as crystallinity
and matrix shear modulus can have an effect on the interfacial adhesion when interfacial
chemistry is kept the same. For the same matrix material, chemical bonding at the fiber-
matrix interface is much more dominant than physical interlocking mechanism with
respect to fiber-matrix adhesion. Whenever covalent bonding is introduced to the fiber-
matrix interface, the interfacial adhesion improved.

Unidirectional natural fiber reinforced polypropylene and epoxy composite were
also fabricated to gain an insight on how the fiber-matrix adhesion affects the composite

mechanical properties. Strong interfacial adhesion usually improves the tensile properties

of the composites. However, for the henequen and polypropylene system studied, the
impact strength can be lowered by the improvement in the interfacial adhesion. Voids and
processing condition can also have an impact on the interfacial adhesion and composite
mechanical performance. Therefore, in selecting surface treatment, an optimized
approach must be taken in conjunction with chemical and physical properties of the

materials.

To my loving family,

and those who believed in me

iv

ACKNOWLEGMENT

First and foremost I would like to thank my advisor Professor L. T. Drzal for all
his encouragement, help and support, without which this work would not have been
possible. I have benefited in many ways from his considerable experience and
knowledge. I would also like to thank my committee members Professor K. Jayaraman
and R. Narayan of the Department of Chemical Engineering and Materials Science,
Professor G. Baker of the Department of Chemistry, and Professor D. P. Kamdem of the
Department of Forestry for their valuable suggestions. I would also like to express my
gratitude to Dr. P. Askeland and Dr. R Schalek of the Composite Materials and
Structures Center at Michigan State University for their assistance and discussions.
Thanks are also due to all the staff and students at the Composite Materials and
Structures Center for their support.

Finally, as the most important people in my life, I acknowledge with love the

support of my wife, Tonhi Bui.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. ix
LIST OF FIGURES ............................................................................................................. x
CHAPTER 1
INTRODUCTION ............................................................................................................... 1
1.1 Motivations ........................................................................................................ 1
1.2 Problem Statement ............................................................................................. 2
1.3 Research Hypothesis .......................................................................................... 4
CHAPTER 2
BACKGROUND ................................................................................................................. 7
2.1 Natural Cellulose Fiber ...................................................................................... 7
2.1.1 PhysicaL Chemical and Mechanical Properties of Natural Fibers ...... 7
2.1.2 Relation between Scale and Properties ............................................... 9
2.1.3 Comparison between Natural Fiber and Other Fibers ...................... 11
2.2 Regenerated Cellulose Fiber ............................................................................ 12
2.3 Natural Fiber Composite .................................................................................. 13
2.4 Modification of Natural Fiber .......................................................................... 14
2.4.1 Physical Methods of Modification .................................................... 15
2.4.2 Chemical Methods of Modification .................................................. 15
2.5 Processing of Natural Fiber Reinforced Plastics ............................................. 16
2.6 Interfacial Adhesion vs. Mechanical Properties of Composite Mataials ...... 18
2.7 Examples of Past Research .............................................................................. 21
2.7.1 Researches on Natural Fiber Surface Characterization .................... 21
2.7.2 Researches on Natural Fiber Reinforced Composites ...................... 22
2.7.3 New Research Challenges ................................................................ 23
CHAPTER 3
EXPERIMENTAL MATERIALS AND TECHNIQUES ................................................. 25
3.1 Experimental Materials .................................................................................... 25
3.1.1 Regenerated Cellulose Fibers — Tencel ............................................. 25
3.1.2 Henequen Fiber (Sisal Hemp) ........................................................... 27
3.1.3 Thermoplastic Polymer Matrices ...................................................... 30
3.1.4 Thermoset Polymer Matrix ............................................................... 31
3.1.5 Silane Coupling Agents .................................................................... 32
3.2 Analytical Techniques ..................................................................................... 33
3.2.1 Environmental Scanning Electron Microscopy (ESEM) .................. 33
3.2.2 X-ray Photoelectron Spectroscopy (XPS) ........................................ 34
3.2.3 Atonric Force Microscopy (AFM) .................................................... 36
3.2.4 Thermogravimetric Analyzer (TGA) ................................................ 38
3.2.5 Micro Video Caliper ......................................................................... 39

vi

3.2.6 Tensile Strength Tester -~ United Test System (UT S) ...................... 40

3.2.7 Microbond Test ................................................................................. 41
3.3 Plasma Treatment of Cellulose Fiber ............................................................... 42
3.3.1 Plasma Treatment Theory ................................................................. 42
3 .3 .2 Plasma Treatment Mechanism .......................................................... 44
3 .3 .3 Oxygen Plasma Treatment ................................................................ 44
3.3.4 Application of Plasma in This Research ........................................... 45
CHAPTER 4
ADHESION STUDY USING TENCEL AND EPOXY ................................................... 47
4.1 Microbond Sample Preparation ....................................................................... 47
4.2 Effect of Plasma Modification of Tencel Fiber ............................................... 48
4.2.1 Oxygen Plasma Treatment Procedures ............................................. 48
4.2.2 Mechanical Properties of Oxygen Plasma Treated Tencel Fiber ..... 49
4.2.3 Surface Topology of Plasma Treated T encel Fiber .......................... 51
4.2.4 Chemical Properties of Plasma Treated Tencel FiberSurface .......... 52
4.2.5 Interfacial Adhesion between Plasma Treated Tencel Fiber and
Epoxy ................................................................................................ 54
4.3 Effect of Silane Treatment of Tencel Fiber ..................................................... 55
4.3.1 Verification of Silane Reaction with Cellulose Fiber ....................... 56
4.3.2 Proposed Silane Reaction Chemistry with Cellulose Fiber .............. 59
4.3.3 Silane Treatment Procedures ............................................................ 60
4.3.4 Mechanical Properties of Silane Treated Fiber ................................. 61
4.3.5 Surface Topology of Silane Treated Tencel Fiber ............................ 61
4.3.6 Chemical Properties of Silane Treated Tencel Fiber Surface ........... 62
4.3.7 Interfacial Adhesion between Silane Treated Tencel Fiber and
Epoxy ................................................................................................ 63
4.3.8 Better Wetting of Silane Treated Tencel Fiber ................................. 64
4.4 Summary .......................................................................................................... 65
CHAPTER 5
ADHESION STUDY USING TENCEL AND POLYPROPYLENE ............................... 66
5.1 Shear-Lag Stress Transfer Model on Interfacial Adhesion .............................. 66
5.2 Microbond Sample Preparation ....................................................................... 67
5.3 Effect of Matrix Crystallinity on the Interfacial Adhesion .............................. 68
5.4 Effect of the Matrix Mechanical Properties on the Interfacial Adhesion ........ 70
5.5 Effect of Reactive PP Matrices on Interfacial Adhesion ................................. 72
5.5.1 Proposed maPP Reaction Chemistry with Cellulose Fiber ............... 72
5.5.2 Interfacial Adhesion between Tencel Fiber and maPP ..................... 73
5.6 Summary .......................................................................................................... 76
CHAPTER 6
ADHESION STUDY USING HENEQUEN AND EPOXY ............................................. 77
6.1 Microbond Sample Preparation and Testing Procedures ................................. 78
6.2 Effect of Plasma Modification of Henequen Fibers ........................................ 81
6.2.1 Plasma Treatment Procedures ........................................................... 81

vii

6.2.2 Adhesion between Plasma Treated Henequen and Epoxy ................ 82

6.2.3 Change in Debond Curves due to Oxygen Plasma Treatment .......... 83
6.2.4 Henequen Fiber Surface Topography vs. Adhesion ......................... 84
6.2.5 Henequen Fiber Surface Chemistry vs. Adhesion ............................ 89
6.2.6 Relationships between Interfacial Adhesion and OH
Concentration and O/C Ratio ............................................................ 90
6.3 Effect of Eliminating Chemical Bonding between Henequen and Epoxy ....... 91
6.3.1 Chemical Treatment Procedures ....................................................... 92
6.3.2 Effect of PO and EA Treatment on Henequen Fiber Surface
Chemistry .......................................................................................... 93
6.3.3 Effect of PO and EA Treatment on Henequen to Epoxy
Adhesion ........................................................................................... 94
6.3.4 Correlation between Adhesion and OH or O/C ................................ 96
6.4 Effect of Silane Treatment of Henequen Fiber ................................................ 97
6.4.1 Silane Treatment ............................................................................... 97
6.4.2 Adhesion between Silane Treated Henequen and Epoxy ................. 98
6.4.3 Change in Debond Curves due to Silane Treatment ......................... 99
6. 5 Plasma-Silane Treatment of Henequen Fiber ................................................ 100
6.5.1 Fiber Chemistry after Plasma-Silane Treatment ............................. 101
6.5.2 Effect of Plasma-Silane Treatment on Henequen to Epoxy
Adhesion ......................................................................................... 102
6.6 Summary ........................................................................................................ 103
CHAPTER 7
PROPERTIES OF NATURAL FIBER COMPOSITES ................................................. 105
7.1 Unidirectional Henequen Reinforced PP Composite ..................................... 105
7.1.1 Sample Preparation Techniques ...................................................... 105
7.1.2 Composite Voids Analysis .............................................................. 106
7.1.3 Efl‘ect of maPP on Henequen/PP Composite .................................. 108
7.1.4 Impact Properties of Henequen and PP Composite ........................ 110
7.1.5 Fracture Surface of the Henequen/PP Composite ........................... 110
7.2 Unidirectional Henequen Reinforced Epoxy Composite ............................... 112
7.2.1 Composite Sample Fabrication ....................................................... 112
7.2.2 Effect of Z6040 Treatment on Henequen-Epoxy Composite
Properties ........................................................................................ 1 13
7.2.3 Effect of 8-minute Plasma Treatment on Composite Properties....113
7.3 Summary ........................................................................................................ 118
CHAPTER 8
CONCLUSIONS ............................................................................................................. l 19
BIBLIOGRAPHY ............................................................................................................ 121

viii

LIST OF TABLES

Table 1 Chemical composition and structural parameters of

some natural fibers [6] ................................................................................. 8
Table 2 Mechanical properties of some natural fibers [6, 7] .................................... 8
Table 3 Literature reports of wetting characteristics of isolated

wood polymer films [43] ........................................................................... 21
Table 4 List of polypropylene matrices and their physical properties .................... 31
Table 5 Silanes used in this research and their chemical structures ....................... 33
Table 6 CDM Silane treatment conditions .............................................................. 56
Table 7 Atomic distribution on the Tencel fiber surface from XPS analysis ......... 62
Table 8 XPS analysis henequen fiber surface before and after

Z6040 Silane treatment .............................................................................. 98
Table 9 XPS analysis of henequen fiber before and after

plasma+ Silane treatment ......................................................................... 101
Table 10 Henequen—PP composite molding conditions .......................................... 106

ix

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8
Figure 9

Figure 10

Figure 11
Figure 12

Figure 13

Figure 14
Figure 15
Figure 16

Figure 17

LIST OF FIGURES

Proposed relation between interfacial adhesion and hydroxyl

concentration ................................................................................................ 6
Chemical structure of cellulose molecules: Poly-[3(1,4)-D-Glucose ........... 9
Relation between Structure, processing, components and modulus

of wood fiber [43] ...................................................................................... 10
TEM picture of cellulose microfibrils generated by

Favier and Co-workers [52] ....................................................................... 10
Comparison of specific tensile strength and modulus of

various materials [51] ................................................................................ 12
TGA analysis of heat drying water soaked Tencel fiber ............................ 17
Relation between the properties of composites and various

rule of mixtures. [59] ................................................................................. 18
Semi-empirical Halpin-Tsai Model [60] .................................................... 20
Wet spinning process diagram [58] ........................................................... 26
Process diagram of Tencel fiber production with amine

oxide recycle [3 8] ...................................................................................... 26
ESEM picture of Tencel N-lOO fiber ......................................................... 27
Picture of 3 year old henequen plant [39] .................................................. 28

Thread-like fibers exposed from the leaves of two species of
monocots in the agave family (Agavaceae): A. Bowstring Hemp

(Sansevieria trifasciata); and B. Giant Yucca (Yucca elephantipes) ......... 28
Picture of sun-drying of henequen fiber [39] ............................................. 29
ESEM Picture of henequen fiber - as received ......................................... 30
Chemical structures of epoxy resin and curing agent ................................ 32
Picture of Model 2020 ESEM from Electro Scan Corp ............................. 34

Figure 18

Figure 19

Figure 20
Figure 21
Figure 22
Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

(a) High energy photon cause surface electron emission.

(b) X-ray photon cause ejection of core electron [40] ............................... 3 5
(a) Workings of a contact mode AFM. (b) MultiMode Scanning

Probe Microscope (SPM) fi'om Digital Instrument ................................... 37
Picture of Perken-Elmer TGA used in this research .................................. 38
Picture of the CUE Micro-300 video caliper setup .................................... 39
Picture of United “SFM-20” Test System ................................................. 40
Schematic setup of microbond testing ....................................................... 41
Schematic of the surface modification of plastic in a gas plasma

reactor ........................................................................................................ 43
Sample oxygen plasma reactions with polymer surface [48] .................... 45

(a) Schematics of nricrbond sample preparation. (b) Picture of
Tencel-epoxy microbond sample ............................................................... 48

(a) Picture of Tencel fibers placed on a u-shape glass support for
plasma treatment. (b) PSOSOO plasma surface treatment system with
variable power and gas flow control from AIRCO .................................... 49

Tencel fiber diameter vs. oxygen plasma treatment time .......................... 49
Tensile strength and modulus of oxygen plasma treated Tencel fiber ....... 50
ESEM image of oxygen plasma treated Tencel fiber surface .................... 51

ESEM image of Tencel showing damage fi'om high energy
electron beam ............................................................................................. 52

Oxygen to carbon ratio (O/C) on the Tencel fiber surface vs.
oxygen plasma treatment time Obtained from XPS analysis
(275W and 0.7511an 02) ......................................................................... 53

Plot of chemical functional group concentration on the Tencel
fiber surface vs. oxygen plasma treatment time; XPS analysis of
CIS peak. (275W and 0.75Umin Oz) ......................................................... 53

Interfacial shear strength of oxygen (02) and nitrogen (N2)
plasma treated Tencel fiber with epoxy ..................................................... 54

Figure 35
Figure 36

Figure 37
Figure 38
Figure 39
Figure 40
Figure 41

Figure 42
Figure 43
Figure 44
Figure 45

Figure 46

Figure 47
Figure 48

Figure 49
Figure 50
Figure 51

Figure 52

Silicon concentration on the Tencel fiber surface after Silane

treatment (before soxhlet extraction) from XPS analysis .......................... 57
Silicon concentration on the Tencel fiber surface after Silane

treatment (after 48 hours of soxhlet extraction) from XPS analysis .......... 58
XPS analysis of 21224 and Z6076 treated Tencel fiber surface ............... 59
Proposed Silane reaction with cellulose fiber [43] ..................................... 60
Mechanical Properties of Silane treated Tencel fiber ................................. 61
ESEM image of Silane treated Tencel fiber .............................................. 62

Interfacial shear stress data of Tencel/epoxy from Microbond test ........... 63

(a) Untreated fiber with epoxy; (b) Z-6040 Silane treated fiber
with epoxy (Angle 0 is the contact angle of epoxy to Tencel fiber) .......... 64

(a) Schematics of nricrbond sample preparation. (b) Picture of
Tencel/PP microbond sample .................................................................... 67

Plot of interfacial shear stress for various PP matrices with different
processing conditions ................................................................................. 69

Typical DSC curve of slowly cooled PP sample showing

crystalline phase transitions ....................................................................... 70

Typical DSC curve of water cooled PP sample without crystalline

transition region ......................................................................................... 7O
8 - (I G

Plot of IF SS and "' "' for six PP matrices ..................................... 71

. . a -(/G .
Plot of Interfacial shear stress vs. the "' '" for vanous

PP matrices ................................................................................................. 72
Possible maPP reaction with cellulose fiber .............................................. 73
Plot of tensile and shear modulus for various PP matrices ........................ 74
Comparison of IFSS between Tencel/PP and Tencel/maPP ...................... 75

(a) Microdrop pullout sample schematic; (b) Picture of
prepared microbond sample ...................................................................... 78

xii

Figure 53

Figure 54
Figure 55
Figure 56

Figure 57

Figure 58

Figure 59

Figure 60

Figure 61

Figure 62

Figure 63

Figure 64

Figure 65

Figure 66

Figure 67

Figure 68

(a) Microbond sample testingtetup and (b) Overview of setup

with Loadcell ............................................................................................. 79
Embedment length measurement using video caliper ............................... 80
Image analysis of fiber-matrix contact perimeters ..................................... 81
Interfacial shear strength between henequen fiber and epoxy ................... 82

Force vs. extension (debond) profiles for untreated henequen
and epoxy ................................................................................................... 83

Force vs. extension (debond) profiles for oxygen plasma
treated henequen and epoxy ....................................................................... 84

AFM height images of (a) untreated, (b) 2 minutes 02 plasma treated,
(c) 4 minutes 02 plasma treated, and (d) 8 minutes 02 plasma
treated henequen fiber surface ................................................................... 85

Roughness factor of imaged henequen fiber surface fi'om AFM
image analysis ............................................................................................ 86

Surface area change of henequen fiber before and after plasma
treatment. Values indicate the difference between measured
surface area and 25 um2 ............................................................................. 86

Plot of interfacial adhesion vs. the changes in fiber surface area
after oxygen plasma treatment ................................................................... 88

Atomic concentrations on henequen fiber surface vs. oxygen
plasma treatment time (XPS) ..................................................................... 89

Surface firnctional group concentration on henequen fiber
surface vs. oxygen plasma treatment time (XPS) ...................................... 90

Correlations between interfacial Shear strength and henequen
fiber surface chemistries afier oxygen plasma treatment ........................... 91

XPS analysis of the henequen fiber surface after plasma and
chemical treatment ..................................................................................... 93

XPS analysis of hydroxyl concentration on the henequen fiber
after various treatments .............................................................................. 94

Interfacial adhesion before and after PO + EA treatment .......................... 95

xiii

Figure 69

Figure 70

Figure 71

Figure 72

Figure 73

Figure 74

Figure 75

Figure 76

Figure 77

Figure 78

Figure 79

Figure 80

Figure 81

Figure 82

Figure 83

Interfacial adhesion comparison before and after chemical treatment
by P0 and EA of plasma treated henequen ............................................... 95

Correlation between interfacial adhesion and henequen fiber surface
chemistries after 02 plasma + PO + EA treatment .................................... 96

Interfacial shear strength comparison between untreated and Silane
treated henequen fiber and epoxy matrix ................................................... 99

Force vs. extension (debond) profiles for Silane treated henequen
and epoxy ................................................................................................. 100

Interfacial shear strength data comparison for different fiber
treatments ................................................................................................. 101

Adhesion comparison between plasma and plasma-silane treated
henequen with epoxy ............................................................................... 103

Fabrication procedure (open mold) of henequen reinforced
polypropylene composite ......................................................................... 106

Determination of void content using image analysis of
composite cross-section ........................................................................... 107

Closed mold processing of henequen-PP composite —
Cross-section view ................................................................................... 108

Tensile strength of Henequen/PP and Henequen/maPP/PP
composites ............................................................................................... 108

Tensile modulus of Henequen/PP and Henequen/maPP/PP

composites ............................................................................................... 109
Impact testing result of henequen/PP and henequen/maPP/PP

composites ............................................................................................... 110
ESEM image of impact fi'acture surface of Henequen/PP composite ..... 111

ESEM image of impact fracture surface of Henequen/4%maPP/PP
composite ................................................................................................ 111

Composite fabrication mold (left) and picture of sample
composite (right) ..................................................................................... 1 13

xiv

Figure 84

Figure 85

Figure 86

Figure 87

Figure 88

Figure 89

Tensile properties of henequen fiber and epoxy composite —
Silane treatment ....................................................................................... 1 13

Typical tensile fracture surfaces of untreated henequen-
epoxy composite ...................................................................................... 114

Typical tensile fracture surfaces of Z6040 Silane treated
henequen-epoxy composite ..................................................................... 115

Tensile properties of henequen and epoxy composite —
Plasma treatment ...................................................................................... 1 16

Tensile fiacture surface of 8min 02 plasma treated henequen
and epoxy composite ................................................................................ 117

Tensile fracture surface of 8min 02 plasma treated henequen
and epoxy composite showing good adhesion between the
fiber and matrix ........................................................................................ 117

CHAPTER 1 INTRODUCTION

 

1.1 Motivation

The reinforced plastics industries essentially began in 1940 with the introduction
of glass fibers as reinforcement. In the last half century, the technology has enabled us to
put fibers into almost all structural plastics. However, most of the traditional reinforcing
fibers such as glass, aramid, and carbon fibers are non-recyclable and non-degradable.
The incineration of these fiber reinforced composites at the end of their lifetime usually
requires more energy and releases large amounts of carbon dioxide that contributes to
global warming. The new directions in solid waste management, rapid changes in
legislation, and growing worldwide interest in degradable products have led material
scientists to increase activities in the deveIOpment and design of environmentally friendly
composite materials. As a result, natural fibers are gaining attention as a reinforcing
phase in polymer matrices [1-4]. Natural fibers possess specific properties such as: (1)
low density, (2) a high energy, polar, relatively reactive surface, and (3) abundance and
low price. Some natural fibers such as ramie, pineapple and hemp have specific
mechanical properties that are close to or even better than widely used glass fiber. In
addition, the cost of recycled plastic is generally higher than their combustion so the use
of organic fillers greatly simplifies the recycling of composite by combustion. When
natural fiber reinforced composites are burned at the end of their lifetime, the amount of
carbon dioxide released to the environment by the fiber is equal to the carbon dioxide
intake during fiber growth, so the net release of carbon dioxide by the natural fiber is zero.
Following the introduction of biodegradable polymers, it’s now possible to produce

composites that are completely biodegradable or at the very least, by combining natural

fibers with synthetic polymers, composites with a substantially higher natural resource
content. However, there are still many technical difficulties to be overcome before the
use of natural fiber in composites can be widely accepted.

According to the theory of composite material, mechanical performance of a fiber
reinforced plastic composite is primarily dependent on three factors: (1) strength and
modulus of the fiber, (2) strength and chemical stability of the resin, and (3) effectiveness
of the bond between resin and fiber in transferring stress across the interface. Since
natural fibers are hydrophilic in nature, fiber dispersion and interfacial compatibility with
hydrophobic matrices are two of the most important factors in the design and
manufacture of cellulosic fiber reinforced composites. These challenges have lead to
chemically modification of natural fibers or matrix materials in order to improve fiber-
matrix compatibility and final composite properties. However, much of the past research
in this area has been focusing on improving the composite properties without truly
understand the physical and chemical factors that govern those properties. Therefore,
fundamental studies are needed to understand the underlying chemical and physical fiber-
matrix adhesion mechanisms in order to produce usable correlations and theoretical

predictive models for natural fiber composites.

1.2 Problem Statement

The theoretical modeling of composite properties is fairly well defined for glass,
carbon and other conventional fiber reinforced composites. However, with the increase
interest in natural fiber reinforced composite, a good predictive model is becoming

crucial for the development of optimized natural fiber composites. Most of the current

research published deals with the role of compatibilizers and their contribution to the
improvement in the mechanical properties of natural fiber reinforced composites.
However, the fiber-matrix bonding mechanism and contribution of physical and chemical
factors to interfacial adhesion are still not well understood.

An important aspect with respect to optimal mechanical performance of natural
fiber reinforced composites in general and durability in particular is the optimization of
the interfacial adhesion between fiber and matrix. Since the fibers and matrices are
chemically different, strong adhesion at their interface is needed for effective transfer of
stress throughout an interface. Numerous research have shown the mechanical properties
of natural fiber reinforced polymer composite properties could be improved significantly
by the pretreatment of the fiber or matrix materials or by incorporating a compatibilizer
in the fabrication process. However, it’s still unclear if the improvement in composite
mechanical properties is due to the material properties of the compatibilizer present at the
interface, the better wetting or mixing of the fiber, or due to the interfacial bonding
between the natural fiber and the matrix material. Therefore, research needs to be done to
quantify the contribution of interfacial chemistry to the fiber-matrix adhesion and how
much this adhesion can improve the composite mechanical properties.

It is the goal of this research to understand the interfacial characteristics of natural
fiber reinforced composites and quantify the contribution of chemical bonding at the
interface to the fiber-matrix adhesion. This is especially important due to the relative
reactivity of the cellulose fiber and large amount of hydrogen bonding present at the fiber
matrix interface. In order to accomplish this objective, a series of natural fiber reinforced

composites will be produced having an incremental change in the level of bonding across

the fiber-matrix interface. The interfacial strength will be varied from basic van der
Waals force to covalent bonding between the cellulose fiber and the polymer matrix.
Unidirectional natural fiber reinforced polymer composites will also be fabricated for
each fiber treatment to see how the change in the interfacial adhesion affect the
mechanical properties of the composite. Based on the interfacial adhesion and composite
mechanical property data, conclusions will be made on the effect of introducing chemical

bonding to the fiber-matrix interface.

1.3 Research Hypothesis

According to the preliminary experimental results on the interfacial shear Strength
between natural fiber and polymer matrix, the interfacial adhesion can be improved by
the introduction of chemical bonding across the fiber-matrix interfaces. Based on the
interfacial adhesion data between a synthetic cellulose (Tencel) fiber and a polypropylene
matrix, if no chemical reaction occurs at the fiber matrix interface, then the interfacial
adhesion can be modeled using H. L. Cox’s model and the interfacial shear stress is
linearly dependent on the product of the matrix strain and the square root of matrix shear
modulus [42]. When a chemical reaction is introduced at the interface, the adhesion value
can be increased dramatically. These results lead to the hypothesis that for a given fiber
and matrix system the interfacial adhesion is governed by two major factors: (1) Physical
properties of the fiber and the matrix material; (2) The amount of chemical bonding
including hydrogen bonding across the fiber matrix interface. Therefore, to model the
interfacial adhesion more accurately for a natural fiber reinforced polymer composite

system, the chemical information must be taken into account.

In order to study the effect of interfacial chemistry on the fiber-matrix adhesion,
the fiber and matrix properties should be kept constant. Then any change in the adhesion
after chemical modification of the fiber surface, must be due to the amount of the
chemical bonds formed at the fiber-matrix interface. For a natural fiber, the basic
building block is cellulose and its chemical structure suggests that the most active
reaction site is the primary hydroxyl group which is positioned off the ring structure. If
we firrther assume the concentration of this hydroxyl group on the fiber surface is
constant for the untreated fiber, then any change in the amount of reactive hydroxyl
group (due to chemical surface modification) will likely alter the interfacial adhesion
between the natural fiber and the polymer matrix, provided that the polymer can react
with the hydroxyl groups on the fiber. Then it will be possible to relate the change in the
interfacial adhesion to the amount of reactive hydroxyl group by the following simple
mathematical relation:

A 1' = c - [0H]

where c is a conversion constant.

To verify this hypothesis, it’s required to change the amount of hydroxyl group on
the fiber surface by means of plasma treatment, Silane reaction, or other chemical
reactions that will selectively react with hydroxyl group so the amount of this hydroxyl
group can be controlled at the fiber-matrix interface. If the interfacial shear strength is
measured for several levels of [OH] concentration, then the conversion constant could be

determined by plotting adhesion data against [OH] concentration (See Figure l).

Ar

Slope = c

 

 

V

[011]

Figure 1. Proposed relation between interfacial adhesion and hydroxyl concentration

If this relation holds true, then the interfacial adhesion of a natural fiber reinforced
composite system can be modeled according to the interfacial chemistry (i.e., the amount
of reactive functional group on the fiber surface). Base on this hypothesis, if the
interfacial chemistry is known for a natural fiber reinforced polymer composite system, it
will be possible to predict the fiber-matrix interaction/adhesion from the chemical
information. Furthermore, the improvement in the fiber to matrix adhesion can also
improve the tensile properties of the natural fiber reinforced composites. The increase
will be the result of better stress transfer from the matrix phase to the reinforcing fiber

phase due to higher interfacial adhesion.

CHAPTER 2 BACKGROUND

 

In this chapter, some of the basic definitions and concepts related to natural fiber
composites will be introduced, followed by summaries of past research related to natural
fiber modification, composite fabrication techniques, and commercial applications of
natural fiber composites. Then some of the problems and concerns that still need to be
addressed will be identified along with the attempts that have been made to resolve some

of these problems.

2.] Natural Cellulose Fiber

Depending on the origin, natural plant fibers can be grouped as bast (jute, banana,
flax, hemp, kenaf, mesta), leaf (pineapple, sisal, screw pine), and seed or fruit fiber (coir,
cotton, oil palm). A single fiber of all plant based natural fibers consists of several cells.
These cells are formed out of crystalline microfibrils based on cellulose, which are
connected to a complete layer, by amorphouse lignin and hemicellulose. Multiple of such
cellulose-lignin/hemicellulose layers stick together to form a multilayer composite, the
cell wall [43]. These cell walls differ in their composition and in the orientation of the
cellulose microfibrils (See Table 1).
2.1.1 Physical, Chemical and Mechanical Properties of Natural Fibers

Physical properties of natural fibers are basically influenced by their chemical
structure such as cellulose content, degree of polymerization, orientation and crystallinity,
which are affected by the plant genetic makeup, conditions during growth of plants as
well as extraction methods used. As a result, there is an enormous variability in fiber

properties (See Table 2) depending upon which part of the plant the fibers came fi'onr, the

quality of the plant and its location [5]. When compared to E-Glass fiber, some natural
cellulose fibers such as Pineapple and Ramine have very comparable mechanical

properties. However, most of the other fibers do not compare well with glass fiber.

Table 1. Chemical Composition and Structural Parameters of Some Natural Fibers [6]

 

 

Heml- Mlcro— Holsture

Fiber Fiber Cellulose ngnln mllulose Pectin Wax flbrlllerlSplral Content
Type Nana m was m m m Anglemgg) Vim
Bast Flax 71 2.2 18.6-20.6 23 1.7 10.0 10.0
Hemp 70.2-74.4 3.75.7 17.9-22.4 0.9 0.8 6.2 10.8
Jute 61-71.5 12-13 13.6-20.4 0.2 0.5 8.0 12.6
Kenaf 31-39 15-19 21.5 - - - -
Ramie 68.6-76.2 0.6-0.7 13.1-16.7 1.9 0.3 7.5 8.0
Led Sisal 67-78 8-11 10.0-14.2 10 2.0 20 11.0
PALF 70-82 5-12 - - - 14.0 11.8
Seed Cotton 82.7 - 5-7 - 0.6 - -
Fruit Coir 3643 41-45 0.15-0.25 3—4 - 41-45 8.0

 

 

 

 

TableLMechanicalPropertiesofSomeNatuaniberslGJ]

 

 

 

TonsWTle xura
Fiber Modulus Strength Elongation at Modulus

(GPa) (MP1!) break (36) (MPa)
Banana 7-20 54-754 1-4 2-5
Coir 4-6 131 -1 75 15-40 -
Cotton - 200-400 6—7 0.03-0.10
Flax 27.6 780 2-4 0.18-0.25
Hemp - 690 1.6 -
Jute 18 226 1.3 0.3-0.5
Mesta - - 1-2 0.35-0.65
Palmyrah 4-6 180-215 7-15 -
Pineapple 34-82 413-1627 0.8-1 0.24-0.40
Ramie 61 .4-128 400-938 12-3.8 -
Sisal 9-2 568-640 3-7 12.5-17.5
Sunhemp - 760 2-4 12.5-17.5
E-Glass 72 2000-3600 2.6

 

 

 

 

The chemical composition and mechanical properties of some natural fibers are
shown in Table 1 [6] and Table 2 [6, 7]. As one can see fiom the chemical composition
table, cellulose is the main component of almost all natural fibers. The elementary unit
of a cellulose macromolecule is anhydro-D-glucose, which contain three hydroxyl (OH)
groups (See Figure 2). These hydroxyl groups form hydrogen bonds inside the
macromolecules itself (intramolecular) and between other cellulose molecules

(intermolecular). Therefore all natural fibers are hydrophilic in nature.

 

 

 

Figure 2. Chemical structure dcelulose molecules: Poly-8(l,4)-D-Glucose

2.1.2 Relation between Scale and Properties

Even for the same natural fiber, such as wood fiber, the mechanical properties can
depend greatly on the scale and structure. The large wood fiber can have a tensile
modulus on the order of 10 GPa, but a crystallites that fi'om the same wood fiber can have
a tensile modulus of around 250 GPa As shown in Figure 3, the properties of natural
fiber also depend on the scale and structure of the fiber. Smaller structures usually lead to

more regular composition and less defects and hence a better mechanical properties.

Structure Process Component Modulus

10 GPa

Pulping —-—>

U

Single Pulp Fiber 40 GPa
Hydrolysis

followed by —

mechanical

disintegration "
Mlcrofibrils 70 GPa

No existing ___.,
technology

 

ii

Crysmllites 250 GPa

 

    

- _1"/‘ 5' I ,

. ., l./,‘ f W. '*>--.. H
, » ‘3 »"/\\\\' ~ ‘. .4} WOT.
s ,\- v --.-'7 . .

  
  
 
 
  

;
\\

     
     
 

    

' ;.‘"r 1.. . ', - . -—1 I
1.. x, ‘ "I \‘ '. ‘ ‘ , . ‘1'
- ' ~ \I N» V .
’ ~ .2 -~-~a‘~ j ‘ , .
‘ , ./ I «I 1
. ‘ l
,\

 

.
/ \ ~/
.\ \\ .
a x .' \
‘ t / . ‘ - l
g. A}: ,,/ '
'. I ‘. / 1 l
a _ ' /

Figure 2. TEM picture of cellulose microfibrils generated by Favier and Co-worlters [52].

However. the existing technologv can onlv take advantage of cellulose fibers in single
fiber or nricrofibril form. Even the nricrofibril from is hard to obtain due to the complex
extraction procedures and high cost of fabrication. Commercial source of microfibrils
from natural fibers are not available at this time and all the research works related to
microfibrils were done with in-house processing. A TEM picture of cellulose rrricrofibrils
extracted from tunicate by Favier et al is shown in Figure 4. Research into nano-scale
cellulose whiskers (microfibrils) reinforced composites is already taking place and could
make the natural cellulose nanofiber even more attractive as a reinforcing phase for
structural polymer matrices [53-57].
2.1.3 Comparison between Natural Fiber and Other Fibers

When comparing the mechanical properties of the cellulose fibers to other
traditional fillers, it’s easy to see why cellulose fibers are gaining more attention. As
shown in Figure 5, the mechanical properties of natural fiber are usually within the
shaded region, and due to their low densities they usually have specific tensile properties
that are close to E-glass fiber. In addition, natural fibers are becoming more favorable
because of their biodegradability and zero carbon dioxide (C02) emission during
incineration. If processing technologies advances can be achieved, the cellulose
nanowhiskers can be made available with mechanical properties surpassing some of the

traditional fiber fillers.

ll

1 0.000

1 .000

Specific
tensile
strength

(MP3) 100

10

Figure 5. Comparison ofspecific tensile strength and modulus ofvarious materials [51]

 

. Graphite

 

'sac
Ashe t
3 cs '. Kevlar
'5 @355 'Colulose
" crystalllte
e .Flamie
Linerboard 3 Spruce
Filled PEK\ 'Al
- 'Steel
0W0"
0
.pp PET OCast iron
. ‘HOPE
LDPE
1 r I r
1 10 100 1.000

Specific tensile modulus (GPe)

2.2 Regenerated Cellulose Fiber

Different natural fibers have different defects such as micro compression, pits, or
cracks. So one way to achieve the same flexibility using natural fiber is to regenerate the
cellulose and make it more uniform by dissolving the microfibrils in solvent and then
precipitating under controlled conditions. The solubility of textile fibers and the
properties of solutions of fibers are important, not only in the production process as such,
but also in the control and investigation of these processes. The ease with which a
cellulose fiber dissolves depends on the chemical nature of the groups that hold the

molecules and the closeness of the molecules to each other such as the crystallinity of the

fiber.

12

Regenerated cellulose fibers are crystalline and the degree of crystallinity depends
on the origin of the material. This crystallinity not only influences physical properties,
but also affects solubility. Cellulose and its derivatives such as cellulose acetate, provide
good examples of the influence of crystallinity on solubility. The large number of
hydroxyl groups in cellulose increases the aflinity to water. The hydroxyl groups in the
amorphous and crystalline region combine with water to give the relatively high moisture
absorption of cellulose fiber. The most satisfactory solvents for cellulose are alkaline and

a common solvent is cuprammonium solution [8].

2.3 Natural Fiber Composite

The combination of a polymer matrix and reinforcing fibers gives rise to
composite materials having the best properties of each component. Several types of
polymers have been used as matrices for natural fiber composites [9-17]. The most
commonly used are therrnoset polymers such as polyesters, epoxies and phenolics.
Thermoplastics like polyethylene (PE), polystyrene (PS), and polypropylene (PP) have
also been used. Since the polymer matrix is sofi, flexible and light weight in comparison
to fibers, their combination provides a high strength-to-weight ratio for the resulting
composite.

The properties of composites also depend on those of the individual components
and on their interfacial compatibility. It is well known in the fiber composite technology
community that the fiber-matrix interfaces give fiber composite their structural integrity.
The interface consists of the bond between fiber and matrix and the immediate region

adjacent to this bond. The interface is usually considered to be zero thickness for

13

analysis purpose. At least three types of bonding are thought to exist at the interface:
chemical, electrical, and mechanical. A composite with weak fiber matrix interface will
not be able to transfer the load fiom the matrix to the reinforcing fiber and usually leads
to poor composite strength.

A major disadvantage of cellulose fibers is their highly polar nature, which makes
them incompatible with non-polar polymers. This incompatibility usually leads to poor
dispersion of the fibers in the matrix material, poor interfacial strength and lower
mechanical performance than glass fiber composite. In addition, the poor resistance to
moisture absorption makes the use of natural fibers less attractive for exterior
applications or applications where they are exposed to a moist environment. In order to
overcome these problems, many researchers have been devising ways to modify the
natural fibers to improve compatibility with polymer matrices and to minimize water

absorption.

2.4 Modification of Natural Fiber

The stress transfer at the interface between two different phases is determined by
the degree of adhesion. Strong adhesion at the interface is needed for effective transfer of
stress and load distribution throughout the composite. This situation calls for the
development of strategies for the modification of the cellulose fiber surface, thereby
gaining control over the fiber-polymer interface. Fiber surface modification can also be
done by physical means such a fibrillation and plasma. Alternatively, in order to improve

the mechanical properties of the composite, a coating can also be applied, which

14

generally consists of coupling agents or compatibilizing agents that introduce chemical
bonds between the fiber and matrix.
2.4.1 Physical Methods of Modification

Physical methods involve surface fibrillation, electrical discharge such as corona
or plasma [18, 19] etc. Physical treatments change the chemical, structural and surface
properties of the fiber surface and thereby influence the mechanical bonding with the
matrix polymer. Surface modification by discharge treatment such as cold plasma,
sputtering and corona discharge is of great interest because of their environment friendly
nature. Plasma treatment causes mainly chemical implantation, etching, polymerization,
fi'ee radical formation and crystallization, whereas sputter etching bring about mainly
physical changes such as increases in surface roughness. Low temperature plasma is a
usefirl technique to improve the surface characteristics of the fiber and polymeric
materials through the action of the electrons, ions, radicals and excited molecules
produced by the electrical discharge. Low temperature plasma can even be generated
under atmospheric conditions in the presence of helium [19]. The action of these plasmas
involves abstraction of protons and creation of unstable radicals that convert and produce
functional groups such as alcohols, aldehydes, ketones and carboxylic acids on the fiber
surface [7].
2.4.2 Chemical Methods of Modification

Strongly polarized cellulose fibers are not inherently compatible with
hydrophobic polymers. The compatibility and dispersability of fiber and matrix can be
improved by developing a hydrophobic coating of compatible polymer on the surface of

the fiber before being mixed with polyma matrix. Pretreatment of fibers by encapsulated

15

coating with coupling agents also provides better dispersion by reducing the fiber-fiber
interaction with the formation of coating on the fiber surface.

Generally, coupling agents facilitate the optimum stress transfer at the interface
between fiber and matrix. Coupling agents are molecules possessing two firnctions. The
first is to react with hydroxyl groups of cellulose and the second is to react with
functional groups of the matrix. The most common coupling agents are Silane, isocyanate
and titanate based compounds, the chemical composition of which allows them to react
with both the fiber surface and the matrix polymer. In the case of cellulosic fiber
composite, isocyanates were found to be reliable. It is expected that the formation of a
primary type of bonds between cellulose and isocyanate and a secondary type of weak
bond between thermoplastics and isocyanate improve the mechanical properties of
natural fiber filled thermoplastics. Maleic anhydride is another well know compatbilizing
agent for the natural fiber composite. The application of maleic anhydride is usually in

the form of copolymer with the matrix material such as maleated polypropylene.

2.5 Processing of Natural Fiber Reinforced Plastics

Drying of the natural fiber is a major step in the processing of natural fiber
reinforced composites, because waster on the fiber surface acts like a separating agent in
the fiber-matrix interface [43]. Additionally, the evaporation of water during the
fabrication process can cause voids in the final composite (most of the thermosets and
thermoplastics have a processing temperature over 100 C). Both phenomena can lead to a
decrease of mechanical properties of natural fiber reinforced composite if natural fibers

were not dried before processing. According to the TGA analysis performed, the drying

16

of cellulose fiber usually takes about one hour at a temperature of 104 C (See Figure 6).
This time also includes the preheating step of heating the oven fi'om room temperature to

100 C. When large amount of fiber is to be dried in a convection oven, a 2-hour drying

 

 

 

 

 

 

 

time is used.
Drying of Cellulose Fiber -'I’GA
180 120
‘3 160 . r 100 g
E .2
v 140 ~ ~— 80
s E
.9 o.
§ 120 + ~— 60 g
2 E
g- 100 - ~4 40 g
Q
a ——Weigllt E
— _- r-
80 —Temperature 20
60 I T l I l l . O
0 20 40 60 80 100 120 140
Drying Tine (nin)

 

 

 

Ir‘igrreaTGAuualysisofheutdryiugwutcrsoukedTeI-celfiber.

To improve the performance of the natural fiber composites, physical or chemical
treatment are often applied before processing. After the fiber drying and treatment,
natural fiber composites are usually processed by compression molding or injection
molding. For the compression molding process, SMC (sheet molding compounds) and
BMC (bulk molding compounds) methods are most commonly used. The injection
molding process is usually used for thermoplastic system. Chopped natural fibers are
mixed with thermoplastic polymers in the extruder and then pelletized. These pellets are
then injection molded to usable parts in the injection molder. However, in comparison

with glass or carbon fiber, the thermal stability of natural fibers (~ 200 C) limits the

17

number of thermoplastics to be considered as matrix materials. The most common
thermoplastics matrices are polyethylene (PE) and Polypropylene (PP). Furthermore, due
to the discontinuous nature of the natural fiber, pultrusion and longer fiber conrposite are

not possible with natural fibers.

2.6 Interfacial Adhesion vs. Mechanical Properties of Composite Materials

A method for the estimation of composite material performance fiom the
characteristics of the fillers and matrices and from the configuration of the filler is
generally called the rule of mixtures. In the basic form of rule of mixtures, some
characteristics of the composite material are represented as a function of characteristics of

constituent components and their volume fraction, as shown in Figure 7.

 

    

s)
K >0.K<0 I

  

XA -----------

 

 

 

ism—re

Figure 7. Relation between the properties of composites and various nrlc of mixtures [59).

18

For a composite material (characteristics: 1C ) that consists of component A
(characteristics: [A , volume fraction: ¢A) and component B (characteristics: 13 ,
volume fraction: 413 ), the basic formulas for rule of mixtures are as follows:

Zc : ZA¢A +XB¢B

(parallel model, linear rule of mixtures, curve 1 in Figure 7) and

1 _¢.r +¢B
ZC IA 13

 

(series model, curve 2 in Figure 7). The two curves exhibit theoretical upper and lower
limits, respectively, based on a simple composite effect in general.

The rule of mixtures described above is valid for simple unidirectional composite
system with well-known structure in which the rule of additivity holds. However, it’s
natural to consider that in practice, any interaction will occur in the interface due to the
contact between A and B. Then, considering the creation of interfacial phase C, different

fi'om component A and B, the following equation can be represented:

ZC = ZA¢A + 2:13:93 +k¢A¢B

The above curve represents a quadratic curve with a maximum (k>0) or minimum (k<0)
depending the sign and the value of k (curve 4 in Figure 7). The parameter k involves an
interaction between component A and B and provides an expression of the interfacial
effect. However, there is rarely any practical application of the quadratic model because
the composite properties usually don’t exceed the prediction of the basic rule of mixtures.

A more realistic mechanical model is the Halpin-Tsai model. This model also has
an adjustable interfacial adhesion parameter E, but it has a theoretical maximum equal to

that of the basic rule of mixtures (See Figure 8).

19

1+ V
Halpin-Tsai Model: P, = P", [i]

1_ZVf
P —P
where I:[T’fL;—5§L],
m

P, , Pm, Pf : the property of composite, matrix and fiber, respectively

Vf: Volume fiaction of fiber
As shown above, the mechanical property such as tensile strength and modulus of the
natural fiber reinforced unidirectional composite could be predicted if the interfacial

adhesion factor E is known.

 

 

 

 

 

 

 

 

 

 

 

p
P
‘ "R5053" HW—
0 0.2 0.4 V3 0.6 0.8 1

Figure 8. Semi-empirical Bulpiu—Tsui Model [60].

These models clearly suggest that the properties of the interfacial phase must be

improved to obtain an excellent composite material.

20

2.7 Examples of Past Research
2.7 .1 Researches on Natural Fiber Surface Characterization

Because of the chemical and morphological complexity of the natural fiber
surface, some investigators have reported to characterized surface of cast films of isolated
natural fiber or wood polymers. In spite of the linrited applicability of these results to the
real surface of the natural fibers, the use of the isolated polymers is necessary to estimate
the individual wetting characteristics of the various natural fiber components. AS shown
in Table 3, the contact angle of water with wood polymer and critical surface tension data

of several wood polymers were reported by various researchers [43].

Table 3. literature reports of wetting characteristics of isolated wood polymer films [43]

 

 

 

 

Polymer “M12231 “I. ”finzmtm) References
Cellulose 34 35.5 Lunar at al. (1969)
33 - Borg‘n (1959)
27.8 (g 66% RH) - ' Borgin (1959)
Hernlcellulose Luner et al. (1969)
Arabinogalectan - 33
Galactoglucomannan - 36.5
Hardwood xylan - 33436.5
Soltwood xyIan - 35
Llnglnl Lee et al (1972)
Hardwood Kraft 60 36
Sollwood Kraft 58 37

 

 

 

Toussaint et al [44] reported a rapid decrease of the contact angle with water with
time for cellulose films, while for other test liquids such as glycerol, ethylene glycol and
diiodomethane a constant contact angle was obtained after 2-5minutes. Since water has a
higher polarity than the other liquids, Toussiant et a] discussed the possibility that the

decrease in the contact angle with time is due to specific interactions between water and

21

the cellulose surface allowing water to penetrate into the cellulose. Therefore, to
minimize water interaction and absorption of water in natural fiber reinforced composite,
natural fiber must be treated to stop this specific interaction.

Liu et al [45] used dynamic contact angle analysis to characterize the surface
energy of heat treated and acetylated rayon, cotton, and wood fibers. The investigation
showed a lower surface energy for the heat treated fiber than for the untreated ones. The
surface energy of acetylated fiber was 40% higher than that of heat treated fibers. In the
case of acetylated fibers the increase in surface energy is assigned to the acetyl group on
the fiber surface.

Felix et al [46] characterized the surface energy of untreated and maleated
polypropylene (maPP) treated cellulose fibers. As expected, the maPP treatment resulted
a distinctly lower polar part of the he surface energy with 4.9 — 8.4 mJ/mz, depends on
the molecular weight of maPP used. For the untreated cellulose fiber the surface energy
was around 42.2 mJ/mz.

2.7.2 Researches on Natural Fiber Reinforced Composites

It was reported that sisal/LDPE composite showed a better reinforcing effect
because of high matrix ductility and high strength/modulus ratio of sisal as compared to
that of LDPE matrix [1 l]. Nair et al. [9] reported the tensile properties of polystyrene
reinforced with short sisal fiber and benzyoylated sisal fiber. The benzylation of sisal
fiber was found to enhance the tensile properties of resulting composite. However, the
incorporation of sisal fiber considerably reduced the glass transition temperature of the
polystyrene. It was reported by Joseph et al. [11] that among polyester, epoxy and phenol

formaldehyde composites of sisal fiber, the phenolic type resin performed better as

22

matrix materials than epoxy and polyester resins with respect to tensile and flexural
properties owing to the high interfacial bonding in phenolic composite. Joseph et al. [12]
also found that the reinforcing ability of sisal fiber in PP matrix is low due to poor fiber-
matrix interaction. The flexural behavior of coir, straw and jute fibers in polyester have
been studied and analyzed by various researchers [15-17]. The potential of
sunhemp/polyester composite in terms of tensile and impact properties has been
investigated by Rohatgi and co-workers [23]. Zadorecki and Flodin [24] used cellulose
fiber in the form of paper sheet as reinforcement in unsaturated polyester composites. The
reinforcing efi'ect of sisal, jute and bamboo fiber in epoxy composites is also reported
[25-28]. Several studies have been carried out by Kokta and co-workers using
chemitherrnomechanical pulp in different thermoplastics [29, 30]. Crystallization
kinetics studies were carries out by Chen and Porter [31] on composite made of
polyethylene and kenaf fiber. The environmental performance of flax fiber mat reinforced
polypropylene was investigated by monitoring the moisture absorption and swelling and
measuring the residual mechanical properties of the samples at different moisture levels
[32]. The effect of filler content and size on the mechanical properties of PP/oil palm
wood flour composites were reported by Zainy et al [33]. Efl’ects of coupling agents on
cellulose fiber reinforced thermoplastics composite and their influence on mechanical
properties have been reported [34-36]. Varma et al [37] reported the effect of various
chemical treatments such as organotitanate, zirconate, Silane, N—substituted
mechacrylamide on the properties of sisal fiber reinforced polyester composites.

2.7.3 New Research Challenges

23

As stated in this section, there are many research papers that deal with natural
fiber surface analysis and the natural fiber reinforced composites. However, most of the
research work are application based and do not involve investigation at the microscopic
level. The question remains as to how the surface treatments affect the stress transfer
between the natural fiber and the polymer matrix and the mechanical properties of the
composites. Is the improvement in mechanical properties of the natural fiber composite
caused by better wetting of the fiber surface, the mechanical properties of the
compatibilizer, or the chemical bonding between the natural fiber and the polymer
matrices? Some fundamental research is still needed to answer those questions.

In order to better understand interfacial properties on the microscopic level, a
systematic study must be performed in which the interfacial chemistry is varied gradually.
This incremental change in the interfacial chenristry will lead to a better understanding
just how much the chemical interaction at the fiber-matrix interface can contribute to the
overall adhesion and mechanical performance of the natural fiber reinforced composite

materials.

24

CHAPTER 3 EXPERIMENTAL MATERIALS AND TECHNIQUES

 

There were two kinds of cellulose fibers used in this study. For the preliminary
and surface chemistry study, a regenerated cellulose fiber was used due to its uniformity
in physical and chemical properties. Henequen fiber was used in the later study to prove
that the chemical effect observed for the regenerated cellulose fiber holds true for natural
fiber which has more complex surface chemistry and morphology. The polymer matrices,
surface modifying chemicals and surface treatment techniques such as plasma treatment
will also be introduced. In addition, the surface analysis and adhesion testing techniques

are also described in later part of this chapter.

3.1 Experimental Materials
3.1.1 Regenerated Cellulose Fibers - Tencel

In order to determine the surface chemistry contribution to the interfacial
adhesion and to the final composite mechanical properties, the test fiber must have
uniform surface chemistry, surface topography, and mechanical properties. Physical and
chemical properties of natural fibers can vary greatly depending growing environment,
fiber source and extraction tecluriques. Therefore, the best choice for this part of the
study is to utilize regenerated cellulose fiber.

The regenerated cellulose fiber used in this research was Tencel N-100. This is a
regenerated cellulose fiber obtained from Tencel, Inc. This material is produced by first
dissolving purified wood cellulose into N-Methylmorpholine-N-Oxide solution under
heat, and then the cellulose solution is extruded (wet spinning) under controlled condition

(See Figure 9).

25

Wet treat

    
 

Filter “’3"

Wet spinning

Figure 9. Wet spinning process diagram [58].

Unlike other regenerated cellulose fiber production processes, the solvent used in
the Tencel production is recovered after the fiber Spinning process (See Figure 10). The
complete recovery of the solvent represents a significant advancement in the production

of manmade cellulose fiber due to its low impact on the environment.

Wood Pulp

   

Fiber

Figure 10. Process diagram of Tencel fiber production with amine oxide recycle [38].

26

As shown in Figure 11, Tencel fiber has smooth cotton like appearance and it’s
semitransparent. It has an average diameter of 16.5 um and an average tensile strength

and modulus of 530 MPa and 26 GPa, respectively.

 

Figure 11. ESEM picture of Tencel N-100 fiber

3.1.2 Henequen Fiber (Sisal Hemp)

Henequen (also called Sisal hemp) is one of the most important agricultural
products of the Yucatan peninsula in Mexico. Hemps such as sisal and henequen are
strong, stable and versatile fibrous plant materials that are increasingly used by the
automotive industry.

Henequen plant is a squat plant with long, knife-shaped leaves that form a rosette
close to the ground. These fleshy, rigid leaves, from which the henequen fiber is derived,

are usually grayish-green to dark green (See Figure 12). The fiber within is coarse, long

27

and extremely strong. Its color is usually creamy white to a pale yellow, but some

henequen can have a reddish cast.

 

    

l’v xr . xv

Figure 12. Picture of 3 ear old henequen plant [39H

As shown in Figure 13, a close up of the fiber leaf fiom agave family, the leaf of

sisal hemp is basically a composite of leaf fibers and connecting tissue cells.

 

Figure 13. Thread-like fibers exposed from the leaves of two species of monocots in the agave family
(Agavaceae): A. Bowstring Hemp (Sansevieria trifasciata); and B. Giant Yucca (Yucca elephantipes).

28

 

Figure 14. Picture of sun-drying of henequen fiber [39]

The process of separating the fibers from the leaves is labor intensive. The leaves
are crushed and then steamed to extract the henequen fiber from them. Next the fiber is
graded and sorted, then either hung in the sun to dry (See Figure 14) or put into a drying
machine. Then the fibers are beaten or brushed to sofien and separate them and to remove
any traces of leaf tissue. There is a 95% loss in weight from the beginning to the end of
the process. One hundred pounds of leaves yields only five pounds of henequen fiber.

Compared with other natural fibers, henequen fibers have a fairly regular shape
(See Figure 15) and mechanical properties. In addition, due to the steam fiber extraction
process, large amount of the lignin and wax that are naturally present on the fiber surface
were removed and the cellulose content of the henequen fiber is higher. The henequen
fibers used in this research have diameters range from lSOpm to 500nm. The tensile

strength and tensile modulus of henequen fiber are 330 MPa and 24 GPa, respectively.

29

 

Figure 15. ESEM Picture ofheuequen fiber— as received

3.1.3 Thermoplastic Polymer Matrices

Since part of this research is to understand the effect of interfacial bonding
strength on the final composite properties, a relative inert polymer matrix is required that
can be modified by the addition of fimctional groups. It’s also required that the polymer
matrix have a processing temperature below 200C so the cellulose won’t be degraded
during processing.

The reference matrix of choice is polypropylene. There were two types of
polypropylene used in this study, pure and maleaic anhydride modified. The pure
polypropylenes (PP) were the Profax—ph020, profaxl31, profaxlSO, profax6031,
profax6051 and profax6061 which were kindly donated by Besell. The maleaic
anhydride grafied polypropylenes (maPP) were Epolene G3003 and Epolene G3015

obtained from Eastman Chemical Company. All the polypropylenes and malice

30

anhydride modified polypropylenes along with some of their mechanical properties are

listed in Table 4.

Table 4. list of polypropylene matrices and their physicd properties

 

 

 

 

 

 

 

 

 

 

 

 

PP Type Tensile Shear Tm p No of
Modulus (GPa) Modulus (MPa) (C) (glml) MAIChain

Profax PH131 2.04 675 162 .90 O
Profax PH180 1.96 654 160 .90 O
Profax 6601 1.81 597 163 .90 O
Profax 6501 1.53 491 163 .90 O
Profax 6301 1.45 453 166 .90 0
Profax PHOZO 1.61 532 162 .90 O
Epolene (33003 1.59 520 162 .912 3.9
Epolene G3015 1.64 531 162 .913 6.6

 

 

 

 

 

3.1.4 Thermoset Polymer Matrix

Natural fibers can also be used to reinforce thermoset polymers. The commonly
used thermoset polymers are polyesters, epoxies and phenolics. In this research epoxy
matrix was used because its availability and relative compatibility with natural fiber. The
epoxy resin used in the research was DGEBA [Diglycidyl ether of bisphenol-A] with
Jeffamine T-403 (Polyoxypropylenetriamine) as curing agents. The chemical structures
of the epoxy resin and the curing agent are shown in Figure 16. After curing, the tensile

strength and tensile modulus of the epoxy matrix were around 65 MPa and 3 GPa,

respectively.

31

0 0H 0
/ \ 9‘3 l 9‘3 / \
CHZ—CH—CHZ- ('2 o—ag—crr—cm— QO—o—crg—CH—CH,
CH3 11 CH;

n=0, 1,2
DGEBA

flh'lmcflaflx'NHz
Cfizflisz‘Hz'IOCHzCHCHaHy-Nflz
CH2’[0(31"12C1*KCHE;)]z-NHz

x+y+z=~53

JEFF AMINE T403
menaCMmhMMmflnuflmmnuhuflummugm

3.1.5 Silane Coupling Agents

Silane coupling agent has proven to be very effective in increasing interfacial
interaction of glass fiber reinforced composite for many years. In order to study the
impact of interfacial chemistry on the fiber-matrix adhesion between cellulose fiber and
epoxy, silane coupling agents were used to modify the level of interfacial bonding. The
silanes used in the research are shown in Table 5 with their chemical structures.

The silanes were chosen for their chemical structures and functional groups. The
Dow Corning Z-1224 silane (Trimethylchlorosilane) was used to determine the nature of
the chemical reaction between the natural fiber and the silane molecules. Since this
silane only has one reactive group and can’t self crosslink, it will be easy to determine if
any chemical reaction is taking place on the fiber surface by XPS analysis of treated fiber
through detection of the silicon atom. The other silanes were used to modify the chemical
interaction between natural fiber and epoxy matrix. All the silanes were applied in the
form of hydrolyzed solution using acetic acid as pH control. After each of the silane

treatment, the treated cellulose fiber was examined by XPS (after soxhlet solvent

32

extraction) to see if any Silicon atoms were present on the fiber surface. Then by
comparing the XPS results of various silane treated cellulose fiber, a more conclusive

reaction mechanism can be determined base on the silane reaction efficiency.

Table 5. Silanes used in this research and their chemical structures

 

 

 

 

 

 

 

 

 

 

Silane Technical Name Chemical Structure
Dow Corning 3-Glycidoxypropyl- /O\
Z-6040 trimethoxysilane CHz—CH .CH20(CH2)3$i(OCH3)3
Dow Corning Methyl- .
Z-607O trimethoxysilane CH3S‘(OCI 13”
D°w 0°"m‘3 Trimethylchlorosilane (CH3);SiCl
Z-1224
Sigma-Aldrich Chloro—dimethyl- .
39897 methoxysilane C1(CH3)2810CH3
3.2 Analytical Techniques

3.2.1 Environmental Scanning Electron Microscopy (ESEM)

The ESEM represents one of the most exciting breakthroughs in electron
microscopy since the invention of the electron microscope. The ESEM is a practical
SEM, as it can operate at pressures ten thousand times higher than that of standard
scanning electron microscope (SEM). This gives important capabilities such as
examining unprepared and uncoated specimens that are free from surface charging and
from damage caused by preparation or by introduction into a high vacuum environment.
It has the advantage of measuring sample at moderately low pressure and it has the ability
to scan nonconductive sample such as cellulose fibers without a conductive coating.

The ESEM used in this research was a Model 2020 ESEM from Electno Scan

Corporation as shown in Figure 17. This ESEM utilizes a LaB6 electron source and is

33

capable of imaging down to nanometer range. ESEM images of the cellulose fiber will be
used to determine surface topography of the treated vs. as received cellulose fiber. This
information is very important with respect to plasma treatment of the fiber. ESEM will
also be used to observe the fracture surface of the final composite and visually determine
the interfacial bond strength by the measurement of fiber pullout. In addition, the ESEM

images can also be used to determine the void content of a composite.

    

Figure 17. Picture of Model 2020 ESEM from Electra Scan Corp.

3.2.2 X-ray Photoelectron Spectroscopy (XPS)

In this research XPS was used in the analysis of the Tencel fiber surface. Both
heated and as received fibers were analyzed for atomic concentrations and the functional
group changes on the fiber as a result of surface treatment.

XPS is an information rich method. The surface to be analyzed is first placed in a

vacuum environment and then irradiated with photons. For XPS, the photon source is in

34

the X-ray energy range. The atoms comprising surface emit electrons (photoelectrons)
after direct transfer of energy from the photon to the core-level electron (Figure 18).

(a)

 

(b)

 

01s photoejected electron 0 c

Figure 18. (a) High energy photon cause surface electron emission. (1)) X-ray photon cause ejection
of core electron [40].

These emitted electrons are subsequently separated according to energy and counted.

The energy of the photoelectron is related to the atomic and molecular environment from

which they originated. The number of the electrons emitted is related to the

concentration of the emitting atom in the sample. The most basic XPS analysis of a

surface will provide qualitative information on all the elements present (except H and He)

[40]. More sophisticated application of the method such as curve fitting can yield much

35

detailed information about the chemistry, organization, and morphology of a surface. The
probe depth is less than lOnm so this method is very surface sensitive and will be a great
tool to characterize the functional groups on the cellulose fiber surface.

The XPS used n this research was the Perkin-Elmer Physical Electronics PHI
5400 ESCA Spectrometer equipped with a standard magnesium X-Ray source. The
elemental composition of the natural fiber surface was determined both before and after
plasma or chemical treatment. Chemical information indicating changes in the treated
fiber surface was elucidated by curve fitting the carbon ls (Cls) and oxygen ls (Ols)
spectra. The Cls and Ols curves were fitted with a Lorentzian-Gaussian mix Voigt
profile using a nonlinear least-squares curve-fitting program. The resulting curve fit had a
level of experimental error around 5-10%. All peaks were referenced to adventitious
carbon at 284.6 eV.

3.2.3 Atomic Force Microscopy (AFM)

In order to better understand of the physical effect of the fiber treatment (such as
plasma) on the fiber-matrix adhesion, AFM was used to quantify the change in surface
roughness and surface area after plasma treatment.

The atomic force microscope (AFM) is one of about many types of scanned-
proximity probe microscopes. All of these microscopes work by measuring a local
property - such as height, optical absorption, or magnetism - with a probe or ”tip” placed
very close to the sample. The AFM works by scanning a fine ceramic or semiconductor
tip over a surface much the same way as a phonograph needle scans a record. The tip is
positioned at the end of a cantilever beam shaped much like a diving board. As the tip is

repelled by or attracted to the surface, the cantilever beam deflects. The magnitude of the

36

deflection is captured by a laser that reflects at an oblique angle from the very end of the
cantilever (See Figure 19 (a)). A plot of the laser deflection versus tip position on the
sample surface provides the resolution of the hills and valleys that constitute the
topography of the surface. The AF M can work with the tip touching the sample (contact

mode), or the tip can tap across the surface (tapping mode).

 

Figure 19. (a) Workings of a contact mode AF M. (b) MultiMode Scanning Probe Microscope (SPM)
from Digital Instrument.

In this research, the physical changes of the cellulose fiber surface due to the
plasma treatment were evaluated using a multimode Scanning Probe Microscope (SPM)
Nanoscope IV from Digital Instruments (See Figure 19 (b)). The surface images of the
henequen fiber before and after each treatment were collected using the Contact AFM
mode with a standard silicon nitride probes (spring constant of the probe used is 0.06
N/m). Afier collecting the topographical images of the henequen fiber surfaces, the

images were analyzed for surface roughness and surface area. Using the data gathered by

37

the AFM, it’s then possible to separate out the physical effect (surface roughening) from
the chemical effect (fiber surface chemistry change) with respect to fiber-matrix adhesion.
3.2.4 Thermogravimetric Analyzer (TGA)

TGA measures weight changes in a material as a fimction of temperature (or time)
under a controlled atmosphere. Its principal uses include measurement of a material's
thermal stability and composition. The furnace can heat samples from room temperature
to over 1000 C (depends on the manufacture and model). The balance is sensitive to 0.1
microgram. The TGA is temperature calibrated using the Curie points of alumel, nickel
and iron. From the graph of % Weight versus Temperature, one can calculate Onset
temperatures, and also delta-Y steps. Examples include the monitoring of the heating of

polymers, and flreir decomposition.

 

 

 

 

 

Figure 20. Picture of Perkin-Elmer TGA used in thk research

The TGA used in this research was a Iii—Resolution 2950 Thermogravimetric

Analyzer from TA Instruments Corporation (See Figure 20). The sample handling,

38

experimental control and data acquisition of this instrument are all automated with the
Thermal Advantage Software.

The main application of this technique to the research was to determine the water
content of the cellulose fiber, drying rate of the cellulose under isothermal condition and
to determine the degradation temperature of all the cellulose fiber used in this research.
3.2.5 Micro Video Caliper

CUE Micro-300 video caliper from Olympus Corp was used in the measurement
of fiber diameter and in the microbond testing. The video caliper is connected to an
optical microscope via a video camera capable of giving out diameter readings base on
the observed image in um. The sample needed to be measured is place under microscope
and its magnified image is displayed on the video screen. Using a calibration slide, one
can determined the planner dimension of the sample by reading the readout from the

video caliper. A picture of the CUE Micro-300 video caliper is shown in Figure 21 .

Video Image
of Sam le

  

 

/’ "‘“CUE Micro-300
‘ 1 Video Caliper

Figure 21. Picture of the CUE Micro-300 video caliper setup

     
   

39

The main purpose of this instrument was to measure natural fiber diameter and to
measure the polymer drop size in the Microbond testing. The video caliper was also used
in the interfacial adhesion measurement of henequen fiber and epoxy matrix. Since the
henequen fiber’s cross section is not circular, it was required to measure the interfacial
contact area between the henequen fiber and epoxy matrix.

3.2.6 Tensile Strength Tester — United Test System (UTS)

The tensile strength testing system has the ability to applied a tensile stress to a
sample at its ends and measure the stress and the strain at the same time. The tensile
stress is measure via a load cell and the strain can be measure by a laser or strain gage.
The sample is loaded between two mechanical grips that capable moving in opposite
direction at a controlled speed determined by the user. The load and displacement
information are recorded by a computer and the mechanical properties of the sample can
then be calculated base on the resulting data. The UTS used in this study is the United

Model SFM-20 Test System from United Calibration Corporation (See Figure 22).

 

Figure 22. Picture of United “SFM-20” Test System

40

The UTS was used to measure the mechanical properties of polymer matrix
materials and also treated and as received fiber samples. In addition, the tensile and
flexural strength and modulus of all the composite samples with difference fiber-matrix
interfaces were also measured using UTS.

As shown in Chapter 4, an increase in the interfacial adhesion can dramatically
affect the mechanical behavior of the composite material. It usually happens that ‘good’
interfacial property increases the tensile and flexural strength and modulus of the
composite but decreases the fracture toughness. If the change in the interfacial adhesion
(measured by a microbond test) can be determined for a particular surface or matrix
treatment, then the mechanical properties of the composite due to the same treatment
could then be related to the interfacial adhesion result and a correlation between the two
sets of data might be possible.

3.2.7 Microbond Test

Microbond test is an experimental method commonly used to measure the

interfacial shear stress between a single fiber and a drop of matrix material that surrounds

the fiber.

 

Fibe' Blade

 

 

Micro Blade

 

 

 

 

Figure 23. Schematic setup of microbond testing

4l

As shown in Figure 23, a pair of fine micro blades is used to grip the matrix drop on the
fiber and the force applied, to completely debond the matrix drop from the fiber, is
recorded for the calculation of interfacial shear stress.

The calculation of interfacial shear stress between the fiber and matrix material is
as follow:

Interfacial Shear Stress, r — Fm“

-7r-d-I

 

where Fmax is the maximum load recorded by the load cell, dis the diameter of the fiber,
and I is the embedment length of the matrix drop.
The Microbond test was used to determine the interfacial adhesion between

Tencel fiber and propylene or epoxy matrix.

3.3 Plasma Treatment of Cellulose Fiber
3.3.1 Plasma Treatment Theory

Plasma treatment is a very interesting technique in the treatment of natural fibers
to improve compatibility with polymer matrices. This treatment method is fast,
environmentally fiiendly and can be done under atmospheric condition [19].

Plasma is a partially-ionized gas containing ions, electrons and various neutral
species at different levels of excitement. The free electrons gain energy from an imposed
electric field, colliding with neutral gas molecules and transferring energy. The collisions
and transfer of energy form free radicals, atoms, and ions. These particles interact with
solid surfaces places in the plasma. This leads to drastic modification of the molecular
structure and providing desired surface properties (See Figure 24). The action plasma

takes on a polymer is determined by the chemistry of the reaction, but the plasma process

42

causes changes only to the surface of the material. In addition, plasma treatment of
surfaces is confined to a layer several molecular layers deep. The type of surface change

that occurs depends on the composition of the surface and the gas used.

Process Gas Inlet

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I l I Plasma Chamber
Excited Gas Species
- atom \‘ -.
i; - molecules Photons {-
i - ions
- electrons /
. free radicals
- metastables Chemically .,
. / Modified Sites ’
i i i A i i i Polymer Substrate “ hi
7 Process Gas 7
RF Power Source vacuum Outlet Electrode

Figure 24. Schematic of the surface modification of plastic in a gas plasma reactor.

Gases or mixtures of gases used for plasma treatment of polymers include
nitrogen, argon, oxygen, nitrous oxide, helium, tetrafluoromethane, water and ammonia.
Each gas produces an unique plasma chemistry and the selection of the process gas
determines how the plasma will alter the polymer surface. Very aggressive plasma can be
created from relative benign gases. For example, a tetrafluoromethane plasma contains
free radicals of fluorine. Oxidation by fluorine free radicals is known to be as effective
as oxidation by the strongest mineral acid solutions, with one important difference: the

plasma by-products do not require special handling. As soon as the plasma is shutoff or

43

the excited species exit the RF field, the species recombine to their original stable and
nonreactive form within few seconds [47].
3.3.2 Plasma Treatment Mechanism

There are three competing molecular reactions alter the polymer surface
simultaneously: ablation, crosslinking and activation. Ablation is an evaporation process.
The plasma breaks covalent bonds of the polymer backbone by bombardment with high-
energy particles, i.e., fi'ee radicals, electrons and ions. As long molecules become shorter,
their volatile oligomers and monomers boil off and are swept away with the exhaust.
Crosslinking is usually done with inert process gas such as argon or helium. The bond-
breaking occurs on the polymer surface just as with any other plasma. Since there are no
free radical scavengers in the plasma, the radicals produced can recombine, joint with
neighbor radicals, or react with an adjoining free radical on the same chain to form
double bonds. In the activation process, surface polymer groups are replaced with atoms
or chemical groups from the plasma. During this process, the plasma breaks the
polymer’s backbone or pendant atoms or groups from the backbone, creating free radicals
on the surface. The free radical being thermodynamically unstable will quickly react
with radicals in the plasma to form covalent bonded groups. The new functional groups
on the polymer surface can alter the surface energy and its adhesion characteristics.
The extent of each depends on the chemistry and process variables.
3.3.3 Oxygen Plasma Treatment

One of the more common plasma processes to enhance adhesion is treatment in a
cold oxygen plasma. Oxygen plasma is aggressive, and forms numerous components.

Within an oxygen plasma one can find 0+, 0', 02', O, 03, ionized ozone, metastably-

excited 02, and the electrons. As the components recombine, they release energy and
photons, emitting a faint blue glow and much UV radiation. The photons in the UV
region also have enough energy to break the polymer’s carbon-carbon and carbon-
hydrogen bonds (Sample oxygen plasma reactions are shown in Figure 25)

RH + O 9 R- + OH-

1104- O-) R’O+ CO,CO,

R - + 20 -) R0,-

RH 4» UV -) R- + Ho

R0; + R”’- 9 R02R’”

R’R”-OH + O -) R'=R” + H20
Figure 25. Sample oxygen plasma reactions with polymer nrface [48]

All of the reactive species react with the polymer, in addition to bombardment by
photons, ions, and neutral particles. Surface energy can be increased very quickly by
plasma induced oxidation. For example, oxygen plasma induced oxidations of
polypropylene increase the initial surface energy of 29 dynes/cm to well over 73
dynes/cm in just a few seconds. At 73 dynes/cm, the polypropylene surface is completely
water wettable [49]. The by-products are C02, CO, H20 and low molecular weight
hydrocarbons that are readily removed by the vacuum system. Even though these
molecules can be excited by the RF field, their effect on the reaction appears insignificant
when compared with oxygen plasma.

3.3.4 Application of Plasma in This Research
For this research, a low pressure cold plasma system was used to modify the

cellulose fiber surface to produce different level of interfacial adhesion with polymer

45

matrices. The fibers were modified with the PSOSOO plasma surface treatment system
from AIRCO using oxygen, argon, or nitrogen as plasma gas. The cellulose fibers were
treated with oxygen plasma under various treatment condition and treatment time. The
cellulose fiber was also treated with argon and nitrogen plasma to compare the effect of
different plasma gas on the interfacial adhesion. After each of the plasma treatment, the

cellulose fiber samples were analyzed with XPS for surface chemical composition change.

46

CHAPTER 4 ADHESION STUDY USING TENCEL AND EPOXY

 

As stated in Chapter 2, to gain a better understanding of how interfacial chemistry
affects the fiber-matrix adhesion between cellulose fibers and polymer matrices, a
regenerated cellulose fiber was used for the adhesion study due to its uniform physical
and chemical properties. In order to accomplish the research goals, the Tencel fibers were
modified with plasma and silane reagents to produce different interfacial chemistries with
polymer matrix materials. Alter the surface modifications, the mechanical, morphological
and chemical properties of the Tencel fibers were examined using UTS, ESEM and XPS.
After the fiber treatment and analysis, Microbond tests of Tencel with epoxy matrix were

conducted to evaluate the effect of surface treatment on the interfacial adhesion.

4.1 Microbond Sample Preparation

The Tencel and epoxy Microbond samples were prepared by mixing the Epon 828
(DGEBA) resin with Jefl‘amine T-403 curing agent in a 100 to 50 weight ratio. The epoxy
solution was then degassed under high vacuum for 3 minutes. A disposable syringe was
used to apply the epoxy onto the Tencel fiber as shown in Figure 26(a). In order to
prepare the microbond sample with very small epoxy drop size, the epoxy drop at the tip
of the needle was used to lightly touch the Tencel fiber and then lifi away in a quick
action. The sample was then cured at 80 C for 2 hours and then post curved at 125 C for 2
hours in a convection oven. The heating rate fiom room temperature to 80 C and then to
125 C was set to 15 C/min. Final epoxy drop size ranged from 80 to 400 pm (See Figure
26 (b)). Due the relative small size of the epoxy matrix drops (large surface area for the

curing agent to escape), a higher weight ratio of curing agent was used in the preparation

47

of the epoxy solution. At the recommended DGEBA to Jefl‘amine mix ratio (100 to 45

by weight), the epoxy drops could not be fully cured.

 

—o—e—e—e—-o—e—e—
Heat Curing

(a)

 

Figure 26. (a) Schematies of microbond sample preparation. (b) Picture at Tencel-epoxy microbond
samp e.

4.2 Effect of Plasma Modification of Tencel Fiber
4.2.1 Oxygen Plasma Treatment Procedures

Chopped Tencel N-lOO fibers received from Tencel, Inc. were mounted onto a
glass frame with a width of 3 inch across. The diameter of the glass rod was 0.5cm,
which means that all fibers were suspended 0.5cm from the plasma plate. The sample
layout and the plasma treatment chamber are shown in Figure 27. The Tencel N-lOO
fibers were treated with oxygen plasma in the PSOSOO plasma chamber at 275W for 0, 0.5,
l, 2, 4, 8, and 16 minutes. Oxygen flows were set at 0.75 liter/min. The base pressure of
the plasma chamber was set at 0.05 Torr. For comparative purpose, some Tencel fibers
were also treated with nitrogen plasma at the same condition as oxygen plasma. After the
plasma treatment, the treated Tencel fibers were tested for changes in mechanical

properties and surface chemistries.

48

 

(a)

 

Figure 27. (a) Picture of Tencel fibers placed on a u-shape glass support for plasma treatment. (b)
PSOSOO plasma surface treatment system with variable power and gas flow control from AIRCO.

4.2.2 Mechanical Properties of Oxygen Plasma Treated Tencel Fiber
The first indication that the plasma treatment had a physical effect on the Tencel
fiber property is from the diameter measurement. According to the results shown in

Figure 28, there was a gradual decrease in the fiber diameter as treatment time increased.

 

Tencel Flber Diameter vs. Oxygen Plasma Treatment Time

 

 

 

Fiber Diameter (micronmeter)

0.0 0.5 1.0 2.0 4.0 8.0
Treatment Time (min)

 

 

 

Figure 28. Tencel fiber diameter vs. oxygen plasma treatment time

The change in fiber diameter can be explained by the ablation process of the oxygen
plasma. However, the large decrease in fiber diameter only occurred after 8 minutes of
plasma treatment.

The tensile properties of the oxygen plasma treated Tencel fibers were evaluated
using UTS. As shown in Figure 29, the tensile strength of the cellulose fiber was greatly

affected by the oxygen plasma treatment and it deceased with increasing treatment time.

 

*1

Oxygen Plasma Treatment at 215W and 0.75leln 0; Flow

 

l
l
l
l

1 l A

Tsnalle Strength (MPa)

d
8

 

 

 

 

Plasma Treatment Tine (min)

F

 

VV-onailieismnfigth 7 '* 34inch Medium 7’]

7 ,7J

Figure 29. Tensile strength CTN! moEIu—srof o;ygen plasma thTencel fiber

For a 4min oxygen plasma treatment, the tensile strength of the cellulose fiber decreased
by as much as 40% compare to untreated fiber. The decrease in tensile strength could be
related to the roughening of the fiber surface and the creation of crack initiation sites. The
reduction in tensile strength had lead to premature failure of fiber in the Microbond test,
e.g., the Tencel fiber would break before the epoxy drop could debond from the fiber.

Even though there are some variations in the average tensile modulus values, the

50

difl‘erence are statistically in significant. In addition, modulus is an intrinsic property of
the material and it shouldn’t be affected by plasma treatment.
4.2.3 Surface Topology of Plasma Treated Tencel Fiber

Afier the oxygen plasma treatment, the Tencel fiber surface was analyzed with

ESEM for surface topology changes due to the treatment.

 

Figure 30. ESEM image of ongen plasma treated Tencel fiber surface

According to the ESEM images (Figure 30), the plasma treatment produced very little
changes to the topology of the Tencel fiber from low magnification. However, there
could be submicron level changes that could not be detected with the ESEM due to fiber

damage and decomposition at high magnifications (See Figure 31).

51

Damage from high
magnification of
ESEM

 

 

Figure 31. ESEM image of Tencel showing damage from high energv electron beam.

4.2.4 Chemical Properties of Plasma Treated Tencel Fiber Surface

After the oxygen plasma treatment, the Tencel fibers were analyzed with XPS for
surface chemical composition changes. The main purpose of the XPS analysis was to
quantify the changes in oxygen content and functional groups such as hydroxyl on the
Tencel fiber surface. The effects of plasma treatment on the surface chemistry of the
Tencel fiber are shown in Figure 32 and Figure 33. Due to the time required to analyze
the fiber samples with XPS, most of the samples were only run once and the error bar

represents the worst case of XPS detection error (5%).

52

 

 

 

OIC Atomic Ratio ' V
O
:5

 

 

 

o 4 8 12 16 20'
l Treatment Time (m’n) j

Figure 32. Oxygen to carbon ratio (O/C) on the Tencel fiber surface vs. oxygen plasma treatment
time obtained from XPS analysis (275W and 0.75Umin 01).

 

 

 

 

 

 

 

 

 

 

 

 

f 50
is: 4...
3 a.

3

5 30

'73

a 20*

0

C

E 10

0 Y 1

f O 5 10 15 20
I Treatment Time (min)

 

Figure 3. Plot of chemical functional group concentration on the Tencel fiber surface vs. oxygen
plasma treatment time; XPS analysis of Cls peak. (275W and 0.7SU-in 01)

According to the XPS analysis, oxygen plasma treatment significantly increased
the oxygen atomic concentration on the fiber surface and it reached a maximum O/C ratio

of 0.7 for after 4 minutes. As shown in Figure 33, the main contribution of the increased

53

oxygen content on the fiber surface was due to the formation of new hydroxyl and other
functional groups. The hydroxyl functional group also peaked at about 4 minutes of
plasma treatment time. The increasing oxygen content of the cellulose fiber due to
oxygen plasma treatment is in general agreement with other published research on spectra
[64] or cellulose fiber [65].It’s interesting to note that based on the chemical structure of
cellulose, the O/C ratio of the untreated fiber should be around 0.83 and an O/C ratio of
only 0.26 was obtained fi'om the XPS analysis. However, this value is comparable to the
O/C carbon ratio of 0.3 observed by Chtourou et. al [63] for explosion of hard-wood
a8pen pulp.
4.2.5 Interfacial Adhesion between Plasma Treated Tencel Fiber and Epoxy

The microbond samples were tested according to the setup shown in Figure 23.
The force used to debond epoxy drop from the Tencel fiber was recorded by a voltmeter
(the force was converted from the voltage signal produced by a 500mg load cell). The

traveling speed of the micro-blade, used to debond the epoxy drop, was around 30um/sec.

35

 

 

 

I t

'8

   
 

El

3

 

Interfacial Shear Strength (MPa)

 

 

.5
0|
).

 

Untreated 02 75W 4min N2 75W 4min

Figure 34. Interfacial shear strength of oxygen (02) and nitrogen (N2) plasma treated Tencel fiber
with epoxy

 

 

54

The interfacial shear strength data for Tencel bonded to epoxy matrix are shown
in Figure 34. The samples used for interfacial adhesion study were tested at 75W with
0.75 L/min of oxygen flow rate. The plasma treatment used for the interfacial adhesion
testing was much less aggressive than the previous treatment for XPS analysis. As stated
earlier, plasma treatment at higher power (27 SW) produced significant damage to the
fiber tensile strength. However, even at reduced power, the plasma still produced enough
damage to cause premature failure of the Tencel fiber during microbond testing. The
interfacial adhesion values (for plasma treated samples) shown in the figure represent the
interfacial stress reached at the point of fiber breaking. The actual debond stresses
(interfacial adhesion) should be greater than the values shown, which are indicated by the
upward arrow. Even though the true interfacial shear strength was not obtained, it’s still
possible to see that the interfacial shear strength values of the plasma treated samples are
higher than those of the untreated ones. This result also showed the limitation of the
Tencel fiber used in this study, i.e., it does not posses the tensile strength for all treatment

conditions.

4.3 Effect of Silane Treatment of Tencel Fiber

Organosilanes are the main group of coupling agents for glass fiber reinforced
composites. They have been developed to couple virtually any polymer to the minerals,
which are used in reinforced composites [43]. However, the effectiveness of silane
treatment in natural fiber reinforced composite system has been controversial. According

to some past research, silane treatment is an alternative method to modify cellulose fiber

55

surface. It can introduce firnctional groups to the cellulose fiber without damaging its
mechanical properties [50]. The effect of modification can also be easily detected by
XPS since cellulose fiber does not contain any Silicon (Si) atom. The level of silane
grafting can be related to the amount of Silicon present on the fiber surface after the
silane treatment. However, care must be taken to remove all non-reacted silane coupling
agent after treatment using solvent extraction.
4.3.1 Verification of Silane Reaction with Cellulose Fiber

The effectiveness of silane treatment was verified in our laboratory by reacting
silanes with Tencel fiber and see if any significant silicon atoms remained on the fiber
surface (using XPS) after soxhlet extraction of silane treated fiber sample. If there were
still significant amount of silicon present on the fiber surface after the extraction process,
then silane must have reacted chemically with the cellulose fiber.

For this part of the experiment, a high purity silane fiom Sigma-Aldrich (Chloro-
dimethyl—methoxysilane (CDM)) was used. Different concentration of CDM silane
solutions were prepared by dissolving silane into an acetone-water solution (9Swt%

acetone to 5wt% water). The details of the treatment conditions are shown in Table 6.

Table 6. CDM Silane Treatment Conditions

 

 

 

 

 

 

 

ID Silane Solution Concentration Fiber Wt Treatment Time (hr)
SS1 1.0 wt% Silane1 0.3g 2
$82 2.0 wt% Silane1 0.3g 2
$83 5.0 wt% Silane1 0.39 2
$84 1.0 wt% Silane1 0.39 24
$85 2.0 was Silane1 0.3g 24
886 5.0 M96 Silanet 0.3g 24

 

 

 

 

 

 

Silane1 = cucmhsmcu,

56

These series of treatment conditions were designed to study the effect of silane
concentration and treatment time. 0.5wt% benzyl peroxide radical initiator was also

added to the solution to promote the reaction of silane to cellulose fiber.

Dimethyl-chioro-methoxyailane Treatment of Tencel Fiber

 

 

 

26
SE ITreatrnent Tlme . 2 hours
2' 20 ~ lTreatrnent Tlme = 24 hours
'3
1%
2 15
E
8
<
a 10 ~
. e
8
s 5-
in
0

 

 

 

W

 

 

Silane Concentration in AcetonelWater

 

Figure 35. Silicon concentration on the Tencel fiber surface after silane treatment (before soxhlet
extraction) from XPS analysis.

After the silane treatment, part of the silane treated Tencel fiber was analyzed
with XPS before the soxhlet extraction. According to the XPS results (Figure 35), there
are significant amount of silicon atoms on the fiber surface and the concentration seemed
to relate to the silane solution concentrations. It is interesting to note that longer treatment
time (24-hours treatment) did not produce any advantage over the shorter treatment time
(2 hours). However, the silicon concentration maybe misleading due to week hydrogen
bonding and physisorption of the silane onto the fiber. Therefore solvent extraction was

performed on the silane treated samples to see the true chemically reacted silane.

57

After the initial XPS analysis, the silane treated samples were soxhlet extracted
with pure ethanol for 48 hours and analyzed again with XPS to see the amount of silicon

on the fiber surface.

 

 

 

 

   

 

Dimethyl-chioro-methoxysila ne Treatment at Tencel Fiber
25
g? I Treatment Time = 2 hours
‘6' 20 . I Treatment Time = 24 hours
:3
n
a:
2 15 .
E
0
g
< 10
('3 -
8
g 5 .
:r
U)
0 .
0wt% 1wt% 2wt% 5wt%
Silane Concentration in AeetoneNVater

 

 

 

Figure 36. Silicon concentration on the Tencel fiber surface after silane treatment (after 48 hours of
soxhlet extraction) from XPS analysis.

However, the reaction of silane with cellulose fiber surface seems to be dependent
on the type of silane used as well. With a monofunctional silane, Dow Corning Z-1224
[(CH3)3SiCl], the chemical reaction between the silane and cellulose seemed to be very
small. As shown in Figure 37, the monofunctional silane only produced 0.22% of Si
atomic concentration on the fiber surface after soxhlet extraction with ethanol. At the
same time, the Z6076 used as a control produced much higher Si atomic concentration
under the same treatment condition. This result indicates that the interaction between
monofunctional silane and cellulose is mostly hydrogen bonding and formation of

polysiloxane (Z6076) could be a major step in the silane-fiber reaction process.

58

 

Silane Treatment of Tencel Fiber in Acetone/Water

12

 

    
 
  

I Before Soxhiet Extraction
‘0 A D Alter Soxhiet Extraction for 24 hr

z-1224 = (CH,),SICI
z-eore = cucumsuocng,

Surface Si Atomic Ratio from XPS (‘56)
a

 

 

 

0wt% 2wt%Z1224 2wt%26076
Silane Concentration In Acetone/Water

 

 

 

Figure 37. XPS analysis of Z1224 and Z6076 treated Tencel fiber surface.

4.3.2 Proposed Silane Reaction Chemistry with Cellulose Fiber

Most of the silane coupling agents can be represented by the following formula:

R-(CH2)n-Si(OR’)3

Where n = 0 -— 3, OR’ is the hydrolysable alkoxy group, and R is the functional organic
group. The organic functional group (R) in the coupling agent causes the reaction with
the polymer matrix. This could be a co-polymerization, and/or the formation of an
interpenetrating network. According to past research, the curing reaction of a silane
treated substrate can also enhance the wetting by the resin [43].

The general mechanism of how alkoxysilane form bonds with the fiber surface

which contains hydroxyl groups could be as follows:

59

RSi(OR’)3 Ti T F

 

 

3H20 —’ l -' 3R’OH HO— Si— 0— ii—O_ $5-0...
|
. ’0‘ ‘
RS|(OH)3 I--> H‘ H H: t-l H‘ H
2R8i<OH)3 —p 1 _, 3H20 “'0’ “'0’ “0|”
Fiber Surface
R R R
I | I A l —’ 2H20
HO— Si— O— Si—O- Si—OH R Ft R
' ' ' I I I
OH OH OH HO— Si— O- Si—O— si—QH
+ — I I ..
HO HO HO 0 0 SJ;
l I l _ | I 1
Fiber Surface Fiber Surface

 

 

Figure 38. Proposed silane reaction with cellulose fiber [43]

Alkoxysilanes undergo hydrolysis, condensation and the bond formation stages under
acid or base catalyzed conditions. In addition to this reaction of silanols with hydroxyls
of the cellulose fiber surface, the formation of polysiloxane structure can also take place
(See Figure 38). In fact, from the study with Tencel, the formation of polysiloxane could
even promote the reaction between cellulose fiber and silane.
4.3.3 Silane Treatment Procedures

Silanes were applied in the form of hydrolyzed solutions using acetic acid as pH
control. 0.5 wt% silane solutions with pH value at around 4 were prepared in 200m] of
water for each 10g of Tencel fiber. The Tencel fibers were immersed in the silane
solutions for 45 minutes at around 60C and then dried at 103C for 1 hour. According to
the literature from the silane manufacture, the high temperature helps the condensation
reaction of the silane with the fiber surface. After the heat treatment, the treated fibers
were soxhlet extracted with ethanol to remove any excess silane and then vacuum dried

for 4 hours at around 100C.

60

4.3.4 Mechanical Properties of Silane Treated Fiber
After the silane treatment, the Tencel fibers were tested with UTS to see if the

silane treatment had any effect on the mechanical properties of the silane treated fiber.

 

Mechanical Properties of Silane Treated Tencel Fiber

 

 

 

 

 

700 60

it? ‘6‘
n. 600 «— g_
a 500 50 g
5 ” -» 40 g
2' 400 A» 3
g 30 '6
25 30° " 2°
% 200 —~ ' 20 g
5 100 -— 1o 5
l- i-

0 ~ — 0

Tencel N-100 0.5 wt% 26040 Silane
-Tensile Strength —e—Tensiie Modulus

 

 

 

 

Figure 39. Mechanical Properties of silane treated Tencel fiber

Unlike plasma treatment, silane treatment did not produce any adverse effect to the
mechanical properties of the Tencel fibers (See Figure 39). The slight increase in the
mechanical properties of the Tencel fiber could be caused by the cross linking of the fiber
surface or increased hydrogen bonding due to heat treatment during fiber drying.
4.3.5 Surface Topology of Silane Treated Tencel Fiber

After the silane treatment, the surface of the Tencel fiber was still very smooth
and there was no visible thick silane coating layer form the ESEM image (Figure 40).
The small patches of surface impurities could be dust from drying process or could be

small amount of polysiloxane.

61

Table 7. Atomic distribution on the Tencel fiber surface from XPS analysis

  

——m—

Figure 40. ESEM image of Silane treated Tencel fiber

4.3.6 Chemical Properties of Silane Treated Tencel Fiber Surface

 

Base on the XPS analysis of silane treated Tencel fiber there was strong evidence

that silane was present on the surface of the fiber even after 48 hours soxhlet extraction

 

 

 

 

 

 

 

 

 

Fiber Sample I C (96) O (96) N (96) SI (96)
Untreated 73.8 24.4 1.8 0.0
26070 Silane 73.9 24.2 1.0 0.9
Treated

26°40 Si'ane 73.0 24.6 1.1 1.3
Treated

 

62

According to the atomic ratio data (shown in Table 7), there were as much as 1.3 atomic

% of silicon atom (Si) on the fiber surface. However, the amount of silicon detected by

 

XPS was still very small compare with the available hydroxyl groups to react with the
silane reagent.
4.3.7 Interfacial Adhesion between Silane Treated Tencel Fiber and Epoxy

The Microbond samples of silane treated Tencel fiber and epoxy were prepared
the same way as shown in Figure 26. The curing conditions of the samples were also kept

the same as for the plasma treated samples.

 

 

 

 

     

35
Higher

0. Reactivity
E Unreactive 275
E 30 a Surlace
g 25 9
g .
(I)
a 25 —
or
.:
a)
E
g 20 -
t
.9
.E

15 ~

0.5% 26070 Untreated 0.5% 26040
Treated Treated

 

 

 

Figure 41. Interfacial shear strength data of Tencel/epoxy from Microbond test

The interfacial shear strength obtained from Microbond test is shown in Figure 41.
The Z-6070 treated fiber sample had the lowest value of interfacial shear strength and the
Z—6040 (epoxy functionalized silane) treated samples had the highest average interfacial
shear strength. This result is expected since the Z—6040 silane has an epoxy ring that can
react with the polymer matrix material and form a link between the cellulose fiber and the

epoxy matrix. The Z-6070 on the other hand can prevent any reaction between the

63

cellulose fiber and the matrix due to its nonreactive methyl functionality. Table 5 shows
the chemical structures of silanes used.

However, the difference in the interfacial shear stress values was not large and
this could be caused by limited reaction between the cellulose fiber and the silane
coupling agent. The XPS analysis only showed around 1% of silane on the fiber surface.
According to Matias et al. [41] addition of benzyl peroxide to the silane solution might be
necessary to promote the reaction between the silane and the hydroxyl group present on
the cellulose fiber surface.

4.3.8 Better Wetting of Silane Treated Tencel Fiber
The silane coupling agent not only had an effect on the interfacial adhesion between
Tencel fiber and epoxy matrix, but also improved the wetting behavior of the epoxy

matrix to the cellulose fiber.

   

(a) (b)

Figure 42. (a) Untreated fiber with epoxy; (b) Z-6040 silane treated fiber with epoxy (Angle 0 is the
contact angle of epoxy to Tencel fiber)

As shown in Figure 42, the Z-6040 silane treated Tencel fiber had smaller contact angle

with respect to epoxy which indicates better wetting behavior than untreated fiber.

4.4 Summary

The plasma and silane treatment could both affect the interfacial adhesion
between cellulose fibers and epoxy. For the oxygen plasma treatment, it significantly
altered the cellulose fiber surface chemistry by increasing the oxygen content of the fiber
surface (O/C from 0.25 to 0.7). This increase in oxygen functional group improved the
interfacial adhesion with epoxy by forming covalent bonds with the epoxy matrix.
However, the oxygen plasma also caused degradation to the mechanical properties of
cellulose fiber and process and the extent of the improvement in interfacial adhesion
(from oxygen plasma treatment) could not be determined due to premature fiber breakage
in microbond testing. Process optimization might be necessary for real applications.

As for silane treatment, it was proven that silane reaction with cellulose is
possible and it’s limited by the available reaction sites on the fiber surface. From the
XPS analysis of various silane treated Tencel fiber, the formation of polysiloxane could
be a major step in the reaction with cellulose fiber. The silane reaction with Tencel fiber
also seems to depend on the type of silane used. From the adhesion results, silane
treatment can either lower or improve the interfacial adhesion base on silane reagent
selection. But compared with plasma treatment, the extent of surface modification with
silane was rather limited and a combination of plasma and silane treatment might be

beneficial in the optimization of interfacial adhesion.

65

CHAPTER 5 ADHESION STUDY USING TENCEL AND POLYPROPYLENE

 

In Chapter 4, the effect of cellulose fiber surface chemistry on the fiber-matrix
adhesion was investigated by treating the Tencel fiber with plasma and silane coupling
reagents. Now the questions become how would the matrix mechanical and chemical
properties affect the fiber-matrix adhesion between the cellulose fiber and the polymer
matrix? To determine the effect of the matrix polymer chemistry on the interfacial
adhesion of natural fiber reinforced composite, polypropylene (PP) and maleic anhydride

grafted polypropylene (maPP) were used as matrix materials for this study.

5.1 Shear-Lag Stress Transfer Model on Interfacial Adhesion

According to the Shear-Lag model proposed by H. L. Cox, the interfacial shear
stress is proportional to the matrix strain multiplied by the square root of the shear
modulus of the matrix material and various parameters related to the composite geometry
[42]. The details of the H. L. Cox’s model are shown below.

H. L Cox’s stress transfer mode]:

 

 

0.5
G sinh 0.5L - x
T=Ef8m[2E ln"(1R/r)] fl( L )
f cosh ,6;

where r: Interfacial Shear Stress
Ef: Tensile Modulus of fiber
8m: Strain in the Matrix
R: Interfiber Spacing
r: Radius of fiber
B: scaling factor
L: Length of the fiber fragment
x: Radial distance outward

66

If we can keep all other factors the same in Cox’s model, then the interfacial shear stress
would be directly proportional to the product of matrix strain and the square root of the

matrix shear modulus; e.g.,
Interfacial Shear Stress, 2' 0c Em q/Gm

According to this theory, the chemical effect of the matrix on the interfacial adhesion
could be determined by comparing the interfacial shear stress values of a Tencel—PP and a

Tencel-maPP sample provided two matrices have identical mechanical properties.

5.2 Microbond Sample Preparation
Tencel and polypropylene Microbond samples were prepared by melting the
polypropylene pellets with a fine point soldering iron and the melted polypropylene (on

the tip of the soldering iron) was applied onto the Tencel fiber by hand (See Figure 43).

PP Drop

\

l Tencel Fiber

—e—e—e—e—e—e—o—
Heat Treatment
(a)

Figure 43. (a) Schematics of microbond sample preparation. (b) Picture of Tencel/PP microbond
sample.

 

The Microbond samples were heated in a convection oven for 20 minutes at 185 C to

give the polypropylene drops sufficient time for the droplet surface to melt and contract

67

to the shape of a sphere due to the influence of the polymer surface energy. The heated
samples were then quenched in a cold water bath for a 30 second period in order to
minimize the crystallization of PP drop. The final sizes of the polypropylene drop were

between 75 pm to 400 pm (See Figure 43).

5.3 Effect of Matrix Crystallinity on the Interfacial Adhesion

There have been many studies in the literature which deal with the mechanical
properties of natural fiber reinforced polypropylene composite [68-70]. The general
conclusions have been that surface treatment of the cellulose fiber using maleated PP or
other kinds of coupling agent will improve the overall mechanical properties of the
composites due primarily to chemical bond formation through the maleic acid. However,
according to our study the interfacial shear strength between cellulose fiber and PP
matrix could also be affected by crystallinity of the PP matrix which has an effect on the
PP modulus. As shown in Figure 44, the interfacial shear strength (IFSS) values between
Tencel fiber and six Profax series PPs also depends on the matrix crystallinity. This
conclusion was deduced from the observation that the slowly cooled samples, which were
significantly different in crystallinity and hence modulus, didn’t agree with the lower
crystallinity data in samples obtained by faster cooling process. That is, they did not
follow the prediction by Cox’s model. Furthermore, since there is no possibility of
chemical bond formation between Tencel and the PP matrix, another factor, i.e., matrix
modulus due to crystallinity, must be responsible for higher level of adhesion. According
to the DSC analysis of the PP drops, it was clear that the crystallinity of the matrix

polymer produced the interfacial adhesion mismatch (See Figure 45).

68

 

 

Slowly Cooid Sample

El Water Quenched Sample

interfacial Shear Stress (MPa)

 

 

 

Profax Profax Profax Profax Profax Profax
6301 6501 6601 PH020 PH180 PH131

 

 

 

Figure 44. Plot of interfacial shear stress for various PP matrices with different processing conditions

Figure 44 indicates that the interfacial shear strength values shown on the left for
each PP grade are higher in IFSS than those on the right. The two different matrix
crystallinity conditions were obtained by slowly cooling or water quenching the
microbond samples after heating to 185 C. In order to minimize the effect of matrix
crystallinity, the morphology of the PP matrices were standardized before microbond
testing by first heating them above their melting temperatures and then quenching them
in cold water immediately for 30 seconds. This fast cooling minimized the large crystal
formation in the PP matrices as verified by the DSC analysis shown in Figure 46, in
which it is shown that cold water quenching of the PP matrix eliminated the crystalline
transitions. The DSC analysis was conducted on the PP drops removed from the
microbond samples after heat treatment both with and without cold water quenching

process. The DSC scan rate was set at 5 degree C per minutes.

69

 

DSC Analysis of Slowly Cooled PP Matrix

 

-o.1 ~
-o.2 4
-o.3 ~ I

-o.4 -
-o.5 ~
-0.6 J
-o.7 1

-0-8 I I T l l I I I
30 50 70 90 110 130 150 170 190

Tenperature (C)

Heat Flow (ng)

 

 

 

 

 

 

 

Figure 45. Typical DSC curve of slowly cooled PP sample showing crystalline phase transitions.

 

DSC Analysis of Water Quenched PP Matrix

 

Heat Flow (Wig)

 

 

 

30 50 70 90 110 130 150 170 190
Temperature (C)

 

 

 

Figure 46. Typical DSC curve of water cooled PP sample without crystalline transition region.

5.4 Effect of the Matrix Mechanical Properties on the Interfacial Adhesion
According the mechanical testing result shown in Table 4, the interfacial shear

strength results between Tencel fiber and PP should be different due to mechanical

7O

property difference in the matrix materials. As shown in Figure 47 and Figure 48, the
interfacial shear stress values between Tencel fiber and various pure PP matrices with
different mechanical properties indeed agree with the shear-Lag model proposed by H. L.
Cox. The PP with the highest shear modulus (Profax PH131) produced higher interfacial
shear strength than all the other grade of PP matrices. The interfacial adhesion data also
agrees well with glass fiber and PP composites since the natural fiber has very similar
surface functional group compared with glass. According to the study published by
Bogoeva-Gaceva et al, the interfacial adhesion between a glass fiber and pure

polypropylene should be around 7.2 MPa [66].

 

10 14

 

— Amorphous PP
+em'Gm".5

O
n

Interfacial Shear Strength (MPa)
em'Gm‘5 (kPa*.5)

 

 

 

Profax Profax Profax Profax Profax Profax
6301 6501 6601 PH020 PH180 PH131

 

 

 

E . ’
Figure 47. Plot of IFSS and "‘ G'" for six PP matrices

Taking into account the change in modulus and replotting the data showed earlier
in Figure 44, Figure 47 shows that the IFSS now does follow the expected dependency as
proposed by Cox’s analysis. For the same interfacial chemistry and varying the shear

modulus of the matrix materials, correlation between interfacial shear strength and

71

8’" . G“ was very close to l (R2=0.96) as shown in Figure 48. This result confirms

that the interfacial adhesion (in the absence of chemical bonding across the fiber-matrix
interface) depends on van der Waals interactions and better stress transfer is the result of

higher polymer matrix modulus.

 

10

 

Interfacial Shear Strength (MPa)

 

 

 

4 -
y = 0.80x - 1.76
R2 = 0.96
2 .
0 r r . r
7 8 9 10 11 12
em*(Gm".5) (kPa*.5)

 

 

 

Figure 48. Plot of interfacial shear stress vs. the 8’" ' “G" for various PP matrices

5.5 Effect of Reactive PP Matrices on Interfacial Adhesion

In order to investigate the effect of a reactive PP matrix on the interfacial
properties, maleic anhydride grafted PPs (maPP) were used in the Microbond testing.
The EpoleneTM (maPP) obtained form Eastman Chemical Company were use as matrix
material in the microbond test samples. This will ensure the direct contact of maPP with
the fiber and eliminate any ambiguity about the composition of the interfacial chemistry.

5.5.1 Proposed maPP Reaction Chemistry with Cellulose Fiber

72

According to Felix et a1 [35], treatment of cellulose fibers with maleic anhydride
grafted polypropylene can form covalent bonds across the fiber—matrix interface under
the typical processing temperature encountered during PP processing. The possible
chemical reaction process between maleic anhydride grafted polypropylene and cellulose
fiber is illustrated in Figure 49. The mechanism of the reaction can be divided into two
steps: (1) activation of the maPP copolymer by heating (170 C), and (2) esterification of
the cellulose. If chemical bonds were formed between the cellulose fiber and the maPP

matrix, then the interfacial adhesion between Tencel and maPP would be higher than for

Tencel and PP.
Activation: % 7‘:
I"2 5 o H 5
HO—C — (I: k A u_ c2 E:
HO— — c—c —"—' 0< I + H20
H fi— C—C
.5
H 2
Esterification: —o—c _ ([2 E;

PP Chain

/ _°_ _ 3":
\

+
I T—O 1
WOW

°" 3’ 11-22}?
| a
— —c

m 3

Figure 49. Possible maPP reaction with cellulose fiber

5.5.2 Interfacial Adhesion between Tencel Fiber and maPP

According to our mechanical testing of the various PP and Epolene matrices, the

Profax ph020 had the same shear and tensile modulus as those of EpoleneTM G3003 and

G3015 (See Figure 50). Hence, by comparing the interfacial shear stress data of the
Tencel/Profax ph020 with those of Tencel/Epolene matrix system, the difference must be
due to the difference in matrix chemistry (i.e., the maleic anhydride modified PP will
react with the Tencel fiber surface and produce a stronger fiber-matrix interface). The
tensile and shear modulus values were obtained from testing of standard dumbbell-

shaped PP specimens on the UTS using ASTM standard D638-89.

 

 

 

 

 

 

800 4.0
I“ Shear Modulus
70° 4 +Tensiie Modulus 3'5
E 600 ~ ' 3-0 ’3
= 6
; soo- 2. :
. =
O _
g 400 - 2.0 §
8 a
E 300 — 1.5 g
‘- o
3 s
5 200 « 1.0 r-
100 « 0.5
04 .._.. __ -- >0-o
Profax Profax Epolene Epolene Profax Profax Profax Profax
6301 6501 63003 @015 ph020 6601 phlso ph131

 

 

Figure 50. Plot of tensile and shear modulus for various PP matrices

From the experimental data shown in Figure 51, there are indeed large differences
in the interfacial shear stress values between pure PP and maPP despite the similarities in
their mechanical properties. As the experimental results indicated, the maPP matrices
increased the interfacial adhesion between Tencel and polypropylene significantly. This
is further proof that the interfacial chemistry plays an important role in the fiber matrix
adhesion. It’s also interesting to note that the maPP with higher maleic anhydride content
(Epolene G3015) did not produce any higher interfacial adhesion result than the one with

less maleic anhydride (Epolene G3003). However, the differences in interfacial shear

74

strength between these two samples are statistically insignificant when using a t-test with
an or value of 0.05. The slightly lower average IFSS of the Epolene G3015 sample could
be the result of chain scission of polypropylene due to the high temperature experienced

by the polymer during sample preparation causing a decrease in matrix modulus [62].

 

 

Interfacial Shear Stress (M Pa)
O

 

 

   

 

 

PH020 Epolene G3003 Epolene G3015

 

Figure 51. Comparison of IFSS between Tencel/PP and Tencel/maPP

Base on the interfacial chemistry, the higher interfacial adhesion between the
Tencel fiber and maPP should be primarily due to the chemical bonding at the fiber-
matrix interface. In addition, the magnitude of the adhesion increase due to covalent
bonding between maPP and the cellulose fiber agrees well with the research data
published by Felix et a1 [65] and Bogoeva—Gaceva et a] [66]. In the study published by
Felix et al, the interfacial adhesion between a rayon fiber and polyethylene was improved
by over 160% using oxygen plasma treatment. Bogoeva-Gaceva et al also achieved a
120% increase in the interfacial adhesion between glass fiber and polypropylene using

maleic anhydride modified PP.

75

5.6 Summary

From this investigation of adhesion between Tencel and polypropylene, it’s clear
that matrix physical and chemical properties can have a great effect on the interfacial
adhesion. Due to the semicrystalline nature of the polypropylene matrix, care must be
taken to eliminate the effect of crystallinity when comparing interfacial adhesion values.
From the study of Tencel-PP and Tencel-maPP samples, the change in the interfacial
adhesion due to changes in interfacial chemistry, are most certainly caused by covalent
bonding between the cellulose fiber and the polymer matrix. To further improve the
efficiency of interfacial chemical bonding, an oxygen plasma treatment of the cellulose
surface might be necessary in order to increase the reactive hydroxyl group on the fiber
surface. Interfacial compatibilizers such as maleated polypropylene (maPP) can be used
to effectively improve the natural fiber to polypropylene adhesion. In addition, maPP can
be applied during the fiber-matrix mixing step in the extruder. Just as important however
is the modulus of the matrix in the interface region which can be modified by processing

condition or molecular weight and crystallinity of the matrix.

76

CHAPTER 6 ADHESION STUDY USING HENEQUEN AND EPOXY

 

The fiber-matrix adhesion results presented in previous two chapters were based
on carefully controlled model studies using a regenerated cellulose fiber (Tencel) because
of its topographic regularity, uniform surface chemistry and morphological properties.
However, natural cellulose fibers have much more complex surface chemistries and
morphologies than regenerated cellulose fiber. Even the same kind natural fibers under
different growth conditions can exhibit great differences in their chemical compositions
and mechanical properties. In addition, natural fibers have a much rougher surface and
mechanical interlocking (friction) at the fiber-matrix interface becomes another important
factor in the consideration of natural cellulose fiber to polymer adhesion. Therefore, it’s
important to apply the results of the regenerated cellulose fiber study to a natural
cellulose fiber composite system to determine if the same factors are equally as important.

For this part of the study, the regenerated cellulose fiber was replaced with a
natural fiber - henequen. Henequen fibers were modified with plasma treatment and
silane treatment as done with the Tencel fibers. The fiber surface chemistry and
topography were analyzed using XPS and AFM after surface treatments. A modified
microbond technique was used to determine the interfacial adhesion between surface-
treated henequen fiber and epoxy. The modification of the microbond test was necessary
due to the relative large diameter of the henequen fiber. To further investigate the
existence or absence of chemical bonding at the fiber matrix interface, the plasma treated
henequen fibers were chemical treated with propylene oxide and ethyl amine to eliminate

any possible reactions sites on the fiber surface that can react with the epoxy matrix.

77

These two chemicals were chosen because they have the same organic functionality as

the epoxy.

6.1 Microbond Sample Preparation and Testing Procedures

Due to the larger diameter of the natural henequen fiber, the placement of epoxy
drops on a single henequen fiber is not practical because the size of the epoxy drop is too
large. As a result, a modified sample preparation techniques and adhesion test was used.
The Microbond samples were prepared as shown in Figure 52 (a). A 1 mil thick
aluminum ring was used as a mold for the epoxy matrix that surrounds the henequen fiber.
An epoxy matrix disk is formed by placing a small drop of the epoxy between the inner
edge of the aluminum ring and the henequen fiber. A thin disk of epoxy is formed
between the fiber and ring by the force of surface tension. The disk thickness varies from
much thicker at the fiber matrix interface to thinner at the edge of aluminum ring.
According to the experimental measurements, the thickness of the epoxy disks that

surrounds the henequen fiber was usually between 400um to 750um.

 

Henequen Fiber

Epoxy Matrix

Aluminum Ring
(1 mil thick)

 

 

 

 

 

(a) (b)

Figure 52. (a) Microdrop pullout sample schematic; (b) Picture of prepared microbond sample.

78

Due to the large force required to debond the epoxy from the henequen fiber
(usually 2 to 6 lb depend on the epoxy drop size and bonding condition), the fiber pullout
samples were tested using an Instron tensile tester (UTS). One of the fiber ends was fixed
to a small aluminum tab with cyanoacrylate glue so the tensile tester can grab the fiber
without damaging it (See Figure 52 (b)). The experimental data was collected by pulling
on the aluminum tab while holding the matrix disk in place. The entire testing setup is
shown in Figure 53 and a 20 lb load cell was used for the debonding force measurement.
The testing speed of the microbond test was set to 0.05 in/min, which is the lowest speed

for this particular UTS machine.

   

Figure 53. (a) Microbond sample testing setup and (b) Overview of setup with Load cell.

The apparent fiber-matrix shear strength was determined using the following equation:

DIX

Interfacial Shear Strength (1'), 1' = 2' = l
p x

l

where Tu is the ultimate interfacial shear strength, Fmax is the maximum load recorded by

the load cell, p is the perimeter of the fiber, and l is the thickness of the matrix disk at the

79

fiber-matrix interface. According to the model proposed by Li [61], at short fiber
embedment length, 'r can be assumed to be constant throughout the interface and equal to
the ultimate interfacial shear strength (Tu). In order to determine the two parameters (1 and
P) for the interfacial shear strength calculation, the debonded sample were examined
microscopically and analyzed with image analysis software. After the fiber completely
debonded from the epoxy disk, an image of the epoxy disk and the remaining hole were
taken. The epoxy disk thickness can be measured directly using the video caliper by

assuming the ends of the matrix disk is very flat (See Figure 54).

 

Figure 54. Embedment length measurement using video caliper

However, the parameter of the hole must be measured using image analysis
software. The image analysis of the fiber parameter was necessary because the henequen
fiber is very irregular in its shape and assuming a circular cross-section is simply
inadequate. In this research, the image analysis software used was “ImagePro Plus” and it
can measure the length of the fiber perimeter by tracing a line around the residual hole in
the matrix disk (See Figure 55). After obtaining the length of the fiber perimeter in the

unit of pixels, it was then converted to micrometers using a calibrated conversion factor.

80

~nol4goodlnnpl1/1]

nol4go

 

Figure 55. Image analysis of fiber-matrix contact perimeters.

As one can see from Figure 55, the hole left by the debonded henequen fiber was not
circular at all. If we have assumed a circular shape for the henequen fiber, the interfacial
adhesion result would have been wrong. If an average diameter value were assumed
instead of image analysis, the interfacial shear strength result could be off by as much as

31% for the sample shown.

6.2 Effect of Plasma Modification of Henequen Fibers
6.2.1 Plasma Treatment Procedures

To investigate the effect of fiber surface chemistry on the interfacial adhesion, the
fibers were modified with oxygen plasma using the PSOSOO plasma surface treatment
system from AIRCO. The henequen fibers were treated with oxygen plasma for 0, 2, 4, 8
and 16 rrrinutes at 412.5 W (13.56MHz) and lUmin 02 flow rate with a base pressure of

0.05 Torr. The henequen fiber was also treated with argon plasma at the same condition

81

for 8 minutes to check for the effect of physical changes of the fiber surface on interfacial
adhesion. Argon plasma treatment is known to increase the surface roughness with
minimal changes in the surface chemistry. After the plasma treatment, the treated
henequen fibers were analyzed by XPS and AFM to quantify the changes to their surface
chemistry and surface topology.
6.2.2 Adhesion between Plasma Treated Henequen and Epoxy

The interfacial adhesion between the henequen fiber and the epoxy was measured
after the plasma treatment of the henequen fiber surfaces. As shown in Figure 56, the
interfacial shear strength between henequen and epoxy was significantly improved by the
plasma treatment of the fiber surface. The interfacial shear strength was increased by

over 100% after 8 minutes of oxygen plasma treatment.

 

30

 

25‘

Interfacial Shear Strength (MPa)
a

 

 

 

 

Untreated 2min 02 Mn 02 8min 02 16nin 02 BilinAr

 

Figure 56. Interfacial shear strength between henequen fiber and epoxy.

However, the longer treatment time (16 minutes) did not produce an increase in
interfacial adhesion than the 8-minute treatment. The interfacial shear strength

differences among the oxygen plasma treated samples after 4 minutes are statistically

82

insignificant according to t-test with an or value of 0.05. In addition, the argon plasma
treatment also produced higher interfacial adhesion than the untreated sample.
6.2.3 Change in Debond Curves due to Oxygen Plasma Treatment

Since the microbond test was conducted using the UTS, the debonded curves for
the various tests were recorded. According to the debond profiles, the oxygen plasma
treated henequen samples had much more linear force vs. extension profile (Figure 58)
which verifies the fact that oxygen plasma treated henequen fiber had better adhesion to
the epoxy matrix. The gradual failure mode of the untreated henequen fiber, shown in
Figure 57, could be explained by incremental failure of the interfacial bonding. The
linear profiles for oxygen plasma treated samples (shown in Figure 58) are more

consistent with catastrophic failure of the fiber-matrix interface.

 

3.0

 

P
01
L

       

P
o

1

Debond Force (lb)
;

 

 

 

0.0 0.5 1 .0 1 .5 2.0
Extension (°/o)

 

 

 

Figure 57. Force vs. extension (debond) profiles for untreated henequen and epoxy

83

 

3.0

 

“\ \\\ \ \
X§\
\ S

‘x
\

Debond Force (lb)

 

 

 

 

 

 

 

 

 

 

 

T I

2.0 2.5 3.0
Extension (7.)

 

 

 

 

Figure 58. Force vs. extension (debond) profiles for oxygen plasma treated henequen and epoxy

6.2.4 Henequen Fiber Surface Topography vs. Adhesion

The change in the interfacial adhesion could be attributed to a combination of
physical changes (surface roughening) and chemical changes (addition of surface
functional groups) produced by the plasma. However, to distinguish between the physical
and chemical effects, the fiber surface must be analyzed for topographical changes. The
fiber surface topography of plasma treated henequen fiber was examined using AFM to
correlate the results to the interfacial adhesion. The surface height images of the
henequen fiber before and after each treatment were collected using a multimode
Scanning Probe Microscope (SPM) from Digital Instruments (Contact AFM Mode) with

a standard silicon nitride probes (spring constant of the probe used is 0.06 N/m).

84

5pm x Sum

 

is? 300nm
Figure 59. AFM height images of (a) untreated, (b) 2 minutes 02 plasma treated, (c) 4 minutes 02

plasma treated, and (d) 8 minutes 02 plasma treated henequen fiber surface.

As shown in Figure 59, the fiber surface appeared to be rougher after the plasma
treatment than the untreated case and the roughness was directly related to the treatment
time. This increase in the surface roughness could have contributed to the increase in the
interfacial adhesion between plasma treated henequen fiber and epoxy. However, due to
the roughness of the original fiber surface, the surface alteration by plasma did not
produce any significant change in roughness factor data (See Figure 60). The

morphology changes due to plasma were usually microscopic in size.

85

 

 

 

 

AFM Surface Analysis - Roughness Factor (Ra)
250
200 -
150 -
G
a:
100 -
5° ..
o -
Untreated 1nin 02 min 02 Min 02 Mn 02 aninAr
Piasrm Piasrm Plasrm Plasma Piasrm

 

 

 

Figure 60. Roughness factor of imaged henequen fiber surface from AFM innge analysis.

 

 

8

cl
O
I

-h
M
1

fi
1

 

 

Surface Area increase from 25 p, m2 (%)

O
I

Untreated 1m|n02 2mln02 4min 02 0min 02 8minAr
Plauna Fianna Piauna Plauna Plasma

 

 

 

Figure 61. Surface area change of henequen fiber before and after plasma treatment. Values indicate
the difference between measured surface area and 25 ml.

In order to correlate this increase in surface roughness/area to interfacial adhesion,

the specific surface area of the henequen fiber was obtained through image analysis. As

86

the surface area analysis from the AFM measurement indicated (Figure 61) that the
plasma treatment of the henequen fiber roughened the fiber surface and produced
measurable changes in specific surface area. All the surface area data was evaluated at a
scan size of 5 by 5 micron square (i.e., a scan area of 25 umz). The surface area of the
plasma treated henequen fiber seemed to linearly dependent on the treatment time. Due
the nature of the natural fiber (very rough), the standard deviation of the area results were
very high. However, the results are never the less significant according to the AFM
topography images.

Since the surface roughness can have an effect on the interfacial adhesion value,
this factor must be taken in to account before determining the effect of chemical bonding
on fiber-matrix adhesion. If we adopt the model used by Li [61], Brittle Failure with
Maximum Friction, then the interfacial adhesion value could be corrected by assuming
the frictional sliding stress is constant. e.g.,

1:2",ef +AA%xz'f

Where Trcf is the reference interfacial shear strength (untreated henequen), AA is the
change in surface area and If is the frictional sliding stress.

In order to determine the value of 1;, an 8 minute argon plasma treatment was
performed to roughen the henequen fiber without significantly altering the surface
chemistry. According to the XPS analysis, the surface chemistry of the 8 minute argon
plasma treated sample was almost identical to the 2 minute oxygen plasma treatment (the
oxygen impurity in the argon gas and oxygen residual in the plasma chamber could have

contributed to the chemical alteration of the fiber surface). As a result, the interfacial

87

shear strength difference between the two samples could be attributed primarily to their
surface area difference and the value of If can be determined using the following equation:

__ Interfacial Shear Strength(2 min 02) - Interfacial Shear Strength(8 min Ar)
f Surface Area(2 min 02) — Surface Area(8 min Ar)

 

After the determination of the sliding stress constant If, the interfacial shear stress
values between plasma treated henequen and epoxy was adjusted for the fiber surface
area change. As shown in Figure 62, the interfacial shear strength corrected for the
change in surface roughness (the dotted line), was still far lower than the measured
results. It clearly proves that the change in interfacial adhesion due to plasma treatment
could not be explained by a physical effect alone (roughening) and a chemical interaction
must be responsible. Furthermore, according to the results, the chemical interaction is

not only important but also the dominant factor in the interfacial adhesion improvement.

 

 

 

 

 

 

 

 

30
E 25 .
E.
5
a
S 20 -
35
g 15 — ............. '
§ 2 """"""""
a 1P
5 1° ‘ A 02 Plasma
E - - - -Adheaion=f(AA)
5 i i I ‘fi i
0 2 4 6 8 10 12
Change in Surface Area from Untreated Henequen (%)

 

Figure 62. Plot of interfacial adhesion vs. the changes in fiber surface area after oxygen plasma
treatment.

88

6.2.5 Henequen Fiber Surface Chemistry vs. Adhesion

The chemical effect of the plasma treatment was analyzed using XPS. Through
XPS analysis, the surface atomic concentration of oxygen and carbon was determined for
each plasma treated sample (See Figure 63). The functional groups such as hydroxyl
(OH), carbon double bonded to oxygen (C—20) and carbon triple bonded to oxygen (C-
30) were determined by curve fitting of carbon is and oxygen ls peaks (See Figure 64).
The XPS results indicated that the surface chemistry of the natural henequen fiber was
significantly altered with increasing plasma treatment until the surface chemistry reached
a saturation level at around 8 minutes. Similar to the oxygen plasma modification of the
Tencel fiber, the surface hydroxyl concentration was affected the most. As the results
indicated, oxygen plasma treatment after 8 minutes did not produce a further increase in

hydroxyl concentration.

 

 

 

 

 

 

 

 

 

 

1.0 .- -
4

o
.5 __ I
I: .
Q 0.83
O
8'
.. 0.6 ~
‘5 .-
H RT r
§ :5 s j:
c 0.4 ~
0
O
0
'g 0.2 +010
0 + C
“
< + O

000 I I T I I I I I I

0 2 4 6 8 1O 12 14 16 18
Treatment Time (nin)

 

 

 

Figure 63. Atomic concentrations on henequen fiber surface vs. oxygen plasma treatment time (XPS).

89

 

0.4

 

0.3 .

 

 

Functional Group Fraction

 

 

 

o 2 4 6 s 10 12 14 16 18
Treatment Time (min)

 

 

 

Figure 64. Surface functional group concentration on henequen fiber surface vs. oxygen plasma
treatment time (XPS)

6.2.6 Relationships between Interfacial Adhesion and OH Concentration and O/C Ratio
In order to relate the interfacial adhesion results to the changes in the surface
chemistry, the interfacial shear strength data was plotted vs. the oxygen to carbon ratio
(O/C) and vs. the hydroxyl group (OH) concentration. According to the plots in Figure 65,
the interfacial shear strength results are linearly dependent on the overall oxygen to
carbon ratio on the fiber surface and also the overall hydroxyl group concentration. The
linear relationship between the interfacial adhesion and surface functional groups can be
explained by the number of chemical bonds formed at the finer-matrix interface. More
functional group available on the fiber surface simply increases the number of chemical
bond formed. If we assume each of the chemical bond (between the fiber and the matrix)
have the same energy, then the total energy that it takes to break all the bonds will simply

proportional to the number of chemical bonds present at the interfacial region.

90

 

 

 

 
  
 
  

 

 

 

30 _ _
E 25 ‘ y = 94.60x - 1.92
E 2
V R = 0.91
5 20 -
Oi
:
2
<7) 15 ~ y =17.62x + 7.84
I- 2 _
3 I R - 0.93
25 1o -
Ti
“6
g 5 « a Adhesion vs. OHConcentration
., - Adhesion vs.OICRatlo
E
o I I r r
0.0 0.2 0.4 0.6 0.8 1 .0
0H Concentration or OIC Ratio

 

 

 

Figure 65. Correlations between interfacial shear strength and henequen fiber surface chemktries
after oxygen plasma treatment

The combination of interfacial shear strength data and the XPS analysis clearly
shows that the interfacial chemistry plays an important role in the determination of
natural fiber to polymer matrix adhesion. In addition, when large concentration of
reactive groups such as hydroxyls are available on the fiber surface, the physical
roughness of the fiber only plays a minor role in the determination of the fiber-matrix

adhesion.

6.3 Effect of Eliminating Chemical Bonding between Henequen and Epoxy

To prove that chemical bonds were primarily responsible for the improvement in
interfacial adhesion, further experiments designed to eliminate any chemical reactions
between the henequen fiber surface and the epoxy matrix were conducted. The
monofunctional chemical propylene oxide (PO) and ethyl amine (EA) were reacted with

the plasma modified henequen fiber surface prior to the formation of the fiber—epoxy

91

adhesion specimens to eliminate any chance of chemical reaction with the epoxy matrix.
If the interfacial adhesion improvements were truly the result of chemical bonding
between the cellulose fiber and the matrix polymer, then the elimination of the chemical
reaction sites should decrease the interfacial adhesion back to the level predicted by the
dotted line in Figure 62 (Li’s model).
6.3.1 Chemical Treatment Procedures

After the oxygen plasma treatment, part of the henequen fibers were further
treated with propylene oxide (PO) and then with ethyl amine (EA) to eliminate the
reactive sites on the fiber surface with respect to epoxy. Since PO and EA have the same
reactive organic functionalities as the epoxy, pretreatment with those chemicals should
consume the reactive surface sites and prevent any further chemical reaction of the
henequen fiber with the epoxy matrix. The henequen fibers were treated by immersion in
a diluted solution of PO (66wt% in acetone) for 10 minutes. After the PO treatments, the
henequen fibers were dried at the curing temperature of the epoxy. The curing
temperature of epoxy was used for fiber drying to ensure any reaction between the
henequen fiber and the epoxy resin (DGEBA) will also take 'place between henequen
fiber and PO. Following the PO and the heat treatment, PO treated henequen fiber was
then immersed in a diluted solution of EA (33wt% in acetone) for 10 minutes as in the
case for P0 treatment. The fiber was then dried at the curing condition of epoxy again to
ensure complete reaction between the henequen fiber and EA. After the PO and EA
treatment, the treated henequen fibers were extracted with pure acetone for 4 hours and

then vacuum dried to eliminate any unreacted P0 or BA from the fiber surface. The

92

chemically treated fiber was also analyzed with XPS to quantify the chemical changes on
the fiber surface.
6.3.2 Effect of PO and EA Treatment on Henequen Fiber Surface Chemistry

The XPS analysis of the henequen fiber indicated that the oxygen content of the
fiber decreased to the untreated level after the chemical treatment with propylene oxide
and ethyl amine. As shown in Figure 66, the oxygen to carbon ratio (O/C) of the
henequen fiber surface went down from the plasma treated case to a relative flat level of
0.27. This result indicates the remove of oxygen functionalities from the surface of the

fiber due to chemical treatment.

 

 

  
  

 

 

 

 

1.0 —
§ 0.8 4
1:
:i
(I)
§ 0.6 ~
it
1:
O / \u
3 0.4 , 4; {\\
s is “ “MM—ms
o 0.2
\
0 —e— Oxygen Plasma
0 0 -—a— Oxygen Plasma + P.O. «1» EA.
0 2 4 6 8 10 12 14 16 18
Treatment Time (min)

 

 

Figure 66. XPS analysis of the henequen fiber surface after plasma and chemical treatment

Curve fitting of the Cls and Ols peaks also indicated a major decrease in the
hydroxyl group concentration on the fiber surface. As indicated in Figure 67, the
hydroxyl group on the fiber surface decreased from the initial baseline level of 30% down

to the untreated level of around 15%.

93

 

 

 

 

 

 

0.55
0.30 -
g 0.25 ~
g .
u- 0.20 i
3' .
o 0.15%
h
6
a:
o 0.101
0.05 _ —e—Oxygen Plasma
+ Oxygen Plasma + P.O. + EA.
0.00 I i I I i T i i
0 2 4 6 8 10 12 14 16 18
Treatment Time (min)

 

 

 

Figure 67. XPS analysis of hydroxyl concentration on the henequen fiber after various treatments.

The XPS analysis of the henequen fiber surface clearly indicated the effectiveness
of the chemical treatment in eliminating the reaction functional group and reducing
oxygen content on the fiber surface. However, the chemical reaction occurred here might
not be a terminating reaction. According to the XPS analysis of the fiber surface, no
increase in nitrogen atomic concentration was detected.

6.3.3 Effect of PO and EA Treatment on Henequen to Epoxy Adhesion

After the chemical treatment with P0 and EA, the henequen to epoxy adhesion
was again tested. As shown in Figure 68, the interfacial adhesion after the chemical
treatment indeed decreased to dramatically to a steady level of 15 MPa for plasma treated

henequen samples.

94

 

 

30 _ l 02 Piasrm

El 02 PIasna+PO+EA
25 .
20 -

Interfacial Shear Strength (MPa)

 

 

 

15— i ,
10- f: , .

Untreated 211m 02 Mn 02 Orrin O2 16nin O2 Orrin k

 

 

 

Figure 68. Interfacial adhesion before and after P0 4» EA treatment.

 

8

 

N
a"
r

N
O
r

 

10 - A 02 Plasma
I O2Plasma+PO+EA
- - - -Mhesion=f(AA)

interfacial shear Strength (MPa)

 

 

l

 

I I i i

2 4 6 8 10 12
Change in Surface Area from Untreated Henequen (%)

 

 

 

Figure 69. Interfacial adhesion comparison before and after chemical treatment by P0 and EA of
plasma treated henequen

If we again plot the adhesion vs. change in fiber surface area, the adhesion values
after PO and EA treatment are very close to the values predicted by the earlier linear

relationship (See Figure 69). This result strongly agrees with the assumption that when

95

chemical bonding is eliminated from the fiber-matrix interface, the interfacial facial can
be predicted base on the mechanism of physical interlocking. This experimental result
further validates our assumption that the interfacial chemical bonding is the major factor
that controls the interfacial adhesion between cellulose fiber and epoxy matrix.
6.3.4 Correlation between Adhesion and OH or O/C

However, not all functional group (OH) detected by the XPS can react with the
epoxy matrix. As shown in Figure 70, even though there are around 15% of hydroxyl
group on the fiber surface after the propylene oxide and ethyl amine treatment, the
positive correlation between the interfacial adhesion and the hydroxyl concentration no

longer exists.

 

 

 

164 FLT '1

12~

y = 1.77x + 14.15 y = -1.14x + 14.84
s— m=am m=am

Interfacial Adhesion (M Pa)

a Adhesion vs. OH Concentration
a Adhesion vs. OIC Ratio

0 I I I r I I
0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45

OH Concentration or OIC Ratio

 

 

 

 

 

 

Figure 70. Correlation between interfacial adhesion and henequen fiber surface chemistries after 02
plasma + PO + EA treatment

This remaining hydroxyl content on the fiber surface could be due to steric hindrance,
strong hydrogen bonding, or a surface analysis artifact that XPS can measure and detect

groups to the depths as great as lOnm below the external surface. This result further

96

supports our conclusion that the surface chemical interactions are the primary factor
responsible for adhesion between the natural fiber and the epoxy matrix polymer.
Without the active chemical bonding between the cellulose fiber and the polymer matrix,

the interfacial adhesion no longer correlates to the surface chemistry of the cellulose fiber.

6.4 Effect of Silane Treatment of Henequen Fiber
6.4.1 Silane Treatment

An alternative approach to improving adhesion is to make use of the hydroxyls
which naturally occur on the cellulose surface. The effect of silane treatment on the
interfacial adhesion between henequen fiber and epoxy matrix was studied using an
epoxy functional silane coupling agent - Dow Corning Z-6040 silane (3-
Glycidoxypropyl-trimethoxysilane). This silane agent was chosen based on its ability to
react with both the henequen fiber and the epoxy matrix. Oven dried henequen fiber was
treated with Z-6040 silane for 2 hours solution under acidic condition. The solution was
prepared by adding 2wt% Z-6040 silane to 500g of acetone-water solution (9Swt%
acetone to 5 wt% water). The pH of the solution was approximately equal to 4. After the
treatment, the henequen fiber was dried in a convection oven for 4 hours at 100C.

The silicon atom is present in the silane molecule and therefore can indicate the
presence of silane, coupling agent on the surface after treatment. Base on the XPS
analysis of silane treated henequen fiber there was strong evidence that silane was present

on the surface of the fiber after 24 hours soxhlet extraction with ethanol.

97

Table 8. XPS analysis henequen fiber surface before and after Z-6040 silane treatment.

 

 

 

 

Fiber Sample C (%) O (%) Ca (%) Si (%)
Untreated 78.95 19.03 2.02 0
Z-6040 Silane Treated 79 18.05 1.77 1.18

 

 

 

 

 

 

According to the atomic ratio data (shown in Table 8), there was around 1.2
atomic % of silicon atom (Si) present on the fiber surface. This silicon concentration was
very close to the silicon concentration detected for the 26040 silane treated of Tencel
fiber (~1.3 %). The XPS result again showed the reaction between silane and henequen
fiber does take place but in small amount. This could be caused by not all the hydroxyl
group are accessible to the silane reagent.

6.4.2 Adhesion between Silane Treated Henequen and Epoxy

The pull-out adhesion samples for silane treated henequen and epoxy were
prepared and tested using the procedures described in the previous section. The
microbond results showed that the Z-6040 silane treated henequen fiber had better
adhesion to epoxy than the untreated fiber. As shown in Figure 71, the improvement in
interfacial adhesion due to silane treatment is about 30%. The interfacial adhesion result
for Z-6040 silane treatment agrees well with past research when compared to silane
treatment of glass fiber. According to the study published by Zhao et a1 [67], the
interfacial adhesion between a glass fiber and epoxy was improved by 36% (compared to
water sized) using epoxy-silane treatment which is comparable to this study. However,
the absolute interfacial shear strength value between henequen and epoxy is much lower
than the value published by Zhao et. al, possibly due to the higher modulus epoxy matrix

used by Zhao and coworkers.

98

 

 

 

 

   

 

Interfacial Shear Strength between Henequen Fiber and Epoxy
20

’8
g 16 .
i
5 12 -1
i
3; a -
2
.9
0
s 4 -
o
E

o -

lhtreated lbnequen 2wt%26040 Treated f'bnequen

 

 

 

Figure 71. Interfacial shear strength compar'son between untreated and silane treated henequen
fiber and epoxy matrix.

6.4.3 Change in Debond Curves due to Silane Treatment

The microbond test was conducted using the UTS, which is capable of recording
the entire load-displacement curve while the pull-out test is underway. Inspection of the
debonded curves indicates that the silane treated samples produced much more linear
force vs. extension profiles (See Figure 72) which indicates that silane treated henequen
fiber has better adhesion to the epoxy matrix. According to the debond curves, the silane
treated henequen samples had more linear force vs. extension profiles. However,
comparing to oxygen plasma treatment, the force vs. extension curves is not as linear.
The load vs. extension profiles are also similar to those published by Zhao and coworkers

[67].

99

 

9"
o

 

.N
or

1°
9
1

 

Debond Force (lb)
2 ;

 

 

 

 

 

9
0|
1

\

 

 

 

P
a

I

0.0 0.5 1.0 1.5 2.0 2.5
Extension (%)

 

 

 

Figure 72. Force vs. extension (debond) profiles for Silane treated henequen and epoxy

6.5 Plasma-Silane Treatment of Henequen Fiber

Even though the silane treatment produced higher interfacial adhesion than the
unheated case, the improvement was not as significant as for the plasma treatment alone
(See Figure 73). This could be caused by the limited reaction sites available on the
untreated henequen fiber surface to react with the silane reagent.

To improve the silane reaction with the henequen fiber surface, plasma treatment
was used in combination with the silane treatment. Since the oxygen plasma has been
shown to produce a higher concentration of reactive hydroxyl groups on the henequen

fiber surface, subsequent application of silanes should have a greater chance for reaction

with the fiber surface.

100

 

35

 

30

25*

 

Interfacial Shear Strength (MPa)

  

 

 

Untreated Z-6040 4 min 02 8 min 02 16 min 02 l
Silane Plasma Plasma Plasma
Treated Treated Treated Treated

 

 

Figure 73. Interfacial shear strength data comparison for different fiber treatments

6.5.1 Fiber Chemistry after Plasma-Silane Treatment
XPS analysis of oxygen plasma and silane modified fiber was performed to see if '

the oxygen plasma has produced more reactive site on the fiber surface.

Table 9. XPS analysis of henequen fiber before and after plasma+ silane treatment.

 

 

 

 

 

Fiber Sample c (%) o (%) Ca (%) sr (%)
Untreated 78.95 19.03 2.02 0
Z-6040 Silane Treated 79.00 18.05 1.77 1.18
8 min 02 Plasma + Z-6040 70.73 26.63 0 2.64
8 min 02 Plasma + Z—6070 67.96 27.06 0 4.97

 

 

 

 

 

 

 

As shown in Table 9, the surface silicon atomic concentration was indeed higher for
oxygen plasma and silane treated samples than the silane only treatment sample verifying
the hypothesis that a higher silane grafting concentration is possible by creating more

reactive sites on the henequen fiber surface. However, since the reactive epoxy groups on

101

the silane can form a covalent bond with the epoxy matrix in the same way as the epoxy
matrix reaction with the hydroxyl on the fiber surface, the plasma and silane treatment
probably won’t produce a significant increase in the interfacial adhesion. Furthermore,
since only ~5% of Z-607O grafted to the fiber surface, the decrease in adhesion will not
as great as the propylene oxide (PO) and ethyl amine (EA) treatment shown earlier.
6.5.2 Effect of Plasma-Silane Treatment on Henequen to Epoxy Adhesion

Two type of plasma-silane treated henequen fiber samples were tested for
interfacial shear strength with epoxy matrix. The epoxy functional Z-6040 silane was
designed to couple the natural fiber to epoxy matrix through chemical bonding since the
amine in the matrix can react with the epoxy ring of this silane. The methyl functional Z-
6070 silane was utilized to present any chemical bonding between treated fiber and the
epoxy matrix since methyl group can not react with either the epoxy resin or the curing
amine. The microbond test results are shown in Figure 74. As expected, the interfacial
shear strength of plasma-silane (Z-6040) treated sample is much higher than silane only
treated sample. As expected, the plasma-silane treatment did not produce any measurable
difference from the plasma only case. The interfacial shear strength data of the epoxy
functionalized (Z-6040 Silane) sample was almost the same as the 8 minute oxygen
plasma treated sample. On the other hand, the methyl functionalized (Z-6070 silane)
produced an interfacial shear strength value of 20 MPa which was 20% less than the 8-
minute 02 plasma treated samples. Since the Z-6070 silane can not react with the epoxy
matrix, this result further proves the importance of interfacial bonding. However, the Z-
607 0 treated henequen fiber sample did not reduce the interfacial adhesion all the way

down to the reference level due to the limited reactivity of this particular silane with the

102

henequen fiber. The higher than expected interfacial shear strength could also be
explained by the roughening of the fiber surface by the oxygen plasma treatment. It was
indicated in the earlier section that an increase in fiber surface area can also lead to higher

interfacial adhesion.

 

l
l Interfacial Shear Strength Between Henequen Fiber and Epoxy
l

 

  

36
3040 = CchHCH10(CH2),Si(OCH,); ‘
so . V
23070 = CH38KOCH3);
25 1 -.
20

‘3] II

Untreated 8 win 02 Plan'- 8rlin 02 Plan- 8nin 02 Plan:
+ Silane 28040 + Silane 26070

Interfacial Shear Strength (MPa)

 

 

 

 

Figure 74. Adhesion comparison between plasma and plasma-silane treated henequen with epoxy.

6.6 Summary

From the various experimental results, it can be concluded that the fiber-matrix
adhesion for a natural fiber reinforced composite system can be improved by the
introduction of chemical bonding at the interface. It’s also shown that the effect of
increasing the fiber surface area using oxygen plasma treatment, hence increasing the
interfacial friction, was not as significant compared to the effect of increasing chemical
bonds across the fiber-matrix interface. It can be inferred that the interfacial adhesion is
directly related to the amount of reactive functional groups, in this particular case the

hydroxyl content on the henequen fiber surface. Base on the PO and EA treatment results,

103

when the possibility of chemical bonding is removed fiom the fiber matrix interface, the
interfacial adhesion no longer relates to the interfacial chemistry and becomes a pure
mechanical phenomenon. The lowering of interfacial adhesion due to PO and EA
treatment was probably caused by the removal of the oxygen functionalities from the
fiber surface.

The effect of silane treatment on the interfacial adhesion is more silane reagent
dependent and the reaction with cellulose fiber is rather limited. For a silane that is
fimctionalized for the matrix material (such as Z-6040), the interfacial adhesion can be
improved but not as significant as other treatment methods. When the combination of
plasma and silane treatment were used, the treatment produced the same level of (maybe
a little better) interfacial adhesion as plasma only treatment (when a compatible silane
was used). In addition, the pretreatment with oxygen plasma also significantly improved

the silane reaction efficiency with the henequen fiber.

104

CHAPTER 7 PROPERTIES OF NATURAL FIBER COMPOSITES

 

The effects of interfacial adhesion on the mechanical properties of natural fiber
reinforced composite were studied by testing unidirectional henequen fiber reinforced
polypropylene (PP) and epoxy composites. The goal of composite testing was to see what
kind of effect that an increase in fiber-matrix adhesion has on the mechanical properties
of the composites. A unidirectional composite was chosen because it’s the easiest type of
fiber reinforced polymer composite to be modeled and the analysis of the composite

properties is relatively straight forward.

7.1 Unidirectional Henequen Reinforced PP Composite

The effect of fiber-matrix adhesion on the mechanical properties of natural fiber
reinforced thermoplastic composite was studied using henequen fiber as reinforcement
and polypropylene as matrix. Unidirectional henequen fiber (both as received and treated)
reinforced polypropylene composites of different fiber volume fraction were fabricated in
the laboratory for this purpose.
7.1.1 Sample Preparation Techniques

Henequen fibers were first brushed and cleaned to get ride of any impurities that
were left over from the fiber extraction process from the henequen farm. After the initial
cleaning, the henequen fiber was compressed under heat and pressure to straighten the
fiber under a laboratory press. These fibers were then dried in a convection at 102 C for 2
hours to remove any absorbed moisture. The straightened and dried henequen fibers were

then cut to specific length and mixed with polymer matrix in an open mold and

105

compressed into composite under vacuum. The sample layout/fabrication procedure is

illustrated in Figure 75.

 

 

‘{ Base Steel Plate
Top Steel Plate
Unidirectional

Henequen + PP
Powder

 

Steel Open Mold

 

 

Vacuum Bag

 

 

\ Med

Vacuum Port

 

 

 

Figure 75. Fabrication procedure (open mold) of henequen reinforced polypropylene composite

The compression molding conditions used are listed in Table 10. Vacuum was used in the
compression molding process to reduce void in the composite. After the compression
molding, the composites block was cut into tensile coupons and tested for tensile and

impact properties using the UTS and the Izod impact tester.

Table 10. Henequen-PP Composite Molding Conditions

 

 

 

 

Steps Time (min) Temperature (C) Pressure (Psi)
Loading/Preheating 5 190 0
Compression Molding 15 190 320
Cooling 20 20 320

 

 

 

 

 

 

7.1.2 Composite Voids Analysis
During the fabrication a high content of voids can be a problem and processing

conditions need to be adjusted to minimize void formation. In order to quantify the voids

106

 

content of the fabricated composite, the composite was sectioned and the cross-sectional
image of the composites was analyzed with image analysis software. As shown in Figure
76, the void content of the composite was obtained by mapping the voids area and the

composite cross sectional area, and then the ratio of the two areas was calculated.

 

Figure 76. Determination of void content using image analysis of composite cross-section.

According to the image analysis of the composite cross-sections, the void content for the
“open mold” compression molding process can be as high as 6%. Therefore, a better
composite fabrication technique was sought to reduce the void content of the composite.
After some experimentation, it was determined that a “closed mold” process (See Figure
77) follow by compression molding can reduce the void content down to less than 0.5%.
As one can see from the “closed mold” design, the pressure from the press can compress

the voids volume and squeeze out the voids flour a PP rich composite.

107

fl Pressure

Mold: Top

 

Composite

 

Henequen-PP

T] Mold: Bottom
Pressure

Figure 77. Closed mold processing of henequen-PP composite — Cross-section View.

7.1.3 Effect of maPP on Henequen/PP Composite

To control the degree of fiber matrix adhesion, a small amount of Epolene G3015

(maPP) were mixed into the PP matrix during the fabrication of the henequen/PP

composite. The determination of the composite tensile strength and tensile modulus for

Henequen/maPP/PP and Henequen/PP composite is shown in Figure 78 and Figure 79.

 

A Tonsil; Strength (MPa)

1
l
1
l

 

 

 

Tendle Strength ve. Volume Fraction of Henequen Fiber

250

 

o 4% Epolene G3015 Treated

e Untreated PPNhtrix — lbference Sarrples .
------- Ruleofoture I
— i-taipin-Tsai Moder (2:4) ,, ‘
————— l‘hlpin-Tsai Moder (2:14) ./'

 

 

0% 20% 40% 60% 80%

W ”PEEP"? 199),, ,

 

100%

 

Figure 78. Tensile strength of Henequen/PP and Henequen/maPP/PP composites

108

 

Tendie Modulus vsVolume Fraction of Henequen Fiber

 

 

   
  
   
    
  

25 J
o 4%Epoiene63015Treated . -

e Urtreated FPNhtrix - beerence Sanpies

20 - Rileoflvbrture

——-— i-bb‘n—Tsai Mrdel (2:12)

15 — ° -- -i'hbin-Tsaihbdei(z=18)

8.5% Void

 

101 . " 0.37%Void

 

Tensile Modulus (GPa) '

 

 

 

0% 20% 40% 60% 80% 100%
Fiber Volume (%)

 

Figure 79. Tensile modulus of Henequen/PP and Henequen/maPP/PP composites

The mechanical properties of the unidirectional composite systems can be modeled using

Halpin-Tsai model with adjustable interfacial adhesion parameter g (z in the figures).

1+ V
Halpin-Tsai Model: P, = P, i
l- 1V!

Pf—PM
where z: W ,
f in

Pc , Pm, Pf: the property of composite, matrix and fiber, respectively
Vf: Volume fiaction of fiber
The results shown in Figure 78 & 79 clearly indicate that better interfacial
adhesion can improve the tensile properties of the composite. However, the tensile
modulus was very sensitive to the void content of the composite. As indicated in Figure

79, with a void content of 0.37%, the composite modulus is almost identical to the rule of

109

mixtm'es prediction. This result indicates good composite fabrication procedures can be
very important in addition to a good fiber-matrix adhesion.
7.1.4 Impact Properties of Henequen and PP Composite

The impact properties of the henequen fiber reinforced PP composite was tested
with a TMI Impact Tester (Izod). As shown in Figure 80, the improvement in tensile
properties of the compatibilized composite came at the cost of decreased impact
properties. The Epolene G3015 modified henequen and PP composite had much lower
impact strength than the un-compatibilized case. In addition, there was a very good linear

relationship between the impact strength of the composite and the fiber content.

 

 

   
    

 

 

 

izod hnpact Strength (Notched)
900
800 r e Henequen and Untreated PP m
I l'bnequen and 4% Epolene G3015 Treated PP

700 ~
5 .
5 600
5 500 «
6
5 .00.
‘6
E 300 ~

200 ~

100 .

o ‘ l r 1 r 7
0% 10% 20% 30% 40% 50% 60%
Fiber Volume (%)

 

 

 

Figure 80. Impact testing result of henequen/PP and heuequeulmaPP/PP composites.

7.1.5 Fracture Surface of the Henequen/PP Composite

ESEM images of the impact fracture surface of Henequen/PP composite were

taken to help us understand the effect of interfacial adhesion and to explain the cause of

110

 

Figure 82. ESEM image of impact fracture surface of Henequen/4%maPP/PP composite

lll

improved mechanical properties. As shown in Figure 81 & 82, the fracture surface of
unmodified Henequen/PP composite had very long fiber pullout and the 4wt% Epolene
(03015) modified composite had almost no fiber pullout. This result clearly indicated
that better interfacial adhesion was achieved with the maPP modification of the
composite interface. However, this result in fracture rather than pull—out of the natural
cellulose fiber from the matrix. The fracture process consumes less energy and therefore

the composite energy absorption decreases (i.e., loss of impact strength).

7.2 Unidirectional Henequen Reinforced Epoxy Composite

The effect of silane and plasma treatment on the mechanical properties of the
henequen and epoxy composite was also investigated. Due to the large amount of fiber
treatment variation studies, only the 2wt% Z-6040 silane and the 8-minute oxygen
plasma treated henequen fibers were used for composite fabrication and testing.
7.2.1 Composite Sample Fabrication

For the unidirectional henequen reinforced epoxy composite, a simpler fabrication
technique using a silicone mold was adopted (See Figure 83). Plasma or silane treated
fiber were dried in a convection oven for 2 hours at 102 C and then cut to the length of
the mold. For each tensile sample, 2 gram of fiber was weighted out for each mold cavity
and the fiber was laid into the mold by hand. An epoxy solution in the ratio of 100 parts
DGEBA to 45 parts of Jeffamine T-403 was prepared and poured into the mold over the
fibers. The entire mold was then degassed in a vacuum oven at a temperature of 80 C for
3 minutes. After degassing the sample, a Teflon sheet was used to cover the top of the

mold (held in place by a stationary weight) and then cured in a programmable oven. The

112

 

curing cycle of the epoxy used was 85 C for 2 hours and then 125 C for 2 hours. After the

curing process, the samples were tested for tensile properties using UTS.

 

Figure 83. Composite fabrication mold (left) and picture of sample composite (right).

7.2.2 Effect of Z-6040 Treatment on Henequen-Epoxy Composite Properties

 

Tensile Properties of Henequen Fiber Reinforced Epoxy

 

Composite (16 wt% fiber)
140 10
I=lTendle Strength 122 124

A ‘20 “ —e—Tendle Modulus +
a ; 4’ 8 E
a 100— ' g

i u
g 80 — 3 4- 6 g
E ~‘" 3
a, so - ”a: I 4 a
2 ~ g
4°- g
i- 20 *- 2 i-

0 l i — 0

 

 

    

 

 

 

 

 

Untreated Fiber 2wt% z-eo4o Treated Rule of Mixtures
Fiber

 

 

 

Figure 84. Tensile properties of henequen fiber and epoxy composite - Silane treatment.

113

Tensile samples were made with both silane treated and untreated henequen fiber
and epoxy. The experimental results showed that silane treated henequen fiber composite
had better tensile strength and relatively no change in tensile modulus (See Figure 84).
The tensile testing result indicates that the pretreatment of henequen fiber with
compatibilizing silane can improves the stress transfer between the fiber and the matrix.
However, the improvement in fiber matrix adhesion didn’t increase the composite
modulus possibility due to the indirect coupling between the fiber and the epoxy matrix.
In addition, the improvement in interfacial adhesion due to silane treatment was not as
significant as oxygen plasma treatment.

In contrast to the thermoplastic PP results, the fracture surface of silane treated
henequen-epoxy samples didn’t show significant difference in fiber pullout than the

untreated henequen-epoxy sample (See Figure 85 & 86).

 

Figure 85. Typical tensile fracture surfaces of untreated henequen-epoxy composite

114

 

Figure 86. Typical tensile fracture surfaces of Z-6040 silane treated henequen-epoxy composite

As can be seen from the ESEM images, the fiber pull-out of the untreated henequen-
composite samples is only slightly longer than the silane treated samples. It’s possible to
conclude that the adhesion between natural fiber (henequen) and epoxy is very good
because of the hydrophilic nature of the cellulose surface and subsequent silane treatment
does not lead to a significant improvement. This fact can also be seen with the silane
treated henequen and epoxy system in which Z-6040 silane treatment only led to a 30%
improvement in interfacial adhesion.
7.2.3 Effect of 8-minute Plasma Treatment on Composite Properties

Unidirectional epoxy composites were also fabricated using the 8-minute oxygen
plasma treated henequen as reinforcement. According to the mechanical property
measurements, the oxygen plasma treatment produced both tensile strength and modulus

improvement compared to the untreated henequen and epoxy composites (See Figure 87).

115

 

 

 

 

 

 

 

 

 

 

 

 

 

Tensile Propertles of Henequen Fiber Reinforced Epoxy
Composite (16 wt% fiber)
140 10
CDTendle Strength "5 124
A 120 J- -e—Tendle Modulus 1-
E 97 L ' .- 8 A
100 -- . "
E I 16»: e
._ .. 6 r
5': so « P; ._ 4 g
.3. a
s 4°“ 5
i- -~ 2
20 -_
0 l : - o
Untreated Fiber 8 nin 02 Piasnn Rule of Mixtures
Treated Fiber

 

 

 

Figure 87. Tensile properties of henequen and epoxy composite - Plasma treatment.

Comparing to the Z-6040 silane treatment, the oxygen plasma treatment had much more
significant effect on the composite modulus. The tensile modulus of 8-minute oxygen
plasma treated henequen-epoxy was 6.2 GPa, which is again very close to the rule of
mixtures prediction. However, the good interfacial adhesion between oxygen plasma
treated henequen and epoxy didn’t produce a significant increase in the tensile strength of
the composite, possible due to fiber damage caused by the plasma treatment.

The ESEM images of the tensile fracture surface of the 8-minute oxygen plasma
treated henequen and epoxy composite also confirm better fiber matrix adhesion. As
shown in Figure 88, the fiber pull-out of the composite from tensile testing is almost zero.
Even the few longer pullout fibers show good interfacial adhesion because epoxy are

firmly attached to the fiber surface (See Figure 89)

116

 

 

Figure 89. Tensile fracture surface of 8min 02 plasma treated henequen and epoxy composite
showing good adhesion between the fiber and matrix.

117

7.3 Summary

The mechanical testing of the unidirectional henequen-polypropylene and
henequen-epoxy composite proves that good interfacial adhesion has a positive beneficial
effect on the mechanical properties of the natural fiber reinforced composite.

For the henequen and polypropylene composite, the tensile properties were
significantly improved by the incorporation of maPP as compatibilizer. When a suitable
compatibilizer or coupling agent is used, the tensile modulus of the natural fiber
reinforced polymer composite can reach its maximum value as indicated by agreement
with the rule of mixtures prediction. However, the improvement in tensile properties
fi'om good interfacial adhesion comes at the price of lower impact strength of the
composite material.

The effects of fiber surface treatment on the composite properties were very
different for oxygen plasma and silane treatment. The silane treatment produced large
improvement in the tensile strength of the composite with little improvement in the
tensile modulus. This could be a result of the indirect coupling (presence of silane at the
interface) between the natural fiber and the epoxy matrix. On the other hand, the oxygen
plasma treatment produced tensile modulus data close to that of the rule of mixtures
prediction but damaged the fiber in the process (lower composite tensile strength than
silane treatment).

It’s also important to note that the composite properties can also depend on
various other factors such as void content and processing conditions. Good interfacial
adhesion between the natural fiber and matrix polymer can only be significant when other

major factors are optimized (i.e., low void content and optimum processing conditions).

118

CHAPTER 8 CONCLUSIONS

 

The adhesion studies using Tencel and henequen fibers have clearly demonstrated
that the interfacial chemistry can play a major role in the determination of natural fiber to
polymer matrix adhesion. According to the adhesion results from oxygen plasma treated
henequen fiber and epoxy, there is a linear relationship between the interfacial adhesion
and the concentration of reactive functional groups present at the fiber matrix interface.
Base on this linear relationship, it is possible to predict the fiber-matrix adhesion base on
the polymer chemistry and the surface chemistry of natural fibers, relative to a reference
system (untreated fiber and polymer). The large improvement in the natural fiber to
polymer adhesion from various fiber or matrix modifications all points to the formation
of covalent bonds across the fiber—matrix interface. In addition, coupling agents or
compatibilizers such as silanes and maleated polypropylene can also improve the natural
fiber to polymer matrix adhesion. However, the interfacial chemistry, though dominant,
is not the only factor that controls the fiber to matrix adhesion. Matrix properties such as
shear modulus and crystallinity (for crystalline and semicrystalline polymers) can also
have an impact on natural fiber to polymer adhesion.

The effects of interfacial adhesion on the mechanical properties of natural fiber-
polymer composites were determined from tensile and impact testing of unidirectional
henequen fiber reinforced polypropylene and epoxy composite. It’s concluded that strong
interfacial adhesion in the natural fiber reinforced polymer composites can have a
positive effect on the composite properties. The tensile strength and modulus can be
improved by fiber surface treatment or by adding a compatiblized coupling agent to the

composite during fabrication step. When good interfacial adhesion is achieved for natural

119

fiber reinforced polymer composite, its tensile properties (such as tensile modulus) can
approach to that of the rule of mixtures prediction. However, the improvement in the
tensile strength and tensile modulus due to good interfacial adhesion usually results in
lowering of the composite impact properties. Therefore, a balanced approach must be
taken in the optimization of the composite interface base on the application requirement.
Furthermore, to achieve the ultimate mechanical properties for a natural fiber reinforced

polymer composite, other factors such as processing conditions must also be optimized.

120

BIBLIOGRAPHY

 

10.

11.

12.

13.

14.

15.

16.

17.

C. Klason, J. Kubat, and H. E. Stromvall, Int. J. Polym. Mater., 10, 159 (1984)
P. Zadorecki, A J. Mchell, Polymer Composite, 10, 69 (1989)
D. Maldas, B. B. Kokta, C. Daneault, Int. J. Polym. Mater., 10, 159 (1984)

J. Karlson, J. F. Blachot, A Peguy, P. Gattenholm, Polymer Composite, 17, 300
(1996)

R. R. Mukherjee and T. Radhakrishnan, Text. ngr., 4, 4 (1972)

A. K. Mohanty, Misra, and G. Hinrichsen, Macromol. Mater. Eng, 276/277, 1-24
(2000)

J. George, M. S. Sreekala, and S. Thomas, Polym. Eng. Sci., 41(9), 1471 (2001)
H. S. B Marshall and H. J. Palmer, J. Text. Inst, 53, 141 (1962)
K. C. M. Nair and S. Thomas, J. Appl. Polym. Sci., 60, 1483 (1996)

J. George, N. Prabhaltaran, S. S. Bhagwan and S. Thomas, J. Appl. Polym. Sci., 57,
871, (1995)

K. Joseph, S. Varghese, G. Kalaprasad, S. Thomas, L. Prasannakumari, P. Koshy
and C. Pavithran, Eur. Polym. J ., 32, 1243 (1996)

K. Joseph and S. Thomas, J. Appl. Polym. Sci, (in press)

U. S. Prokop and A. Tanner, J. Sci. Ind. Res, 55, 381 (1996)

W. H. Zhu, B. C. Tobias and R. S. P. Coutts, J. Mater. Sci. 14, 508 (1995)
N. M. White and M. P. Ansell, J. Mater. Sci, 18, 1549, (1983)

S. V. Prasad, C. Pavithran and P. K. Rohatgi, J. Mater. Sci, 18, 1443 (1983)

A. N. Shah and S. C. Lakkad, Fiber. Sci. Technol, 15, 41 (1981)

121

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

M. N. Belgacem, P. Bataille, and S. Sapieha, J. Appl. Polym. Sci, 53, 379 (1994)
T. Walkida and S. Tolcino, Ind. J. Fiber and Test. Res, 21, 69 (1996)

S. Dong, S. Sapieha and H. P. Schreiber., Poly. Eng. Sci., 32, 1734 (1992)

P. Batille, N. Belgacem and S. Sapieha, SPE ANTEC, 39, 325 (1993)

S. Dong and S. Sapieha, J. Appl. Polym. Sci., 37, 1154 (1991)

A. R Sanadi, S. V. Prasad and P. K. Rohatgi, J. Mater. Sci. 21, 4299 (1986)

P. Zadorecki and P. Flodin, Polym. Compos, 7, 170 (1986)

E. T. N. Bisanda and M. P. Ansell, Comp. Sci. T echnol., 41, 165 (1991)

J. Gassan and A. K. Bledzki, Polym. Compos, 18, 179 (1997)

U. C. Jindal, J. Comp. Mater., 20, 19 ( 1986)

J. Gorge, J. Ivens and I. Verpoest, Proceedings of ICCM-12, 5-9 July, Paris (1999)
D. Beshay, B. V. Kota and C. Daneault, Polym. Compos, 6, 261 (1985)

B. V. Kota, C. Daneault and A D. Beshay, Polym. Compos, 7, 251 (1986)

H. L. Chen and R. S. Porter, J. Appl. Polym. Sci., 54, 1781 (1994)

T. Peijs, S. Garkhail, R Heijenrath, M. Oover and H. 303, Macromol. Symp., 127,
193 (1998)

M. J. Zainy, M.Y. A. Fuad, Z. Ismail, M. S. Mansor and J. Mustafah, Polym.
International, 40, 51 (1996)

R G. Raj, B. V. Kokta and C. Daneault, Mcromol. Chem. Macromol. Symp., 28,
187 (1989)

J. M. Felix and P. Gatenholm, J. Appl. Polym. Sci., 42, 609 (1991)

J. George, S. S. Bhagawan and S. Thomas, Comp. Interfaces, 5, 201 (1998)

122

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

B. Singh, M. Gupta and A. Varma, Polym. Compos, 17, 910 (1996)
Tencel Inc. Web reference: http://www.Tence1.com/about.htrnl
Gertjan van Roekel, Web reference: http://home.p1anet.n1/~roeke008/sld001.htm

J. Vickerrnan, Surface Analysis - The Principal Techniques, John Wiley & Sons,
New York (1997)

M. Matias, M. Orden, C. Sanchez, and J. Urreaga, J. Appl. Polym. Sci., 75, 256
(2000)

D. Hull and T. W. Clyne, An Introduction to Composite Materials, 2nd Ed,
Cambridge Univ. Press, NY, 1996, pp. 105-109.

A. K. Bledzki, and J. Gassan, Prog. Polym. Sci, 24, 221 (1999)

A. F. Toussaint and P. Luner, Proc. 10'h Cellulose Conf., Syraccruse, New York,
Vol. 29.5-29.6, 379 (1988)

F. P Liu, M. P. Wolcott, D. J. Gardner and T. G. Rails, Composite Interfaces, 2 (6),
419(1994)

J. M. Felix, P. Gatenholm, H. P. Scheriber, Polym. Composites, 14(6), 449 (1993)
S. L. Kaplan and P. W. Rose, Adhesion ’90, University of Cambridge, UK (1990)

G. P Hansen, S. T. Kaplan, R Rushing, and R Warren, Adhesive ’89, Society of
Manufacturing Engineers, Atlanta Georgia, September 12-14 (1989)

D. M. Brewis and D. Briggs, Polymer, 22, 7 (1981)
J. Gassan, A. K. Bledzki, and A. Die, Macromol. Chem, 236, 129 (1996)

R M. Powell, Proc. Of the 18th Riso Int. Symp. Pm Materials Science, Riso
National Laboratory, Denmark, 127 (1997)

V. Favier, H. Chanzy, J. Y. Cavaille, Macromolecules, 28, 6365 (1995)

W. Helbert, J. Y. Cavaille, and A Dufi'esne, Polymer Composite, 17, 604 (1996).

123

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

65.

66.

67.

68.

69.

Hajji, J. Y. Cavaille, V. Favier, C. Gauthier, and G. Vigier, Polymer Composite, 17,
612 (1996).

V. Favier, G. R. Canova, S. C. Shrivastava, and J. Y. Cavaille, Polymer
Engineering Science, 37, 1732 (1997).

L. Chazeau, J. Y. Cavaille, G. Canova, R Dendievel, B. Bouthem, Journal of
Applied Polymer Science, 71, 1797 (1999)

A. Dufresne, J. Y. Cavaille, W. Helbert, Polymer Composite, 18, 2, 198 (1997)
S. B. Warner, Fiber Science, Prentice Hall, Englewood Cliffs, NJ, 52 (1995)

R. Yosomiya, K. Morimoto, A. Nakajima, Y. Ikada, and T. Suzuki, Adhesion and
Bonding in Composite, Marcel Dekker, Inc.New York and Basel, 1 (1990)

J. C. Halpin and J. L. Kardos, Polym. Eng. Sci., 16, 344 (1976)
J. Li, Journal of Applied Polymer Science, 53, 225 (1994)

R. M. Rowell, A R Sanadi, D. F. Caulfield and R. E. Jacobson, Lignocellulosic-
Plastics Composites, 23 (1997)
Web Reference: http:/lwww.fpl.fs.fed.us/documnts/PDF1997/rowe197d.pdf

H. Chtourou and B. RiedL J. Adhesion Sci. Technol., 9(5), 551 ( 1995)
D. A. Biro, G. Pleizer, and Y. Deslandes, J. Applied Polym. Sci, 47, 883 (1993)

J. M. Felix, C. M. Gilbert Carlsson and P. Gatenholm, J. Adhesion Sci. Technol. 8
(2), 163 (1994)

G. Bogoeva, A. Janevski and E. Mader, J. adhesion Sci. Technol. 14(3), 363 (2000)

F. M. Zhao and N Takeda, Composites Part A: Applied Science and Manufacturing,
31 (11): 1203 (2000)

K. Van de Velde and P. Kiekens, J. Thermoplastic Composite Materials, 15 (4), 281,
(2002)

K. Van de Velde and P. Kiekens, J. Adhesion Sci. and Technol., 16(8), 999, (2002)

124

70. M. Avella, C. Bozzi, R. Dellerba, B, Focher, A. Marzetti and E. Martuscelli,
Angewandte Makromolekulare Chemie, 233, 149 ( 1995)

125