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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c-JCIRC/DateDuepGS-p. 15

INTERFACIAL ENGINEERING OF THE INTERPHASE BETWEEN CARBON
FIBERS AND VINYL ESTER RESIN

By

Lanhong Xu

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemical Engineering and Materials Science

2003

ABSTRACT

INTERFACIAL ENGINEERING OF THE INTERPHASE
BETWEEN CARBON FIBERS AND VINYL ESTER RESIN

By
Lanhong Xu

Vinyl ester resins have been extensively used for the manufacture of low cost high
performance composites. Carbon ﬁbers are important reinforcement materials. The use
of vinyl ester composites reinforced with carbon ﬁbers requires an improvement in the
ﬁber/matrix adhesion levels. The objectives of this study were to gain an understanding
of the factors controlling interfacial adhesion between carbon ﬁbers and vinyl ester resin;
to model the contributions of the factors controlling ﬁber/matrix adhesion; and to provide
an engineered and optimized interface between carbon ﬁber and vinyl ester for tailoring
structurally efﬁcient carbon ﬁber/vinyl ester composites. This work consists of three
parts.

Part I: a partially cross-linked DGEBA epoxy polymer sizing placed onto carbon ﬁber
surface was found to be a beneﬁcial interphase between the carbon ﬁber and vinyl ester
resin resulting in an increase in ﬁber-matrix adhesion. The adhesion was evaluated as
interfacial shear strength (IF SS) with micro-indentation. Nam-indentation and nano-
scratch technique were used to investigate the gradient between this epoxy sizing and
vinyl ester resin. An optimized thickness of this sizing was found and the mechanism by
which this sizing improved adhesion was also investigated. A set of 2-D non-linear ﬁnite
element models was set up for simulation of the micro—indentation process and consistent

results were found between the experimental data and numerical results. It was found

that the epoxy sizing formed more chemical bonds with the surface of the carbon ﬁber
reinforcement and an interpenetrating interphase with the vinyl ester resin. The resulting
interphase between vinyl ester matrix and epoxy sizing reduced the residual stress caused
by the volume shrinkage of the vinyl ester after curing.

Part 11: since it is known that the carbon ﬁber surface can interfere with the vinyl ester
polymerization, the effects of preferential adsorption of the catalysts and styrene on the
carbon ﬁber surface on ﬁber/matrix adhesion have been investigated. It was found that
although the adsorption of the promoter and accelerator does occur on the carbon ﬁber
surface, it does not substantially affect the ﬁber/matrix adhesion. The vinyl ester
monomers and styrene monomers interact differently in the interphase. Optimum ﬁber-
matrix adhesion can be obtained by selecting the initiator and adjusting the amount of
initiator.

Part III: vinyl ester resins undergo signiﬁcant volume shrinkage upon cure. Cure
volume shrinkage was measured by a lab-made dilatometer. The cure volume shrinkage
was found to reduce ﬁber/matrix adhesion as a result of residual stress. A 3-D non-
linear ﬁnite element model was used to analyze the residual stress distribution at the
ﬁber/matrix interphase. Higher von Mises effective stress was found for large cure

volume shrinkage which is consistent with the experimental results.

Copyright by
Lanhong Xu
2001

To my dear parents: Ruren Xu and Xuelin Zhao

ACKNOWLEDGMENT

I would like to express my sincere gratitude to my advisor, Professor Lawrence T.
Drzal, for his valuable academic guidance, continuous spiritual encouragement and
ﬁnancial support throughout the time taken to complete this work. Without his help, I
would not have reached this point today.

Special thanks also go to Dr. Ahmed Al-ostaz and Dr. Tom Mase for their
intellectual contributions in ﬁnite element analysis and Dr. Richard Schalek for his
assistance in nano-indentation test as well as their friendship.

I am thankful to the members of my committee—Dr. Ronald Averill, Dr. Phillip
Duxbury, Dr. Krishnamurthy Jayaraman, Dr. Andre Lee, and Dr. Dashin Liu for their
careful help in resolving my problems and the time they dedicated to my PhD study.

My thanks are also due to all the faculty, staff and students of the Composite
Materials and Structures Center at Michigan State University who have helped me in
many countless ways during my PhD study, especially, Mike Rich, Kelby Thayler,
Hiroyuki Fukushima.

Finally, I thank my husband Ben Cai for his love and support in enabling me to

complete this work.

vi

TABLE OF CONTENTS

List of Tables
List of Figures
CHAPTER 1 INTRODUCTION

CHAPTER 2 BACKGROUND

2.1 CARBON FIBER REINFORCEMENT

2.1.1 Manufacture Process
2.1.2 Topography and Morphology
2.1.3 The Surface Chemistry and Energetics
2.1.4 Carbon Fiber Surface Treatment
2.2 VINYL ESTER MATRIX RESINS
2.2.1 Introduction to DERAKANE Epoxy Vinyl Ester Resin
2.2.2 Cure Kinetics of Vinyl Ester Resin
2.2.3 Phase Separation and Microgel
2.2.4 Inﬂuence of Styrene Monomer Content
2.2.5 Cure Shrinkage of Vinyl Ester
2.3 CARBON F IBER- VINYL MATRIX INTERPHASE
2.3.1 General Introduction to Fiber/Matrix Interphase
2.3.2 Evaluation the Strength of the Fiber-Matrix Interphase
2.3.3 Low Adhesion between Carbon Fiber and Vinyl Ester Resin

2.3.4 Improvement of Adhesion between Vinyl Ester and Carbon
Fiber

2.4 REFERENCES

vii

iv

10

10

13

15

19

19

22

26

26

29

31

31

35

35

41

42

CHAPTER 3 STATEMENT OF THE PROBLEM
CHAPTER 4 EXPERIEMNTAL METHODS
4.1 FIBER SIZING
4.1.1 Sizing Materials
4.1.2 Sizing Process
4.1.3 Sizing Thickness Measurement (TGA)
4.1.4 Environmental Scanning Electron Microscope (ESEM)

4.2 MEASUREMENTS OF MECHANICAL PROPERTIES OF
MATERIALS

4.2.1 Dynamic Mechanical Thermal Analysis (DMTA)
4.2.2 United Testing System (UTS)
4.2.3 MTS Nam-indentation and Nano-scratch Test

4.2.3.1 Nam-Indentation Test
4.2.3.2 Nano-Scratch Test

4.3 MEASUREMENT OF THE ADHESION BETWEEN CARBON
FIBER AND VINYL ESTER RESIN

4.4 REFERENCES
CHAPTER 5 IMPROVEMENT OF ADHESION BETWEEN
CARBON FIBERS AND VINYL ESTER RESIN WITH
EPOXY SIZING
5.1 INTRODUCTION
5.2 EXPERIMENTAL MATERIALS AND METHODS
5.3 RESULTS AND DISCUSSIONS

5.3.1 The Thickness of Sizing can be Controlled by the
Concentration of Sizing Solution

viii

49

53

53

53

57

57

61

61

61

65

65

66
69

71

75

77

78

80

89

89

5.3.2 Improvement of carbon ﬁber/vinyl ester adhesion with
DGEBA sizing and the optimum sizing thickness for the best
Interfacial Shear Strength (IFSS)

5.3.3 The role DGEBA sizing material played between carbon ﬁber
and vinyl ester

5.4 CONCLUSIONS
5.5 REFERENCES
CHAPTER 6 INFLUENCES OF THE COMPONENT CHEMISTRY TO

THE ADHESION BETWEEN VINYL ESTER AND
CARBON FIBER

6.1 INTRODUCTION

6.2 EXPERIMENTAL MATERIALS AND METHODS

6.3 RESULTS AND DISCUSSIONS

6.3.1 Inﬂuence of the Catalyst Concentration on the Properties of the
Matrix Material

6.3.2 The Effect of Catalyst Concentration on the Interfacial Shear
Strength (IFSS)

6.3.3 The Contribution of Vinyl Ester and Styrene Monomer of the
D411-C50 to the Adhesion

6.4 CONCLUSIONS
6.5 REFERENCES
CHAPTER 7 CURE VOLUME SHRINKAGES OF VINYL ESTER

RESINS AND THEIR INFLUENCES ON THE
ADHESIONS BETWEEN CARBON FIBERS AND
VINYL ESTER MARIX RESINS

7.1 INTRODUCTION

7.2 EXPERIMENTAL MATERIALS AND METHODS

7.3 RESULTS AND DISCUSSIONS

ix

93

94

104

106

108

109

112

116

116

120

125

129

131

133

134

135

145

7.3.1 Cure Shrinkage of Vinyl Ester Matrices

7.3.2 Inﬂuence of Cure Volume Shrinkage of Matrix Resin on the
Adhesion between Carbon Fiber and Vinyl Ester

7.3.3 Interphase between DGEBA Epoxy Sizing and Vinyl Ester of
Different Volume Shrinkage

7.3.4 Finite Element Analysis of the Residual Stress Caused by
Cure Volume Shrinkage

7.4 MODELLING INTERPHASE STRENGTH

7.5 CONCLUSIONS

7.6 REFERENCES
CHAPTER 8 CONCLUSIONS AND RECOMMENDATIONS
APPENDICES:
APPENDIX I EVALUATION OF F IBER/MATRIX ADHESION
APPENDIX 11 C++ PROGRAM FOR NANO-SCRATCH TEST

APPENDIX III RELATIONSHIP BETWEEN LINEAR SHRINKAGE
AND VOLUMETRIC SHRINKAGE

145

152

155

159

162

166

168

170

175

197

202

Table 2.1

Table 2.2

Table 2.3

Table 2.4

Table 4.1

Table 4.2

Table 5.1

Table 6.1

Table 7.1

Table 7.2

Table 7.3

LIST OF TABLES

Selected Properties of AS4 Carbon Fiber

A Comparison of the Selected Properties to the Styrene Monomer
Content for D4] 1 Vinyl Ester Resin

Volume Shrinkage of Vinyl Ester Resins
Selected Properties of D41 1-C50 vinyl ester
Selected Properties of DGEBA-T403

The resolution and the ranges of MTS nano-scrach XP head and
nano-indentation DCM Head

Materials mechanical properties for ﬁnite element model
Cure Formulations

Property comparison of D41 l-CSO to Fuchem 891 vinyl ester

The cure formulations for the investigation of the cure volume
shrinkage

Themomechanical properties for ﬁnite element model

xi

18

28

28

30

55

66

88

115

137

141

144

Figure 2.1

Figure 2.2

Figure 2.3

Figure 2.4

Figure 2.5
Figure 2.6

Figure 2.7

Figure 2.10

Figure 2.11

Figure 2.12

Figure 2.13

Figure 4.1

Figure 4.2

Figure 4.3

LIST OF FIGURES

Manufacturing process of PAN based carbon ﬁber

Schematic draws of a three-dimensional model of a part of
carbon ﬁber

Surface chemistry of AS4 carbon ﬁber

A schematic representation of the electrochemical oxidative
surface treatment of the carbon ﬁbers

Vinyl ester resin are produced from epoxy resin
Catalysts for cure of vinyl ester resin

(a) Chemical structure of typical DGEBA-based VE-ST resin
system components, (b) a schematic representation of the VE-
ST network formation showing the cross-linking capacity of the
difunctional VE monomer, and (c) reactions encountered during
the free radical bulk copolymerization of VE resin systems
(draw not to the scale

Fiber-Matrix Interphase

Schematic diagram of the micro-indentation method for
measuring ﬁber/matrix adhesion

Interfacial shear strength measured with the single ﬁber
fragmentation test over a series of carbon ﬁbers in an epoxy
matrix, Epon 828-mPDA, two vinyl ester matrices, the
Derakane 411-C50 and 8084

Interfacial shear strength of micro-droplet with polymers and
as-received carbon ﬁbers

Molecular of Diglycidy Ether of Bisphenol A (DGEBA) and
J effamine T 403

Using a pre-impregnation machine to simulate the application
of sizing in an industrial process

Using TGA to measure the sizing thickness

xii

11

12

14

20

23

24

32

36

38

39

54

56

59

Figure 4.4

Figure 4.5

Figure 4.6

Figure 4.7

Figure 4.8

Figure 4.9

Figure 5.1

Figure 5.2

Figure 5.3

Figure 5.4

Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

Figure 5.10

Figure 5.11

 

Using DMTA to measure the storage modulus at certain
temperature and the glass transformation temperature

Use UTS to measure the tensile modulus

A hypothetic set of continuous load-displacement data of
nano-indentation

Coefﬁcient of friction (COF)

Interfacial Testing Systems (ITS) used for performing
micro-indentation tests on ﬁber reinforced composites

ITS images for ﬁber before and after test

Effect of ﬁber sizing on the interfacial shear strength and
composite mechanical properties

Molecular Structures for the chemicals used in this
chapter

Cure formulation and cure process

The effect of the tip size to an interface with
inﬁnitesimal wideness

Four-phase model (drawing not to the scale) for ﬁnite
element analysis

The sizing thickness versus the concentration of sizing
solution

Sizing quality assessed with Environmental Scanning
Electron Microscope (ESEM)

The interfacial shear strength (IF SS) related to the sizing
thickness

Potential chemical reactions between carbon ﬁber
surface and epoxy/amine system

Nam-indentation test

Nano-indentation results for the DGEBA/Vinyl ester
interphase

xiii

63

64

67

70

72

73

79

82

83

84

88

90

91

92

95

97

98

Figure 5.12

Figure 5.13

Figure 5.14

Figure 6.1

Figure 6.2

Figure 6.3

Figure 6.4

Figure 6.5

Figure 6.6

Figure 6.7

Figure 6.8

Figure 6.9

Figure 7.1

Figure 7.2

Figure 7.3

Figure 7.4

Nano-scratch test traces for vinyl ester/DGEBA interphase

Comparison of the nano-scratch data between DGEBA
sizing/vinyl ester interphase and reference interphase

Model for the interphase between carbon ﬁber and vinyl ester
resin with epoxy sizing

Comparison of the interfacial shear strength between Panex
carbon ﬁber/Derekane 411-C50 vinyl ester adhesion with

different initiator

Inﬂuence of the amount of CHP-S to the properties of cured
D411-C50 resin

Inﬂuence of the amount of CoNap to the properties of cured
D411-C50 resin

Inﬂuence of the amount of DMA to the properties of cured
D411-C50 resin

The inﬂuence of the amount of CHP-S to the IFSS between
D411-C50 and AS4/Coated AS4 ﬁbers

The inﬂuence of the amount of CoNap to the IF SS between
D411-C50 and AS4/Coated AS4 ﬁbers

The inﬂuence of the amount of DMA to the IFSS between
D411-C50 and AS4/Coated AS4 ﬁbers

Moduli comparison between vinyl ester, D411-C50 and styrene

The IFSS of vinyl ester, D411-C50 and styrene with AS4 and
coated AS4 ﬁbers

The setup used for the cure volume shrinkage measurements

A transversely isotropic array of composite ﬁbers with the
representative volume element indicated by a dashed line

The silicone oil level in the pipette of the lab-mad dilatometer
versus time

The volume shrinkage observed from the lab-made dilatometer
according to time

xiv

99

100

103

111

117

118

119

121

122

123

126

127

138

143

146

147

Figure 7.5

Figure 7.6

Figure 7.7

Figure 7.8

Figure 7.9

Figure 7.10

Figure 7.11

Figure 7.12

Figure 7.13

Inﬂuence of Cure Process to the Cure Volume Shrinkage of
Vinyl Ester

One of the DMTA Measurement for Room Temperature
Cured Fuchem 89]

Inﬂuence of vinyl ester cure volume shrinkage on the bulk
moduli

Inﬂuence of vinyl ester cure volume shrinkage on the
Interfacial shear strength

Nano-scratch trace for the interphase of DGEBA epoxy
sizing/Fuchem 89l-MEKP

Comparison of nano-scratch data for the interphase of
DGEBA epoxy sizing/Fuchem 891-MEKP and the interphase
of DGEBA epoxy sizing/D411-CHP

Indenter force time history

von Mises stress for a point in the matrix at the interphase-

matrix boundary located approximately 18 pm from the free
surface

von Mises Stress Distribution at 5 g Indenter Load

Figure 7.14 The interphase property presented by interfacial shear

strength(IF SS) plot

XV

150

151

153

154

156

157

159

160

161

163

 

 

CHAPTER 1

INTRODUCTION

Advanced polymer composite with carbon ﬁber as reinforcement are extremely
strong and light—around 10 times the strength to weight ratio of most metals. Carbon
ﬁber composites started out in the 19503 and became a mature structural material in the
19805. Vinyl ester resins were introduced in the late 19605, originally were developed
for their superior toughness and chemical resistance in comparison to unsaturated
polyester, now ﬁnd a wide range of applications. Carbon ﬁbers have mainly been used in
aerospace applications together with epoxy or high temperature therrnoplastics, whereas
vinyl ester resins have found use in large-volume and low-cost applications with
primarily glass ﬁber as reinforcement. More recently, the availability of lower cost
carbon ﬁbers is improving the potential for the manufacture and application of carbon
ﬁber reinforced composites in segments previously seemed too expensive, such as

[L21

automobiles and infrastructures However when carbon ﬁbers are combined with

vinyl ester, the mechanical properties of the resulting composites are lower than desirable

comparing with carbon ﬁber reinforced epoxy composite material [3’4].

Low adhesion between carbon ﬁbers and vinyl ester resins was extensively reported

PM”. It is well known that ﬁber/matrix adhesion can be a signiﬁcant reason for lowering

l[l2-15]

the mechanical properties of a composite materia . The interfacial shear strength of

carbon ﬁber/vinyl ester system measured by single ﬁber fragmentation is only 50% of

 

 

that of the carbon/epoxy system [3’4].

Previous study found that the application of a
lightly cross-linked, epoxy polymer onto the carbon ﬁber surface provides a beneﬁcial
interphase between the carbon ﬁber and vinyl ester resin matrix resulting in an increase in
ﬁber-matrix adhesion and mechanical properties of the composites as well [4’16].
However, the exact mechanism by which this sizing improved adhesion remained
unknown and the epoxy-amine sizing need to be optimized for compatibility and
thickness. The ﬁrst part of this study focused on the investigation of the effects of the
thickness and gradient of the epoxy sizing on the adhesion between carbon ﬁber and
vinyl ester resin.

Typical vinyl ester resins contain 35-50wt% styrene monomer as reactive diluents.
The copolymerization between styrene monomer and vinyl ester monomer is a
heterogeneous, free radical, chain-growth, and cross-linking reaction. The wetting
condition to the carbon ﬁber surface of styrene monomer could be different from that of
vinyl ester monomer because the surface energy of styrene monomer would be smaller
than that of vinyl ester resin (refer: 40(epoxy) “7]; 26.7(benzene) [18]). In addition, a

1. “9] investigated the reactions of the

previous investigation by Weitzsacker, Drzal et a
catalyst and/or promoters to determine if they were competing with the vinyl ester matrix
resin for reactive sites at the carbon ﬁber surface. Cobalt was detected at 2.6% at the
surface of the AS4 carbon ﬁber exposed to cobalt naphthenate (CoNap) and a minor
change was also noted in the surface chemistry of the ﬁber exposed to dimethyl aniline

(DMA) which are the promoters and accelerators for the free radical copolymerization

respectively. The inﬂuences of the reaction catalysts and the two components of vinyl

ester resin, the vinyl ester monomer and styrene monomer on the interface properties
between carbon ﬁber and vinyl ester was the second interest of this study.

Residual stress can be another factor that lowers the adhesion if it creates a transverse
tensile stress of sufﬁcient magnitude at the ﬁber-matrix interphase. Since it is known that
vinyl ester resin can undergo as much as 9% volume shrinkage with cure while typical

epoxy system undergo only 3-4% shrinkage during cure [20‘2”

. This shrinkage can induce
signiﬁcant stresses in the composite already before loading. On the other hand,
composite materials have residual stress from the fabrication. In the case of carbon
ﬁbers, the coefﬁcient of thermal expansion (CTE) is quite small and can actually be

[22]

negative . The ﬁber is anisotropic and the radial and longitudinal thermal expansions

can be quite different. The polyester is isotropic but has a CTE a factor of 80x10’6/K [23]

(compare with Epoxy: 60x10'6/K [2”, 127x10'6/K [24]). This disparity becomes
increasingly signiﬁcant as higher processing temperatures are reached with the absolute
difference between the glass temperature and the use of temperature determining the
magnitude of these residual thermal stresses. In addition, the difference of thermal
conductivities of materials can introduce unevenly distributed temperature during process

and application can also result in residual stresses even defects [25]

. The ﬁnal part of this
study focused on the cure volume shrinkage of vinyl ester resin and its inﬂuences on the
adhesion between carbon ﬁber and vinyl ester resin.

Finite element methods have been widely used to predict the effective elastic
properties and the elastic response of ﬁbrous composites under mechanical and/or

thermal loading. Finite elements are particularly useful for detail studies of the stress

distributions in the ﬁber and surrounding matrix as a ﬁmction of the actual ﬁber

arrangement, matrix properties, loading condition and application temperature [23’ 26'29].

Using IDEAS for the pre-processing and ABAQUS for processing and post-processing, a
set of non-linear contact ﬁnite element models were set up to ﬁnd out the inﬂuence of the
thickness of sizing material and the results were compared with the experimental data in
the ﬁrst part of the research. LS-DYNA, one of the most powerful non-linear ﬁnite
element analysis methods, was used for the simulation of the cure volume shrinkage and
the stress brought by the shrinkage in the third part of the research.

The research has been organized into three major sections as described above and is
represented by Chapter] through Chapter 8. Chapter 1 and 2 detail the driving principles
behind the research undertaken in this study. Chapter 3 is the objectives of this study.
The experimental methods are covered in Chapter 4. Chapter 5, Chapter 6, and Chapter 7
provide the experimental works covering the summary, introduction, experimental
procedures, results and discussions of the three major sections of the researches. Chapter
5: the inﬂuence of sizing thickness and gradient on the adhesion between carbon ﬁber
and vinyl ester matrix; Chapter 6: the inﬂuence of the component chemistry on the
adhesion between carbon ﬁber and vinyl ester matrix; Chapter 7: the inﬂuences of resin
cure volume shrinkage on the adhesion between carbon ﬁber and vinyl ester matrix.

Chapter 8 contains the conclusions and the ideas for future works.

REFERNECES:

10.

11.

12.

13.

High Preformance Composites, Vol. 4, pp 11, 1996
Zoltek Companies, Inc., Annual Report, pp 6, 1995

D. Hoyns. and L. T. Drzal, “Hydrotherrno-mechanical Response and Surface
Treated and Sized Carbon Fiber/Vinyl Ester Composites”, ASTM Symposium on
the Fiber Matrix and Interface Properties, Nov., 1994

S. M. Corbin, MS. Thesis, Department of Chemical Engineering, Michigan State
University, 1998

V. Rao and L. T. Drzal, Polym Compos, Vol. 12, No. 48, 1991
L. H. Peebles, Carbon Fibers, CRC Press, Boca Raton, 1990

TH. Yoon, H. M. Kang, M. Bump, and J. S. Rifﬂe, “Effect of solubility and
miscibility on the adhesion behavior of polymer-coated carbon ﬁbers with vinyl
esterresins ”, J. App. Poly. Sci. (USA), vol. 79, no. 6, pp. 1042-1053, 2001

Kim, I.C., Yoon, T.-H., “Enhanced interfacial adhesion of carbon ﬁbers to vinyl
ester resin using poly(arylene ether phosphine oxide) coatings as adhesion
promoters”, Journal of Adhesion Science and Technology (Netherlands), vol. 14,
no. 4, pp. 545-559, Feb. 2000

Derakane 411-45 (Vinyl Ester Resin), Alloy Digest (USA), pp. 2, Aug. 1991

M. J. Reis, M. J., M. Botelho Do Rego, and J. D. Lopes Da Silva, J. Mater.Sci. ,
Vol. 30, No. 118, 1995

M. J. Rich, S. Corbin, and L. T. Drzal, “Adhesion of carbon ﬁbers to vinyl ester
matrices”, Proceedings of the 12'" International Conference of Composite
Materials, 1999

M. J. Pitkethly, J. P. Favre, U. Gaur, J. Jakubowski, S. F. Mudrich, D. L.
Caldwell, L. T. Drzal, M. Nardin, H. D. Wagner and et al., “A Round-robin
Program on Interfacial Test Methods”, Campos. Sci. Technol., ﬂ, 205-214,
1993

Yu. A. Gorbatkina, Adhesive Strength of Fiber-Polymer Systems, Ellis
Horwood, New York, 1992

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

L. T. Drzal and M. Madhukar, "Fiber-Matrix Adhesion and Its Relationship to
Composite Mechanical Properties ” J. Matr. Sci., 28, 569-610, 1993

 

Hoecker, K. Friedrich, H. Blumberg and J. Karger-Kocsis, Comp. Sci. Tech,
Vol.54, pp.317, 1995

L. Xu, L. T. Drzal, “Inﬂuence of interphase chemistry on the adhesion between
vinyl ester resin and carbon ﬁbers”, Proceedings of the 23rd Annual Meeting of
“Adhesion Science for the 21” Century”, 2000

L. T. Drzal, "The Epoxy Interphase in Composites”, review chapter in Advances
in Polymer Science [1, 75, K. Dusek, ed., Spring-Verlag, pp. 16, 1985

L. Xu, Mater Thesis, School of Mechanics and Materials Science, Washington
State University, 1998

L. Weitzsacker, M. Xie and L. T. Drzal, “Using XPS to Investigate F iber/Matrix
Chemical Interactions in Carbon Fiber Reinforced Composites”, Surf. Interf.
Anal. 25, 53-63, 1997

M. B. Launikitis, “Vinyl Ester Resin”, Hand Book of Composite, pp.38-49

L. S. Penn and T. T. Chiao, “Epoxy Resins”, Hand Book of Composite,
pp.57~88

J. M. Whitney and L. T. Drzal "Three Dimensional Stress Distribution Around An
Isolated Fiber Fragment”, ASTM STP 937, Toughened Composites, 179-196,
1987

A. ten Busschen, Ph.D thesis, Delft University of Thechnology, the Netherlands,
1996

D. K. Hadad, “Physical and chemical characterization of epoxy resins”, Epoxy
Resins_Chemistry and Technology, 2nd edition, edited by C. A. May, pp.
1089-1172, 1988

V. Rao and L. T. Drzal, "The Temperature Dependence of Interfacial Shear
Strength for Various Polymeric Matrices Reinforced with Carbon F ibers", J.
Adhesion, 37, 83-95, 1992

 

L. Zhang, Ph. D thesis, Delft University of technology, the Netherlands, 1995

H. Ho and L. T. Drzal, “Non-linear Numerical Study of the Single-Fiber
Fragmentation Test, Part 1: Test Mechanics”, Composites Engineering, i, No.
10-11,1231-1244,1995

28.

29.

H. Ho and L. T. Drzal, “Non-linear Numerical Study of the Single-Fiber
Fragmentation Test, Part II: A Parametric Study”, Composites Engineering, _5,
No. 10-11, 1245-1259, 1995

H. Ho and L. T. Drzal, “Evaluation of Interfacial Mechanical Properties of Fiber
Reinforced Composites Using the Microindentation Method ”, accepted for
publication in Composites, Part A, 1998

CHAPTER 2

BACKGROUND

This chapter details the background of this study in regard of understanding the
chemical and physical nature of carbon ﬁber, vinyl ester, and carbon ﬁber/vinyl ester

interphase.

2.1 CARBON FIBER REINFORCEMENT

The purpose of ﬁber reinforcement is different for different classes of matrix
materials. For ceramics, it is to introduce a measurement of toughness, for metals, it is to
inhibit plastic deformation and for polymers, it is to impart stiffness and strength [1].
Fiber reinforced polymeric composite technology is based on taking advantage of the
high strength and high stiffness of ﬁbers. Graphite or carbon ﬁbers are the most widely
used advanced ﬁber. Graphite ﬁbers normally have at least 99% of carbon while carbon
ﬁbers typically are less than 95% carbon [2]. The term “carbon” instead of “graphite” is
used to denote these ﬁbers, since they are composed of crystalline graphite, and also of

non-crystalline material and areas of crystal misalignment [3 1.

TH \IH / \CH / 2 \CH /
l l
C C C
\\N \\N \\N C \\N

Cyclization
270°C

\
Combination
H H 270°C
/ \ / \

Dehydrogenation N / N
400-600°C \ __

 

Denitrogenation

\ N\ /N 600~1300°c
N N /
\ /

Figure 2.1 Manufacturing process of PAN based carbon ﬁber [4]

2.1 .1 Manufacture Process

Carbon ﬁbers are usually produced by carbonization of organic precursor ﬁbers
followed by graphitization at high temperature, Figure 2.1 [4]. A number of precursors
can be used for making ﬁbers, such as polyacrylonitrile (PAN), rayon (cellulose) and
mesosphere pitch. The majority of the commercial carbon ﬁbers currently produced are
based on PAN [5]. PAN- based carbon ﬁber can be grouped into type A (low modulus)
and type HM (high modulus). Type A ﬁbers were graphited at approximately at 1500°C
while type HM ﬁbers were graphited at 2600°C [6]. AS4 carbon ﬁber (Hercules
Incorporated) is type of A, Surface treated, PAN based carbon ﬁber. The tow size of the

ﬁber used in this study is 12K ﬁlaments.

2.1.2 Topography and Morphology

So-called carbon ﬁber is made of carbon which is a very light element. Carbon can
exist in a variety of crystalline forms. Beside the well-known covalent diamond
structure, graphite structure wherein the carbon atoms are arranged in the form of
hexagonal layers is another main structure of crystalline carbon. Carbon ﬁber is of
graphite structure. The properties of a carbon ﬁber are a direct reﬂection of the structure
of graphite which is highly anisotropic on a nanoscopic scale. The basic structure of the
carbon ﬁbers is the graphite crystallites which are composed of turbostatically layered
basal planes or graphite ribbon {5'8} as shown in Figure 2.2. The ribbons increase in size

with increasing graphitization temperature. The ribbons undulate and twist along the

10

 

_, 3.35.3. ___

 

 

 

 

 

 

 

 

 

0"—
if
L

 

 

 

 

 

 

1 .42A i

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fiber Axis

 

    

 
  

   

\4 ‘i' A
‘\ "(JI
,0 ‘ <=
\
o 0.4:“
e e. 0..
v e We id“
'09... 9": l‘q \

ozij
“0.0 “"0‘0'} /\

Figure 2.2 Schematic draws of a three-dimensional model of a part of
carbon ﬁber

11

O
H _ O Lactone

Carboxylic Acid /OH

Ho 9 o 0m“
Hydroxyl H—N ’_ O Na —\ O PYrone
Imme 9

H\ O 6 0:0
H/H 9.9 Carbonyl

Amme
_:
Nee o

Cynao

Figure 2.3 Surface chemistry of AS4 carbon ﬁber

ﬁber axis and the degree of alignment varies with graphitization temperature. Type A
ﬁber has graphitic basal planes plus edges and corners which are highly reactive
comprising the ﬁber surface while the more graphite type HM ﬁber has a surface
composed mostly of graphitic basal plane. The proportion of the surface composed of
edge and comer areas would change with ﬁber modulus. The high bond strength
between the carbon atoms in the basal plane gives an extremely high modulus along the
ﬁber axis, while the weak van der Waals type of bonding between the neighboring layers
produces a low modulus along the edge plane.

Another aspect of topographical feature of the carbon ﬁber is the surface area. In
ﬁber reinforced composites where the ﬁber diameter is on the order of 5 to 10 microns, at
a 50 volume percent loading, the actual interfacial surface area between ﬁber and matrix

can be on the order of 5,000 to 10,000 square meter per cubic centimeter of compositem].

2.1.3 The Surface Chemistry and Energetics

The quantity and the type of chemical group present on the ﬁber surface is a function
of the type of ﬁber, i.e. its graphitization temperature, and the type of surface treatment

['9]. Boehm et a1 [20] have attempted to

used, i.e. gaseous, liquid or catalytic oxidation
identify the type of chemical function groups present on the carbon surface. They have
shown that the oxygen present on the carbon surface can be in as many as four different
chemical groups. Carboxylic acid, lactone, carbonyl and alcoholic oxygens. It was also

found that the Nitrogen in the form of amine or cyano groups is almost always on the low

heat treatment temperature ﬁber surface [2”. Trace amounts or elements such as silicon,

13

Voltmeter

Ammeter
+ II -
Ha ll

Untreated Carbon ﬁber anode Washing and drying

ﬁber spool l 1
r l

Uptake sp6a

 

    
   

 

 

 

Electrolytic solution

\ \\\\\\\\\\\\\\\\\\\\\\\
\\ ‘~.\\\\\'\.\\\\\\\\\\\\\\“~
\\

 

 

Stainless steel cathode bath

Figure 2.4 A schematic representation of the electrochemical oxidative
surface treatment of the carbon ﬁbers [27]

14

from the earlier polymer ﬁber spinning steps, as being on the ﬁber surface has shown that
the sodium has the ability to diffuse to the ﬁber surface from the bulk of the ﬁber at
moderate elevated temperatures.

A necessary condition for ﬁber and matrix compatibility is thermodynamic wetting.
The thermodynamic criterion for wetting of the ﬁber by the matrix is in general that the
surface free energy of the ﬁber surface must be greater than that of the matrix [23 1. The
surface free energy of AS4 carbon ﬁber has been measured and reported by Drzal “’1 by
contact angle technique based on the method proposed by Kaelble [24]. The total surface
free energy is range from 39 mJ/m2 to 57 mJ/m2 which were contributed by both polar

and dispersive component.

2.1.4 Carbon Fiber Surface Treatment

Carbon ﬁbers are subject to post treatments including surface treatments and/or
application of organic sizing. Surface treatment of carbon ﬁbers can in general be
classiﬁed into oxidative and non-oxidative treatments. Oxidative treatments are further
divided into dry oxidation in the presence of gases, plasma etching and wet oxidation. In
the process of dry oxidation, the carbon surface layers simply burn away unevenly to
create pits in line that coalesce into channel resulting a high surface rigidity. The
fundamental principle of a plasma treatment is to induce the formation of active species
in a gas by a suitable energy transfer. The result of this treatment is that many chemical

bonds on the surface are broken forming very reactive species [28].

15

One of the wet oxidation surface treatments is electrochemical oxidation. The ﬁbers
are the electrical anode in a circuit as they are continuously drawn through a solution of
electrolyte in water as shown in Figure 2.4. The voltage in the circuit is maintained
constant, while the current and the residence time in the electrolytic bath are varied to
change the degree of surface treatment.

The total surface free energy is increased for both type A surface treated (AS) and
type HM surface treated (HMS) by surface treatment. The measured change is primarily
in the polar component of the surface free energy and directly related to Interfacial shear

[14.28.25]

strength between ﬁber and matrix . The dispersive part of surface free energy is

virtually insensitive to surface treatment, probably due to the chemically inert graphitic
basal planes [262930].

In addition to the use of surface treatment which are primarily chemical in natural,
surface ﬁnishes or coatings are also used to affect ﬁber-matrix adhesion. Organic coating
or sizing or ﬁnishes in an amount of 0.5~7 wt% are typically applied in order to improve
their compatibility with the resin, protect the ﬁber surface from damage, protect ﬁber

surface reactivity and/or their “handle ability” [3'32]

. These coatings are applied to both
untreated and surface treated ﬁbers by passing the ﬁbers through a sizing bath ﬁlled with
organic solution. The sizing agents most commonly employed are polyvinyl alcohol,

[3 I]. The mechanism by which sizing operates is fundamentally

epoxy, polyimide and etc
different than the surface treatment mechanism. The sizing layers are hundreds of

nanometers thick. The properties of this sizing layer itself are imparted to the interphase

and can control the adhesion between ﬁber and matrix.

16

AS4 carbon ﬁbers are PAN based, type A carbon ﬁbers which have intermediate axial
tensile modulus. They are surface treated with electrochemical oxidation process which
optimizes the adhesion to epoxy. Firstly, the treatments remove a weak out ﬁber layer
initially present on the ﬁber. Secondly, surface chemical groups are added which
increase the interaction with the epoxy. On the carbon ﬁber surface, Oxygen reactive
groups in the forms of carboxylic acid, lactones, carbonyl and alcoholic hydroxyl and
Nitrogen reactive groups in the form of amine and cyano on are potential chemical
reactive sites to epoxy group and amine group of epoxy resin. Some of the properties of

AS4 carbon ﬁbers are list in Table 2.1.

17

 

AS4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Diameter (106m)

 

 

Properties
E1, GPa (Msi) 241(35)
E2, GPa (Msi) 21 (3)
E3, GPa (Msi) 21 (3)
V12 0.2
V13 0.2
V23 0.25
G12, GPa (Msi) 21 (3)
G13, GPa (Msi) 21 (3)
(323, GPa (M81) 8.3 (1.2)
(11 900°C) -o.11
a2(10'6/°C) 8.5
a3 (104mg 8.5
p (g/ma) 1.77
Surface Energy y (mJ/mz) 39~57
Tow size 12,000
7~8

 

Note: E: Tensile Modulus
v: Poisson’s Ratio
G: Shear Modulus

or: Thermal Expansion Coefﬁcient
Direction 1: Fiber Axis Direction

Direction 2 and 3: Fiber Transverse Direction

Table 2.1 Selected Properties of AS4 Carbon Fiber

 

2.2 Vinyl Ester Matrix Resins

The matrix holds the ﬁber together in a structural unit and protects them from
external damage, transfer and distributes the applied loads to the ﬁbers, and in many
cases contributes some needed property such as ductility, toughness, or electrical

insulation.

In the early 19605 scientists from both Shell Oil and Dow Chemical were attempting
to expand the epoxy market. They discovered that polymeric bisphenol A epoxies could
be reacted with acrylic acids yielding a structure with pendant vinyl groups. These resins
could be reacted with styrene to yield very resilient thermoset structures. This chemistry
has been extended to other epoxy structures to give "vinyl esters" with special properties.
Chemists at ICI Americas soon discovered that the propoylated bisphenol intermediate
used to make the Atlas polyesters could be reacted with an isocyanate and then used to

form a urethane modiﬁed vinyl ester.

2.2.1 Introduction to DERAKANE Epoxy Vinyl Ester Resin

Epoxy vinyl ester resins were commercially introduced by Shell Chemical Company
under the trade name of EPOCRYL Resins in 1965. In 1966 Dow Chemical Company
coined the DERAKANE epoxy vinyl ester resins destination with the introduction of a

similar series of resin for molding applications. DERAKANE 411 epoxy vinyl ester

19

 

O
/ \ I / \
CHz—CH—CHZ R—CHz—CH—CHZ R—CHz—CH—CHZ
I1

Diglycidy Ether Bisophenol A (DGEBA)

 

O R'
II I Catalysts
+ H0 —C —C :C
Heat
Methacrylic Acid or Acrylic Acid
R' o OH OH 0H 0 R'

l II | I I ll |
Cszc—C-O—CHz-CH—CH2{R—CH2—CH-CH2}vR—CHz—CH—CH2-O-C-C=CH2
n

DGEBA Epoxy Based Vinyl Ester

+ Styrene (as reactive diluent) + Inhibitors

CH3

|
where n=0~3, R =O©C O0 , and R' = H or—CH3
I

CH3

Figure 2.5 Vinyl ester resins were produced from epoxy resins

20

resins are produced by the addition of unsaturated carboxylic acids (methacrylic or
acrylic acid) to an epoxide resin, such as Diglycidy Ether of Bisphenol A (DGEBA), as
shown in Figure 2.5 and appropriate diluents, normally styrene, and polymerization
inhibitors are added during or after esteriﬁcation.

DERAKANE epoxy vinyl ester resins have outstanding resistance to corrosion.
Chemical attack on these types resin occurs through hydrolysis of the ester groups or the
splitting of unreacted carbon-to-carbon double bond through actions such as oxidation or
halogenations. The number and arrangement of ester linkage and carbon-to-carbon
double bond on DERAKANE epoxy vinyl ester resins make them less polar than

polyesterl33 ]

, see Figure 2.7. In cured bisphenol A fumaric acid polyester and isophthalic
polyesters, the ester linkages occur throughout the molecular chain, making the chain
more susceptible to attack by hydrolysis. In addition to ester linkage, carbon-to-carbon
double bonds also occur randomly throughout the bisphenol A fumaric acid polyester and
isophthalic polyester chains, and not all of those double bonds react during
polymerization. Those unreacted double bond are susceptible to chemical attack. In
DERAKANE epoxy vinyl ester resins, the carbon-to-carbon double bonds are at the ends
of the molecular chain that make them exceptional active. As a result, DERAKANE
epoxy vinyl ester resins cure rapidly and consistently to give fast green strength and
superior creep resistance. The second hydroxyl groups on the vinyl ester molecule also
have a beneﬁcial effect on the adhesion to the glass ﬁbers which is rich in hydroxyl
groups on their surfaces. DERAKANE epoxy vinyl ester resins can be used to fabricate

a wide range of corrosion-resistant ﬁber reinforced polymeric (FRP) applications by all

conventional fabricating techniques. For example, DERAKANE 411 resins are medium

21

viscosity materials widely used for contact molding, pultrusion, matched die molding,
continuous laminating, and ﬁlament winding. DERAKANE 411-C50 resins are lower
viscosity versions, primarily used for resin transfer molding, centrifugal casting, and

other applications requiring extremely fast wet-out.

2.2.2 Cure Kinetics of Vinyl Ester Resin

In thermosetting polymers, the liquid resins are converted into hard brittle solids by
chemical cross-linking, curing, which leads to the formation of a tightly bound three-
dimensional networks of polymer chains. Curing can be achieved at room temperature
but it is usual to use a cure schedule, process, which involves heating at one or more time
temperatures to achieve optimum cross-linking and hence optimum properties [37]. A
relatively high temperature ﬁnal post-cure treatment is often given to minimize any
further cure and change in properties during service. The gel time can also be varied with
careful control of the amount of catalysts such as initiator, promoter and/or accelerator.
Initiators are chemicals to initiate the chemical reaction that cause the resin to cross-link
and cure. Commonly employed initiators include organic peroxides and organic hydro
peroxides. Promoters and accelerators are used to speed up and enhance the cross-link
reaction. The basic molecular structure for the initiators, promoters and accelerators are
shown in Figure 2.6.

DERAKANE 411-C50 epoxy vinyl ester resin (D411-C50) is a mixture of DGEBA

) [33-35]

based vinyl ester monomer (VB) and styrene monomer (ST Using nuclear

magnetic resonance (N MR), Palmese and et al found that the average molecular weight

22

 

CZH5 C2H5

HO—O——-C——O—O——C———O—OH
(3sz (3sz

Methyl Ethyl Ketone Peroxide (MEKP)

CH3

@4001

CH3

Cumene Hydroperoxide (CHP)

Benzoyl Peroxide (BPO)

(Initiators)

2
CO+ 4 >

 

 

Cobalt Napthanate (CoNap)

(Promoter)

Dimethyl Aniline (DMA)

(Accelerator)

Figure 2.6 Catalyst for cure of vinyl ester resin

23

CH3 0 OH OH H3

C
II | l I II |
CH2: —C-O—CH2-CH—CH2{R—CH2—CH-CH2}VR—CHz-CH—CHz-O-C-C=CH2
n

Derekane 411 DGEBA Epoxy Based Vinyl Ester (VE)

C=CH2

fl”.
0 whereR= O.C O O
|

CH3

Styrene (ST) (3)

H3 . .
W 1

Cf a.
egg: . .- a.

(b)
I —-> 2R0 Initiator decay
R0 + VE —) VE- Radicals attack monomer

R0 + ST —) ST-

VE- VE —) (VB-VB). VE homopolymerization
VE- ST —-> (V E-ST)0 VE, ST copolymerization
ST. VE -> (ST-VB). VE, ST copolymerization
VS- ST —) (ST-ST). ST homopolymerization

(c)

Figure 2.7 (a) Chemical structure of typical DGEBA-based VE-ST resin
system components, (b) a schematic representation of the VE-ST network
formation showing the cross-linking capacity of the difunctional VE monomer,
and (c) reactions encountered durin the free radical bulk copolymerization
of VE resin systems‘ 1( draw not to the scale)

24

of VB is 908g/mol (compare the molecular weight of ST: 104g/mol) and the resin was

found to contain 45.0wt% of styrene [37]

. The mixture of vinyl ester and styrene contain
double bonds that react and cross-link into network structure. During network formation,
the vinyl-ester monomer provides cross-linking capacity and branch points for the
network, while the styrene monomer provides linear chain extension. Upon cure, the two
reactive vinyl end groups, while styrene has only one are all gone. Thus, the vinyl ester
monomer styrenated vinyl ester resin exhibits excellent resistance to acids, bases, and
solvents. The reaction between styrene (ST) and vinyl ester (VE) is a heterogeneous
free-radical chain cross-linking copolymerization. As shown in Figure 2.7(a) [36], the
vinyl-ester monomer has provides a cross-linking capacity and branch points for the
formation of a network, as shown schematically in Figure 2.7(b), while the styrene
monomer provides linear chain extension during cure. Figure 2.7(c), reactions that
contribute to the cross-linked network include homopolymerization of vinyl ester
monomer, homopolymerization of styrene, and copolymerization of vinyl-ester with
styrene.

The reaction competition between VB and ST monomer was investigated by Palmese
and et al as well [36]. The results indicate that the rate of fractional conversion of styrene
double bonds is initially less than that of vinyl-ester vinyl groups. However, styrene
monomer continues to react after conversion of vinyl ester double bonds has ceased. In

addition, the overall extent of conversion was found to increase with increasing

isothermal cure temperature.

25

2.2.3 Phase Separation and Microgel

An important feature of free-radical crosslinking copolymerization of styrene and
unsaturated polyester (UP) is the formation of a heterogeneous through strong
intermolecular reactions and phase separation. The cure of unsaturated polyester resins
has been found to result in heterogeneous morphologies characterized by the presence of
spherical structures ranging in size from 10 to 200 nm [3640]. Brill found closely packed
microgels on the order of 100 nm in diameter for VE resin cured at 90°C and washed
with acetone, which attests to the formation of microgel structures in VB systemslm.
Many investigators have shown that curing of UP resins results in the formation of

[4247]. They deﬁne microgels as localized

microgels during the initial stages of reaction
microregions with a higher crosslink density than the surrounding medium. It has also
been suggested that the size and number of microgels depend on cure temperature and
resin formulation (i.e., styrene content and degree of unsaturation), and that these factors

[45-47]

may affect material behavior . Such structure affects not only the cure behavior and

morphological changes of the resin but also the physical properties of the ﬁnal products.

2.2.4 Inﬂuence of Styrene Monomer Content

The mechanical properties of thermosetting polymer depend on the molecular units
making up the network and on the length and density of the cross-link. The former is

mined by the initials, promoters and accelerator are used and the later by the control of

26

the cross-link processes which are involved in the cure, especially cure temperature. It
was found that the initial cure temperature signiﬁcantly affects the mechanical behavior
of vinyl ester resin systems in particular, value of strength and fracture toughness for
125°C postcured samples initially cured isothermally at 30°C are much higher than those

obtained for sample initially cured isothermally at 90°C [37].

Thermosetting polymers are
normally considered as brittle. Cross-linking makes sliding of molecules past one and
another difﬁcult thus making the polymer strong and rigid but often do not have much
elongation before break. It was also found that low temperature cured, 30°C cured,

D411-C50 vinyl ester resin exhibits ductile behavior, other case such as 125°C postcured

after low temperature curing, 90°C cured and 125°C postcured after 90°C curing,

however, show typical brittle behavior [37].

The properties of cured vinyl ester also
depend strongly on the concentration of styrene monomer, as reactive diluents, in the

original resin, see Table 2.2.

27

 

Styrene Content (wt%) 0 15 50

 

 

 

Heat Distortion Temp. (°C) 149 138 1 16
Tensile Modulus GPa (Mpsi) 4.1 (0.59) 3.7 (0.53) 3.3 (0.48)
Elongation (°/o) 2.0 2.3 4.5

 

 

 

 

 

 

Table 2.2 A Comparison of the Selected Properties to the
Styrene Monomer Content for D411 Vinyl Ester Resin [‘81

 

 

 

 

 

 

 

 

Double Volume Shrinkage (%)
R inliliislsmi gwrfnet 3°“ F Li id
es n .0 ecu ar 0': en Content From Gel Point to ram qu
“'9'" W (sq/1009) Full Cure ”‘3': F“"
500a 0 0.40 4.3 5.4
5008 25 0.55 6.8 7.6
500° 50 0.70 9.1 10.3
900b 50 0.61 - 7.0
1200° 50 0.58 - 5.4

 

 

 

 

 

Note: a: EPOCRYL Resin 12; b: EPOCRYL Resin 321; c: EPOCRYL Resin 322.

Table 2.3 Volume Shrinkage of Vinyl Ester Resin [‘8‘

28

 

2.2.5 Cure Shrinkage of Vinyl Ester

Upon cure, the volume shrinkage of vinyl ester resin is generally lower than that of
unsaturated polyester resin but higher than that of their parent epoxy resin. Vinyl ester
resin can undergo as much as 5-10% volume shrinkage with cure depends on the
molecular weight of vinyl ester monomer and the content of styrene, see Table 2.3 [48],
while typical epoxy system undergo only 3-4% shrinkage during cure [48‘49]. Shrinkage
can induce signiﬁcant stresses in the composite already before loading. While when the
material in the liquid state, no stress will be build up. Even when the material in one
region is solidiﬁed, the surrounding material can still ﬂow and shrinkage induced stress is
limited. Stress caused by volume shrinkage begins to build up at a certain point of
volume shrink. The cure shrinkage of some of the vinyl ester resin are listed Table 2.3.

The curing stage and volume shrinkage rate are not yet available, the determination of
the so-called certain point of cure volume shrink is till a problem. According to Ten

[50], among the total volume shrinkage of 9.3%, about 2.7% occurs in the solid

Busschen
state. Also when cured vinyl ester experience elevated temperature, it will expense just
like other resins. The most characteristic property of thermosetting resin is in response
to heat since, unlike thermoplastics, they do not melt on heating. However, they lose
their stiffness properties at the distortion temperature and this deﬁnes an effective upper

limit for the use in structure components. Some of the properties of D411-C50 vinyl

ester resin are list in Table 2.4.

29

 

Properties D411-C50

 

 

 

Tension Modulus, E, GPa (Msi) 3.38 (0.49)
Poisson Ratio, v12 0.36
Shear Modulus, G, GPa (Msi) 1.25 (0.181).

 

Coefficient of Thermal Expansion 80m]
(CTE), or, (109°C)

 

Density, p, (g/m3) 1.12

 

 

 

Surface Enemy, y, (mJ/mz) 40(epoxy)I43]"; 26.7(benzene)"7] "

 

 

Note: *Measured by Cobin “81 for and the Poisson ratio is based on the
measurement

“Surface Energy of vinyl ester should be close to that of epoxy while the
surface energy of styrene closes to that of benzene

Table 2.4 Selected Properties of D411-050 vinyl ester

30

Most recently, Shanghai F uchem has developed a new-type epoxy vinyl ester resin
featured with very low cure volumeshrinkage together with an improved elongation rate

& impact resistance performance, refer Chapter 7 for details.

2.3 Carbon Fiber— Vinyl Matrix Interphase

2.3.1 General Introduction to Fiber/Matrix Interphase

Topography and Morphology of Interphase: the combination of ﬁber and matrix

creates inevitable interface or interphase, Figure 2.8. The concept of interface thickness

[52-54] [57)

keeps developing from 0 (interface) to 500nm , up to 1000nm , a three

dimensional interphase [52’5“

. The boundaries of the interphase are deﬁned as extending
from the point in the matrix where the local properties start to change from the bulk
properties to the point in the ﬁber where the local properties are those of the bulk matrix.
This region includes matrix that may have chemical and morphological features different
from the bulk matrix. It can include impurities, unreacted polymer components, non-
polymerized matrix additives, sometimes polymer transcrystallites [5845”, layer of sizing
and etc. On the ﬁber side, topographic, morphological and chemical features can be
different from the bulk. At the interface, not only can there be chemical and physical
interactions between ﬁber and matrix, but also voids, adsorbed gas and surface chemical
groups can be concentrated.

Strength of the Interphase: Adhesion between ﬁber and matrix can also be considered

as the strength of the interphase. The interphase is where the ﬁber and matrix bonded

31

POLYMER RESIN

    
 

Variable Crosslink Density
Segregation
Transcrystallinity

W Trace Impurities
THERMAL Sizing/Finishes
CHEMICAL VolatilesNoids
MECHANICAL Fiber Topography
Surface Chemistry
Fiber Morphology

CARBON FIBER

Figure 2.8 Fiber-Matrix Interphase [Drzal]

32

matrix, or van der Waals force, electrostatic attraction, acid-base interaction including H-
together. The bond strength often refers to adhesion which can be evaluated by the work
of adhesion, WA, refer to Appendix I. For ﬁbers without sizing, the contributions to the
adhesion could be chemical reaction or inter-diffusion of element between ﬁber and
bond, or mechanical interlock. For ﬁber with organic coating the situation is rather
complex. Beside the interactions mentioned above, the entanglement between the
molecules of the polymer sizing and polymer matrix and the inter-diffusionl62'65] between
compatible sizing and matrix polymers could be one of the important contributions to the
adhesion. On the other side, voids, contaminates, etc., and thermal residual stress could
be negative contribution to the adhesion.

Thermal Residual Stress at Interphase: In addition to chemical and structure
consideration, the state of stresses, which result from the process of the composite
material itself, can inﬂuence the degree of ﬁber-matrix adhesion. The coefficient of
thermal expansion (CTE) of carbon ﬁber is quite small and can actually be negative while

the matrix has a CTE of thirty times larger [27'30‘53'83’84]

. The ﬁber is anisotropic and the
radial and longitudinal thermal expansions can be quite different. The polyester is
isotropic but has a CTE a factor of 80x10'6/K [50] (compare with Epoxy: 60x10'6/K [49],
127x106/K [66]). This disparity become increasingly signiﬁcant as higher processing
temperatures are reached with the absolute difference between the glass temperature and
the use of temperature determining the magnitude of these residual thermal stresses. In
addition, the difference of thermal conductivities of materials can introduce unevenly
distributed temperature during process and application can also result in residual stresses

even defects [67].

33

Interphase thermodynamics: a necessary condition for ﬁber and matrix bond is
thermodynamic wetting. Wetting can be described thermodynamically as the creation of
an interface whose free energy decreases. If contact angles are used as a measure of
wettability this means that the contact angles should be less than 90° for wetting to occur.
This happens when the surface energy of the wetting liquid is less than that of surface
which it is placed. Carbon ﬁbers, as mentioned in Section 2.1.3, generally are high in
surface energy and the porosity and micro-topography present on their surface aid the
wetting of the substrate. In the case of carbon ﬁber with D411-C50 resin which is a
mixture of vinyl ester resin and styrene monomer, the Vinyl ester resins have a surface
free energy high than the styrene monomer. This might imply that styrene potentially
predominate on the surface of the carbon ﬁber.

Interphase Engineering with Fiber Sizing: With ﬁber sizing, the interphase could be
controlled and designed. Usually these sizings, sometimes called coatings or ﬁnishes, are
100-500 nm thick [54]. Some of the properties of the sizing materials were imparted on
the ﬁber surface after sizing process. The sizings protect the ﬁber surface from damage,
improve the wetting of ﬁber by matrix and protect ﬁber surface reactivity [54]. The sizing
should be designed to adhere well to the ﬁber surface and should partially diffuse into the
matrix resin during cure. The interphase design could be design of the sizing chemistry,
sizing molecular structure and/or sizing thickness. Sizing could increase the strength of

the interphase by introducing more chemical reactive site and/or more surface area [23’ 66‘

68-70]

34

2.3.2 Evaluation the Strength of the Fiber-Matrix Interphase

Direct measurement of the interphase strength, normally refers to adhesion, is very
difﬁcult. Therefore, the level of adhesion is assessed practically by the interphase
strength value, the interfacial shear strength (IFSS), obtained from destructive mechanical
tests. These tests could be classiﬁed into three groups, ﬁber pullout test, single ﬁber

fragmentation test and microindentation test [79]

, shown in Figure 2.9, for details see
Appendix I. The microindentation test has attracted much attention because it is an in
situ testing method conducted on a real composite, thus allowing for evaluation of the
processing or environmental exposure encountered either during manufacturing or in
service. The interfacial shear strength value obtained by these tests for a given system is

consistent with the mechanical properties of macrocomposites ”I'm. A strong correlation

has also been found between the interfacial shear strength and such thermodynamic

[75-77] [77.78]

parameters as surface free energy or speciﬁc enthalpy of adsorption
Therefore, it seems to be possible to obtain information about ﬁber-matrix adhesion from

micro-mechanical experiments.

2.3.3 Low Adhesion between Carbon Fiber and Vinyl Ester Resin

Low adhesion was extensively reported between vinyl ester and carbon ﬁber [5 L 80'8”.
One example was measured by single ﬁber fragmentation shown in Figure 2.10””. The

carbon ﬁbers used were Panex 33 ﬁber —100% electrochemical surface treated. The

35

r:>

   

Single-Fiber Fragmentation

 

 

View. , mid. F It ,
L r ;. l H L . r”

(1) (2) (3)
Single-Fiber Pull-Out

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

m v d ELLE"
E PE‘
(1) A (2) (3)

Microindentation

1:1 Matrix :1 Fiber

Figure 2.9 Schematic diagram of the microindentation method for
measuring ﬁber-matrix adhesion [79]

36

interfacial shear strength of a carbon ﬁber/vinyl ester system, measured with single ﬁber
fragmentation method, possessed only 50% of that of a carbon ﬁber/epoxy system. Not
only when compared with epoxy, another measurement from drop pull-out tests shown in

1m], carbon ﬁber/ vinyl ester adhesion was tested to be fairly lower than all the

Figure 2.1
other tested polymer matrixes.

One of the reasons lows adhesion between vinyl ester and carbon ﬁbers could be the
chemical nature of the constituents of vinyl ester polymer, the catalysts and the
monomers. A previous investigation by Weitzsacker, Drzal et al. [88] has investigated the
reactions of the catalyst and/or promoters to determine if they were competing with the
vinyl ester matrix resin for reactive sites at the carbon ﬁber surface. Drzal and et al found
that as little as 3% of chemical bonding accounts for a 25% increase in the Interfacial
Shear Stress (IF SS) [89]. Cobalt was detected at 2.6% at the surface of the AS4 carbon
ﬁber exposed to cobalt naphthenate (CoNap) and a minor change was also noted in the
surface chemistry of the ﬁber exposed to dimethyl aniline (DMA) which are the
promoters and accelerators for the free radical copolymerization respectively.
Adsorption of catalysts on the ﬁber surface resulting less reactive site to form strong
chemical bonding could be one of the reason that lowers the adhesion between carbon
ﬁber and vinyl ester. Besides, the preferential adsorption of any of the catalysts would
change the stoichiometry. Further more the ﬁber/matrix interphase also could be
changed since removal of even a small amount of the initiator or promoter from the vinyl

ester to the carbon ﬁber surface has the potential to greatly affect the free radical

copolymerization, and further more effects such as micro-gel, micro-structure, and phase

37

 

A
O

37.0

00
01

00
O

18.4

N
O

17.0

 

 

Interfacial Shear Strength
(M Pa)
I; I)"

 

 

 

 

 

 

 

_x
O

 

 

Epon828- D8084 Vinyl D411-050
mPDA Ester Vinyl Ester

Figure 2.10 Interfacial shear strength measured with the single ﬁber
fragmentation test over a series of carbon ﬁbers in an epoxy
matrix, Epon 828-mPDA, two vinyl ester matrices, the
Derakane 411-050 and 8084 [5”

38

\l (D
O O

O)
O

A
O

Interfacial Shear Strength
(M Pa)
8 8

N
O

 

 

 

 

 

 

 

+ 29.1

 

 

 

 

 

 

 

 

PEPO Udel PHE PHEA Ultem Vinyl

Ester

Figure 2.11 Interfacial shear strength of micro-drlgatlilet with polymers and as-

received carbon ﬁbers

39

separation of the vinyl ester material close to ﬁber surface were mentioned in section 2.2.
Typical vinyl ester resins contain 35-50wt% styrene monomer as reactive diluents.

The copolymerization between styrene monomer and vinyl ester monomer is a
heterogeneous, free radical, chain-growth, and cross-linking reaction. The wetting
condition to the carbon ﬁber surface of styrene monomer could be different from that of
vinyl ester monomer because the surface energy of styrene monomer would be smaller
than that of vinyl ester resin (refer: 40(epoxy) [90]; 26.7(benzene) [9”). A potential styrene
rich layer on the ﬁber surface could be another reason that lows the ﬁber/matrix
adhesion.

Residual stress caused by signiﬁcant cure volume shrinkage of vinyl ester can be
another factor that lowers adhesion. Refer to Table 2.3, vinyl ester resin can undergo as
much as 5-10% volume shrinkage with cure depends on the molecular weight of vinyl
ester monomer and the content of styrene, while typical epoxy system undergo only 3-4%

shrinkage during cure [52-54].

On the other hand, composite materials have thermal
residual stress from the fabrication. In the case of carbon ﬁbers, the coefﬁcient of
thermal expansion (CTE) is quite small and can actually be negative [92]. The ﬁber is
anisotropic and the radial and longitudinal thermal expansions can be quite different. The
polyester is isotropic but has a CTE a factor of 80x10'6/K [50] (compare with Epoxy:
60x10'6/K [49], 127x10'6/K [66]). This disparity become increasingly signiﬁcant as higher
processing temperatures are reached with the absolute difference between the glass

temperature and the use of temperature determining the magnitude of these residual

thermal stresses. In addition, the difference of thermal conductivities of materials can

40

introduce unevenly distributed temperature during process and application can also result

in residual stresses, even defects which would lower the ﬁber/matrix adhesion [67].

2.3.4 Improvement of Adhesion between Vinyl Ester and Carbon Fiber

One strategy to improve adhesion is to interfacially engineer the interphase through
the use of a ﬁber sizing. Sizing could increase the adhesion by introducing more

[23’ 66' 68'701. Previous data has shown that

chemical reactive site and/or more surface area
the application of a lightly cross-linked amine-cured epoxy sizing to the carbon ﬁber
surface creates a beneﬁcial interphase between the carbon ﬁber and vinyl ester resin
matrix resulting in a substantial increase in ﬁber-matrix adhesion and the mechanical

93]. Corbin found that with

properties of carbon ﬁber/vinyl ester composites as well [51‘87‘
epoxy sizing, the interfacial shear strength measured with single ﬁber fragmentation
increased about 30%. The mechanical properties of the composite material, ﬂexural
strength and interlaminate shear strength, also increased 34% and 25% respectively [5 "87].
This results showed that the use of ﬁber epoxy sizing to interfacially engineer the
interphase offers a potential avenue for improve the carbon ﬁber/vinyl ester adhesion. It
also should be noted that, the exact mechanism by which this sizing improved adhesion is not

known and the interaction between the epoxy sizing and the vinyl ester has to be

optimized for compatibility and thickness.

41

2.4 REFERENCES

1.

10.

ll.

12.

13.

14.

K. Ashbee, Fundamental Principle of fiber reinforced composites, 2cd edtion,
Technomic, pp. Xiii, 1993

R.F.Gibson, Principle of Composite Material Mechanics, McGraw-Hill, Inc.,
New York, pp8, 1994

W. D. Callister, Jr., Materials Science and Engineering An Introduction, 4th
edition, John Wiley and sons, Inc., 1997

K. K. Chawla, Composite Materials Science and Engineering, 2nd Edition,
Springer, New York, pp20, 1998

J. Kim and Y. Mai, Engineered Interfaces in Fiber Reinforced Composites,
Elsevier, pp.l83, 1998

L. T. Drzal and G. Hammer, “Graphite Fiber Surface Analysis Through XPS and
Polar/Dispersion Free Energy Analysis”, Appl. Surf. Sci., 4, 340-355, 1980

N. Sugiura, M. Kobuke, H. Asai and L. T. Drzal, “Topography of High Modulus
PAN-based Carbon Fiber Surfaces ”, Proceedings of the 4th Japan-SAMPE,
1995

S. C. Bennett and D. J. Johnson, Proc. 5'” London Carbon and Graphite
Conference, Vol. 1, Society for Chemical Industry, London, pp377, 1978

L. T. Drzal, " The Epoxy Interphase in Composites”, review chapter in Advances
in Polymer Science 11, 15, K. Dusek, ed., Spring-Verlag, pp. 16, 1985

L. Weitzsacker, M. Xie and L. T. Drzal, “Using XPS to Investigate F iber/Matrix
Chemical Interactions in Carbon Fiber Reinforced Composites”, Surf. Interf.
Anal. 2;, 53-63, 1997

L. T. Drzal, "The Surface Composition and Energetics of Type A Graphite
Fibers" Carbon, 1_5_, 129-138, 1977

L. T. Drzal, J. Mescher and D. Hall, "The Surface Composition and Energetics of
Type HM Graphite Fibers", Carbon, 1_7_, 375-382, 1979

L. T. Drzal, "The Surface Composition and Energetics of Graphite Fiber
Surfaces”, in Treatise on Adhesion and Adhesives, 2, R. Patrick, editor, 1981

L. T. Drzal, M. Madhukar and M. C. Waterbury, ”Adhesion to Carbon Fiber

Surfaces: Surface Chemical and Energetic Effects ", Advances in Fiber Reinforced
Composites Technology, Elsevier Science Publishers, Ltd., London, 1992

42

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

G. Fisher and LT Drzal, ”The Effect of Surface Chemistry on Adhesion of
Carbon Fibers to a Polyphenylene Sulﬁde (PPS) Resin", 4th International
Conference on Composite Interfaces, Cleveland, OH, 1992

L. T. Drzal, M. Madhukar and M. C. Waterbury, "Surface Chemical and Surface
Energetic Alteration of the 1M6 Carbon Fiber Surface and Its Effect on

Composite Properties ", 14th Annual Meeting of the Adhesion Society, Hilton
Head, SC, 1992

L. T. Drzal, M. Madhukar and M. C. Waterbury, "Adhesion to Carbon Fiber
Surfaces: Surface Chemical and Energetic Effects ", Composite Structures, _2_7,
65-72, 1993

CL. Weitzsacker, and LT. Drzal, “Analysis of the F iber/Matrix Interface in
Carbon Fiber Reinforced Polymer Composites by XPS”, ANTEC ‘96:
Proceedings of the 54th Annual Technical Conference of the Society of
Plastics Engineers, 3, 3656-3660, 1996

J. B. Donnet, Carbon, Vol.6, pp161, 1968

H. P. Boehm, E. Diehl,W. Hech and R. Sappock, Agnew. Chem., Vol. 3, pp.669,
1964

F. Hopfgarten, Fiber Sci. Tech., Vol. 11, pp. 67, 1978
D. M. Brewis et a1, Fiber Sci. Tech., Vol. 12, pp. 41, 1979

L. T. Drzal, "The Effect of Polymeric Matrix Mechanical Properties on
F iber-Matrix Interfacial Shear Strength", Mat. Sci. Engr., A126, 289-293, 1990

D. H. Kaelble, P. J. Dynes and E. H. Cirlin, J. Adhesion, Vol. 6, pp.23, 1974

L. T. Drzal, "T he Role of the F iber/Matrix Interphase on Composite Properties ",
Vacuum, 4_1_, 1615-1618, 1990

G. Krekel, K. J. Huttinger, W. P. Hofﬁnan, “The relevance of the surface
structure and surface chemistry of carbon fibers in their adhesion to high

temperature thermoplastics Part 11 Surface Chemistry”, J. Mater. Sci., Vol.
29(13), pp.3461-3468, 1994

V. K. Raghavendran, PhD Dissertation, Department of Material Science and
Mechanics, Michigan State University, 1998

L. T. Drzal, M. Richand and P. Lloyd, "Adhesion of Graphite Fibers to Epoxy
Matrices. I. The Role of Fiber Surface Treatment”, J. Adhesion, E, 1-30, 1983

43

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

M. Desaeger, M. J. Reis and et a1, “Surface characterization of poly(acrylonitrile)
based intermediate modulus carbon fibers ”, J. Mater. Sci., Vol. 31(23), pp.6305-
6315, 1996

K. E. Atkinson, G. J. Farrow and C. Jones, “Study of the interaction of carbon
fiber surfaces with monofunctional epoxy resin ”, Composites Part A, Vol. 27(9),
pp. 799-804, 1996

Composites—Engineered Materials Handbook. Vol. 1, ASM International,
pp.122, 1987

L. T. Drzal, M. Rich, M. Koenigand and P. Lloyd, "Adhesion of Graphite Fibers
to Epoxy Matrices. II. The Effect of Fiber Finish”, J. Adhesion, 1_6_, 133-152,
1983

DOW Plastics, DERAKANE Epoxy Vinyl Ester Resins_Chemical Resistance
and Engineering Guide, Revised Edition, 1999

DOW Plastics, DERAKANE Resins_Technical Product Information

DOW Plastics, Fabricating Tips_DERAKANE Epoxy Vinyl Ester Resins,
Revised Edition, 1994

Brill, R. P. and Palmese, G. R., “An investigation of vinyl ester-styrene bulk
copolymerization cure kinetics using Fourier Transform Inﬁared Spectroscopy”,
J Appli. Polym Sci, Part B, Vol. 76, No. 6, pp. 1572-1582, 2000

Ziaee, S. and Palmese, G. R., “Effect of temperature on cure kinetics and
mechanical properties of Vinyl-Ester Resins ”, J Polym Sci, Part B: Polym Phys,
Vol. 37, No. 7, pp. 725-744,1999

C. P. Hsu and L. J Lee, “Free-radical crosslinking copolymerization of
styrene/unsaturated polyester resins: 1. Phase seperation and microgel
formation ” Polymer, Vol. 34, pp. 4496-4506, 1993

C. P. Hsu and L. J Lee, “Free-radical crosslinking copolymerization of
styrene/unsaturated polyester resins: 2. Electronic spin resonance study”
Polymer, Vol. 34, pp. 4507-4515, 1993

C. P. Hsu and L. J Lee, “Free-radical crosslinking copolymerization of
styrene/unsaturated polyester resins: 3. Kinetics-gelation mechanism ” Polymer,
Vol. 34, pp. 4516-4523, 1993

R. P. Brill, R. L. McCullough, and G. R. Palmese, Proc Am Soc Compos, pp576-
583, 1996

44

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

C. P. Hsu and L. J Lee, “Free-radical crosslinking copolymerization of
styrene/unsaturated polyester resins: 1. Phase seperation and microgel
formation ” Polymer, Vol. 34, pp. 4496-4506, 1993

C. P. Hsu and L. J Lee, “Free-radical crosslinking copolymerization of

styrene/unsaturated polyester resins: 2. Electronic spin resonance study”
Polymer, Vol. 34, pp. 4507-4515, 1993

C. P. Hsu and L. J Lee, “Free-radical crosslinking copolymerization of
styrene/unsaturated polyester resins: 3. Kinetics-gelation mechanism” Polymer,
Vol. 34, pp. 4516-4523, 1993

Y. S. Yang and L. J. Lee, “Microstructure formation in the cure of unsaturated
polyester resins ”, Polymer, Vol. 29, pp. 1793-1800, 1988

S. B. Liu, J. L Liu and T. L. Yu, “Microgelation in the curing of unsaturated
polyester Resins”, J Polym Sci, Part B: Polym Phys, Vol. 53, pp. 1165-1177,
1994

Y. J. Huang and C. J. Chen, “Curing of unsaturated polyester Resins: Eﬂect of
comonomer Composition. 1] High teperature reactions ”, J Polym Sci, Part B:
Polym Phys, Vol. 47, pp. 1533-1549, 1993

M. B. Launikitis, “Vinyl Ester Resin”, Hand Book of Composite, pp.38-49

L. S. Penn and T. T. Chiao, “Epoxy Resins”, Hand Book of Composite,
pp.57~88

A. ten Busschen, Ph.D thesis, Delﬁ University of Thechnology, the Netherlands,
1996

S. M. Corbin, MS. Thesis, Department of Chemical Engineering, Michigan State
University, 1998

L. T. Drzal, "Composite Interphase Characterization”, SAMPE Journal, 12,
7-13, 1983

L. T. Drzal, M. Richand and P. Lloyd, "Adhesion of Graphite Fibers to Epoxy
Matrices. I. The Role of Fiber Surface Treatment”, J. Adhesion, _1__6, 1-30, 1983

L. T. Drzal, M. Rich, M. Koenigand and P. Lloyd, "Adhesion of Graphite Fibers

to Epoxy Matrices. II. The Effect of Fiber Finish”, J. Adhesion, L6, 133-152,
1983

45

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

L. T. Drzal, M. Richand and M. Koenig, "Moisture Induced Interfacial Effects on
Graphite Fiber-Epoxy Interfacial Shear Strength”, Paper 4-G, 38th Annual
Technical Conference, Reinforced Plastic/Composites Institute, SP], 1983

T. Chaudhry, H. Ho, L. T. Drzal, M. Harris and R. M. Laine, “Silicon
Oxycarbide Coatings on Graphite Fibers. II. Adhesion, Processing, and
Interfacial Properties”, Matr. Sci. Engr., A195, 23 7-249, 1995

K. N. E. Verghese, N. S. Broyles, J. J. Lesko, R. M. Davis and J. S. Rifﬂe,
“Durability of polymer composites for inﬁastructure: the role of the interphase ”,
Proceeding’s of the 23rd Annual Meeting of the Adhesion Society, 2000

Teishev and G. Marow, “The effect of transcrystallinity on the transverse
mechanical properties of single-polymer polyethylene composites ”, J. Applied
Polymer Science, Vol. 56, pp. 959-966, 1995

Lustiger, “Morphological aspects of the interface in the PEEK-carbon fiber
system”, Polymer Composites, Vol. 13, No. 5, pp. 408-412, 1992

Moon, "The effect of interfacial microstructure on the interfacial strength of glass
ﬁber/polypropylene resin composites”, J. Applied Polymer Science, Vol. 54, pp.
73-82, 1994

E. Chen and B. Hsiao, “The eﬂect of transcrystalline interface in advanced
polymer composites ”, Polymer Engineering and Science, Vol. 32, No. 4, pp.

280-286

E. Jabbari and N. A. Peppas, “A model for interdiffusion at interfaces of polymer
with dissimilar pysical properties”, Polymer, Vol. 36, No. 3, pp.575-586, 1995

E. Helfand, “Theory of unsymmetric polymer-polymer interfaces”, J. Chemical
Physics, Vol. 62, No. 4, pp.l327, 1975

P. G. Gennes, “Dynamics of ﬂuctuation and spinodal decomposition in polymer
blends”, J. Chem. Phys. Vol. 72, No. 9, pp4756, 1980

P. Pincus, “Dynamics of ﬂuctuation and spinodal decomposition in polymer
blends. II”, J. Chem. Phys. Vol. 75, No. 4, pp1996, 1980

D. K. Hadad, “Physical and chemical characterization of epoxy resins ”, Epoxy
Resins_Chemistry and Technology, 2"d edition, edited by C. A. May, pp.
1089-1172, 1988

V. Rao and L. T. Drzal, "T he Temperature Dependence of Interfacial Shear
Strength for Various Polymeric Matrices Reinforced with Carbon Fibers ”, J.
Adhesion, _3_Z, 83-95, 1992

46

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

L. E. Nielen, Mechanical Properties of Polymers and Composites, Mercel
Dekker, New York, 1974

V. K. Raghavendran, M. C. Waterbury, V. Rao and L. T. Drzal , “Inﬂuence of
Matrix Molecular Weight and Processing Conditions on the Interfacial Adhesion
in Bisphenol-A Polycarbonate/Carbon Fiber Composites”, J. Adhes. Sci. Tech.,
11,No.12,1501-1512,1997

A. Kelly and W. R. Tyson, “Tensile properties of fiber reinforced metals:
Copper/Tungsten and Copper/Molybdenum ”, J. Mech. Phys. Of Solid, Vol. 13,
pp.329-350, 1965

M. J. Pitkethly, J. P. Favre, U. Gaur, J. Jakubowski, S. F. Mudrich, D. L.
Caldwell, L. T. Drzal, M. Nardin, H. D. Wagner and et al., “A Round-robin
Program on Interfacial Test Methods ”, Campos. Sci. Technol., ﬂ, 205-214,
1993

Yu. A. Gorbatkina, Adhesive Strength of Fiber-Polymer Systems, Ellis
Horwood, New York, 1992

L. T. Drzal and M. Madhukar, "Fiber-Matrix Adhesion and Its Relationship to
Composite Mechanical Properties " J. Matr. Sci., 28, 569-610, 1993

Hoecker, K. Friedrich, H. Blumberg and J. Karger-Kocsis, Comp. Sci. Tech.,
Vol.54, pp.317, 1995

H. D. Wagner, H. E. Gallis and E. Wiesel, J Mater. Sci., Vol.28, pp.2238, 1993
G. S. Sheu and S. S. Shyu, J. Adhesion Sci. Tech., Vol. 8, pp.1027, 1994

H. J. Jacobasch, K. Gruncke, P. Uhlmann, F. Simon and E. Mader, Composite
Interfaces, Vol. 3, pp. 293, 1996

M. Nardin and J. Schultz, Composite Interfaces, Vol. 1, pp.l77, 1993

P. Herrera-Franco and L. T. Drzal, "Comparison of Methods for the Measurement
of F iber-Matrix Adhesion in Composites ", Composites, 23, 2-27, 1992

D. Hoyns, and LT. Drzal, “Hydrothermomechanical Response and Surface
Treated and Sized Carbon F iber/Vinyl Ester Composites”, ASTM Symposium
on the Fiber Matrix and Interface Properties, 1994

V. Rao and L. T. Drzal, Polym Compos, Vol. 12, No. 48, 1991

L. H. Peebles, Carbon Fibers, CRC Press, Boca Raton, 1990

47

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

TH. Yoon, H. M. Kang, M. Bump, and J. S. Riffle, “Effect of solubility and
miscibility on the adhesion behavior of polymer-coated carbon ﬁbers with vinyl
esterresins ”, J. App. Poly. Sci. (USA), vol. 79, no. 6, pp. 1042-1053, 2001

Kim, I.C., Yoon, T.-H., “Enhanced interfacial adhesion of carbon ﬁbers to vinyl
ester resin using poly(arylene ether phosphine oxide) coatings as adhesion
promoters”, Journal of Adhesion Science and Technology (Netherlands), vol. 14,
no. 4, pp. 545-559, Feb. 2000

Derakane 41 1-45 (Vinyl Ester Resin), Alloy Digest (USA), pp. 2, Aug. 1991

M. J. Reis, M. J ., M. Botelho Do Rego, and J. D. Lopes Da Silva, J. Mater.Sci ,
Vol. 30, No. 118, 1995

M. J. Rich, S. Corbin, and L. T. Drzal, “Adhesion of carbon fibers to vinyl ester
matrices", Proceedings of the 12th International Conference of Composite
Materials, 1999

L. Weitzsacker, M. Xie and L. T. Drzal, “Using XPS to Investigate F iber/Matrix
Chemical Interactions in Carbon Fiber Reinforced Composites”, Surf. Interf.
Anal. Vol. 25, pp. 53-63, 1997

L. T. Drzal, N. Sugiura and D. Hook, “Effect of interfacial chemical bonding and
surface topography on adhesion in carbon fiber/epoxy composites”, Proceeding
of the lith Annual Meeting of the American Society for Composites, 1994

L. T. Drzal, "The Epoxy Interphase in Composites”, review chapter in Advances
in Polymer Science II, E, K. Dusek, ed., Spring-Verlag, pp. 16, 1985

L. Xu, Mater Thesis, School of Mechanics and Materials Science, Washington
State University, 1998

J. M. Whitney and L. T. Drzal "Three Dimensional Stress Distribution Around An
Isolated Fiber Fragment", ASTM STP 937, Toughened Composites, 179-196,
1987

L. Xu, L. T. Drzal, “Inﬂuence of interphase chemistry on the adhesion between
vinyl ester resin and carbon ﬁbers”, Proceedings of the 23rd Annual Meeting of
“Adhesion Science for the 21Sit Century”, 2000

48

CHAPTER 3

STATEMENT OF THE PROBLEM

The main objective of this study was to gain an understanding of the factors
controlling interfacial adhesion between carbon ﬁbers and vinyl ester resin; to model the
contributions of the factors controlling the ﬁber/matrix adhesion; and furthermore to
provide an engineered and optimized interface between carbon ﬁber and vinyl ester for
tailoring a structurally efﬁcient carbon ﬁber/vinyl ester composites. The adhesion was
evaluated as interfacial shear strength (IF SS) measured by micro-indentation. Previous
data has shown that the application of a lightly cross-linked amine-cured epoxy sizing to
the carbon ﬁber surface creates a beneﬁcial interphase between the carbon ﬁber and vinyl
ester resin matrix resulting in a substantial increase in ﬁber-matrix adhesion and an
improvement in the mechanical properties of carbon ﬁber/vinyl ester composites as well
“‘41. But the exact mechanism by which the use of this sizing improved adhesion is not
known and the interaction between the epoxy sizing and the vinyl ester has to be
optimized for compatibility and thickness. The ﬁrst unknown to be investigated in this
study is the inﬂuence of the thickness of the ﬁber epoxy sizing on the adhesion between
carbon ﬁber and vinyl ester resin and ﬁnd out the optimum thickness.

The vinyl ester utilized as a base system for this study was DERAKANE 411-C50

epoxy vinyl ester which contains 50% styrene as reactive diluents. The copolymerization

between styrene and vinyl ester undergoes a heterogeneous, free radical, chain-growth,

49

crosslinking reaction. The polymerization catalysts of vinyl ester resin were found to be
adsorbed on the carbon ﬁber surface. Cobalt was detected at 2.6% at the surface of the
carbon ﬁber exposed to cobalt naphthenate (CoNap) and a minor change was also noted
in the surface chemistry of the ﬁber exposed to dimethyl aniline (DMA), which are the
promoters and accelerators for the free radical copolymerization respectively [5 1.
Considering the “controlled and designed” epoxy sizing of optimum thickness as the
engineered interphase between carbon ﬁber and vinyl ester resin, it should isolate the
carbon ﬁbers from contact with the vinyl ester resin and hence eliminate the adsorption of
the catalysts onto the carbon ﬁbers surface. Comparing the micro-indentation data of the
composites with an engineered interphase to that of the composites made with “as-
received” carbon ﬁbers, the inﬂuence of the adsorbed catalysts on the adhesion between
carbon ﬁbers and vinyl ester resins can be found. On the other hand, compared to the
pure epoxy vinyl ester resin, the pure styrene monomer has a much lower surface energy.
As a result of the surface energy forces, there is a potential to create a styrene rich layer
on the surface of reinforced carbon ﬁber in the composites.

Vinyl ester resin can undergo as much as 9% volume shrinkage with cure while a
typical epoxy system undergoes only 3-4% shrinkage during cure [6‘7]. This shrinkage
can induce signiﬁcant stresses in the composite even before loading. A new type of vinyl
ester resin having lower cure volume shrinkage was found [8]. One of the investigations
of this part of the research is to use the low shrinkage DGEBA-T403 carbon ﬁber sizing,
the “controlled and designed” interphase, to isolate the ﬁbers from the vinyl ester resin
and then to observe the inﬂuence of the cure volume shrinkage on the IFSS values

measured by micro-indentation. The other strategy is to ﬁnd of the inﬂuence of matrix

50

resin cure volume shrinkage on the ﬁber/matrix adhesion by comparing the IFSS data of
vinyl ester matrices of different volume shrinkage.

Once the physics and chemistry of the interphase and the interaction of the epoxy
sizing with the vinyl ester matrix is known, ﬁnite element model would be used to
understand the experimental results. Using IDEAS for the pre-processing and ABAQUS
for processing and post-processing, a set of non-linear contact ﬁnite element models were
set up to determine the inﬂuence of the thickness of the sizing material and the results
were compared with the experimental data. LS-DYNA, one of the most powerful non-
linear ﬁnite element analysis methods, was used for the simulation of the cure volume
shrinkage and the stress brought by the shrinkage in the third part of the research.

The completed research has been organized into three major sections as described
above. Section 1: Use of a controlled and designed interphase to improve the adhesion
between carbon ﬁber and vinyl ester matrix which focused on the inﬂuences of the sizing
thickness and sizing property gradient on the interfacial shear stress. Section 2: The
inﬂuence of the component chemistry on the adhesion between carbon ﬁber and vinyl
ester matrix was investigated. This section focused on the effect of the constituents of
the vinyl ester resin system on the interfacial shear stress, the effects of the catalysts, the
effects of the styrene monomers and the DGEBA vinyl ester monomers. Section 3.
Thermal characterization of the interphase between carbon ﬁber and vinyl ester matrix
which focused on effects of the cure volume shrinkage of vinyl ester and the interfacial

thermal residual stress on adhesion between carbon ﬁber and vinyl ester resin.

51

REFERENCES:

1. S. M. Corbin, MS. Thesis, Department of Chemical Engineering, Michigan State
University, 1998

2. D. Hoyns, and LT. Drzal, “Hydrothermomechanical Response and Surface
Treated and Sized Carbon F iber/ Vinyl Ester Composites”, ASTM Symposium
on the Fiber Matrix and Interface Properties, 1994

3. M. J. Rich, S. Corbin, and L. T. Drzal, “Adhesion of carbon fibers to vinyl ester
matrices ”, Proceedings of the 12th International Conference of Composite
Materials, 1999

4. L. Xu, L. T. Drzal, “Inﬂuence of interphase chemistry on the adhesion between
vinyl ester resin and carbon ﬁbers”, Proceedings of the 23rd Annual Meeting of
“Adhesion Science for the 21”t Century”, 2000

5. L. Weitzsacker, M. Xie and L. T. Drzal, “Using XPS to Investigate F iber/Matrix
Chemical Interactions in Carbon Fiber Reinforced Composites Surf. Interf.
Anal. 25, 53-63, 1997

6. M. B. Launikitis, “Vinyl Ester Resin”, Hand Book of Composite, pp.38-49

7. L. S. Penn and T. T. Chiao, “Epoxy Resins”, Hand Book of Composite,
pp.57~88

8. WWW.FUC11'E1\4.COI\4

 

52

CHAPTER 4

EXPERIMENTAL METHODS

4.1 FIBER SIZING

4.1.1 Sizing Materials

Diglycidyl Ether of Bisphenol A (DGEBA) and an aliphatic Polyether Triamine
(Jeffamine T-403) were used as a carbon ﬁber sizing and curing agent respectively.
Diglycidyl Ether of Bisphenol A (DGEBA) was widely used as an organic sizing
material. The molecular structure of DGEBA and Jeffamine T-403 are shown as Figure
4.1. DGEBA can be formulated to accommodate most of limitations presented by
speciﬁc curing requirements. Jeffamine T-403 was selected to be the curing agent for
DGEBA considering its aliphatic amine nature, amine functional group and molecular
structure. Aliphatic amine—cured epoxy is room temperature cured which can keep the

thermal stresses to a minimum [M].

This system cannot be used at high temperature.
Table 4.1 presents the selected properties of DGEBA cured with J effamine T-403.
In a mixture of epoxy resin and curing agent, competitive adsorption may take place.

The amines have a surface free energy higher than epoxy. The epoxy may be adsorbed

on the basal plane of carbon ﬁbers [5'7]. The adsorbed epoxy may be displaced by a

53

O OH O
/ \ I / \
CHz—CH-CHZ R—CHz—CH—CHZ R—CHz—CH—CHZ
n

CH3

I
where n=0~3, R =OOC GO
|

CH3

Diglycidy Ether Bisphenol A (DGEBA)

(OCHZClDH)x—NH2

CH3
CH30H2 ———(OCH2CH)y——NH2
c‘m,
(OCHZCH)Z -—NH2
EH 3

where, x+y+z ~5.3

Aliphatic Polyether Triamine (JEFFAMINE T-403)

Figure 4.1 Molecular of Diglycidy Ether of Bisphenol A (DGEBA) and
Jeffamine T 403

54

 

 

 

 

 

 

 

 

 

 

Properties DGEBA-T403
Tension Modulus, E, GPa (Msi) 3.24 (0.47)
Poisson Ratio, v12 0.354
Shear Modulus, G, GPa (Msi) 1.27 (0.184)
Coefﬁcient of Thermal Expansion (CTE), a, 60
(10’5/°C)
Heat Distortion Temperature (°C) 62
Density, p, (g/m3) 1.16
Surface Energy, y, (mJ/mz) 40‘5]

 

Table 4.1 Selected Properties of DGEBA-T403 [21

55

 

 

 

 

Pro-impregnation Machine

 

 

 

Heating Control

/----

Control Desk

 

 

 

 

Fiber Roll

 

 

 

 

 

 

- Tension 0.6 . Drum Carriage 25 . Rotation 15rpm

Figure 4.2 Schematic of the pre-impregnation machine to apply sizing
to the carbon ﬁbers

56

curing agent molecule resulting a region where the local composition at the interface is

different from the bulk [7].

4.1.2 Sizing Process

A stoichiometric mixture of DGEBA and T-403 was mixed and let cure for 30
minutes, then added to acetone to form the sizing solution for the carbon ﬁber. Fibers
were sized with a Pre-impregnation machine, Figure 4.2. AS4 carbon ﬁbers were
allowed to go through the bath ﬁlled with sizing solution and left to dry for 24 hours. The
ﬁbers were cut into 6 in length and kept in tubes sealed with tube caps. The Pre-

impregnation machine was set to Tension at 0.6. Drum carriage at 25 and Rotation at 15.

4.1.3 Sizing Thickness Measurement (TGA)

Thermal Gravimetric Analysis (TGA) measures changes of mass of a sample as a
function of time isothermally, or as a function of temperature, from ambient to 1000°C in
a controlled gaseous environment. It can provide the percentage of weight loss after
coated ﬁbers have been heated up to 400°C for 3 hours to burn off the organic materials.
Assuming all the ﬁbers are evenly coated, the sizing thickness can be calculated based on
the densities of both the organic materials and the ﬁber. A typical TGA data was shown
in Figure 4.3. A 2950 TGA HR instrument of Mode: TGA 1000°C was used. The typical

experimental parameters were set up as following:

57

Method Segments:
1. Ramp 20.00°C/min res 40°C to 400 0C
2. Isothermal for 120 mins
3. Ramp 1000°C/min to 600°C
Sample Flow 40
Reference Flow 60
Advanced Parameters:
Data Sampling interval: 2.0sec/pt
Data Compression Threshold Values
Temperature: 0.05 0C
Signal A: 0.0000mg

Post Test Parameters:

Futnace: Open And Unlosd
Air Cool for 6.00min.

58

 

 

101% «w

 

 

 

 

 

 

 

100% ‘

(I)
U)
3 99%
E .
0') Weight Loss(WL) = 0.3994
'6 98% WL Pecentage W=4.1341

97% -,

96%

0 100 200 300 400 500 600 700

Temperature (°C)

Sizing Material

Carbon ﬁber

 

Carbon ﬁber and the sizing materials

Figure 4.3 Using TGA to measure the sizing thickness

59

As shown in Figure 4.3, assume the total weight loss percentage is WL, known the

density of carbon ﬁber is pf and the density of the sizing material is Ps, then,

2 2
R _
WL= 7“ r )p, 4.1

”W —r2>p. +r2pf1

 

The sizing thickness t=R-r, then

 

 

 

R: WL pf r2+r2 4.2
l—WL ps
So,
t=R—r= ILP—fr2+r2—r 4.3
l—WL p5

For a TGA measurement as Figure 4.3, refer to Table 2.1 and Table 4.1, the density
of the AS4 carbon ﬁber and DGEBA-T403 were 1.77gcm'3and 1.16gcm'3. According to
equation 4.3, t= 113.3nm for AS4 carbon ﬁber with a diameter of 7pm. With this
calculation, the sizing has been assumed to be even through out the whole ﬁber tow.
Therefore, the thickness of sizing calculated from the weight loss measured by TGA is an

average number.

60

4.1.4 Environmental Scanning Electron Microscope (ESEM)

Environmental Scanning Electron Microscope (ESEM) is a new technique in
scanning electron microscopy speciﬁcally designed to study insulating materials.
Polymers, biological cells, bacteria, concrete, etc can be observed in the ESEM without
prior specimen preparation or conductive coatings which are necessary for a traditional
Scanning Electron Microscope. Environmental SEM provides information regarding the
topography, morphology, and microstructure of a specimen by the gaseous detection
device which utilizes the ionization of the water vapor for the detection of secondary
electrons from the specimen surface. The real sizing topographies on the coated ﬁber
surface were observed by ESEM. A model 2020 ElectroScan ESEM was used to observe

ﬁber sizing.

4.2 Measurements of Mechanical Properties of Materials

4.2.1 Dynamic Mechanical Thermal Analysis (DMTA)

Dynamic Mechanical Thermal Analysis (DMTA) is an analytical technique, which
measures the modulus (stiffness) and damping (energy dissipation) properties of
materials as those materials are deformed under periodic (oscillatory) stress as a function

of temperature.

In DMTA, a sample is subject to a low-strain sinusoidal deformation, and its response

to the deformation is measured. The resultant measurements obtained are the storage

61

modulus E’, which is related to the stiffness of the material, and the loss modulus E”,
which determines the viscous loss properties. A third parameter, tan 5, is calculated by
dividing E” by B”. As the temperature is increased from subambient to the melt-state the
tan 8 curve exhibits a series of peaks. The most signiﬁcant peak is found in the region of
the glass transition temperature (Tg). Previous studies revealed not only the glass
transition temperature of the matrix material, but also a peak in the tan 5 curve where the
Tg of the coating material was detected [8]. This indicates that the DMTA is not only a
good technique for the measurement of mechanical properties of materials but also a
useful tool for characterization of the composite interphase. A typical DMTA plot is

shown in Figure 4.4.

A 2980 DMTA instrument and Multi-Frequency-Single Cantilever was used. From
the curves in Figure 4.4 storage modulus at certain temperature such as 30°C was
measured and glass transformation temperature was measured from the tan 6 curve. The
typical experimental parameters were set up as following:

Method Segments: 1.Ramp 3.00 0C/min to 170"
Amplitude: 50pm
Frequency Table (117.): 1.000

Advanced Parameters: Data sampling interval: 4.030c/pt

Post Test Parameters
Return to temperature range: 25.00 0C to 30.00 “C

Furnace: Open
Clamp: Lock

62

2500

2000

Storage Modulus (MPa)

-500

StcrageM 01101... —; repeat.
14

 

 

1’
29.97 C

‘ zoeBMPa

123.13C

f 1.2
,, 1.0
5 0.8
- i 0.6
‘ 0.4

(12

 

f 0.0

 

 

20

70 120 170

Temperature (°C)

Figure 4.4 Using DMTA to measure the storage modulus at certain
temperature and the glass transformation temperature

63

21190 "9.1.

Tensile Stress

(Lbs/Sqinches)

 

 

 

  

ATM—Lidulus=VStreslextensi0ri

 

 

1 2 3
Extension (°/o)

Figure 4.5 Use UTS to measure the tensile modulus

64

 

4.2.2 United Testing System (UTS)

Standard Test Method for Tensile Properties of Plastic— Test Method D683 was
carried out by a United Testing System to measure the elastic modulus of the vinyl ester.
This test method covers the determination of the tensile properties of unreinforced and
reinforced plastics in the form of standard dumbbell-shaped test specimens when tested
under deﬁned conditions of pretreatment, temperature, humidity, and machine speed.
The typical experimental parameters were set up as following:

Load Cell Capacity (Lbs): 1000
Cross Head Speed (in/min): 0.1
Preload Value (Lbs): 5

One typical load-extension curve given by UTS is as Figure 4.5. The modulus of

elasticity is calculated by extending the initial linear portion of the load-extension curve

and dividing the difference in stress corresponding to any segment of section on this

straight line by the corresponding difference strain.

4.2.3. MTS Nam-indentation and Nano-scratch Test

The nano-indentation method was originally designed to for the purpose of probing
the mechanical properties of very small volumes of materials. It is ideal for mechanically
characterizing thin ﬁlms, coatings and surface layers including those modiﬁed by ion
implantation, because the layers does not have to be removed from its substrate [9]. Most

recently, it was found that the nano-scratch test could be successfully employed in the

65

investigation of composite interphases “0]. A MTS nanoindentation machine was used to
investigate both the mechanical properties of the bulk material and the properties of the
interphases between materials. The resolution and ranges of the instrument is listed in

Table 4.2.

Table 4.2 The resolution and the ranges of MTS nano-scrach XP head and nano-
indentation DCM Head

 

 

 

 

 

 

 

 

Resolutions XP Head DCM Head
Loading Resolution 1nN 1nN
Displacement Resolution 0.4nm 0.0002nm
Indenter Range 1mm 20pm
Load Ran e 500mN 10mN

 

4.2.3.1 Nano-Indentation Test

Generally the indentation process is described as three steps. 1. As the indenter is
driven into the material, both elastic and plastic deformation cause the formation of a
hardness impression conforming to the shape of the indenter to some contact depth, be. 2.
Then the machine holds position for a short time at a constant load in the indenter. 3. As
the indenter is withdrawn, only the elastic portion of the displacement is recovered. It is
the recovery in the third step which allows one to determine the elastic properties of a

material.

66

 

 

I'll

Lo_a_d ile

   
 
   
  
 

: ime Dependent

 

 

Slope =Elastic Stifness

Loading

llllllll'llllllll1lllllJlrllllllll‘llllll

 

Unloading

 

l J l
Displacement (nm)

 

 

Figure 4.6 A hypothetical set of continuous load-displacement data
collected during the Nana-Indentation experiment”)

67

The indenter tip is a DCM type with a Berkovich pyramid shape. A hydrothetical set
of continuous load-displacement data is present in Figure 4.6. [9] Note that the slope has
dimensions of force per unit distance, and it is also known as elastic stiffness, S, of the
contact. The reduced elastic modulus, E,, accounts for bi-directional displacements in

both the indenter and the sample which is related to S as:

_1_./Es

E, = —— 4.4

rﬂ 2 A

where A is a factor related to contact area, for a conical tip with contact radius, a ,
az(1t/A)”2; The factor 7 corrects for these approximations and depends on the strain
imposed by the indenter and the Poisson’s ratio pf the test material; B=1 for any tip that

creates a circular contact, B=1.012 for a Vickers pyramid, B=1.034 for a Berkovich

pyramid, and

 

—= ‘ + ’ 4.5

where E, and v, are the elastic modulus and Poisson’s ration for the sample. E, and v,
are the same properties for the indenter. For diamond, E, =1141 GPa and v, =0.07. For
most engineering material have a Poisson’s ratio between 0.15 and 0.35, always less than

0.5. With an input number of vs, the machine can calculate the Es.

68

4.2.3.2 Nano-Scratch Test

The nano- scratch test involves moving the diamond indenter tip across a sample
surface at a ﬁxed depth . The tip used in this scrapping experiment is also Berkovich
pyramid shape but is of XP type. The scratch tip was oriented with the sharp edge into
the direction of motion. The normal force is maintained at a constant value and the
lateral force is measured from the deﬂection of the shaft and the lateral displacement.
The ratio of these two forces is the coefﬁcient of friction (COF) between the material of
the indenter and that of the scratched material which is different from the one in physics
concept, see Figure 4.7. The depth of the indenter is also recorded, thus indicating the
hardness of the surface being scratched.

The interphase transition information can be inﬂuenced by two factors. One is the
stress ﬁeld and corresponding plastic zone, the other is the tip size. When the indenter is
pressed into the sample material it creates a corresponding stress ﬁeld and associated

“H31. Thermosetting polymers are materials for which the coefﬁcient of

zone of plastic
work hardening is very small. The stress and plasticity interaction effect should be
negligibly small when a scratch is created in the regions of polymer materials. However,
when the indenter tip is in close proximity to the interface near a carbon ﬁber where the
elastic modulus is at least two orders of magnitude above that of the polymer, the effect

could be signiﬁcant. Therefore, a scratch extending from the stiffer carbon ﬁber to the

softer polymer is more preferred than a scratch from polymer to carbon ﬁber.

69

Fr
————>

FN Motion of Indenter

 

Indenter

 

 

 

Scratch Friction Coefficient = f3 = FTIFN
Actual Friction Coefﬁcient = fa = 1:16

 

Figure 4.7 Coefﬁcient of friction (COF)[9]

70

 

The depth of the scratch depends on the indenter size. The indenter size could
enlarge an inﬁnitesimally small interface into a much wider region creating an artifact in
the data . As mentioned before, materials of different hardness will have different depth
of scratch. This scratch tip size and depth inﬂuence is compensated by a C++ program

which includes calculations based on some reasonable assumptions, see Appendix II.

4.3 Measurement of the Adhesion between Carbon Fiber and Vinyl Ester Resin

The Interfacial Testing System (ITS) apparatus was used to conduct micro-
indentation measurements to determine the interfacial shear stress (IFSS) “5]. The
measured interfacial strength is given by a generalized empirical equation (ITS shear
equation) which is embedded in the data reduction soﬁware of the apparatus. The ITS
apparatus indentation test is an in situ interface test for real composites and has the
advantage of reﬂecting actual process conditions. It can determine the interphase
strength due to fatigue or environmental exposure, or possibly monitor the interphase
properties of parts in service. In this method a sample of actual composite is tested.
Selected single ﬁbers perpendicular to a cut and polished surface are compressively
loaded to produce debonding or ﬁber slippage. Interphase bonding is monitored
microscopically between steps, until debonding is observed, see Figure 4.8 and Figure
4.9. ITS micro-indentation tests were performed until 10-15 good failures were observed
for each ﬁber/matrix system. The standard deviation of measured IF SS values for each

system could be very large. Normally a standard deviation of 10~15% of IFSS would be

71

 

Video

Camera \

Computer
Objective Lens

0.. 1 l

Indenter l 1:]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Specimen

 

 

Load Ce 1

<-—>
E —l_. XY controller

1— 2 controller

 

 

 

 

 

 

 

 

 

 

 

 

 

Optical
Microscope

 

 

 

 

IFSS: 181 1001: 0.875696 -—-—Gm — 0.018626n(—dn )- 0.026496 4.6
2 E d
df f f

fg = Load on Fiber

Gm= Shear Modulus of Matrix

E = Tensile Modulus of Fiber

dn= Distance between the Nearest Fiber
df= Diameter of the Fiber

A= Conversion Factor

Figure 4.8 Interfacial Testing Systems (ITS) used for performing
micro-indentation tests on ﬁber reinforced composites‘ 4]

72

    

Fiber before test

 

Fiber after test

Figure 4.9 ITS images for ﬁber before and after test

73

acceptable. It is important to point out that the ﬁber selection for the ITS test was critical
to obtaining reproducible results. The ITS shear equation is derived based on a
maximum shear stress criterion using the results of a linear axi-symmetric ﬁnite element

[15,16]

analysis and generalized by empirical data for samples of various ﬁber and matrix

[15] [17]

combinations . According to Ho , the ITS shear equation, equation 4.6, agrees well
with the nonlinear ﬁnite element method in deriving the maximum interfacial shear stress
when the ﬁber volume fraction=30~50%. To control the ﬁber volume fraction condition
for the testing ﬁber, the distance between the selected test ﬁber and its nearest neighbor
must exceed 2 micron meters and the distance between the selected test ﬁber and the at
least three nearest neighbor ﬁbers must be within a distance of half of diameter of the

selected test ﬁber. In addition, the selected test ﬁber and the ﬁbers near it must be free of

defects such as voids, scratched, chips, and cracks.

74

4.4 REFERENCES

10.

11.

12.

13.

S. M. Corbin, M.S. Thesis, Department of Chemical Engineering, Michigan State
University, 1998

L. S. Penn and T. T. Chiao, “Epoxy Resins”, Hand Book of Composite,
pp.57~88

M. A. F. Robertson, M. B. Bump, K. E. Verghese, S. R. McCartney, J. J. Lesko,
and J. S. Rifﬂe, “Designed Interphase Regions in Carbon Fiber Reinforced Vinyl
Ester Matrix Composites”, J. Adhesion, mg7802, SAS98V01. 74, 1999

Bergeret, N. Bina and A. Crespy, “Epoxy resins as surface treatments for mica
ﬂakes in propylene/ethylene copolymer composite"” Composite Interfaces,
Vol.5 No. 6 pp. 529-542, 1998

L. T. Drzal, "T he Epoxy Interphase in Composites”, review chapter in Advances
in Polymer Science II, E, K. Dusek, ed., Spring-Verlag, pp. 16, 1985

K. Horie, H. Murai and I. Mita, Fiber Sci. Tech., Vol 9, pp253, 1976

J. L. Racich, and J. A. Koutsky, Boundary Layers in Thermosets, Chemistry
and properties of Crosslinked Polymers, ed. S. Labana, pp 303, Academic
Press, 1977

“Advanced Techniques for Polymer Characterization” Beijing University of
Aeronautics and Astronautics (BUAA), 1990

J, Hay, “Mechanical testing by indentation ", Applied Nano Metrix, Inc, 1999
Hodzic, Z. H. Staehurski, J. K. Kim, “Nam-indentation of polymer-glass
interfaces Part 1. Experimental and mechanical analysis”, Polymer, Vol. 41, pp.
6895-6905, 2000

K. L. Johnston, Contact Mechanics, Cambridge University Press, 1987

B. D. Beake and G. J. Leggett, “Nanoindentation and nanoscratch testing of
uniaxially and biaxially drawn poly(ethylene terephthaslate ﬁlm”, Polymer, Vol.
43, pp. 319-327, 2001

S. Gao and E. Mader, “Characterization of interphase nanoscale property

variation in glass fiber reinforced polypropylene and epoxy resin composites ”,
Composites, 2001

75

14.

15.

16.

17.

V. K. Raghavendran, PhD Dissertation, Department of Material Science and
Mechanics, Michigan State University, 1998

The Interfacial Testing System (ITS), The Dow Chemical Company, Freeport,
Texas

J. F. Mandel], D. H. Grande, T. H. Tsiang and F. J. McGarry, “Modified
microdebonding test for direct in situ fiber/matrix bond strength determination in
fiber composites”, Composite Materials: Testing and Design, ASTM STP893,
American Society for Testing and Materials, pp. 87-108, 1986

H. Ho and L. T. Drzal, “Evaluation of Interfacial Mechanical Properties of Fiber

Reinforced Composites Using the Microindentation Method ”, accepted for
publication in Composites, Part A, 1998

76

CHAPTER 5

IMPROVEMENT OF ADHESION BETWEEN CARBON
FIBERS AND VINYL ESTER RESIN WITH EPOXY SIZING

Vinyl ester resins have been extensively used for the manufacture of low cost high
performance composites. Carbon ﬁbers are important reinforcement materials for the
production of high stiffness and strength composites. However when carbon ﬁbers are
combined with vinyl ester, the mechanical properties of the resulting composites are
lower than desirable because of widely reported low adhesion between carbon ﬁber and

vinyl ester “'91

. Previous data has shown that the application of a lightly cross-linked
epoxy polymer onto the carbon ﬁber surface provides a beneﬁcial interphase between the
carbon ﬁbers and vinyl ester resin matrices in promoting the interfacial shear strength
(IFSS) and the mechanical properties of carbon ﬁber reinforced vinyl ester matrix

[mm]. The interaction between the epoxy sizing and the vinyl ester resin has

composites
to be optimized not only for compatibility but also sizing thickness if further
improvements are to be realized. This chapter focuses on the understanding the adhesion
dependence on epoxy sizing thickness between carbon ﬁbers and vinyl ester resin;
determination of the optimized ﬁber sizing thickness; and investigation of the mechanism
by which an epoxy sizing inﬂuences carbon ﬁber/vinyl ester adhesion. The ﬁber/matrix

adhesion was evaluated as interfacial shear strength measured with micro-indentation.

The compatibility between the epoxy sizing and the vinyl ester matrix resin was

77

investigated with nano-indentation and nano-scratch characterization experiments as
well. A set of non-linear contact ﬁnite element models was also set up to simulate the
micro-indentation process. It was found that the epoxy sizing resulted in an increase in
ﬁber/matrix adhesion of about 30%. Apparently a stronger interphase was formed when
the epoxy sizing forms chemical bonds with the carbon ﬁbers and an interpenetrating
network with the vinyl ester resin. It was concluded that the sizing thickness has an
optimum value around 90mm for the best adhesion. The ﬁnite element analysis agrees

with these results.

5.1 INTRODUCTION

Vinyl ester resins are now widely used in large-volume and low-cost applications
primarily with glass ﬁber as reinforcement. The manufacture and availability of heavy
tow (i.e. large number of carbon ﬁber ﬁlaments in each tow) carbon ﬁbers has lead to

price reductions without a loss in strength or stiffness “1].

Because of their superior
speciﬁc strength and stiffness, the low price carbon ﬁbers could be economically
competitive with glass ﬁbers for use in markets previously deemed too expensive, such as
infrastructure, automobiles and etc. Additional economic advantage would be enjoyed if
carbon ﬁbers could be substituted for glass ﬁbers used in existing composite
manufacturing methods, such as resin transfer molding with vinyl ester matrices.
However, composites of carbon ﬁbers in vinyl ester polymers possess unacceptably low

[1.2]

mechanical properties It is well known that ﬁber/matrix adhesion can be a

signiﬁcant reason for lowering the mechanical properties of a composite material. Low

78

 

U1
0

 

 

 

 

 

 

 

 

 

   

 

 

45 E .
40 ' El"as-rece1ved"
5 :
'5) 35 : epoxy-treated
: L
e :2 »
H L
m 20 E 19
10 I .'.:'. 33551:.
’5 1"? L- 5
TO 0' _ n- m 2)
(U V h v L (U
m <19 5 3 .c E .. n.
E) 'c '53 (D ‘5) E ‘13 0
*‘ m 5 1T. C b a v
U) a) (D

Figure 5.1 Effect of ﬁber sizing on the interfacial shear strength and
composite mechanical properties [2]

79

adhesion was extensively reported between vinyl ester and carbon ﬁber “'91. The
interfacial shear strength of a common carbon ﬁber/vinyl ester system possessed only

50% of that of a carbon ﬁber/epoxy system “‘2’”.

The use of vinyl ester matrices
reinforced with carbon ﬁbers requires an improvement in ﬁber-matrix adhesion levels.
DGEBA Epoxy (Epon 828) cured with Jeffamine T403 (T403) was applied as a
sizing for Panex 157% carbon ﬁbers [2]. Panex 157% carbon ﬁbers are PAN-based
carbon ﬁbers with an electrochemical surface treatment. A toughened ﬁber/matrix
interphase region was achieved and properties of composite were recorded by single ﬁber
fragmentation (SFFT), three points ﬂex tests and Iosipescu out-of—plane shear test.
Corbin found that with an epoxy sizing, the interfacial shear strength measured with
single ﬁber fragmentation increased about 30%. The mechanical properties of the
composite material, ﬂexural strength and interlaminate shear strength, also increased 34%
and 25% respectively, as shown in Figure 5.1”]. This chapter focuses on the
understanding the inﬂuences of epoxy sizing thickness on the adhesion between carbon
ﬁbers and vinyl ester resin; ﬁnding out the optimized ﬁber sizing thickness, and

investigating the mechanism of the carbon ﬁber/vinyl ester adhesion improvement with

epoxy sizing on the ﬁber surface.

5.2 EXPERIMENTAL MATERIALS AND METHODS

Experimental Materials: The ﬁber was an AS4 carbon ﬁber from Hexcel, Inc. It is a
type A, electrochemical surface treated (S), about 7-micron diameter, PAN based carbon
ﬁber. The ﬁber tow size is 12,000 ﬁlaments. The matrix resin was Derakane 411-C50

vinyl ester resin from Dow Chemical which contain 50wt% of Epoxy-based vinyl ester

80

and 50wt% of styrene monomer. The catalysts of D411-C50 were CHP-5 (diluted
cumene hydroperoxide) from Witco Chemical used as the initiator and CoNap and DMA
were both from Aldrich Chemicals used as promoters and accelerators respectively.
Diglycidyl Ether of Bisphenol A (DGEBA) and Aliphatic Polyether Triamine (J effamine
T403) were from Shell and Huntsman respectively. Figure 5.2 shows the molecular

structures of all the chemicals used in these experiments.

Experimental Methods: A stoichiometric mixture of Diglycidyl Ether of Bisphenol A
(DGEBA) and Aliphatic Polyether Triamine (Jeffamine T—403) was prepared and held for
one half hour before dilution in acetone to form a sizing solution. Fiber sizing was
carried out using a pre-impregnation machine. An ElectroScan Environmental Scanning
Electronic Microsc0py (ESEM) was used to investigate the uniformity of the sizing.
Thermal Gravimetric Analysis (TGA) was used to measure the thickness of the sizing. A
digitally controlled, programmable oven is used for specimen curing to make sure all the
samples were processed in the same way. Figure 5.3 is the catalyst formulation and the
time-temperature schedule for the cure process used in this chapter. Adhesion was
evaluated as an interfacial shear stress (IFSS) measured with a micro-indentation system,
Interfacial Testing System (ITS). Nano-indentation and nano-scratch techniques were

used to proﬁle the gradient of the sizing/matrix interphase.

81

CH3 0 OH OH OH 0 CH3
I

| I | I II |
CH2:C —C-O—CH2-CH-CH2{R —CH2—CH—CH2 }r R—CHz—CH—CHz—O—C —C =CH2
n

C : CH2
DGEBA Epoxy Based Vinyl Ester

C o 0.. o
/\ /

I \
Styrene (as reactive diluent) CH2 —-CH—CH2{ R—CHz—CH—CHZ‘}R—CH2—CH —— CH2
n

9H3 Diglycidy Ether Bisophenol A (DGEBA)
1H3 (OCHzcle)x—-NH2
CH3
Cumene Hydroperoxide (CHP) CH3CH2 ———(OCHZCH)y-—NH2
I
+2 'O—C C—O' (OCHZCH)Z —NH2
C0 i (IZH 3

 

0O where, x+y+z ~5.3
J

Aliphatic Polyether Triamine (JEFFAMINE T-403)
Cobalt Napthanate (CoNap)

N
\ CH3 where R — 0'? ‘0
CH3
Dimethyl Aniline (DMA)

Figure 5.2 Molecular Structures for the chemicals used in this chapter

82

165

.I. .3 .s
o N R
U! U1 01

Temperature (°C)

45

25

65>

 

D411-050 CURE FORMULATION

 

 

 

 

 

CATALYSTS D41 1-C50
CHP-5 2.00%
CoNap 0.30%

DMA 0.10%

 

 

 

 

 

 

125°C 1.5 hour

90°C 1 hour / \

- é Cool

down

Naturally

- 91' 1

1 ﬂ ;
0 50 100 150 200 250

Time (min)

Figure 5.3 Cure formulation and cure process

83

 

 

e
m
f
r
e
t
n
I

Scratch Indenter

 

 

 

 

 

 

 

 

 

 

0.65

O
6.
0

5 0 5
5. 5. A
0 0 0

230580 c262“.

0
4.
0

0.35

Scrach Distance

Figure 5.4 The effect of the tip size to an interface with inﬁnitesimal wideness

84

Effect of Indentation Tip Size of the Nano-Indentation Scratch Test: As shown in Figure

5.4, the indenter size can create an artifact by enlarging an inﬁnitesimally small interface formed
when the sizing meets the vinyl ester matrix. The interface line will be a region in the measured
value of the coefﬁcient of friction in the scratch test. As shown in the ﬁgure, the region can be
identiﬁed when the indenter begins to make contact with the interface and ends when the indenter
is totally out of the interface. The indenter size depends on the depth of the scratch. The depth is
determined by the hardness of the test material and the vertical load applied during the scratch
experiment. The role of the scratch tip size and depth was compensated through the use of a C++
program simulation, refer Appendix 11. The calculation was based on some reasonable
assumptions and allowed for a more quantiﬁable determination of the interphase region

compensating for the shape of the indenter.

Finite element analysis: A two-dimensional axisymmetric ﬁnite element model which
utilizes a two-dimensional scheme was set up based on a four phase cylinder model as
shown in Figure 5.5. The boundary condition between the ﬁber and the indenter is
contact which brought nonlinearities to this model. Other boundaries such as the
boundary between ﬁber and interphase, the boundary between interphase and matrix, and

the boundary between matrix and composite were all set to be continuous which assumed

perfect bonds between the two neighbor materials.

All the materials used in the model were assumed to be homogeneous isotropic,
linear elastic material which is rarely the real case. No residual stresses were considered.
The width and length of the 2-dimensional model is 20pm and 100nm respectively.
Fiber diameter, df, was an input number measured from the ITS test. The interphase was

divided into 8 layers. The thickness of the interphase is 8x0.02um=0.16pm. For

85

 

 

 

 

 

 

 

 

Indenter
\\\\\j
\ Matrix
— Composites
so...

 

Figure 5.5 Four—phase model (drawing not to the scale) for ﬁnite element
analysis

86

interphase thickness equals to 0.02pm, one layer of the interphase materials properties
were set to be the same as DGEBA epoxy. The other seven layers were set to be the
same properties as vinyl ester. A similar set of boundary conditions was also used for
interphase thicknesses equal to 0.1pm, ﬁve layers of the interphase materials properties
were set to be the same as DGEBA epoxy while the other three layers were set to be the
same properties as vinyl ester.

According to Ho and Drzalllz'm, the ITS empirical equation agrees well with the
nonlinear ﬁnite element method in deriving the maximum interfacial shear stress as the
interfacial shear strength when the ﬁber volume fraction=30~50%. The best agreement
can be obtained Vr=36%. Neglect the interphase, Figure 5.3, rta2 /rc(a+b)2=0.36, then
a/(a+b)=0.6, because a=7/2, then a+b=11.667/2. Thickness of the matrix=b=2.33um.
Matrix thickness also can be determined by the ﬁber arrangement type and volume
fraction of the ﬁber content of the composite. The width of the composite is (20-2.33-
0.16'df)/2.

The material mechanical properties used in the models are listed in Table 1.
Composite properties were calculated base on the “rule of mixture” of a transversely

isotropic composite from Chamis “5].

The models were meshed using IDEAS.
Axisymmetric four-node elements were used. The boundary conditions were set up using

IDEAS. Contact loading and calculation were carried out using ABAQUS.

87

 

 

 

 

 

Fiber Interphase Matrix Composite
Tensile Modulus, E,
(GPa) 241 2.1 3.38 85
Shear Modulus, G,
(GPa) 96.5 0.92 1.24 32
Poisson Ratio, v 0.250 0.356 0.356 0.318

 

 

 

 

 

Table 5.1 Materials Mechanical Properties for Finite Element Model

88

 

5.3 RESULTS AND DISCUSSIONS

5.3.1 The Thickness of Sizing can be Controlled by the Concentration of Sizing

Solution

Fibers were sized with a pre-impregnation machine. The machine was set to Tension
at 0.6, Drum carriage at 25 and Rotation at 15. A stoichiometric mixture of Diglycidyl
Ether of Bisphenol A (DGEBA) and Aliphatic Polyether Triamine (Jeffamine T403) was
mixed and left standing for 30 minutes, before adding acetone to form the sizing solution.
Fiber tows were immersed and pulled through a bath containing the sizing solution and
then dried at room temperature for 24hours. The ﬁber tows with totally dried DGEBA-
T403 sizing materials were cut into 6 inch pieces and sealed in glass tube to keep the
sizing from deteriorating with storage. The thickness of ﬁber sizing was calculated from
the weight loss measured by TGA which was an average value of the sizing thickness.
The sizing thickness was linearly related to the concentration of the DGEBA-
T403/Actone sizing solution as shown in Figure 5.6. The thickness of sizing can also be
inﬂuenced by the tension of the ﬁber tow during process and the rotation speed and the
carriage speed of the drum as well.

The uniformity of the sizing was investigated by Environmental Scanning Electron
Microscope (ESEM). ESEM examination of the sized ﬁbers shows that the sizing
material was distributed unevenly as shown in Figure 5.7. From the ESEM observation,

the 5% sizing solution sized ﬁber gives the most uniform coating. For the 1% sizing

89

 

 

 

 

 

 

 

160 m...
140 f
E . y=1497.6x
5 120 R2= 0.99 ,
g 100 4 , ,, ,
i" 80 f or
1‘5 ,.
t» 60 ,j
.s I _ _ -__.__, ”A“
1% 4O “’ '9-thickness
20 f o .— L'Peﬂthickpgss)
0 f

0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10%11%
Concentration of DGEBA-T403/ACETONE sizing solution

Figure 5.6 The sizing thickness versus the concentration of sizing solution

90

 

 

Sizing bridges between ﬁbersl

 

Fiber as received

Figure 5.7 Sizing Quality Assessed with Environmental Scanning Electron
Microscope (ESEM)

91

(D
O
O)
O

 

 

 

75
5 ,
g) 70 5.0
g 65 ......
(t) >
L A 60 4.0
a m i z
5 g 55 j g
9 v 50 ' 3.0
.92
0 45 ~ ~ ~ ~
g + IFSS Measured by ITS
.93 40 -A-Finite Element Results ' 2'0
E 35 - 0 -Maximum Load,“

30 1.0

-20 0 20 40 60 80 100 120 140 160
Sizing Thickness (nm)

Figure 5.8 The interfacial shear strength (IFSS) related to the sizing
thickness

92

Bunse J. 10; pee-I ulnngew

solution sized ﬁbers, the ﬁber surface was not totally covered by the sizing material even
through the calculation based on TGA data tells that the sizing thickness is 20nm. The
sizing could not be sufﬁcient when sized with a sizing solution of 1% concentration.
This might indicate that the sizing material, which is slightly cross-linked, stays at the
ﬁber surface with a thickness of more than 20nm after drying. The sizing applied from
the 7.5% and 10% sizing solution created ﬁbers which have “sizing bridges” between
them. There is no other detectable topographic phenomenon between the different sizing

solution sized ﬁbers except the so-called “sizing bridges”.

5.3.2 Improvement of carbon ﬁber/vinyl ester adhesion with DGEBA sizing and the

optimum sizing thickness for the best Interfacial Shear Strength (IFSS)

Test data form Interfacial Testing System (ITS) showed that with Diglycidyl Ether of
Bisphenol A (DGEBA)/ Aliphatic Polyether Triamine (J effamine T403) epoxy sizing, the
Interfacial Shear Strengths (IF SS) increased up to 34.3%, Figure 5 .8. ITS tests also
showed that ﬁbers with sizing thickness around 90nm give the maximum value of IFSS.
The IF SS of ﬁber with 90nm sizing thickness increased 30% when compared with that of
ﬁbers without sizing. The IFSS value had little increase after the sizing thickness greater
than 100nm considering the error bars. The ﬁnite element model was a simulation of the
real ITS test. The load was applied on the ﬁber as a point contact as the indenter did in
the real test. The load was the maximum shown in Figure 5.8 with respect to the
thickness of sizing. The results of the ﬁnite element model was compatible to the results

of the ITS tests for maximum IFSS. The IFSS decreases when the sizing thickness is

93

greater than 100nm. This is because the input moduli of the pure epoxy sizing material
layers are only 2.1GPa. Comparing with modulus of 3.38GPa of vinyl ester matrices, the
epoxy interphase in the model by itself is softer than the matrix material. Therefore, it is
reasonable that as the sizing becomes thicker, its inﬂuence increases. Too much sizing
makes the IFSS decrease in the ﬁnite element analysis which neglects inter-diffusion
between the ﬁber sizing and matrix resin. In the real case, the interphase formed with
vinyl ester inter-diffusing with the epoxy sizing is stronger than the pure epoxy sizing
material itself. The experimental data indicate that a 5% sizing solution can provide
sufﬁcient sizing material on the ﬁber surface to form strong bonding between the ﬁbers

and sizing material.

5.3.3 The Role Of DGEBA Sizing Material Between Carbon Fiber And Vinyl Ester

The previous discussions show that the ﬁber sizing creates a beneﬁcial interphase
between the carbon ﬁber and vinyl ester resin matrix resulting in a substantial increase in
ﬁber-matrix adhesion regardless of the magnitude of the matrix cure shrinkage. As
discussed in chapter 2.1, the AS-type carbon ﬁbers receive an electrochemical oxidation
surface treatment during production which has been reported to promote adhesion to

“6"”. Firstly, the treatments

epoxy matrix materials through a two-step mechanism
remove a weak out ﬁber layer initially present on the ﬁber. Secondly, surface chemical
groups are added which increase the interaction with the epoxy. XPS investigation was

carried out by Weitzsacker and etc. which monitored the oxygen/carbon and

nitrogen/carbon ratio “6"”. The changes in the O/C and N/C ratio indicated that

94

H H

 

 

 

.1 O H E
' / \ ' I __ __ _ —
_,. HIE—(434042 _._,’ N CH2 C CHz R
O H T
I“ I
—> “—‘C—CPCHr C—Crb—R or —c—o—CHz—<,:—cn-tz—R
0H OH
O H H
----’ I: l ‘
’—C—N—R or ’C—N—R
O H H
-—= I: I I
-> ‘—C—-N *R or ——C_N-.R
I 1

The reaction groups are joined by the same kind of lines. The respective products of
the reactions are shoed at the end of the respective kind of arrows

Figure 5.9 Potential chemical reactions between carbon ﬁber surface and
epoxy/amine system

95

reactions between the epoxy and carbon ﬁber surface had occurred. As mentioned in
chapter 2.1, the carbon ﬁber surface chemistry includes hydroxyl, carbonyl and carboxyl
functional groups; however not all of these functional groups react with either the epoxy
or the amine. Figure 5.9 illustrates possible reactions between carbon ﬁber surface and
epoxy/amine compounds. Comparing the chemistry of vinyl ester with amine-cured
epoxy, the amine-cured epoxy sizing material has a greater possibility of forming a
chemical bond with the carbon ﬁber surface. Previous studies have found that as little as
3% of chemical bonding accounts for a 25% increase in the Interfacial Shear Stress
(IFSS) “81

Derekane 411 vinyl ester is made from Diglycidyl Ether of Bisphenol A (DGEBA)
epoxy. According to the rule of miscibility, two materials with similar chemistry should
be compatible. Investigations had been carried out to proﬁle the interphase between
DGEBA and D411-C50 “920] and determine the extent of miscibility. An interphase
specimen was made by fabricating a 2 layered specimen consisting of a DGEBA-T403
layer made from sizing solution coupled to a layer of D411-C50 resin mixed with its
catalysts. The specimen was cured aﬁer both components were brought into contact, and
then cured according to standard process. In Figure 5.10, nanoindentation tests were
conducted from the DGEBA-T403 side to the D411-C50 side of the interphase. For test
#1, 40 indents, tilt 25°, spacing 20pm and test #2 was 30 indents, tilt 45°, spacing 10pm.
Elastic moduli calculated from the indentation release curve are shown in Fig.14, where.
Distance 0 is about the location of the interphase. It was interesting to see that the value
of the moduli in the interphase region was a value between those of the DGEBA and

D411 bulk materials. Test #1 gave very consist moduli for both the DGEBA and the

96

Vinyl Ester

 

Figure 5.10 Nana-indentation test

97

Moduli (GPa)

4.3

4.2 .

4.1
4.0

3.9 -
3.8 ~

3.7
3.6
3.5
3.4
3.3
3.2

 

 

IEELLU IESIZ. ¥

I

I

I

I

I

I

I

.3 - 0 - , -e
i

1

I

I

I

I 1

1 L l

-35 0 35 70 105 140 175

Lateral Distance to Interface (pm)
(0 is the estimated interface location)

210 245

Figure 5.11 Nana-indentation results for the DGEBANinyl ester

interphase

98

C Interphase

 

Figure 5.12 Nana-scratch test traces for vinyl ester/DGEBA interphase

99

Coefficient of Friction (COF)

0.65

 

 

 

 

 

A Sizing/Resin Interphase
0'63 . A Reference Interphase A “45‘ ‘.‘z‘
0.61 f —C++ Simulation , M‘A‘A‘ 75), A2 AA
0.59 “Map? A... i H
A ‘A A
0.57 ‘
A“

0.55 ‘
0.53 A
0.51 WAAA A ‘ 5 A

I M. A A M
0.49

15 20 25 30

Scratch Distance (um)

Figure 5.13 Comparison of the nano-scratch data between DGEBA
sizing/vinyl ester interphase and reference interphase

100

D411 side, about 3.45GPa and 3.8GPa respectively. The scatter in measured values of
test #2 was greater because of the stress ﬁeld overlap caused by the plastic deformations
of the resins after indentation. This was because the spaces between each indentation
were too small compared with the tip size of the indenter in test #2. To avoid the overlap

of the stress ﬁeld during the nanoindentation test, the nano-scratch technique was used.

Unlike the nano-indentation test, the nano—scratch test gives continuous information
using an index called the coefﬁcient of friction (COF) between the material of the
indenter and that of the scratched materials. According to the published literature for
nano-scratch experiments conducted with polymers, when the same diamond scratch tip
is used, the harder substrate material produces a lower COF. A reference interphase
sample with no inter-diffusion was fabricated from the sizing solution after complete
oven drying. The vinyl ester resin was poured on this hardened layer and cured. The
scratch trace and scratch data were shown in Figure 12 and Figure 13. The sample
surfaces were microtoned with a diamond knife to avoid the introduction of artifacts as a
result of polishing. The scratch distance is 50 pm, the scratch depth is around 700nm.
The scratch data were related to the tip size as mentioned in the Experimental Methods
Section and the tip size is related to the depth of scratch. For Figure 5.13, the line is the
representation of the effect of the tip size, the open triangles are the COF for the
reference interphase which was assumed to be quite narrow. The solid triangles are the
COF of the sizing /matrix interphase. The COF of the reference interphase was consistent
with the simulation line. It was very clear that the width of sizing/matrix interphase was
much larger than that of the reference interphase, with additional data showing that the

width could be 0.5~1.5 pm.

101

Summarizing the discussion presented above, it could be concluded that the epoxy
sizing material provided both more chemical bonding opportunities to the ﬁber surface as
well as becoming a suitable interphase allowing inter-diffusion of the monomers in vinyl
ester resin. Comparing the IFSS value from micro-indentation and ﬁnite element
simulation of the micro-indentation process, a stronger interphase has been formed with
the inter-diffusion between epoxy sizing and vinyl ester matrix resin in the real
composites. This situation could be described as illustrated in Figure 5.14. Let, t=0 be
the time the ﬁber emerged in the resin and t=oo represent the ﬁnal composite material.
The interphase formed between the carbon ﬁber and vinyl ester with epoxy ﬁber sizing
would form chemical bonds between carbon ﬁber and epoxy sizing and interpenetrating

networks between the epoxy sizing and vinyl ester matrix resin.

102

Epoxy Sizing Chemical Bonding Swelled Epoxy Siz n

_\ ................... I

     
   
 
 

 

:I:I:‘EiEViﬁiiiiEetéri09.019133?

Carbon ﬁber

t=0 t=oo

Figure 5.14 Model for the interphase between carbon ﬁber and vinyl ester
resin with epoxy sizing

103

5.4 CONCLUSIONS

The application of a slightly cross-linked epoxy-amine polymer as a sizing on the
carbon ﬁber surface has been found to be an effective way to increase the adhesion
between carbon ﬁber and vinyl ester resin. The interfacial shear stress measured by
microindentation increases up to 34%. From this study the following conclusions can be

made:

1. The sizing thickness on the carbon ﬁber surface can be controlled by the
concentration of sizing solution. The sizing material was unevenly distributed on
the ﬁber surface after dry. 5wt% of DGEBA-T403 could provide sufﬁcient sizing
material on the AS4 carbon ﬁber surface to form strong bonding between the
ﬁbers and sizing material, in this case when the ﬁber diameter is 7~8 um. Too
much sizing material or to high of a concentration of the sizing solution could
produce so called “sizing bridges” between ﬁbers after drying the sizing material.

2. Slightly cross-linked DGEBA-T403 provides a beneﬁcial interphase between
carbon ﬁber and vinyl ester resin resulting in a substantial increase in the value of
interfacial shear strength, more than 30%. The optimum sizing thickness is
90~100nm.

3. Finite element analysis results are compatible with the micro-indentation result.
The maximum value of IFSS is attained when the sizing thickness is about
90~100nm. When the thickness is larger than this value, the modulus of the pure

epoxy sizing material is lower than the vinyl ester matrix material. The IFSS

104

decreases as the modulus of the matrix I the interphase decreases. The inter-
diffusion between DGEBA sizing material and vinyl ester could bring an
interphase which is stronger than the epoxy sizing material itself.

. DGEBA epoxy sizing provides a beneﬁcial interphase between the carbon ﬁber
and vinyl ester resin. First of all it provides more possibilities to form chemical
bonds between the composite matrix resin and the surface of the carbon ﬁber
reinforcement. Second, the epoxy vinyl ester was made from DGEBA epoxy
which should be compatible with the DGEBA sizing material. Third styrene
monomers are very small which could penetrate the slightly cross-linked epoxy
sizing material. Nano-scratch measurement found that the inter-diffusion distance
could be up to 1.5 microns. The optimum sizing thickness is 90~100nm. Nano-
indentation and nano-indentation scratch tests have been helpful in determining

the interphase proﬁle.

105

5.5 REFERENCES

10.

11.

12.

13.

D. Hoyns, and LT. Drzal, “Hydrothermomechanical Response and Surface
Treated and Sized Carbon F iber/ Vinyl Ester Composites”, ASTM Symposium
on the Fiber Matrix and Interface Properties, 1994

S. M. Corbin, MS. Thesis, Department of Chemical Engineering, Michigan State
University, 1998

V. Rao and L. T. Drzal, Polym Compos, Vol. 12, No. 48, 1991
L. H. Peebles, Carbon Fibers, CRC Press, Boca Raton, 1990

TH. Yoon, H. M. Kang, M. Bump, and J. S. Rifﬂe, “Effect of solubility and
miscibility on the adhesion behavior of polymer-coated carbon fibers with vinyl
esterresins ”, J. App. Poly. Sci. (USA), vol. 79, no. 6, pp. 1042-1053, 2001

Kim, I.C., Yoon, T.-H., “Enhanced interfacial adhesion of carbon ﬁbers to vinyl
ester resin using poly(arylene ether phosphine oxide) coatings as adhesion

promoters”, Journal of Adhesion Science and Technology (Netherlands), vol. 14,
no. 4, pp. 545-559, Feb. 2000

Derakane 41 1-45 (Vinyl Ester Resin), Alloy Digest (USA), pp. 2, Aug. 1991

M. J. Reis, M. J ., M. Botelho Do Rego, and J. D. Lopes Da Silva, J. Mater.Sci. ,
Vol.30, No. 118, 1995

M. J. Rich, S. Corbin, and L. T. Drzal, “Adhesion of carbon fibers to vinyl ester
matrices”, Proceedings of the 12th International Conference of Composite
Materials, 1999

L. Xu, L. T. Drzal, “Inﬂuence of interphase chemistry on the adhesion between
vinyl ester resin and carbon fibers”, Proceedings of the 23rd Annual Meeting of
“Adhesion Science for the 21St Century”, 2000

Zoltek Corp., St. Louis MO.

H. Ho and L. T. Drzal, “Non-linear Numerical Study of the Single-Fiber
Fragmentation Test, Part 1: Test Mechanics", Composites Engineering, _5, No.
10-11,1231-1244, 1995

H. Ho and L. T. Drzal, “Non-linear Numerical Study of the Single-Fiber

Fragmentation Test, Part II: A Parametric Study”, Composites Engineering, _5,
No. 10-11, 1245-1259, 1995

106

14.

15.

16.

17.

18.

19.

20.

H. Ho and L. T. Drzal, “Evaluation of Interfacial Mechanical Properties of Fiber
Reinforced Composites Using the Microindentation Method ”, accepted for
publication in Composites, Part A, 1998

C. C. Chamis, NASA Tech. Memo. 83320, 1983

L. T. Drzal, M. Rich, M. Koenigand and P. Lloyd, "Adhesion of Graphite Fibers
to Epoxy Matrices. II. The Effect of Fiber Finish”, J. Adhesion, m, 133-152,
1983

L. T. Drzal, M. Richand and M. Koenig, "Moisture Induced Interfacial Effects on
Graphite F iber-Epoxy Interfacial Shear Strength”, Paper 4-G, 38th Annual
Technical Conference, Reinforced Plastic/Composites Institute, SP], 1983

L. T. Drzal, N. Sugiura and D. Hook, “Effect of interfacial chemical bonding and
surface to ography on adhesion in carbon fiber/epoxy composites ”, Proceedings
of the 10t Annual Meeting of the American Society for Composites, 1994

L. Xu, L. T. Drzal, A. Al-Ostas, , and R. Schalek, “Improvement 0f Adhesion
Between Vinyl Ester Resin and Carbon Fibers by A Controlled and Designed
Interphase”, Proceeding of the 2002 Adhesion Society/WCARP-II Meeting,
2002

L. Xu, T. Mase, and L. T. Drzal, “Influence 0f Matrix Cure Volume Shrinkage on

the Adhesion between Vinyl Ester and Cardon Fiber”, Proceedings of the 14tll
International Conference for Composite Materials, 2003

107

CHAPTER 6

INFLUENCE OF THE COMPONENT CHEMISTRY ON
THE ADHESION BETWEEN VINYL ESTER AND
CARBON FIBER

Understanding the mechanism of adhesion between carbon ﬁber and vinyl ester is
critical important to tailoring structurally efﬁcient composites. Typical vinyl ester resins
contain 35-50% styrene as reactive diluents which could potentially be separated from the

matrix resin at the ﬁber/matrix interphase “”1

. Cure catalysts of vinyl ester system were
found to be adsorbed on the carbon ﬁber surface [5] which could signiﬁcantly change both
the chemistry and the mechanical properties of the ﬁber/matrix interphase. As discussed
in CHAPER 5, slightly cross-linked epoxy sizing, Diglycidyl Ether of Bisphenol A
(DGEBA) epoxy cured with Aliphatic Polyether Triamine (Jeffamine T403) amine,
provides a beneﬁt interphase between carbon ﬁber and vinyl ester resin and the optimum
sizing thickness is 90~100nm. Using epoxy ﬁber sizing to isolate the carbon ﬁber
surface from the vinyl eater resin, the effect of the vinyl ester, styrene and catalysts on the
adhesion between AS4 carbon ﬁber and DERAKANE 411-C50 vinyl ester was
investigated in this chapter. It was found that that the adsorption of the promoter and
accelerator on the carbon ﬁber surface does not substantially affect the ﬁber/matrix

adhesion. Optimum ﬁber-matrix adhesion can be obtained by properly selecting initiator

and adjusting the amount of initiator. It was also concluded that application of a lightly

108

cross-linked amine-cured epoxy polymer to the carbon ﬁber surface creates a beneﬁcial
interphase between the carbon ﬁber and vinyl ester resin matrix resulting in a substantial
increase in ﬁber-matrix adhesion. The vinyl ester and styrene components interact

differently with the DGEBA-T403 sizing at the carbon ﬁber and vinyl ester interphase.

6.1 INTRODUCTION

Typical vinyl ester resins contain 35-50% styrene as reactive diluents. The
copolymerization between styrene and vinyl ester is a heterogeneous, free radical, chain-
growth, and cross-linking reactions. Corbin found that among the three initiators, methyl
ethyl ketone peroxide (MEKP), cumene hydroperoxide (CHP), and benzoyl peroxide
(BPO), CHP provided the best result for the Panex carbon ﬁber/Derekane 411-C50 vinyl
ester adhesion [6], shown in Figure 6.1. Previous research by Weitzsacker, Drzal et al.[5]
have investigated the reactions of the catalyst and/or promoters to determine if they were
competing with the vinyl ester matrix resin for reactive sites at the carbon ﬁber surface.
Cobalt was detected at 2.6% at the surface of the AS4 carbon ﬁber exposed to cobalt
naphthenate (CoNap) and a minor change was also noted in the surface chemistry of the
ﬁber exposed to dimethyl aniline (DMA) which are the promoters and accelerators for
the free radical copolymerization respectively. The adsorption of catalysts on the carbon
ﬁber surface could induce two conditions: 1). The matrix material could be off-
stoichiometric that could affect the mechanical properties of the material; further more
affect the strength of the interphase. Because different amount of initiators, promoters

and accelerators produce different heats of reaction, crosslink density and degrees of

109

microgel formation and/or phase separation “‘41 as well as a different interphase
microstructure between the ﬁbers and the matrix. 2). The adsorption of the catalysts ions
on the carbon ﬁber surface normally would occupy the reactive sites of the carbon ﬁber
surface; carbon ﬁber surface would be less reactive site after adsorption which could
reduce the amount of chemical bonding between carbon ﬁber and vinyl ester resin. It
was reported that even as little as 3% of chemical bonding accounts for a 25% increase in
the Interfacial Shear Stress (IFSS) [71. Since removal of even a small amount of the
initiator or promoter from the vinyl ester to the carbon ﬁber surface has the potential to
greatly affect the polymerization reaction, one strategy is to prevent these constituents
from getting to the ﬁber surface by applying a ‘coating’ or ‘sizing’ to isolate the ﬁber
surface from the vinyl ester and cure catalysts. As discussed in CHAPTER 5, it was
found that that the application of a lightly cross-linked epoxy polymer, Diglycidyl Ether
of Bisphenol A (DGEBA) epoxy cured with Aliphatic Polyether Triamine (Jeffamine
T403) onto the carbon ﬁber surface provides a beneﬁcial interphase between the carbon
ﬁber and vinyl ester resin matrix resulting in an increase in ﬁber-matrix adhesion while at
the same time preventing the ﬁber from adsorbing the initiator and catalyst [6‘ 8"0]. The
optimum sizing thickness was found to be 90~100nm when 5% DGEBA-T403/Acetone

sizing solution was used for the sizing process [9].

Another issue in this study is the
wetting condition of the vinyl ester at the carbon ﬁber surface. As mentioned above,
typical vinyl esters use styrene as the reactive dilute. Compared with pure vinyl ester

resin, pure styrene monomer has a lower surface energy that would be preferred by the

carbon ﬁber surface. A rich styrene layer potentially formed on the carbon ﬁber surface

110

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

% 80 -—~-—-

5 7° 65 51 58 5

5:.) 60 49 52 '

L 50

“3 ”a? 40

c

a.

(4,3) g 30

a. 20

'23 10 »

it: 0

“ °\°\ °\°\ o\°\ o\o\ °\o\
q\ Q8 0\ Q0 {3
0% o .52 so $0

Panex Carbon Fiber
( the percentages are the levels of surface treatment)

Figure 6.1 Comparison of the interfacial shear strength between Panex carbon
ﬁber/Derekane 411-C50 vinyl ester adhesion with different initiator

lll

of a cured composite could be one of the reasons that lower the ﬁber/matrix adhesion.
Using DGEBA-T403 epoxy ﬁber sizing to isolate the ﬁbers from the matrices, the effect
of the reaction catalyst and the vinyl ester and styrene, have been investigated and are

reported in this research.

6.2 EXPERIEMTAL MATERIALS AND METHODS

Materials: the ﬁber is an AS4 carbon ﬁber from Hexcel, Inc. It is a type A, surface
treated (S), PAN based carbon ﬁber. The ﬁber tow size is 12K ﬁlaments. The matrix
resin is Derakane 411-C50 vinyl ester resin from Dow Chemical which contain 50wt% of
Bisphenol A Epoxy-based Vinyl Ester and 50wt% of Styrene. CHP-5 (diluted cumene
hydroperoxide (CHP)) from Witco Chemical is used as the initiator because it provided
the best IFSS among the three initiators, CHP, MEKP (methyl ethyl ketone peroxide) and
BPO (butanone peroxide). Both CoNap and DMA are from Aldrich Chemicals which are
used as promoters and accelerators respectively. Diglycidyl ether of bisphenol A
(DGEBA or Epon828) and Aliphatic Polyether Triamine (JEFFAMINE T403) are from
Shell and Huntsman respectively. Pure styrene monomer and pure bisphenol A epoxy

based vinyl ester were kindly donated by Huntsman and Reichhold respectively.

Methods: a digitally controlled, programmable oven was used for specimen curing to
make sure all the samples were processed in the same way, refer Figure 5.3 for the time-
temperature schedule. Dynamic Mechanical Thermal Analysis (DMTA), DMA 2980

from TA Instruments, and physical testing, United Testing System (UTS), of the

112

composite samples was conducted to measure the mechanical properties of the matrix
resin formulations. Adhesion was evaluated as an interfacial shear stress (IF SS) measured
with a microindentation system. Interfacial Testing System (ITS) apparatus was used for
microindentation measurement to measure the interfacial shear stress (IFS S). A thorough
discussion of the apparatus and the empirical equations for calculation of IF SS are
included in CHAPTER 4. The nanoindentation technique was used to test the mechanical

property of materials.

Controlled and Designed Interphase: as discussed previously, a lightly crosslinked
DGEBA epoxy sizing provides a beneﬁcial interphase between the carbon ﬁber and vinyl
ester resin. The optimum sizing thickness is 90~100nm corresponding to the thickness
obtained when using a 5% concentration sizing solution. This 90~100nm thick of epoxy
sizing on the ﬁber surface should completely coat the ﬁber surface with this slightly
cross-linked epoxy. This epoxy coating would isolate the functional groups on the
carbon ﬁber surface from contacting the vinyl ester matrix resin and its catalysts. It is
also known that the epoxy sizing material provided more opportunities to form chemical
bonds to the ﬁber surface. Furthermore, since the epoxy is only lightly crosslinked, the
monomers in the vinyl ester resin can diffusion into the sizing material to form a stronger
ﬁber/matrix interphase in the real composite material. The thickness of this interphase
can be controlled by the properly selecting the concentration of sizing solution and the

parameters of sizing processing pre-impregnation machines.

113

Sample Preparation: two kinds and three sets of each sample were made. The two kinds
of samples refer to two kinds of ﬁbers, one made from AS4 carbon ﬁber “as-received”,
named as “AS4”, the other were made from a 5% sizing solution sized AS4 carbon ﬁber,
named as “CAS4”. The three sets refer to three sets of vinyl ester matrix cure recipes.
Since it was shown that adsorption of the catalysts of the D411-C50 resin could take
place on the carbon ﬁber surfaces, it was expected that the vinyl ester in the interphase
region would not have the same stoichiometry as the bulk. Therefore, several off-
stoichiometric compositions were prepared and their physical properties were measured.
The curing recipes are listed in Table 6.1. Data Set 1 refers to those samples cured with
recipe 1-1, 1-11 and l-III which refers to a reduction in the concentration of CHP-5. Data
Set 2, 2-1, 2-11 and 2-III is a composition with a reduction in the amount of CoNap and
Data set 3, 3-1, 3-11 and 3-III with a reduction of DMA respectively. The IFSS between
carbon ﬁber and D411-C50 with depletion of cure catalysts, corresponding to the recipe
listed in Table 1 2-111 (without CoNap, the promoter) and 3-III (without DMA, the

accelerator), were also measured.

114

 

 

 

 

 

 

I II III
2.0%CHP-5 1.5%CHP-5 1.0%CHP-5
0.3%CoNap 0.3%CoNap 0.3%CoNap

0.1%DMA 0.1%DMA 0.1%DMA
2.0%CHP-5 2.0%CHP-5 2.0%CHP-5
0.2%CoNap 0.1%CoNap

0.1%DMA 0.1%DMA 0.1%DMA
2.0%CHP-5 2.0%CHP-5 2.0%CHP-5
0.3%CoNap 0.3%CoNap 0.3%CoNap
0.075%DMA 0.5%DMA

 

 

 

Table 6.1 Cure formulations

 

6.3 RESULTS AND DISCUSSIONS

6.3.1 Inﬂuence Of The Catalyst Concentration On The Properties Of The Matrix

Material

The data plotted as vertical bars in Figure 6.2, Figure 6.3 and Figure 6.4 were the
amounts of CHP-5, CoNap and DMA respectively, indicated by percentages, used in the
various formulations. For an isotropic material, Gm=Em/[2(l+v)], where Em is the
Young’s modulus and v is the Poisson ratio of the matrix material respectively. The
Poisson ratio of each formulation was assumed to be 0.36 (refer to reference 6) and
insensitive to the amount change of the CHP-5. The shear modulus and the Young’s
modulus then are related through the relationship Gm=Em/2.7. Figure 6.2, Figure 6.3 and
Figure 6.4 show that a small change of the catalyst could result in a detectable change in
the glass transition temperature, but the change of the mechanical properties such as
tensile modulus and storage modulus from DMTA measurement are rather small even in
two extreme situations (no CoNap and no DMA), especially for the calculated shear
modulus which is about 1/3 of the tensile modulus. Also the change of glass transition
temperature is not proportional to the change of the mechanical properties. Equation 6.1
is the empirical equation for the calculation of the interfacial shear strength (IF SS) from
the ITS. Equation 6.1 was based on a ﬁnite element analysis and adjusted by
experimental data “"121. The change in the shear moduli, Gm, of the vinyl ester cured
with different formulation was so small that it is assumed that the depletion of catalyst

from the matrix material near the carbon ﬁber surface does not affect the mechanical

116

Moduli (GPa)

‘_ £7.3toragieuMOdUIl-lsi from DLTAJGPa)

 

 

 

 

Figure 6.2 Inﬂuence of the amount of CHP-5 to the properties of cured
D411-C50 resin

117

Q
5 m
(n
at
s:-
4 a
l’.
a:
3 8
f6
2 3
(I)
H
.‘P.
1 E
m
0
H 1-l| H”
[:3 CHP-5 Amount (4=2%) -<> -Glass Transition Temperature (centigrade)
‘-l-Tensile Modulus from UTS (GPa) -0 -Shear Modulus from calaulation (GPa)

Moduli (GPa)

CoNap Amount (3=0.3%)
+Tensile Modulus from UTS (GPa)
-7A -§torage Modulus from DMTAQPa)

    

 

 

 

1-1 2-1 2-11 2-III

 

140

135

130

125

115

(Do)
eimeieduie; uolllsuen 98219

Figure 6.3 Inﬂuence of the amount of CoNap to the properties of

cured D411-C50 resin

118

-E>i-Glass Tgsition Temperatuie (centigfade).
-0 -Shear Modulus from calaulation (GPa)

 

 

 

 

8 -» 126
1

‘ G)
7 _
125 a
(I)
6 2r
“5 124 I“
0. 5 a
g ’5 '33
:5 4 123 0 g
v r-o-
'° (D
E 3 3
122 'D
(‘D
2 B
121 C
1 6

0 ' " ‘ 120

, 1-l 3-l W 3—11 7 3-lll L, ,
DMA amont (8=0.1%) -0 -Glass Transition Temperature (centigrade)
+Tensile Modulus from UTS (GPa) -0 -Shear Modulus from calaulation (GPa)

- Ari-Storage Modulus from DMTA (GPa) 7.

Figure 6.4 Inﬂuence of the amount of DMA to the properties of cured
D411-C50 resin

119

property of the material. Any changes of the Interfacial Shear Strength (IFSS) are the
result of change of the chemical or physical bonding between the ﬁber and the matrix, not

the mechanical properties of the matrix resin.

 

 

IFSS = A :1} {0.875696 Igﬂ — 0.018626ln(::” )— 0.026496} 6.1[l I-12]
f

2
f f

where,

fg = Load on Fiber

Gm: Shear Modulus of Matrix

E = Tensile Modulus of Fiber

d,,= Distance between the Nearest Fiber
df= Diameter of the Fiber

A= Conversion Factor

6.3.2 The effect of catalyst concentration on the Interfacial Shear Strength (IFSS)

As adsorption of catalysts on the carbon ﬁber surface was found. The adsorbed
catalyst ions could compete for the reactive sites of the carbon ﬁber surface with the
matrix polymer to reduce the amount of chemical bonding which could lower down the
adhesion between ﬁber and matrix. It was found and previously reported that that 3%

increase of chemical bonding could account for a 25% of increase of adhesion [“3 1.

Since the sizing provided by 5wt% of DGEBA-T403/ACETONE sizing solution was

effective at isolating the carbon ﬁbers from contacting

120

Interfacial Shear Strength IFSS

(MPa)

120

80

60

40

20

 

 

 

    

 

 

 

 

 

 

 

 

72 .63
66160 ;
49.44 I ‘
I
I
1 5%. . 1% 1%
1-l/AS4 1-lI/AS4 1-III/AS4 1-I/CAS4 1-Il/CAS4 1-lll/CAS4

H: the D411~050 is cured according to recipe H in table1
CAS4: 5% EPON828-T403/ACETONE sizing solution coated AS4

Figure 6.5 The inﬂuence of the amount of CHP-5 to the IFSS
between D411-050 and AS4/Coated AS4 ﬁbers

121

Interfacial Shear Stength IFFS

(MPa)

49.79

I
I

46.13 52;.83

 

0.2% 0.1%

 

 

 

 

 

 

 

1-l/AS4 2-l/AS4 2-lI/AS4 2-IIIIAS4

 

 

[ 84.63
66.89 I
57.31
0.1% 0%

 

 

 

 

 

1-I/CAS4 2-l/CAS4 2-lI/CAS4 2-lll/CAS4

CAS4: 5%EPON-T403/ACETONE sizing solution coated AS4
2-lll: the D411-C50 is cure according to recipe 2-lll in table 1

Figure 6.6 The influence of the amount of CoNap to the IFSS between
D411-050 and AS4/Coated AS4 ﬁbers

122

Interfacial Shear Stength IFFS

 

 

53.43 I

 

 

 

60 I 55 09
53:43 50*68 1

(MPa)
8

20 0079.40.05% 0% 0.075% 0.05% 0%

    

 

 

 

 

 

 

 

 

 

‘1-l/AVS4 3-l/AS4 3—ll/AS4 3-lll/AS4 1-I/CAS4 3-l/CA84 3-lI/CAS4 3-lll/CAS4

CAS4: 5%EPON-T403/ACETONE sizing solution coated AS4
3-lll: the D411-C50 is cured accroding to recipe 3-lll in table 1

Figure 6.7 The inﬂuence of the amount of DMA to the IFSS between D411-
050 and AS4/Coated AS4 ﬁbers

123

the vinyl ester and the catalysts. The sizing would prevent the carbon ﬁber from
interacting chemically with the constituents of the vinyl ester matrix. The set of bars on
the left side in these ﬁgures correspond to the samples made of ‘as-received’ ﬁbers,
indicated as AS4, and the data on the right side refer to ﬁbers ‘with epoxy sizing’,
indicated as CAS4, which were sized with the 5wt% of DGEBA-T403/ACETONE sizing
solution. The percent numbers in Figure 6.5, Figure 6.6 and Figure 6.7 are the amounts
of CHP-5, CoNap and DMA in each vinyl ester formulation respectively. The ﬁbers with
epoxy sizing overall produce greater interfacial shear strength (IF SS) than the ‘as-
received’ ﬁbers. For ‘as received’ ﬁbers, it is clear that the IFSS of the samples with
different amount of CoNap and DMA are almost the same, even for those without CoNap
and DMA, see Figure 6.6 and Figure 6.7. Figure 6.6 might indicate that although cobalt
adsorbs on the carbon ﬁber surface, it does not result in a signiﬁcant change in adhesion.
The same can be said for DMA in Figure 6.7. It is interesting to see that slight changes
in the amount of the initiator cause a signiﬁcant change of IFSS (Figure 6.5). At a
concentration of 1.5% of CHP-5, the greatest value of IF SS was measured. Changes in
the initiator concentration are responsible for a change in the number of reactive species
in the vinyl ester system. It appears that at a certain. concentration, the vinyl ester
copolymer has a microstructure (phase separation or micro-gel formation) which

produces good adhesion.

124

6.3.3 The Contribution of Pure Vinyl Ester and Styrene Monomer of the D411—C50

Vinyl Ester Resin to the Adhesion

Further investigation has been carried out to ﬁnd the roles that pure styrene and pure
vinyl ester play in forming the interphase between AS4 carbon ﬁber and D411-C50 vinyl
ester resin. New composite specimens were prepared with the AS4 carbon ﬁbers in
either pure vinyl ester or pure styrene where the catalyst concentration was adjusted to
maintain the same double bond concentration based on the standard recipe, recipe 1-1, in
table 1. As an input parameter for the ITS test, as shown in Equation 6.1, the shear
moduli of both pure materials are critical to get the value number of interfacial shear
strength (IFSS). Pure bisphenol A epoxy-based vinyl ester is a solid and styrene
monomer has a very high vapor pressure which makes neither one very easy to process at
room temperature. To obtain the in-situ mechanical properties, a nano-indentation test
was used with the corresponding ITS sample. The nanoindentation testing area is far
away from the ﬁber area to avoid other affects. Literature citations of mechanical
properties (D411-C50 vinyl ester was provided by the manufacturer) are also used as

references, Figure 6.8.

The ITS test results are very interesting, (Figure 6.9). The data on the left side in
F ig.12 corresponds to samples made with ‘as-received’ ﬁbers and the data on the right
side are for ﬁbers sized with DGEBA-T403/ACETONE sizing solution. Again, for ‘as-
received’ ﬁbers, there are slight IF SS differences between vinyl ester (~+20%), and

styrene (~-10%) compared to the D411-C50 vinyl ester. Styrene gives the lowest IFSS

125

 

 

moduli (GPa)

 

 

 

   

 

 

 

 

 

 

6.0 L LLLALI__L,_
I 4_9 .EEureLVEDﬂI-Péo £13433 renal
5'0 39 4.1
' ' I 4.1 4_o
4.0 : I
3.0 i '
2.0 E
1.0 E
0.0
Nanoindent Literature
Data Source

Figure 6.8 Moduli comparison between vinyl ester, D411-C50 and
styrene

126

Interfacial Shear Strength IFSS

 

 

 

 

 

 

   

 

 

 

 

 

 

120 PWT ei~ .__ ~
89.91
100 I
80 86.88 I
58.21
E 49.79
E 60 1 44.43
. _. T 1
20
O I
VE/AS4 D411/AS4 ST/AS4 VE/CAS4 D411/GAS ST/CAS4

VE: vinyl ester; D411: Derakane 411-C50; ST: styrene;
CAS4: 5% EPON828~T403IACETONE sizing solution coated

Figure 6.9 The IFSS of vinyl ester, D411-050 and styrene with AS4 and
coated AS4 ﬁbers

127

because it is the soﬁest material among these three. D41 l-C50 gives an IFSS value in between.
As pure styrene resin has lower surface energy (compare with the surface energy of Benzene:
26.7mJ/m2lm) than pure bisphenol A epoxy-based vinyl ester (compare with the surface energy of
epoxy resin: 40mJ/mzl'5'). Theoretically the carbon ﬁber surface would prefer styrene rather than
bisphenol A epoxy-based vinyl ester. But the IFSS number above does not indicate that there is a
styrene rich layer at the interphase between ﬁber and D41 l-CSO mixture. In contrast, the carbon
ﬁber with the lightly cross—linked sizing has the highest adhesion in the pure styrene matrix.
According to Rouse model “6' the overall mobility and diffusion coefﬁcient D is proportional to
M'l and the overall mobility and diffusion coefﬁcient D is proportional to M'2 according to
Reptation Model “7]. This suggests that the styrene monomers (monomer molecular weight, M, is
104g/mole while vinyl ester is about 900g/mole) diffuse into lightly cross-linked DGEBA-T403
sizing and bring about a big change to the ﬁber/matrix interphase. It might create a stronger
material; smoother gradient at the interphase compared to pure bisphenol A epoxy-based vinyl

ester resin.

128

 

 

6.4 CONCLUSIONS

Derakane 41 l-C 50 vinyl ester resin is a mixture of vinyl ester and styrene monomer,

crosslinked by a free radical polymerization. The interface between this resin and AS4

carbon ﬁber is much more complex than that between epoxy and AS4 carbon ﬁber. From

this study, several conclusions can be made.

1.

Slight change of catalyst concentration does not make a major difference in the
mechanical properties of the cured D411-C50 resin. The glass transition
temperatures of the cured resin appear to be inﬂuenced by the catalyst
concentration. The glass transition temperature is not necessarily related to the
mechanical properties of the materials for this system.

The adsorption of the catalysts, neither the promoter (CoNap) nor the accelerator
(DMA) signiﬁcantly inﬂuences the level of adhesion between AS4 carbon ﬁber
and D411-C50 vinyl ester resin. Changes in the initiator (CHP-5) concentration
cause changes in the polymerization of the vinyl ester system. Optimum ﬁber-
matrix adhesion could be obtained by properly select the initiator and adjusting

the amount of initiator.

. A sizing selected for isolating the carbon ﬁber surface from the reacting polymer

appears to be useful in improving adhesion. With DGEBA-T403 sizing, the
interfacial shear strength between D411-C50 and AS4 carbon ﬁber can be
increased indicating a strong interphase formed.

The contributions of styrene and vinyl ester in D411-C50 to the adhesion between

the ﬁber and matrix are different especially for ﬁbers with the lightly cross-linked

129

 

 

DGEBA-T403 sizing. The styrene monomers diffuse into the DGEBA-T403
sizing material and provide a much smoother interphase than only pure vinyl
ester.

5. Neither the adsorption of catalysts on the carbon ﬁber surface nor the potential of

 

phase separation between styrene monomers and vinyl ester molecules are the
reasons of lower adhesion between carbon ﬁber and vinyl ester resin. Other

factors such as resin cure shrinkage should also be considered.

 

 

130

 

6.5

1.

10.

ll.

12.

REFERENCES

Y. S. Yang and L. J. Lee, “Microstructure formation in the cure of unsaturated
polyester resins”, Polymer, Vol. 29, pp. 1793—1800, 1988

S. B. Lin, J. L Liu and T. L. Yu, “Microgelation in the curing of unsaturated
polyester Resins”, J Polym Sci, Part B: Polym Phys, Vol. 53, pp. 1165-1177, 1994

Y. J. Huang and C. J. Chen, “Curing of unsaturated polyester Resins: Eﬂect of
comonomer Composition. [1 High teperature reactions”, J Polym Sci, Part B:
Polym Phys, Vol. 47. Pp. 1533-1549, 1993

R. P. Brill and G. R. Palmese, “An investigation of vinyl-ester——styrene
copolymerization cure kinetics using Fourier Transform Inﬁared Spectroscopy”, J.
App. Poly. Sci., Vol. 76, pp1573-1582, 2000

L. Weitzsacker, M. Xie and L. T. Drzal, “Using XPS to Investigate Fiber/Matrix
Chemical Interactions in Carbon Fiber Reinforced Composites”, Surf. Interf. Anal.
25, 53-63 1997

S. M. Corbin, M.S. Thesis, Department of Chemical Engineering, Michigan State
University, 1998

L. T. Drzal, N. Sugiura and D. Hook, “Effect of interfacial chemical bonding and
surface topography on adhesion in carbon fiber/epoxy composites”, Proceedings of
the 10‘h Annual Meeting of the American Society for Composites, 1994

L. Xu, L. T. Drzal, “Inﬂuence of interphase chemistry on the adhesion between vinyl
ester resin and carbon fibers”, Proceedings of the 23rd Annual Meeting of
“Adhesion Science for the 21St Century”, 2000

L. Xu, L. T. Drzal, A. Al-Ostas, , and R. Schalek, “Improvement 0f Adhesion
Between Vinyl Ester Resin and Carbon Fibers by A Controlled and Designed
Interphase”, Proceeding of the 2002 Adhesion Society/WCARP-II Meeting, 2002

L. Xu, T. Mase, and L. T. Drzal, “Influence 0f Matrix Cure Volume Shrinkage on the
Adhesion between Vinyl Ester and Cardon Fiber”, Proceedings of the 14"I
International Conference for Composite Materials, 2003

The Interfacial Testing System (ITS), The Dow Chemical Company, Freeport,
Texas

J. F. Mandell, D. H. Grande, T. H. Tsiang and F. J. McGarry, “Modified
microdebonding test for direct in situ fiber/matrix bond strength determination in
fiber composites”, Composite Materials: Testing and Design, ASTM STP893,
American Society for Testing and Materials, pp. 87-108, 1986

I31

 

 

 

13. K. J. Hook, R. K. Agrawal and L. T. Drzal, "Effects of microwave processing on
ﬁber-matrix adhesion. II. Enhanced chemical bonding of epoxy to carbon ﬁbers ”, J.
Adhesion, Vol. 32, pp. 157-170, 1990

14. L. Xu, Mater Thesis, School of Mechanics and Materials Science, Washington State
University, 1998

15. L. T. Drzal, "T he Epoxy Interphase in Composites”, review chapter in Advances in
Polymer Science II, Z_5_, K. Dusek, ed., Spring-Verlag, pp. 16, 1985

16. P. C. Hiemenz, Polymer Chemistry: The Basic Concept, Dekker, New York, pp.
185-190, 1984

17. J, Esmaiel and RA. Nikolaos, “Polymer-polymer interdiffusion and adhesion”
J.M.S.-RREV. MACROMOL.CHEM. PHYS., C34(2), pp.205-241,1994

132

 

 

CHAPTER 7

CURE VOLUME SHRINKAGE OF VINYL
ESTER RESINS AND THEIR INFLUENCE ON
ADHESION BETWEEN CARBON FIBERS AND
VINYL ESTER MARIX RESINS

Composites of carbon ﬁbers and vinyl ester polymers possess unexpected low

mechanical properties due to low ﬁber-matrix adhesion “'31.

Vinyl ester resin can
undergo as much as 10% volume shrinkage with cure while typical epoxy system
undergoes only 3-4% shrinkage during cure [4'6]. The cure volume shrinkage could have
induced signiﬁcant stresses in the ﬁber/matrix interphase already before composite
loading. This could also be an important factor that lowers the adhesion between carbon
ﬁber and vinyl ester resin. In this study, the inﬂuence of the matrix cure volume
shrinkage on the adhesion between carbon ﬁber and vinyl ester resin was investigated.
Cure volume shrinkage of vinyl ester resins were measured with a dilatometer. Adhesion
was evaluated as interfacial shear strength (IFSS) measured with micro-indentation
experiments. Finite element analyses were used to simulate the stress at the interphase
and matrix and how it would change with matrix shrinkage. It was found that cure
volume shrinkage of vinyl ester could introduce a residual interfacial radial tensile stress
which would decrease the adhesion between ﬁber and matrix. The greater the shrinkage

the more signiﬁcant the reduction in adhesion. The cure volume shrinkage was

determined by the molecular weight of the vinyl ester monomer and styrene content and

133

 

 

 

also is related to the cure process and catalysts for polymerization. It was also found that
an epoxy sizing applied to the ﬁber, which swelled as a result of exposure to styrene
could counteract the cure volume shrinkage of the matrix. The results from ﬁnite
element analyses were consistent with the experimental results that larger shrinkage

produced a higher von Mises effective stress.

7.1 INTRODUCTION

Free radical cured thermosetting vinyl ester resins, typically containing 35~50% of
styrene monomer as reactive diluents, have superior toughness and chemical resistance
compared to unsaturated polyester. But composites of carbon ﬁbers and vinyl ester
polymers possess unexpected low mechanical properties due to low ﬁber-matrix adhesion
[1’3]. Upon cure, the volume shrinkage of vinyl ester resin is generally lower than that of
unsaturated polyester resin but higher than that of epoxy resin. Vinyl ester resin can
undergo as much as 5-10% volume shrinkage with cure depending on the molecular
weight of the vinyl ester monomer and the content of styrene (see Table 2.3). The cure
volume shrinkage can induce signiﬁcant residual stresses in the ﬁber/matrix interphase
before loading. This could be the one of the most important factors that lowers the
adhesion between carbon ﬁber and vinyl ester resin. As discussed in CHAPTER 5, the
application of a lightly cross-linked Diglycidyl Ether of Bisphenol A (DGEBA) cured
with Aliphatic Polyether Triamine (Jeffamine T-403) epoxy sizing to the carbon ﬁber

surface creates a beneﬁcial interphase between the carbon ﬁber and vinyl ester resin

matrix resulting in a substantial increase in ﬁber-matrix adhesion and the sizing has a

134

 

optimum thickness [2‘3’7‘8].

Vinyl ester resin can undergo as much as 10% volume
shrinkage with cure while typical epoxy systems undergo only 3-4% shrinkage during
cure [4'6]. One of the strategies of this research is to use the low shrinkage DGEBA-T403
carbon ﬁber sizing to isolate the carbon ﬁbers from the vinyl ester resin to observe the
inﬂuence of the cure volume shrinkage on the IFSS values measured by micro-
indentation. The other strategy is to ﬁnd of the inﬂuence of matrix resin cure volume
shrinkage on the ﬁber/matrix adhesion by comparing the IFSS data of different vinyl
ester matrices of different volume shrinkage. Most recently, Shanghai Fuchem Chemical
Company has developed a new-type epoxy vinyl ester resin which was claimed “free” of
cure volume shrinkage [9]. In summary, the goal of this study is to gain an understanding

of the matrix cure volume shrinkage on the adhesion between carbon ﬁber and vinyl ester

matrix resin.

7.2 EXPERIMENTAL MATERIALS AND METHODS

Experimental Materials: the ﬁbers used in this study are AS4 carbon ﬁbers from Hexcel,
Inc. The matrix resins are Derakane 411-C50 epoxy vinyl ester resin (D411-C50) from
Dow Chemical and Fuchem 891 epoxy vinyl ester resin (Fuchem891) from Shanghai
Fuchem Chemicals Co. respectively. CHP-5 (diluted cumene hydroperoxide) from Witco
Chemical and MEKP (methyl ethylketone peroxide) from Aldrich chemicals was used as
the initiator. Both CoNap and DMA from Aldrich Chemicals were used as promoters
and accelerators respectively. Diglycidyl ether of bisphenol A (DGEBA epoxy) and

Aliphatic Polyether Triamine (JEFFAMINE T403) were from Shell and Huntsman

135

 

 

respectively. Refer Figure 2.6, Figure 2.7 and Figure 5.2 for the molecular structures of
the epoxy vinyl ester resins, the sizing epoxy material and the cure catalysts. Molecular
structure of D411-C50 and Fuchem 891 are both bisphenol A epoxy based vinyl ester but
different molecular weights, about 900g/mol for D411-C50 and 1500~2000g/mol for
Fuchem 891 respectively. Both resin use styrene as a reactive diluent. The styrene
content of D41 1-C50 and F uchem891 are 50% and 35% respectively.

Fuchem 891 vinyl ester resin was developed by Shanghai Fuchen Chemicals Co as a
new-type of epoxy vinyl ester resin featuring low shrinkage together with an improved
elongation rate & impact resistance performance. Fuchem 891 vinyl ester resin is also a
mixture of DGEBA epoxy vinyl ester (refer Figure 2.7) and styrene. The styrene content
is about 35%. It was claimed that the result from tests shows that the pure vinyl ester
resin has a 0.015% linear shrinkage after curing in ambient temperature, 0.160% after 2
hours of post-curing at 80°C. Refer Appendix III for the relationship between linear
shrinkage and volumetric shrinkage. Considering D411-C50 vinyl ester resin has large
volume shrinkage with cure, Fuchem 891 was use in this research to quantify the
inﬂuence of the cure volume shrinkage to the adhesion between carbon ﬁber and vinyl
ester resin. The property comparison of the D411-C50 and Fuchem 891 are listed in

Table 7.1.

136

 

 

 

 

 

 

 

 

 

 

Properties D411-C50 Fuchem 891
DGEBA Vin | Ester Molecular Wei ht
Y (g/mol) 9 907 1500-2000
Styrene Content 50% 35%
Tension Modulus, E, (GPa)_ 3.38 3.12
Tension Strength, (MPa) 79.3 75
Elongation, % 5~6 4.0
Barcol Hardness 35 33
Density, pJg/ma’) 1.12
Heat Distortion Temp. 99~105°C 92i2°C

 

 

 

Table 7.1 Property comparison of D411-050 to Fuchem 891 vinyl ester

137

 

 

 

Pipette (smallest read 0.01ml)

Teﬂon faced screw cap
(Teﬂon seal pad inside)

 

Silicon Oil

Glass tube

Clams

Polymer resin with cure agents.

Stand

 

 

 

Figure 7.1 The dilatometer apparatus used to measure the volumetric
change with cure

138

 

Experimental Methods: 8 lab-made dilatometer was used to measure cure volume
shrinkage of vinyl ester resin from liquid resin to full cure. Dilatometry is a technique
that uses the volume decrease (density increase) that occurs upon polymerization to
follow that conversion time. Figure 7.1 depicts the simple dilatometer used in this study.
The design depends mainly on the expected volume change after resin curing. This is a
glass culture tube with a Teﬂon-faced screw cap equipped with a 2-ml pipette in which
the liquid (silicone oil) level can be read, the smallest read was 0.01ml. The screw cap
was drilled through, the hole diameter is almost the same as that of the pipette. A Teﬂon
sealant pad whose diameter is that same as the screw cap was drilled through with a much
small diameter so that it can vacuum seal the gap between the pipette and the screw cap.
Silicone oil which is very stable under 120°C is used as the indicator liquid.

A digitally controlled, programmable oven was used for specimen curing to make
sure all the samples were processed in the same way. The cure processes were the same
for all samples: room temperature for 1 hour, 90°C for 1 hour and 125°C for 1.5 hours.
Two types of recipe systems were used for the samples for the investigation of inﬂuence
of cure shrinkage on the adhesion between carbon ﬁber and vinyl ester resin. Two
different systems were used with different initiators. One is CHP-5 and the other one is
MEKP. See Figure 2.6 for the molecular structures. The CHP-5 cure system was based
on the cure recipe for D411-C50 recommended by the manufacturer and the MEKP cure
system is based on the cure recipe for Fuchem 891 recommended by the manufacturer,
shown in the gray columns of Table 8.2. For the same initiator, the corresponding
recipes, the white columns of Table 8.2, were calculated from the recommended recipes,

the gray columns, based on the concentration of the carbon-carbon double bond, C=C.

139

 

 

 

 

The mixture of DGEBA and J EFAMINE T-403 was prepared and held for one half
hour then added to acetone to form a 5wt% sizing solution. Fiber sizing was carried out
by using a pre-impregnation machine. Dynamic Mechanical Thermal Analysis (DMTA),
DMA 2980 from TA Instruments, and physical testing, was conducted to measure the
mechanical properties of the matrix resin. A MTS nano-indentation instrument was used
to measure the bulk moduli of the vinyl ester resins. Nano-scratch was also carried out
by the MTS nano-indentation machine to quantify the gradient of the modulus between
the sizing and the DGEBA epoxy and D411-C50 vinyl ester resin matrices. A RMC MT-
7 ultramicrotome was used for preparing the nano-scratch sample surface with a diamond
knife. Adhesion was evaluated as interfacial shear strength (IFSS) measured with a
micro-indentation system, Interfacial Testing System (ITS). Finite element analyses were
used to simulate the stress distribution around the ﬁber-sizing-matrix interphases during
volume shrinking of the matrix resin and the micro-indentation process as well.

Finite Element Analysis: consider the composite as an orthotropic material with
regularly spaced ﬁbers as shown in Figure 7.2. A single ﬁber is singled out for
simulation, and furthermore, only a quarter of this ﬁber is modeled because of symmetry.
The dashed line indicated the representative volume element (RVE) that is simulated and
is considered the ﬁber the micro-indenter would contact. This RVE is extended
downward a length of 20 times the ﬁber diameter in the Z direction to form a 3-
dimensional RVE. The two sides of this dashed line intersecting at the ﬁber center
represent symmetry planes for this repeating cell Figure 7.2. Also shown in this ﬁgure is

an interphase region between the ﬁber and matrix shown as a gray ring around the ﬁber.

140

 

 

 

 

 

 

 

 

 

 

Initiator Fuchem 891 D411-C50
CHP-5 1.40% 2.00%
CHP's CoNap 0.21% 0.30%
MEKP 2.00% 2.85%
MEKP CoNap 0.10% 0.14%

 

 

 

 

 

Table 7.2 The catalyst recipe for the investigation of the cure
volume shrinkage in vinyl ester matrices

141

A commercial ﬁnite element code, LS-DYNA “0], was used to model the

matrix/interphase shrinkage and indenter contact. The explicit aspect of the code was
used to model the matrix/interphase shrinkage followed by contact with the indenter. A
thermal excursion was used to model the matrix/interphase shrinkage. A linear elastic,
orthotropic, temperature dependent material was used. This material model is time-
independent. Standard orthotropic elastic constants are E, Eb, EC, vba, vca, vcb, Gab, Gbc,
and Gm. The therrnoelastic constitutive behavior is deﬁned by three, orthotropic

coefﬁcients of thermal expansion, ad, (1,, ac, that produce normal strain components

deﬁned by e, = aIAT , where T is the temperature. All of the therrnoelastic parameters

used are given in Table 7.3. Note the coefﬁcient of thermal expansion in the ﬁber
directions (b-direction) are set to zero. This was done so the surface the tup impacted
remained planar. When these coefﬁcients were not zero the matrix and interphase
contracted considerably relative to the ﬁber making a domed surface for indenting. For
the interphase two values are listed that represent the CTE for the 7.18% and 1.73%
shrinkage respectively. The higher CTE value is for the lower voltune shrinkage because

the temperature excursion was less in this case.

142

   

Symmetry—ﬂ
I
I

Symmetry

 

Figure 7.2 A transversely isotropic array of composite ﬁbers with the
representative volume element indicated by a dashed line [1.

143

 

 

 

 

 

 

 

 

 

 

 

 

 

Fiber (GPa) (GPa) (°K- 101‘)

E. 21 v... 0.256 G... 10.0 a. 8.5

E. 241 v... 0.33 G.. 10.0 a, 0

E. 21 11.. 0.256 G.. 8.3 a. 8.5
Interphase

E. 3.24 v... 0.356 G... 1.195 .1. 826/343

E. 3.24 v... 0.356 G... 1.195 a, 0

E. 3.24 1.. 0.356 G... 1.195 0.. 8.26843
Matrix

E. 3.38 v... 0.356 G... 1.246 a, 80

E. 3.38 v... 0.356 G.. 1.246 a, 0

E. 3.38 v... 0.356 G... 1.246 a. 80

 

 

 

 

 

 

 

 

Table 7.3 Themomechanical properties for ﬁnite element model

 

 

 

7.3 RESULTS AND DISCUSIONS

7.3.1 Cure Shrinkage of Vinyl Ester Matrices

The effectiveness of the lab-made dilatometer could be conﬁrmed by the data shown
in Figure 7.3. There were three of the lab-made dilatometers were set up at the same time
for the sample measurements. The Y-axis was the reading of the silicone oil level
directly obtained from the lab-made dilatometer. The X-axis was the time. Time 0 was
the time of completion of the adding of the curing agents to the liquid resin. The mixing
took about 2 minutes after the cure agents added in. The initial silicone oil levels of the
three lab-made dilatometers were different though those of the measurement 1 and
measurement 2 were very close. Since the difference in readings was the important
factor, the initial differences did not matter.

The results of the cure volume shrinkage measured by the lab-made dilatometer are
shown in Figure 7.4. The data shows that D411-C50 produces a volume shrinkage of
7~8% upon curing. Whereas the Fuchem 891 vinyl ester resin cured with the
recommended recipe, 2% of MEKP and 0.1% of CoNap, produced a very small cure
volume shrinkage, only 1.73%. (It is interesting that when Fuchem 891 vinyl ester resins
were cured with CHP-5 as initiator using a concentration the same as that of D411-C50
recommended by the manufacturer, the cure volume shrinkage is 5.85%, not a small

number!)

145

 

 

 

Pipette Read (ml)

—<> Measurementg1 + Measuremean + MekasuregmentB

 

0.5 .

 

 

 

 

500 1000 1500 2000 2500 3000 3500
Time (seconds)

Figure 7.3 The silicone oil level in the pipette of the lab-mad dilatometer

versus time

146

 

 

 

 

[0 Measurement1 - Measurement2 4 Measurement 3I
4.5%

 

 

 

3.5%

2.5%

1.5%

 

0.5%

Volume Shrinkage

 

 

e
#9
J”
-0.5% i

0 500 1000 1500 2000 2500 3000 3500 4000
Time (seconds)

 

Figure 7.4 The volume shrinkage observed from the lab-made
dilatometer according to time

147

 

F uchem 891 DGEBA epoxy vinyl ester resin was claimed by the manufacturer as
having a 0.015% linear shrinkage after curing in ambient temperature, 0.160% after 2
hours of post-curing. The relationship between linear shrinkage to volumetric shrinkage

can be described as following Equation 7.1 refer Appendix 111.

—-= —- 7.1

where AV/Vo is the volumetric shrinkage and AL/Lo is the linear shrinkage. 1.73% of
volumetric shrinkage equals to 0.54% of linear shrinkage which is much bigger than
0.015%, which was claimed by the manufacturer. The linear shrinkage was measured
with a stainless steel mold which was a semi-circle of 2.22cm of internal diameter and
2.86cm of external diameter and 25cm in length. The measurement technique was close
to ASTM D2566. Refer to A. T. Busschenll '1 in chapter 2, this kind of measurements has
shown an anisotropic cure shrinkage even for isotropic material depending on the
surroundings. Another set of tests was carried out by the manufacturer which the same
method was used but the samples were post cure at 125°C for 2hour. The result was
reported as: “0.35% linear shrinkage, and the volume shrinkage rate is 1%”.

This differences are most probably due to the test method used. The method using
semi-circle stainless steel mold measured the linear direction which is the most
constrained direction, in another word, the smallest shrinkage shown direction but most
signiﬁcant residual stress direction. While the lab-made dilatometer used in this study
measured the three dimensional shrinkage even through the tube still provides constraint

to the vinyl ester.

148

 

 

The cure volume shrinkage of vinyl ester could be changed not only depending on the
catalysts but also the cure process. Figure 7.5 shows that room temperature cured vinyl
esters resulted in lower cure volume shrinkages. But DMTA tests found that room
temperature cured vinyl ester was not ﬁrlly cured. Figure 7.6 illustrates a typical DMTA
measurement. A room temperature cured Fuchem 891 sample was cured with a
temperature cycle from room temperature to 160°C twice at the ramp of 6°C/minute (cure
recipe was recommended by the manufacturer, see Table2). In the second run of the
DMTA test, a much higher storage modulus and high glass transition temperature (Tg)
was detected indicating that full cure was not achieved in the ﬁrst DMTA run. Room

temperature cure is not sufﬁcient for this free radical polymerization.

149

10.0%

8.0%

2.0%

Cure Volume Shrinkage

0.0%

6.0% '

4.0% .

 

11 Fuchem-MEKP,
1:1 D411-CHP-5

7.18%

    

 

 

 

Room Temperature Cured 125 °C Cured

Figure 7.5 Inﬂuence of Cure Process to the Cure Volume Shrinkage of

Vinyl Ester

150

 

3500

 

 

3000
2500 9..
2000 g
1500 5

1000

Storage Moduli (MPa)

500 ‘

 

 

 

 

 

 

 

25 45 65 85 105 125 145

Temperature (°C)

Figure 7.6 One of the DMTA Measurement for Room Temperature Cured
Fuchem 891

151

7 .3.2 Inﬂuence of Cure Volume Shrinkage of Matrix Resin on the Adhesion between

Carbon Fiber and Vinyl Ester

The mechanical properties of matrix, which are very important for ITS interfacial
shear strength measurements, were measured by nano-indentation. During the cure
process, it was found that styrene vaporization was signifcant. The vaporization would
be different for samples made with open sample molds from these sample whichs were
made with less open molds and signiﬁcantly different from the samples made with the
totally closed lab-made dilatometer. To avoid the inﬂuence of styrene vaporization,
nano-indentation was used to measure the bulk mechanical properties of the material. It
was also found that the bulk elastic moduli measured by nano-indentation were
compatible to the tensile moduli measured. The inﬂuence of the cure volume shrinkage
of the vinyl ester matrix materials on their bulk moduli were shown in Figure 7.7. The
bulk moduli of the vinyl ester materials increased with the increase of the cure volume
shrinkage of the materials. It might indicate that the larger the cure shrinkage the more
compact morphology the polymer chains achieve resulting in an increase of the material

mechanical property.

Two sets of composite samples were tested by the ITS to ﬁnd out the inﬂuence of the
cure volume shrinkage of the matrix material on the adhesion between ﬁber and matrix.
One set of composite samples was the vinyl esters having different cure volume
shrinkage reinforced by AS4 carbon ﬁber ‘as received’ and the other set was the vinyl
esters having different cure volume shrinkage reinforced by AS4 carbon ﬁber with 5% of

DGEBA-Jeffamine T-403/Acetone sizing solution. The test results are shown in

152

 

 

 

1.27 _L--_.___-----

 

y = 9.47x2 + 0.16x + 3.83

 

 

 

 

 

g 1.26 R2=0.9739 W
9 .L- __-_.
=3,
8 1.25 I, W
2
16
(l)
.C
(D 1.24 9W
1.23
0% 2% 4% 6% 8% 10%

Cure Volume Shrinkage

Figure 7.7 Inﬂuence of vinyl ester cure volume shrinkage on the bulk moduli

153

 

 

 

 

U with AS4
. -° With5%DQ§BA-A$f1

(D
O

(D
O

I
I l

N
O
__.l_

Interfacial Sgear Strength
(M Pa)

 

 

50
1.0% 3.0% 5.0% 7.0% 9.0%

Cure Volume Shrinkage

 

Figure 7.8 Inﬂuence of vinyl ester cure volume shrinkage on the
Interfacial shear strength

154

 

Figure7.8. The ITS test results are very interesting, for the samples made with AS4
carbon ﬁber as received, the interfacial shear strength showed a progressive decrease
with the cure volume shrinkage of the matrix material, the larger the cure volume
shrinkage --- the more signiﬁcant the decrease of the interfacial shear strength. In
contrast, for the samples made with carbon ﬁber with DGEBA-T403 sizing, the value of
the interfacial shear strength underwent little change even for big cure volume shrinkage
of 8.2%. This suggested that the cure volume shrinkage would produce a residual stress
which had a negative effect on the adhesion between carbon ﬁber and vinyl ester resin.
And it was also suggested that the sizing material plays a very important role in

elimination of the residual stresses caused by the cure volume shrinkage.

7.3.3 Interphase between DGEBA Epoxy Sizing and Vinyl Ester of Different Volume

Shrinkage

Further investigation was carried out to ﬁnd the inﬂuence of the vinyl ester cure
volume shrinkage on the interphase between the DGEBA epoxy sizing. Figure 7.9 is the
nano-scratch trace. The scratch data which was averaged by scratch distance of 300nm is
shown in Figure 7.10 where for Sizing/D411-CHP interphase the D411-CHP vinyl ester
had a volumetric shrinkage of 7.18% and for Sizing/Fuchem-MEKP interphase the
Fuchem-MEKP had a volumetric shrinkage of 1.73%. Both data of Sizing/D411-CHP
interphase and Sizigg/Fuchem-MEKP interphase are off from the reference interphase
which is consisting to the C++ simulation. But the scratch data between Sizing/D411-

CHP interphase and Sizing/Fuchem-MEKP interphase are barely observable.

155

 

Figure 7.9 Nano-scratch trace for the interphase of DGEBA epoxy
sizing/Fuchem 891-MEKP

156

Coefﬁcient of Friction (COF)

 

0.65 _ ___ _____ 2-
—°— Sizing/D411-CHP

0.63 - « Reference Interphase
— C++ Simulation
061 ~ w -- Sizing/Fuchem-MEKP

0.59
0.57
0.55

 

0.53

 

0.51 .'.

 

0.49 i
20 21 22 23 24 25 26 27 28 29 3o

Scratch Distance (um)

 

Figure 7.10 Comparison of nano-scratch data for the interphase of DGEBA
epoxy sizing/Fuchem 891-MEKP and the interphase of DGEBA epoxy
sizing/D41 1-CHP

157

7.3.4 Finite Element Analysis of the Residual Stress Caused by Cure Volume

Shrinkage

The residual stress caused by cure volume shrinkage was studied with ﬁnite element
analysis. Two situations were selected for this study. One was the Derakane 411-C50
vinyl ester cured with CHP-5 which undergoes a volume shrinkage of 7.18% upon cure
and the other was Fuchem 891 vinyl ester cured with MEKP. As mentioned in the
experimental methods section, a thermal excursion was used to model the
matrix/interphase shrinkage. Once the temperature change associated with 1.73% and
7.18% matrix shrinkage was determined the indenter loading of the ﬁber was modeled.
The indenter was modeled by a spherical contact entity (*CONTACT_ENTITY) which
was given a prescribed displacement. A displacement function deﬁned in three,
piecewise linear portions was prescribed. Initially, while the temperature was decreasing
the indenter remains ﬁxed. Once the temperature has lowered to the required value, the
indenter moved in a bi-linear fashion. The indenter rate for the ﬁrst half of the loading is
reduced to avoid impacting the ﬁber too harshly.

The nature of the indenter displacement function may be seen by examining the
indenter force (Figure 7.11). Notice the difference between these two plots between
times 0.01 and 0.025 ms where the 7.18% case is loading. What is happening is the
contraction of the matrix, interphase and ﬁber is lengthening the RVE enough to contact
the tup before it starts to move. Ultimately, the two cases match up during the loading
phase of the ﬁber, so the early difference was ignored.

A good measure of the distortional strain energy (shearing energy) is the von Mises

effective stress. Figure 7.12 shows the von Mises stress for a matrix element adjacent to

158

 

 

 

 

1.2E+O7

 

1.0E+07 f
8.0E+06

6.0E+06 _

Force (pg)

4.0E+06

2.0E+06

 

0.0E+00
0.0E+OO

- 07.18% shrink
-— 1.73%. shrink

 

 

 

2.0E-02 4.05-02 6.0E-02
Time

Figure 7.11 Indenter force time history

159

8.0E-02

 

 

1.4E+05
1.2E+05
1.0E+05
8.0E+04
6.0E+04

4.0E+04

von Mises Stress

2.0E+04 »

 

+ 1 .i3%ﬁshriﬁr{k
- <>-7._18%shrink1

9.0.0-0-0-0’

 

 

 

0.0E+00 <2
0.0E+00

Figure 7.12 Von Mises stress for a point in the matrix at the interphase-matrix

2.0E-02 4.0E-02
Time

6.0E-02

8.0E-02

boundary located approximately 18 pm from the free surface.

160

 

 

 

1.2E+05

1.0E+05 ; ,0.

f"

‘0

‘-<>-.<>_.o 0
" ~-<>--o--<>--<>--<>--<>--<>--<>
8.0E+O4

 

60504 '- 557.18% shrink“.

4.0E+O4 ‘ l"'.'1,;73% shrink

 

2.0E+04

von Mises Stress (kPa)

 

 

0.0E+OO

 

o 5 1o 15 20 25 30
Distance from free surface (mm)

Figure 7.13 von Mises Stress Distribution at 5 g Indenter Load

161

 

 

 

the interphase and approximately 18 pm from the free surface in the cases of 1.73% and
7.18% volume shrinkage. (The maximum stress occurs in this vicinity.) Again the
nature of the prescribed tup displacement is apparent in the von Mises stress curves. The
stress evolutions from the tup loading are similar, but the volume shrinkage has these
starting at different value. Thus, the matrix shrinkage has a considerable inﬂuence on the

matrix stress.

7.4 MODELLING INTERPHASE STRENGTH

During this work, the ﬁber/matrix adhesion was evaluated as interfacial shear strength
(IFSS) measured by Interfacial Testing System (ITS) micro-indentation apparatus given
by a generalized empirical equation (ITS shear equation, equation 7.2), which is

embedded in the data reduction software of the apparatus.

G
IFSS = AL B —m — CLn[-:—) — E 7.2

02 Ef

 

 

Where:
IF SS=Interfacial Shear Strength (psi)
P=Maximum Load (g)
D=Fiber Diameter (um)
Gm=Matrix Shear Modulus (psi)
Ef=Fiber Tensile Modulus (psi)
d=Distance to Nearest Neighbor (um)
A, B, C, E =Constants;
A=1,81100 B=0.875696 C=0.018626 E=0.026496

162

 

 

 

Ideal
Interphase

_ ' i L ' " " f Chemical Bonds and/or
. l Strengthen Materials ~
L ‘ -w~-~—— =ﬁ=====

 

(D
O
0
g.
:7
01
a?
U
G)
E
s
23

 

 

 

 

O)

O
I I Y Y ‘I
l

 

 

\I
O
T
:6

Residual Stress

 

 

Interfacial Shear Strength
(MPa)

 

 

 

 

L - c
60 M51518)? + 79.9x + 73.9 ____ .
50 i

 

1.0% 3.0% 5.0% 7.0% 9.0%
Cure Volume Shrinkage

Figure 7.14 The interphase property presented by interfacial shear
strength(lFSS) plot

163

for details. Accounting for the ﬁber Er, matrix Gm, morphology of the in situ composites
dn/df, this equation ignores the ﬁber/matrix interphase. Actually the interphase property
was tested by the Load on Fiber for Debonding, fg and presented as part of the IFSS
value. See Figure 7.14, sample A, B, C, and D were all made from the AS4 carbon ﬁbers
and vinyl ester matrices of almost the same shear modulus. Because a very stringent
ﬁber selection criterion was applied to all the tests, differences between dn/df were very
small. The differences between the IFSS values should represent the differences of the
interphases. For example, comparing the IFSS of sample A to that of the sample B,
sample A has a stronger interphase which could be the result of the change in chemical
bonding and/or the material at the interphase; Comparing Sample D to sample C, sample
D has the stronger interphase which could be because of the change in chemical bonding
plus the relaxation of the residual stress; and sample A and sample D could have
interphase of same strength even they have different vinyl ester matrices.

Based on data in Figure 7.14, prediction of the interfacial shear strength of a
ﬁber/matrix system could be accomplished through use of equation 7.3:

l

G 2
IFSSoc(a—v%)2* ,6 m —/iLn(-g-]-5 +O'CB 7,3

Ef

 

 

164

Where:

lFSS=Interfacial Shear Strength

a=adj ust parameter related to cure volume shrinkage
v%=percentage of cure volume shrinkage
Gm=Matrix Shear Modulus

Ef=Fiber Tensile Modulus

D=Fiber Diameter

d=Distance to Nearest Neighbor

0'c13=Strength provided by additional chemical bonding

,6, 2. and 5 are Constants;

165

7.5 CONCLUSIONS

1.

Cure volume shrinkage of vinyl ester resins are generally determined by the
average molecular weight of the vinyl ester monomer and the content of styrene
monomer. Normally larger average molecular weight of the vinyl ester monomer
and lower the content of styrene monomer result in small cure volume shrinkage
as noted in the literature”). The shrinkages were also related to the catalyst of the
free radical polymerization and time-temperature schedule of the process.
Different cure temperature and catalysts produce different heats of reaction, cross-

link density and degrees of micro-gel formation and/or micro-phase-separation “2'

14].

The cure volume shrinkage of vinyl ester could introduce residual interfacial
tensile stress which would decrease interfacial shear strength (IFSS) values
suggesting a negative effect on the adhesion between ﬁber and matrix. The
greater the shrinkage the more the adhesion is reduced. The bulk moduli of the
vinyl ester resin measured by nano-indentation increase as the cure volume

shrinkages increase resulting in an increase of the material mechanical properties.

The epoxy sizing on the ﬁber surface is very important for the relaxation of
thermal residual stress caused by cure volume shrinkage. Epoxy-sizing swelling

could counteract the cure volume shrinkage of the matrix.

Finite element analyses was very useful for analyzing the interfacial shear stress

distribution during matrix shrinkage. The results from ﬁnite element analyses

166

were consistent with the experimental results that larger shrinkage produces

higher von Mises effective stress.

. DGEBA epoxy sizing provides a beneﬁcial interphase between the carbon ﬁber
and vinyl ester resin. First of all it provided the ability to form chemical bonds
with the composite matrix resin and the surface of carbon ﬁber reinforcement.
Second, the epoxy vinyl ester was made from DGEBA epoxy which should be
compatible to the DGEBA sizing material. The small styrene monomers can
interdiffuse into the slightly cross-linked epoxy sizing material before the matrix
is fully cured. The swelling of the sizing with styrene diffusion may be critical for
counteracting the matrix volume shrinkage and further more eliminate the residual
stress. The optimum sizing thickness is 90~100nm. Nano-indentation test and
nano-indentation scratch test are useful techniques for measurement of interphase

proﬁle.

. Modiﬁed interphase strength could be qualitatively predicted. By comparing the
ITS interfacial shear strength (IFSS) values, more accurate interphase properties

can be obtained.

167

7.5 REFERENCES

10.

11.

12.

. D. Hoyns. and L. T. Drzal, “Hydrothermo-Mechanical Response and Surface

Treated and Sized Carbon F iber/ Vinyl Ester Composites”, ASTM Symposium
on the Fiber Matrix and Interface Properties, 1994

S. M. Corbin, MS. Thesis, Department of Chemical Engineering, Michigan State
University, 1998

Rich, M., Corbin, S., and Drzal, L., “Adhesion of Carbon Fibers to Vinyl Ester
Matrices”, Proceedings of the 12'" International Conference of Composite
Materials, Paris, ICCM12, 1999

M. B. Launikitis, “Vinyl Ester Resin”, Hand Book of Composite, pp.38-49

L. S. Penn and T. T. Chiao, “Epoxy Resins”, Hand Book of Composite,
pp.57~88

L. Xu and L. T. Drzal, “Inﬂuence 0f Matrix Cure Volume Shrinkage on the
Adhesion between Vinyl Ester and Cardon Fiber”, Proceedings of the 2003
Adhesion Society, pp.415-417, 2003

Xu, L. and Drzal, L., “Adhesion Improvement between Vinyl Ester and Carbon
Fiber”, Proceedings of the 13"I International Conference of Composite
Materials, Beijing, ICCM13, 2001.

Xu, L., Drzal, L., Al—Ostas, A., and Schalek, R., “Improvement 0f Adhesion
Between Vinyl Ester Resin and Carbon Fibers by A Controlled and Designed
Interphase”, Proceeding of the 2002 Adhesion Society/WCARP-II Meeting, 2002

http://www.fuchem.com

Livermore Software Technology Corporation, LS—DYNA Keyword User ’s
Manual, Version 940, 1997

A. Ten Busschen, Ph.D thesis, Delft University of Technology, the Netherlands,
1996

Weitzsacker, L., Xie, M. and L. T. Drzal, “Using XPS to Investigate F iber/Matrix

Chemical Interactions in Carbon Fiber Reinforced Composites”, Surf. Interf.
Anal. 2_5, 53-63, 1997

168

 

13. R. P. Brill and G. R. Palmese, “An investigation of vinyl-ester—styrene

copolymerization cure kinetics using Fourier Transform Inﬁared Spectrosc0py”,
J. App. Poly. Sci, Vol. 76, pp1573-1582, 2000

14. C. P. Hsu and L. J Lee, “Free-radical crosslinking copolymerization of

styrene/unsaturated polyester resins: 3. Kinetics-gelation mechanism” Polymer,
Vol. 34, pp. 4516-4523, 1993

 

169

 

CHAPTER 8

CONCLUSIONS AND RECOMMENDATIONS

8.1 CONCLUSIONS

Derakane 411-C50 vinyl ester resin is a mixture of vinyl ester and styrene monomer
which polymerizes by a free radical polymerization. The interface between this resin and
AS4 carbon ﬁber is much more complex than that between epoxy and AS4 carbon ﬁber.
From this study, several conclusions can be made.

Application of a slightly cross-linked epoxy-amine polymer on the carbon ﬁber
surface is found as an effective way to increase the adhesion between carbon ﬁber and
vinyl ester resin. The sizing thickness on the carbon ﬁber surface can be controlled by
the concentration of sizing solution. The optimum sizing thickness is 90~100nm that
brought a substantial increase in the value of interfacial shear strength (IFSS) measured
by micro-indentation more than 30%. Finite element analysis result was compatible with
the micro-indentation result. The maximum value of IFSS happens when the sizing
thickness is about 90~100nm. At thicknesses greater than this value, no improvement is
detected because of the low modulus of the epoxy sizing. Inter-diffusion of the styrene
into the epoxy sizing creates an interphase which is stronger than the epoxy sizing

material itself.

170

DGEBA epoxy sizing provides a beneﬁcial interphase between the carbon ﬁber and
vinyl ester resin. First of all it provided much more possibilities to form chemical
bonding to the surface of carbon ﬁber reinforcement surface. Second, the epoxy vinyl
ester was made from DGEBA epoxy which should be compatible to the DGEBA sizing
material. Third styrene monomers are very small which could penetrate the slightly
cross-linked epoxy sizing material. Nano-scratch measurement found that the inter-
diffusion distance could be up to 1.5 microns. Last but not the least, the epoxy sizing on
the ﬁber surface is very important for the relaxation of thermal residual stress caused by
cure volume shrinkage.

Cure volume shrinkage of vinyl ester resins are generally determined by the average
molecular weight of the vinyl ester monomer and the content of styrene monomer.
Normally larger average molecular weight of the vinyl ester monomer and lower the
content of styrene monomer result in small cure volume shrinkage. The shrinkages were
also related to the catalysts of the free radical polymerization and time-temperature
schedule of the process. Different cure temperature and catalysts of produce different
heats of reaction, cross-link density and degrees of micro-gel formation and/or micro-
phase-separation resulting in different cure volume shrinkage. The cure volume
shrinkage of vinyl ester could introduce residual interfacial tensile stress which would
decrease interfacial shear strength (IFSS) values The greater the shrinkage the more
signiﬁcant the effect. While the bulk moduli of the vinyl ester resin measured by nano-
indentation increase as the cure volume shrinkages increase, the material mechanical
properties increase as well. Finite element analyses was very useful for analyzing the

interfacial shear stress distribution during matrix shrink. The results from ﬁnite element

171

analyses were consistent with the experimental results that larger shrinkage produced a
higher von Mises effective stress.

Preferential adsorption of the components of the vinyl ester system, the promoter
(CoNap), the accelerator (DMA), or the styrene monomer does not make a major
difference on the interfacial shear strength value measured. However, changes in the
initiator and the concentration of initiator (CHP-5) cause changes in the polymerization
of the vinyl ester system. Optimum ﬁber-matrix adhesion could be obtained by properly
select the initiator and adjusting the amount of initiator. The contributions of styrene and
vinyl ester in D411-C50 to the adhesion between the ﬁber and matrix are different
especially for ﬁbers with the lightly cross-linked DGEBA-T403 sizing. The small
styrene monomers diffuse into the DGEBA-T403 sizing material and provide a
strengthened interphase.

Modiﬁed interphase strength could be qualitatively predicted. By comparing the ITS
interfacial shear strength (IFSS) values, more accurate interphase properties can be

obtained.

8.2 RECOMMENDATIONS

While most of the experimental work presented in this study is completed, the
mechanical properties of the composite material with engineered interphase are not
provided here. Though using three point ﬂexural tests increased strength were found for
composites with an engineered interphase and composites with low shrinkage vinyl ester.
The error in these measurements was high as a result of deﬁcient fabrication conditions.

Once these sources of error are brought under control, (e.g. ﬁber volume percentage, the

172

control of the ﬁber distribution, and etc.), more accurate mechanical properties of the

composites could be collected in the future.

Nano-indentation and nano-scratch techniques were introduced to proﬁle the
interphase gradient between epoxy sizing and vinyl ester resins. The effects of tip size,
both for indentation and scratch were found. Especially for nano-scratch, the effect of the
indenter size were accurately simulated by a C++ program. The inter-difusion between
polymers was measured though it was only a range at this point. For example, the inter-
diffusion distance between vinyl ester and slightly cross-linked epoxy in this study would

be 391 nm~1 521nm. According to Einstein’s equation,

[2:201

where l is the inter-diffusion distance, D is inter-diffusion coefﬁcient and t is the
diffusion time. Assume diffusion time r—‘geltime=20mins=12005econds, D=(12/2t), D
would be 6.37x10"3~9.64x10"2cm2/s. According to literature the inter-diffusion
coefﬁcient, D, for low molecular weight polymers and oligomers could be in the range of
10‘5~10'10cm2/s “3 I] while normal polymer could be in the range of 10'8~10"7cm2/s “3”.
Although the inter-diffusion could be proﬁled by the nano-scratch technique, the large
range of values indicated further improvement is required. The scratch size of the
Berkovich scratch indenter would be smaller if a smaller load was applied but the smaller
the load the more noisy the data. More work needs to be done until an accurate

measurement of the inter-diffusion distance can be measured.

The modeling of the interphase strength presented in CHAPTER 7.4 is a ﬁrst step to

incorporate interphase properties, such as chemical bonding and residual stress at the

173

interphase, into the interfacial shear strength. To solve the whole puzzle, more
generalized empirical data for samples of various ﬁber and matrix combinations are

required.

 

174

 

Appendix I

EVALUATION OF FIBER/MATRIX ADHESION

The interphase is where the ﬁber and matrix bonded together. The bond strength
often refers to ﬁber/matrix adhesion. For ﬁbers without sizing, the contributions to the
adhesion could be chemical reaction or interdiffusion of element between ﬁber and
matrix, or van der Waals force, electrostatic attraction, acid-base interaction including H-
bond, or mechanical interlock. For ﬁber with organic coating the situation is rather
complex. Beside the interactions mentioned above, the entanglement between the
molecules of the sizing and matrix and the interdiffusion “”1 between compatible sizing
and matrix polymers could be one of the important contribution to the adhesion. On the
other side, voids, contaminates, etc., and thermal residual stress could be negative
contribution to the adhesion.

Direct measurement of the interphase strength, normally refers to adhesion, is very
difﬁcult. Therefore, the level of adhesion is assessed practically by the interphase
strength value, the interfacial shear strength (IF SS), obtained from destructive mechanical
tests. These tests could be classiﬁed into three groups, ﬁber pullout test, single ﬁber
fragmentation test and microindentation test. The microindentation test has attracted
much attention because it is an in situ testing method conducted on a real composite, thus
allowing for evaluation of the processing or environmental exposure encountered either

during manufacturing or in service. The interfacial shear strength value obtained by these

175

 

 

tests for a given system is consistent with the mechanical properties of macrocomposites
[5'8]. A strong correlation has also been found between the interfacial shear strength and

[9.11]

such thermodynamic parameters as surface free energy or speciﬁc enthalpy of

“”21. Therefore, it seems to be possible to obtain information about ﬁber-

adsorption
matrix adhesion from micromechanical experiments. However, there is a problem
concerned with correct treatment of experimental data. The attempts made to compare

bond strength value obtained by different micromechanical techniques even for the same

polymer-ﬁber pair turned out to be unsatisfactory because the large scatter in the results.

I-l From Stress Transfer Point of View

For ﬁber reinforced polymeric composites, a high-modulus ﬁber embedded in a low
modulus matrix. As mentioned in Chapter 2, the matrix holds the ﬁber together and
transmits the applied load to the ﬁber through a “layer” between ﬁber and matrix—the
interphase. The topic of load transfer from the matrix to the ﬁber has been treated by a
number of researchers. These range from simpliﬁed physical model such as Kelly-Tyson
model ['3] to a number of models in terms of thermo-mechanical properties and
microfailure mechanics.

a) Kelly-Tyson Equation “3"61

Assumptions

1. Fiber is linear elastic and isotropic surrounded by rigid plastic isotropic matrix.

176

 

2. Perfect bonding between ﬁbers and matrix.
3. The transfer of tensile stress from matrix to ﬁber by means of interfacial shear stress
ti which is a constant equal to the matrix yield stress in shear.

4. Each ﬁber has two ends which are free from tensile stress.

Fig. 1.1 Variation of interfacial shear stress 1 (b) and ﬁber normal stress 0';
(c) with distance along the ﬁber of a representative volume element (RVE)
(a) according to Kelly—Tyson model "6'

 

 

 

 

 

 

 

 

 

(b) *—

 

 

 

 

 

O' ______________
(C) fmax

 

 

A simple idea of force balance gives the solution of interfacial shear stress, see Fig.

1.1, as:

7rdL zﬂdz Ufmax
C

2' 1.1
y 4 LC/2

 

So,

177

T = —— 1.2

where 2LC is the “critical transfer length” and Gym, is the maximum stress in the

ﬁber center. To achieve the maximum stress in the ﬁber center, the ﬁber length should at

least not be shorter than the critical transfer length.

b) Shear Lag Model and Modiﬁed Shear Lag Model “6"“

Assumptions

1. Both ﬁber and matrix are linear elastic and isotropic.

2. Perfect bonding between ﬁbers and matrix, the transfer of tensile stress from matrix
to ﬁber by means of interfacial shear stress.

3. The lateral stiffness of the ﬁber and matrix are the same. In other words, identical
axial strain in the ﬁber and matrix.

4. Each ﬁber has two ends which are free from tensile stress.

Because both ﬁber and matrix are linear elastic, so that the interfacial shear stress is
proportional to the different between it and v, where u is the axial displace ment at a point
the same in the ﬁber and v is the axial displacement the matrix would have at the same
point in the represented volume element (RVE) with no ﬁber present. The rate of change

of the ﬁber axial load P is given by,

178

—:Hu—V [.3
dx ( )

where H is a propotioality constant to be determined from geometrical and material

property data. Differentiating Eq. 1.3, then

 

2
dx dx AfEf

 

. du
where the expressmn — =

d is taken from elementary mechanics of materials
it .

A}. f
m

dv . . . . .
and dx- =8 18 the matrix strain With no ﬁber present. Solve this second order

differential equation, the interfacial shear stress I at point x is

}_.

Gm sinh ,6(0.5L — x)

D) cosh ,60.5L
/

1.5

 

 

T=Ef8m

 

 

where Ef‘ is the tensile modulus of the ﬁber, Gm is the shear modulus of the matrix,

2mm
AfEf ln(D/d)'

 

andﬂ2 =

179

The solution can be decried by the following ﬁgure, Fig. 1.2.

Fig. 1.2 Variation of interfacial shear stress Ti (b) and ﬁber normal

stress or Z(c) with distance along the ﬁber of a RVE (a) according to
Cox’s shear lag model "6'

 

 

 

 

 

 

Ti Lc i
a» k *——-:

 

 

 

()‘fZ .......

W.

Dow modiﬁed Cox’s model, assuming that the matrix axial displacement is not

 

constant as opposed to the original assumption of Cox. Rosen further reﬁned the model
by considering that the matrix encapsulating the ﬁber is surrounded by a material having
the average properties of the composites. Using the same idea employed by Rosen,
Whitney and Drzal “71 also proposed a two-dimensional thermo-mechanics model. The
signiﬁcance of Rosen’s work lies in the attempt of quantifying the efﬁciency of the stress
transfer across the interface with respect to the ﬁber length by introducing the concept of
“ineffective length”. While the great wok done by Drzal et a1. is to use Weibull modulus

to evaluate the statistic nature of the critical transfer length.

c) Stress Analysis Model “7]

180

Whitney and Drzal also proposed a two-dimensional thermo-mechanics model of
stress transfer based on superposition of solutions for two axisymmetric problems of the
exact far ﬁeld solution and the approximate local transient solution. The model is based

on a single broken ﬁber sorrounded by an unbounded matrix as Fig. 1.3.

Fig. 1.3 Stress analysis model "7'

 

 

Assumptions

1. Both ﬁber and matrix are linear elastic and the matrix is isotropic while the ﬁber is
transversely isotropic.

2. Perfect bonding between ﬁbers and matrix, the transfer of tensile stress from matrix
to ﬁber by means of interfacial shear stress.

3. Axisymmetric behaviors for both ﬁber and matrix.

4. Each ﬁber has two ends which are free from tensile stress.

The model is based on superposition of the solutions to two axisymmetric problems,

an exact far ﬁeld solution and an approximate transient solution. Axisymmetric behavior

181

is assume with the origin of an x, r axis system at the broken end of the ﬁber (Fig. 1.3).
The results are summarized here. The axial normal stress, 03,, in the ﬁber is independent

of d and is of form

a, = [1 — (4.75; +1)e‘4-75x]A150 1.6

where 2 =x/Lc, so is the far ﬁeld strain, and A1 is a constant dependent on material
properties, the expension strain and the far ﬁeld axial strain. The interfacial shear stress

is given by the relationship

 

 

1,, = —4.75}p4150 )e‘4~75x 1.7
where
G
,u = , m 1.8
Elf - 4V12me

I-2 From Physical-Chemical Interactions Point of View

The modern view of the origin of intermolecular forces is the presence of both polar
and non-polar interactions between different molecules. The non-polar one is well
accepted as London dispersion force and Fowkes and co-workers “8'2” have led the
argument that polar forces (which include hydrogen bonds) are actually Lewis acid-base

interactions. So

182

where W is the energy of adhesion and superscript D and ab denotes dispersion and acid-

base interaction respectively.
a) Schultz’s Tentative Model

A tentative model correlating the total work of adhesion with the interfacial shear
stress was proposed by Schultz et al. First, a linear relationship between interfacial shear
strength 1: and the thermodynamic work of adhesion, WA was found for poorly polar

k [24,25]

matrix ﬁber reinforced composites from experimental wor . Then for more polar

matrixes, a linear relationship between interfacial shear strength I measured by

fragmentation and the speciﬁc interaction parameter, A, was found [26’2"

. The speciﬁc
interaction parameter A can be considered as being a speciﬁc enthalpy of Lewis acid-
base interaction, -AHab, at the ﬁber-matrix interface. And according to Fowkes, the work

of adhesion can be deﬁned in the follow way [28]:

WA=WD+Wab

l 1.10
=2(y?y£>2 —fAH"”n"”

where 7 refers to surface free energy, subscribe f and m denotes ﬁber and matrix
respectively, superscript D corresponds to dispersion interaction, f is a correction factor
(near unity) to transform enthalpy value to free energy values, and nab is the number of

acid-base site per unite surface area of interface which cannot be directly measured.

183

After applying equation 1.10 with an estimated value of nab: 6x10'6 mol/m2 [28] to the
result of linear relationship between T and A, a linear relationship between t and WA

came out again. It is therefore possible to write:

T = (IWA 1.1 l
-1.
E f 2
The previous work “03' gave a relation:— = — . Comparing with Kelly-
c Em
0f
Tyson equation. I y = — —‘— , Schultz and et al approximate that:
c
E I
2
m

where Ef is the tensile modulus of the ﬁber, Gm is the shear modulus of the matrix.

b) Gutowski’s Theoretical Model

The work of adhesion, WA, equals to the work required to separate the two materials

at their interface is given by Dupre equation:

WA=YI +Y2'Y12 1.13

184

Where y is the surface free energy, subscribe 1, 2 and 12 refers to materials 1,

materials 2 and interface respectively. According to Fowkes, W A 2W” +W"” and

l
D
WD=2(r1Drz )2,so

l
ab D D —
W =71+72_712‘2(71 72 )2 1.14
Then use Schultz’s semi-experimental expression [29] for acid-base interaction,

3 — 1 ) 1.15

71/72 w12(71/72)2

 

 

-U12 =71(

Where —U12 is the energy of interaction which is equal to the work of adhesion WA

here and w1=(yc(1)/y.)m . The free energy of any material is given by [30],

7— 2 2 e
327! ro

Gutowski began with the interaction force, F12, between two materials. F 12 =-dU12/dr,
put equation [.14 into equation 1.15 and differential, then put equation 1.16 back get the
relationship between the interaction force and the surface free energies if the materials as

below:

185

 

3 _ 2
(71/72)2 #11201 W)3

 

F12 0C2’1 1.17

l-3 Measurement of the Interfacial Shear Strength [3 "40'

A number of testing methods was developed for measuring ﬁber-matrix adhesion
using single or groups of ﬁbers. For each method, ﬁrst of all, a mathematics relation was
set up. Most of them were based on the theoretical consideration of micromechanics
models discussed in section H, the Kelly-Tyson model and/or shear lag model or other
model extended there from. Then a fracture mode is necessary for a failure criterion.
Among the three typical failure modes shown at Fig. 1.4, the interphase fracture is
normally considered as a combination of mode I and mode 11. Even stress intensity factor
criterion (ch, the critical stress intensity factor for mode I, Kuc is for mode 11 and so on)
and strain energy release rate criterion (Glc, the critical strain energy release rate for mode
I, GHC is for mode 11 and so on) in general deals with a more fundamental aspect of the
interface debond problem, the tensile fracture strength criterion or the maximum shear
stress criterion has an important advantage in that the interfacial shear strength, whether
for bonded, partially debonded or debonded region, can be directly determined from the

experimental results of the tests.

186

FigI.4 Fracture Mechanics

@ \%

Mode I, ch, GIC Mode II, Knc, Guc Mode III, ch, Gmc

Cleavage Shear Torsion

a) Fiber Pull-out Technique

In the ﬁber pull out experiments, Fig.1.5, one end of a single ﬁber is embedded in a
block of matrix. The free end is gripped and an increasing load is applied as the ﬁber is

pulled out of the matrix while the load and displacement are measured.

Fig. 1.5 Schematic diagram of the single-ﬁber pull-out method for
measuring ﬁber-matrix adhesion “07'

id I I d d
"- [ﬁll—>1: _ 'i-,; , . F

-1.T 4LT LT

(1 ) (2) (3)
I: Matrix m Fiber

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

187

From each load and displacement curve the force at debonding , F, and the embedded
length, L, were derived and the interfacial shear stress, I, is calculated using the simple

Kelly-Tyson Equation, equation 1.18.

F

2':—
did.

1.18

The interfacial shear stress, I, measured here reﬂects a combination of adhesion and

friction due to coulomb force (Coulomb force also can be considered as one of the

interaction force contributed to adhesion).

b) Single Fiber Fragmentation Technique

Fig. 1.6 Schematic diagram of the single-fiber fragrﬂgptation

method for measuring ﬁber-matrix adhesion

q

   

El Matrix E: Fiber

In single ﬁber fragmentation test, Fig.1.6, the ﬁber is totally encapsulated in a matrix
coupon, a tensile load is applied to the coupon. The interfacial shear stress transfers the
coupon tensile to the encapsulated ﬁber through the interface. As the load is increased on
the specimen, shear forces are transmitted to the ﬁber tensile. The ﬁber tensile stress

increases to the point where the fracture strength is exceed and the ﬁber breaks inside the

188

matrix. This fragmentation process is repeated as the sample strain is increased
producing shorter and shorter ﬁber fragments within the coupon until the remaining
fragment lengths are no longer sufﬁcient in size to produce further fracture through this
stress-transfer mechanism. The ﬁnal ﬁber fragment length, also called critical length, is
measured. The interfacial shear stress can be calculated based on the ﬁnal ﬁber fragment

length using either Kelly-Tyson model or shear lag model.

In practice, there is a distribution of the ﬁnal ﬁber fragment lengths. Based on Kelly-

Tyson model, Drzal et al use Weibull statistic to ﬁt the data as shown

I = (2)1(1— i) 1.19
2,6 a

where or and B are the scale and shape parameters from the Weibull distribution, F is

the gamma function.

189

c) Micro-indentation

Fig. 1.7 Schematic diagram of the microindentation method for
measuring ﬁber-matrix adhesion '38

 

 

 

ll

m

Illllll"lii' . 4)
LEE: p

i“

:Mi

 

 

 

 

 

 

(1) (2) , (3)
l:l Matrix Fiber

The microindentation technique (or push-out test as opposed to the pull-out test) is a
single ﬁber test capable of examining ﬁbers embedded in the actual composite. It can
determine the interphase strength due to fatigue or environmental exposure, or possibly
monitor the interphase properties of parts in service. The test utilizes a microhardness
indenter with various tip shape and sizes to apply a compressive force to push against a
ﬁber end into metallographically polished surface of a matrix block. In the ﬁrst
approach, Fig. 1.7 (1), a load is applied continuously to compress the ﬁber into the
specimen surface. The embedded ﬁber length and the specimen thickness are delicately
controlled. In the second approach, Fig. 1.7 (2), very thin slice specimens of known
embedded ﬁber lengths are employed to distinguish debonding and post debonding
frictional push-out in a continuous loading. This technique has become most popular in
recent years among various of specimen geometry and loading method. In the third

version of test, Fig. 1.7 (3), a selected ﬁber is loaded using spherical indenters in steps of

190

 

increasing force, and the interface bonding is monitored microscopically between steps,
until debonding is observed.

The interfacial shear stress is calculated from ”0‘4”:

Tmax
Tdeb = 0 adj 1.20

0"” FEM

 

where tdcbis the maximum shear stress for debonding, Cadj is the adjusted compressive
stress applied to the ﬁber end at debonding and capp is the applied stress, (rmax/oapp)FEM is

the ratio of the maximum interface shear stress to the applied stress from the linear

axisymmetric ﬁnite element.

The ﬁnite element result are compared with those obtained from an analytical
solution of the Hertz contact problem of a point load on a half-space with imaginary

boundaries. Good agreement was found:

2
{:3 in [1) 1.21

191

where Ef is the tensile modulus of the ﬁber, Gm is the shear modulus of the matrix, I

is the interfacial shear stress, 0 is the stress applied to the ﬁber end, (1 is the diameter of

the ﬁber and Tm is the distance to the nearest neighbor.

REFERENCES

10

'9)

H. Ho and L. T. Drzal, “Non-linear Numerical Study of the Single-Fiber
Fragmentation Test, Part 1: Test Mechanics”, Composites Engineering, _5, No. 10-
11, 1231-1244 (1995).

H. Ho and L. T. Drzal, “Non-linear Numerical Study of the Single-Fiber

Fragmentation Test, Part II: A Parametric Study”, Composites Engineering, 5

_9

No. 10-11, 1245-1259 (1995).

H. Ho and L. T. Drzal, “Evaluation of Interfacial Mechanical Properties of Fiber
Reinforced Composites Using the Microindentation Method ”, accepted for
publication in Composites, Part A (1998)

J. Rich, S. Corbin, and L. T. Drzal, “Adhesion of Carbon Fibers to Vinyl Ester
Matrices”, Proceedings of International Conference on Composite Materials,
ICCM-12, Paris, France, July 5-10, 1999.

Hoecker, K. Friedrich, H. Blumberg and J. Karger-Kocsis, Comp. Sci. Tech., Vol.54,
pp.317, 1995

H. D. Wagner, H. E. Gallis and E. Wiesel, J Mater. Sci., Vol.28, pp.2238, 1993

G. S. Sheu and S. S. Shyu, J. Adhesion Sci. Tech., Vol. 8, pp.1027, 1994

192

8.

10.

l6.

17.

18.

19.

H. J. Jacobasch, K. Gruncke, P. Uhlmann, F. Simon and E. Mader, Composite
Interfaces, Vol. 3, pp. 293, 1996

H. Simon, Ph.D thesis, University de Haute Alsace, Mulhouse, France, 1984

J. Schultz, L. Lavielle and H. Simon, “Science and new applications of carbon

ﬁbers”, Proc. Tntern. Symp., pp. 125, 1984

. J. Schultz and L. Lavielle and C. Martin, J. Adhesion, Vol. 23, pp. 45, 1987

. J. Schultz and L. Lavielle, “Inverse Gas Chromatography”, Amer. Chem. Soc.

Symp. Ser., Vol391, pp.185, 1989

. H. L. Cox, “The elasticity and strength of paper and other ﬁbrous materials”, British

J. App. Phys, Vol. 3, pp.72-79, 1952

. P. K. Mallick, Fiber-Reinforced Composites, Marcel Dekker, Inc., New York, 1993

. D. Hull and T. W. Clyne, T.W., An Introduction to Composite Materials, 2nd

edition, Cambridge university press, New York 1996

L. T. Drzal, "The Epoxy Interphase in Composites”, review chapter in Advances in

 

Polymer Science 11, 75, K. Dusek, ed., Spring-Verlag, 1985

J. M. Whitney and L. T. Drzal "Three Dimensional Stress Distribution Around An
Isolated Fiber Fragment ", ASTM STP 937, Toughened Composites, 179-196 ,1987
F. M. F owkes, “Acid-base Interaction in Polymer adhesion”, Physical-chemical
Aspect of polymer surfaces, (Mittal ed), Plenum Press, New York 1983

W. B. Jensen, “Acids, Bases, and Adhesion.” Chemtech. v12, 755 1982

. V. Gutmann, The Donor-Acceptor Approach to Molecular Interaction. Plenum

Press, New York 1978

193

21

k)
10

is.)
b)

31.

R. G. Pearson, “Hard and Soft Acids and Bases”, Am. Chem. Educ. v64(7) 563

1987

. R. S. Drago, G. C. Vogel and T. E. Needham, “A Four-parameter Equation for

Predicting Enthalpies of Adduct Formation”, J. Am. Chem. Soc. v93, 6014, 1971

. L. H. Lee, “The Chemistry and Physics of Solid Adhesion”, Fundamentals of

Adhesion, Plenum Press, New York 1991

. H. Simon, Ph.D thesis, University de Haute Alsace, Mulhouse, France, 1984

. J. Schultz, L. Lavielle and H. Simon, “Science and new applications of carbon

ﬁbers”, Proc. Tntern. Symp., pp. 125, 1984

. J. Schultz and L. Lavielle and C. Martin, J. Adhesion, Vol. 23, pp. 45, 1987

. J. Schultz and E. Lavielle, “Inverse Gas Chromatography”, Amer. Chem. Soc.

Symp. Ser., Vol391, pp.185, 1989

. F. M. Fowkes, J. Adhesion Sci. Tech., Vol.1, pp7, 1987

. L. T. Drzal, M. Madhukar and M. C. Waterbury, "Surface Chemical and Surface

Energetic Alteration of the 1M6 Carbon Fiber Surface and Its Effect on Composite
Properties", 14th Annual Meeting of the Adhesion Society, Hilton Head, SC

(1992).

. L. T. Drzal, "T he Role of the F iber/Matrix Interphase on Composite Properties ",

Vacuum, 41, 1615-1618, 1990
L. T. Drzal and M. Madhukar, "Fiber-Matrix Adhesion and Its Relationship to
Composite Mechanical Properties " J. Matr. Sci., 3, 569-610, 1993

G. S. Sheu and S. S. Shyu, J. Adhesion Sci. Tech., Vol. 8, pp.1027, 1994

194

33. L. T. Drzal and P. Herrera-Franco, "Composite F iber-Matrix Bond Tests ", Chapter
in the Engineered Materials Handbook: Adhesives and Sealants, 3, 392-405,
1990

34. M. C. Waterbury and L. T. Drzal, "On the Determination of Fiber Strengths by In-
Situ Fiber Strength T esting", J. Comp. Tech. Res., Q, 22-28 , 1991

35. V. Rao, P. Herrera-Franco, A. D. Ozzello and L. T. Drzal, "A Direct Comparison of
the Fragmentation test and the Microbond Pull-Out Test for Determining the
Interfacial Shear Strength ", J. Adhesion, ﬂ, 65-77 , 1991

36. P. Herrera-Franco and L. T. Drzal, "Comparison of Methods for the Measurement of
F iber-Matrix Adhesion in Composites ", Composites, 23, 2-27, 1992

37. P. Herrera-Franco, W-L. Wu, M. Madhukar and L. T. Drzal, "Contemporary Methods
for the Measurement of F iber-Matrix Interfacial Shear Strength ", Polymer, 1991

38. P. Herrera-Franco, V. Rao, L. T. Drzal and M.Y.M.Chang, "Bond Strength
Measurement in Composites-Analysis of Experimental Techniques ", Composites
Engr, 2, 31-45, 1992

39. P. J. Herrera-Franco, M. Chiang and L. T. Drzal, "T he Microbond Technique for
Measurement of F iber-Matrix Interfacial Shear Strength -A Theoretical Analysis-",
Proceedings of ANTEC '92, Detroit, MI, 1992

40. The Interfacial Testing System (ITS), The Dow Chemical Company, Freeport,
Texas

41. J. F. Mandell, D. H. Grande, T. H. Tsiang and F. J. McGarry, “”Modiﬁed

microdebonding test for direct in situ ﬁber/matrix bond strength determination in

195

ﬁber composites”, Composite Materials:Testing and Design, ASTM STP893,

American Society for Testing and Materials, 1986, pp. 87-108

196

Appendix 11

C++ PROGRAM FOR NANO-SCRATCH TEST

#include <stdio.h>
#include <math.h>
#include <stdlib.h>

#define NULL 0
#define FALSE 0
#define TRUE 1
#define PIE 3.1415926

float Ydisplacement_offset, Sdis_offset, rad;
float Ydis, Sdis, Los, Ylf, Ycof, Thcof,Plcof, P2cof,
int Step, Count, AveCount, Condition;

Shratio;

float SumYdis, Sudeis, SumLos,SumYlf, SumYcof, BaseYdis;

float AveLos, AveYlf, AveYcof, AveYdis, AveSdis;
FILE *fp_r, *fp_w;
bool First;

float function_of_d(int Con, float d, float 3)

float theta, beta;

switch (Con)
{
case 1:
{
return 0;
break;
}
case 2:
{
if(d<=Ydisplacement_offset)
return Plcof;
else if (d >= (sqrtl3)/2
*Shratio*fabs(s)+Ydisplacement_offset))
return P2cof;

else return (Plcof * ((s*Shratio)*(s*Shratio)-
4.0/3.0*(d—Ydisplacement_offset)*(d-Ydisplacement_offset))+P2cof *

4.0/3.0* (d-Ydisplacement_offset)*(d—
Ydisplacement_offset))/((s*Shratio)*(s*Shratio));

197

//break;

case 3:

{
s=s-Sdis_offset;

if(d<=Ydisplacement_offset)
return Plcof;

else if ( d > Ydisplacement_offset && d<= (sqrt(3)/2
*Shratio*fabs(s)+Ydisplacement_offset))
{
theta= PIE/6—asin((Ydisplacement_offset+rad -
d)/(2*rad));

return
4/(sqrt(3)*(s*Shratio)*(s*Shratio))*(Plcof*(sqrt(3)/4*(s*Shratio)*(s*Sh
ratio)-theta*rad*rad+sin(theta)*cos(theta)*rad*rad-
sqrt(3)*rad*rad*sin(theta)*sin(theta))+
P2cof*(theta*rad*rad-
rad*rad*sin(theta)*cosltheta)+sqrt(3)*rad*rad*sin(theta)*sin(theta)));

}

else if( d > (sqrt(3)/2
*Shratio*fabs(s)+Ydisplacement_offset) && d <= Ydisplacement_offset +
rad — 2*rad*sin(PIE/6-asin(Shratio*fabs(s)/(2*rad))))
i
theta: PIE/6-asin((Ydisplacement_offset+rad -
d)/(2*rad));
beta=acos(l—(d—sqrt(3)/2*Shratio*fabs(s)—
Ydisplacement_offset)/rad);

return
4/(sqrt(3)*(s*Shratio)*(s*Shratio))*(Plcof*(sqrt(3)/4*(s*Shratio)*(s*Sh
ratio)—theta*rad*rad+rad*(Ydisplacement_offset+rad-d)*sin(theta)
+beta*rad*rad —rad*rad*cos(beta)*sin(beta))+
P2cof*(theta*rad*rad-
rad*(Ydisplacement_offset+rad -d)*sin(theta)-
beta*rad*rad+rad*rad*cos(beta)*sin(beta)));

}

else if ( d > Ydisplacement_offset + rad - 2*rad*sin(PIE/6-
asin(Shratio*fabs(s)/(2*rad))) && d <= Ydisplacement_offset + 2*rad )
return P2cof;

else if ( d> Ydisplacement_offset + 2*rad && d <= sqrt(3)/2
*Shratio*fabs(s)+Ydisplacement_offset + rad + sqrt(rad* rad - 0.25 *
Shratio*fabs(s) * Shratio*fabs(s)))

{

//theta = acos((d—Ydisplacement_offset-rad)/(l+rad));

198

theta=asin((d-Ydisplacement_offset —rad)/(2*rad))-
1.0/6.0*PIE;

return
4/(sqrt(3)*(s*Shratio)*(s*Shratio))*(Plcof*(sqrt(3)/4*(s*Shratio)*(s*Sh
ratio) —(theta*rad*rad-
rad*rad*sin(theta)*cos(theta)+0.5*(2*rad*sin(theta)+Shratio*fabs(s))*(s
qrt(3)/2*Shratio*fabs(s)-

sqrt(3)*rad*sin(theta))))+P2cof*(theta*rad*rad-
rad*rad*sin(theta)*cosltheta)+0.5*(2*rad*sin(theta)+Shratio*fabs(s))*(s
qrt(3)/2*Shratio*fabs(s)-sqrt(3)*rad*sin(theta))));

}

else if ( d > sqrt(3)/2
*Shratio*fabs(s)+Ydisplacement_offset + rad + sqrt(rad* rad - 0.25 *
Shratio*fabs(s) * Shratio*fabs(s)) && d <= Ydisplacement_offset + 2*rad
+ sqrt(3)/2 *Shratio*fabs(s))
I
theta = acos(ld-Ydisplacement_offset-rad -
sqrt(3)/2*Shratio*fabs(s))/rad);

return
4/(sqrt(3)*(s*Shratio)*(s*Shratio))*(Plcof*(sqrt(3)/4*(s*Shratio)*(5*Sh
ratio)—theta*rad*rad
+rad*rad*sin(theta)*Cosltheta))+P2cof*(theta*rad*rad-
rad*rad*sin(theta)*cosltheta)));

return Plcof;

default:
return 0;
//break;

printf("Condition = "l:
scanf("%d", &Condition);

printf("Ydisplacement_offset = ");
scanf("%f", &Ydisplacement_offset);

printf("Sdis_offset = ");
scanf("%f", &Sdis_offset);

printf("Radius = ");
scanf("%f", &rad);

printf("Shratio = ");
scanf("%f", &Shratio);

199

printf("P1cof = ");
scanf("%f", &P1cof);

printf("P2cof = ");
scanf("%f", &92cof);

printf("\nStep = ");
scanf( "%d", &Step);

if((fp_r=fopen("original.txt", "r"))==NULL)
{
printf("Error Open Read!\n");
exit(-1);

}

if((fp_w=fopen("output.txt", "W"))==NULL)
{
Printf("Error Open Write!\n");
exitl-l);

}
First=TRUE;

SumYdis=0;
Sudeis=0;
SumLos=0;
SumYlf=0;
SumYcof=0;
Count=1;

AveCount=0;

whilelfscanflfp_r, "%f %f %f %f %f", &Ydis, &Sdis, &Los, &Ylf,
&Ycof)!=EOF)
{
if(First)
{
First=FALSE;
BaseYdis=Ydis;
}

Sdis=Sdis/1000;

if( (Ydis-BaseYdis) <= (Step * 0.001) * Count)
{

SumYdis += Ydis;

Sudeis += Sdis;

SumLos += Los;

SumYlf += Ylf;

SumYcof += Ycof;

AveCount++;

else //

if(AveCount!=0)

200

AveYdis=SumYdis/AveCount;
AveSdis=Sudeis/AveCount;
AveLos =SumLos/AveCount;
AveYlf =SumYlf/AveCount;
AveYcof=SumYcof/AveCount;

Thcof=function_of_d(Condition, AveYdis,
AveSdis);
// Thcof=0;

fprintf(fp_w,"%11.6f %11.6f %11.6f %11.6f
911.6f %11.6f %ll.6f\n", AveYdis, AveSdis, AveLos, AveYlf, AveYcof,
Thcof, AveYcof-Thcof);

Count=(int)((Ydis-BaseYdis)*1000/Step)+1;

SumYdis=Ydis;
Sudeis = Sdis;
SumLos = Los;
SumYlf = Ylf;
SumYcof = Ycof;

AveCount=1;

} //else

} //whi1e

if(AveCount!=0)

{
AveYdis=SumYdis/AveCount;
AveSdis=Sudeis/AveCount;
AveLos =SumLos/AveCount;
AveYlf =SumYlf/AveCount;
AveYcof=SumYcof/AveCount;

Thcof=function_of_d(Condition, AveYdis, AveSdis);

fprintf(fp_w,"%ll.6f %ll.6f %1l.6f %ll.6f %ll.6f %ll.6f

%11.6f\n", AveYdis, AveSdis, AveLos, AveYlf, AveYcof, Thcof, AveYcof—

Thcof);
}

fclose(fp_r);
fclose(fp_W);

return 0;

201

Appendix III

RELATIONSHIP BETWEEN LINEAR SHRINKAGE AND
VOLUMETRIC SHRINKAGE‘”

Shrinkage can induce signiﬁcant stresses in the composite already before loading.
While when the material in the liquid state, no stress will be build up. Even when the
material in one region is solidiﬁed, the surrounding material can still flow and shrinkage
induced stress is limited. The main shrinkage caused residual stress is caused by the

volume shrinkage after the main domain become a kind of solid. Here we refer this
volume shrinkage as AV.

For a speciﬁc change of volume AV/V0 (V0 is a volume at which the stress caused by
volume shrinkage begin to build up) and a deformation state, for an arbitrarily large

deformation:
AV/VO =J1+J2 +J3

Where, J 1+1 2+1 3 are the strain invariants, and

J1=8x+8y+82

J2 = exey +8yez + 8x82

202

l l
8x jyxy 57x2

1
J3“§7xy 8y '2‘7yz

 

 

l 1
‘2' 7x2 E 7322 82
For a small strain,

AV/VO le

J1=8x +8}, +82 =(8x0+£xr)+(8y0+8yr)+(80 +8”)

8x, 2 5y, = 8,, = 0 and 8x0 = eyo = .920 = 80

So: AV/V0 = 3.s,=3sS

where as is the linear strain caused by cure shrinkage.

REFERENCES

1. A. Ten Busschen, Ph.D thesis, Delﬁ University of Technology, the Netherlands,
1996

203