I 91.5".

n.

 

 

m‘a-

»

x
l,

 

 

 

 

'1
1".

 

 

 

 

 

 

 

 

 

u .-u.; n..~1v

 

 

 

 

 

  

 

     

 

 

 

   

 

 

 

 

   

 

 

 

"HERE;

lllllllllllllllllllll

This is to certify that the

thesis entitled

Characterization of the Interphase and its
Influence on the Behavior of Glass
Fiber-Reinforced Epoxy Composites

presented by

Edward Kent Drown

has been accepted towards fulfillment
of the requirements for

M.S. Chemical Engineering

degree in

ééwwmf,

Major professor

 

 

 

Date 8/9/91

 

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

[TLWW-Aav-n‘wv. mn-.--Pv-z:‘ '- a; -._. . A “ta-n,“ -5 ‘3‘

g L :5 g: 53A a if l
‘ M . S55 was «at r“?- i i? “5.. 5
; “3 t5 gar-an £5; at; j

l‘.

- 5r ‘ '1 ' ‘ k I .
i Univ-eta !
l .

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
c:\circ\ddedue.pm3—p. 1

CHARACTERIZATION OF THE INTERPHASE AND ITS INFLUENCE ON THE
BEHAVIOR OF GLASS FIBER-REINFORCED EPOXY COMPOSITES
By

Edward Kent Drown

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

MASTER OF SCIENCE

Department of Chemical Engineering

1991

ABSTRACT

CHARACTERIZATION OF THE INTERPHASE AND ITS INFLUENCE ON THE
BEHAVIOR OF GLASS FIBER—REINFORCED EPOXY COMPOSITES

By

Edward Kent Drown

Research was conducted to determine to what degree an epoxy-compatible glass—ﬁber
sizing system alters ﬁber-matrix adhesion and its effect on the mechanical properties of
a glass ﬁber-reinforced epoxy-matrix composite. Blends of the sizing system and bulk
matrix were used to model the ﬁber—matrix interphase formed when the sizing system
interacts with the matrix during fabrication of the composite. In this case, the sizing
system’s interaction with the matrix produced a material with an increased modulus and
tensile strength but lower toughness and glass transition temperature. The presence of
the sizing system increased the level of ﬁber-matrix adhesion and shifted the failure locus
from the interface to the matrix. Mechanical testing of the composite revealed that the
presence Of the sizing system affects the failure behavior of interphase insensitive tests

and affects both the failure behavior and properties of interphase sensitive tests.

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES .................................... V
LIST OF FIGURES .................................... vi
CHAPTER 1 INTRODUCTION
1.1 INTRODUCTION ........................... 1
1.2 BACKGROUND ............................ 2
1.2.1 GLASS FIBERS ....................... 2
1.2.2 SILANE COUPLING AGENTS .............. 2
1.2.3 GLASS FIBER SIZING SYSTEMS ........... 4
1.2.4 ADHESION AT THE FIBER-MATIX
INTERFACE ......................... 5
1.3 RESEARCH OBJECTIVE ...................... 8
CHAPTER 2 EXPERIMENTAL PROCEDURE
2.1 INTRODUCTION ........................... 9
2.2 MATERIALS ............................. 9
2.2.1 SYSTEM CONSTITUENT S ................ 9
2.2.2 COMPOSITE FABRICATION ............... 10
2.2.3 LAMINATE CHARACTERIZATION .......... 11
2.3 FIBER TENSILE STRENGTH ................... 14
2.4 BLEND MECHANICAL PROPERTIES ............. 15
2.4.1 DYNAMIC MECHANICAL ANALYSIS ........ 15
2.4.2 TENSILE TESTING ..................... 16
2.5 INTERFACIAL CHARACTERIZATION ............ 16
2.5.1 EMBEDDED SINGLE FIBER TEST ........... 16
2.5.2 INTERFACIAL TESTING SYSTEM ........... 19
2.6 COMPOSITE MECHANTCAL PROPERTIES .......... 23
2.6.1 SHORT BEAM SHEAR TESTING ............ 23
2.6.2 FLEXURE TESTING .................... 23
2.6.3 TENSILE TESTING ..................... 24
2.7 FRACTOGRAPHY .......................... 24

 

CHAPTER 3 RESULTS and DISCUSSION
3.1 INTRODUCTION ........................... 25
3.2 EFFECT OF THE SIZING ON FIBER TENSILE

iii

STRENGTH ............................. 25

 

 

 

 

3.3 BLEND PHYSICAL PROPERTIES ................ 27
3.3.1 GLASS TRANSITION TEMPERATURE ........ 27
3.3.2 MECHANICAL PROPERTIES .............. 28

3.4 EFFECT OF THE SIZING ON INTERFACIAL
BEHAVIOR .............................. 31
3.4.1 FIBER-MATRIX FAILURE LOCUS ........... 31
3.4.2 FIBER—MATRIX ADHESION ............... 35

3.5 EFFECT OF THE SIZING ON COMPOSITE
PROPERTIES ............................. 37
3.5.1 SHORT BEAM SHEAR ................... 37
3.5.2 90° FLEXURE ........................ 38
3.5.3 0" FLEXURE ......................... 39
3.5.4 90° TENSILE ......................... 43
3.5.5 0° TENSILE .......................... 43

3.6 EFFECT OF THE SIZING ON COMPOSITE
FRACTOGRAPY ........................... 44
3.6.1 SHORT BEAM SHEAR ................... 44
3.6.2 0° FLEXURE ......................... 46
3.6.3 90° TENSILE ......................... 49
3.6.4 0" TENSILE .......................... 51

CHAPTER 4 CONCLUSIONS ............................ 53
LIST OF REFERENCES ................................ 55

Table 1.

Table 2.
Table 3.
Table 4.
Table 5.
Table 6.

Table 7.
Table 8.
Table 9.
Table 10.
Table 11.

Table 12.

LIST OF TABLES

Elemental Composition of E—Glass Fibers at the Surface

and in the Bulk ...........................

Typical Components of a Glass-Fiber Sizing System . . . .

Laminate Characteristics .....................

ITS Specimen Polishing Protocol ................
Mechanical Properties of D.E.R. 383/DACH/Sizing Blends
Embedded Single Fiber Results for D.E.R. 383/

DACH/Bare E-Glass .......................

ITS Results for D.E.R. 383/DACH/E-Glass Composites
SBS Results for D.E.R. 383/DACH/E-Glass Laminates . .
90° Flexural Properties of D.E.R. 383/DACH/E—Glass

Laminates ..............................

0° Flexural Properties of D.E.R. 383/DACH/E-Glass

Laminates ..............................

90° Tensile Properties of D.E.R. 383/DACH/E—Glass

Laminates ..............................

0° Tensile Properties of D.E.R. 383/DACH/E-Glass

Laminates ..............................

...2
5
...14
...22

..29

...35

....36

...37

...38

...39

...43

...44

Figure 1.
Figure 2.
Figure 3.

Figure 4.
Figure 5.
Figure 6.
Figure 7.
Figure 8.
Figure 9.

Figure 10.
Figure 11.

Figure 12.
Figure 13.

Figure 14.
Figure 15.
Figure 16.
Figure 17.
Figure 18.
Figure 19.
Figure 20.
Figure 21.
Figure 22.

Figure 23.

Figure 24.

LIST OF FIGURES

Structure of a Silane Coupling Agent ................ 2
Hydrolysis of a Silane Coupling Agent ............... 3
Adsorption and Reaction of a Silane Coupling Agent

with Surface Silanol Groups ...................... 4
Chemical Bonding Theory of Fiber-Matrix Adhesion ....... 6
Interphase Theory of Fiber—Matrix Adhesion ............ 7
Polished Laminate Cross Section ................... 12
Polished and Etched Laminate Cross Section ............ 13
ITS Microindenter and Composite Specimen ............ 20
The Interfacial Testing System .................... 21
E-Glass Fiber Strength Versus Gage Length ............ 27
Tg Versus Sizing System Concentration in

D.E.R. 383/DACH ........................... 28
Blend Physical Properties Versus Sizing Concentration ..... 30
Fiber Break at Low Strain, D.E.R. 383/DACH/Bare

E—Glass, X67 ............................... 32
Fiber Break at High Strain, D.E.R. 383/DACH/Bare

E-Glass, X67 ............................... 33
Fiber Break at Low Strain, D.E.R. 383/DACH/Sized

E—Glass, X67 ............................... 34
0° Flexure Specimens, Tensile Side, D.E.R. 383/

DACH/Bare E—Glass .......................... 40
0° Flexure Specimens, Edge View, D.E.R. 383/

DACH/Bare E-Glass .......................... 40

0° Flexure Specimens, Tensile Side, D.E.R. 383/
DACH/Sized E—Glass .......................... 41
0° Flexure Specimens, Edge View, D.E.R. 383/

DACH/Bare E-Glass .......................... 42
SBS Failure Surface, D.E.R. 383/DACH/Bare E—Glass ..... 45
SBS Failure Surface, D.E.R. 383/DACH/Sized E-Glass ..... 46
0° Flexure Failure Surface, D.E.R. 383/DACH/Bare

E-Glass .................................. 47
0° Flexure Failure Surface, D.E.R. 383/DACH/Sized

E-Glass .................................. 48
0° Flexure Failure Surface, Compressive Side, D.E.R.
383/DACH/Sized E-Glass ....................... 48

vi

Figure 25.
Figure 26.
Figure 27.
Figure 28.

Figure 29.

0° Flexure Failure Surface, Tensile Side, D.E.R. 383/

DACH/Sized E-Glass .......................... 49

90° Tensile Failure Surface, D.E.R. 383/DACH/Bare

E—Glass .................................. 50

90° Tensile Failure Surface, D.E.R. 383/DACH/Sized

E—Glass .................................. 50
0° Tensile Failure Surface, D.E.R. 383/DACH/Bare

E—Glass .................................. 52
0° Tensile Failure Surface, D.E.R. 383/DACH/Sized

E—Glass .................................. 52

vii

CHAPTER 1

INTRODUCTION

1.1 INTRODUCTION

Glass ﬁber-reinforced polymer matrix composites have been in continuous use for
over forty years largely because of the development of silane coupling agents. These
coupling agents are incorporated into protective coatings for glass ﬁber surfaces called
"sizes" which contain silane coupling agents as well as processing aids. These sizes are
applied to the surface of the glass ﬁbers shortly after ﬁber manufacture and before
composite fabrication. These sizes increase the level of adhesion between the matrix
polymer and glass reinforcement and aid in the long-term retention of composite
mechanical properties. While the chemistry of silane coupling agents and their
interactions with both the glass ﬁber’s surface and the polymer matrix has been
extensively studied,“11 little fundamental information of a predictive nature attesting to
the relationships between sizing application and composite mechanical properties has been
published. Contemporary art views the application of a size as necessary for ease of
processing and for acceptable retention of ﬁber strength during handling and composite
fabrication. This study was undertaken to investigate the glass ﬁber-sizing system—
polymer matrix interphase to better understand the role of the interphase in the

performance of glass ﬁber-reinforced epoxy-matrix composites.

1.2 BACKGROUND

1.2.1 GLASS FIBERS. The glass ﬁbers used as reinforcements are complex mixtures
of oxides, a typical elemental composition for E—glass ﬁbers in the bulk and at the

surface is given in Table 1.

Table 1. Elemental Composition of E—Glass Fibers at the Surface and in the Bulk1

 

Location B Ca O F Mg Al Si
Bulk 4.5 6.4 61.8 0.3 2.3 5.8 18.8
Surface 3.4 1.9 61.2 1.4 0.6 7.0 24.3

 

 

 

 

 

 

 

 

 

 

 

 

The oxide groups present on the ﬁber surface react readily with atmospheric water to

form hydroxyl groups of the type M-OH where M = silicon or aluminum.2

1.2.2 SILANE COUPLING AGENTS. The structure of a typical silane coupling agent

is shown in Figure 1.

R
9
RO-si—

O
R

Figure 1. Structure of a Silane Coupling Agent

CH2—CH2—CH2—R’

The coupling agent consists of a central silicon atom connected to an organofunctional
group (R’) via a propyl chain. Three methoxy or ethoxy groups (OR) are also attached
to the silicon atom. The organofunctional group can be selected to be reactive with the

matrix, in the case of epoxy, R’ can be an epoxide group or a functional group found in

3

the hardener. Before being applied to the ﬁbers, the coupling agent is converted to the

silanol form by hydrolysis, shown in Figure 2.

 

R
9
RO-Si—CHZ—CH2—CH2—R’ + excess H20
0
R
H
9
HO—Si—CHZ—CHZ—CHZ—R’ + 3 ROH + mo
0
H

Figure 2. Hydrolysis of a Silane Coupling Agent

The hydrolyzed form of the coupling agent can then undergo further elimination reactions
between silanol groups to form Oligomers and a three dimensional polysiloxane
networkf”4

The structure Of physisorbed and chemisorbed silane coupling agents at silica
($09 and glass surfaces has been investigated via FTIR,5 NMR,°~7 and XPS.8 The
adsorption and reaction of silane coupling agents with the ﬁber surface, illustrated in
Figure 3, is a multistep process. The coupling agent physisorbs onto the ﬁber surface
through hydrogen bonding of the hydroxyl groups on the coupling agent and ﬁber
surface. The coupling agents have been found to then form covalent oxane bonds via an
elimination reaction between a silanol group on the silane coupling agent and a hydroxyl
group on the glass surface.9 Silane coupling agents also form Oligomers in solution and
may adsorb as such. This interaction extends beyond the monolayer range and evidence

indicates that the silane coupling agent adsorbs in multilayers when possible'o'“

 

 

 

 

(HO)3— Si—CHZ— CH2— CH2— R’

Si-- 0H Silane Coupling Agent

'1 .'Si.;_\ 30 O

- . o\ \—Si—CH2—CH2—CH2—R’
‘ , .9 O

' F, H

5 I .

.‘B7

E H20

TR

. , H
O
Si-j-O—Sli — CH2—CH2—CH2— R’

O

Sif—O—gi—CHz—CHZ—CHZ—R’
’ ' H

 

 

 

Figure 3. Adsorption and Reaction of a Silane Coupling Agent with Surface

Silanol Groups

1.2.3 GLASS FIBER SIZING SYSTEMS. The composition of a glass-ﬁber sizing
system is complex, with the coupling agent comprising a relatively small portion of the
material applied to the ﬁbers. Most commercial sizing systems, especially those used in
glass ﬁber-reinforced epoxy composites are complex proprietary compositions. For
example, Table 2 lists the general proportion of components in a commercial sizing
system excluding the solvent or carrier used to apply the mixture. The ﬁlm-forming
resin aids in the distribution of the sizing over the ﬁbers and acts as a binder providing
integrity to the ﬁber tow. Antistatic agents prevent the build-up of static electricity

during handling. Lubricants decrease friction between the ﬁbers, thereby reducing

5

Table 2. Typical Components of a Glass-Fiber Sizing System12

 

 

 

 

 

 

 

 

Component Percent

Film-Forming Resin 1 ~ 5
Antistatic Agent 0.1 ~ 0.2
Lubricant 0.1 ~ 0.2
Coupling Agent 0.1 ~ 0.5

 

abrasive damage. The remaining fraction of the sizing system is the silane coupling
agent, the optimum amount has been found to be 10 to 20 monolayer equivalents on the
surface.13

The constituents of typical sizing systems are compatible and can interact with
each other. The resulting distribution of these components in the interphase region is not
well understood, with the formulation of effective sizing systems being an empirical
undertaking for the most part. The processing aids may change the interphase properties

in a manner similar or counter to that of the silane coupling agent acting alone.

1.2.4 ADHESION AT THE FIBER-MATRIX INTERFACE. The two most common
concepts of adhesion across the ﬁber-matrix interface are the chemical bonding and
interphase theories. The chemical bonding theory, Figure 4, holds that approximately
a monolayer of the silane coupling agent acts as a bridge between the glass surface and
the polymeric matrix. The presence of covalent bonding between the coupling agent and
the ﬁber surface has bolstered the chemical bonding concept. However, chemical
bonding cannot account for the increase in adhesion experienced between non-reactive

matrices such as polyoleﬁns and silane coupling agent-treated inorganic reinforcements

 

 

 

'if,»'si'—~o—Si—-
¢ . l
, , ‘13
HI I. T
.B' Si--O—Si— R
E | I
(I) X
‘I’
Si4-O—Ti—
o

 

 

 

 

Figure 4. Chemical Bonding Theory of Fiber-Matrix Adhesion

in which chemical bonds will not be formed.14 This observation, among others, leads
to an alternative proposal that an "interphase" composed of various constituents forms
surrounding the reinforcement. This third phase (i.e. interphase) in the composite is
possibly formed through interdiffusion of physisorbed silane coupling agent and matrix
molecules in the interphase and perhaps preferential adsorption of both matrix
components as well as silane coupling agents on the reinforcement surface. '5 The
interphase thus formed may appear somewhat as depicted in Figure 5, having a gradation
in the properties of the interphase until equivalence to the matrix is reached some
distance away from the ﬁber. Alteration of the interphase mechanical properties alone

near the reinforcement surface16 can explain many of the phenomena associated with

 

73111me

 

 

 

 

AAA/v = Silane Coupling Agent

 

= Sizing System and Matrix Components

Figure 5. Interphase Theory of Fiber—Matrix Adhesion

silane coupling agents. The increased hydrolytic stability of coupling agent treated
composites likewise has been proposed to occur through hydrolyzable bonds between the
silane coupling agent and reinforcement surface that can break and reform easily enough
to relieve stress at the interface and preserve the integrity of the composite. This type
of argument is supported by an interphase model17 in which the molecular structure of
the interphase is most often considered to be an interpenetrating network between the

coupling agent and matrix resin.

1.3 RESEARCH OBJECTIVE

The objective of this study was to investigate the interaction of a commercial
sizing system with the matrix in a glass ﬁber-reinforced epoxy composite and its
inﬂuence on composite behavior. Interphase physical properties were modelled using
epoxy/ sizing system blends in dynamic mechanical and tensile tests. Effects of the sizing
system on the mechanical response of the ﬁber—matrix interface (interfacial shear
strength, failure mode) were characterized using the embedded single ﬁber test and the
Interfacial Testing System. Mechanical testing coupled with fractography were used to
characterize the effect of the sizing system on composite mechanical properties and

failure modes.

CHAPTER 2

EXPERIMENTAL PROCEDURE

2.1 INTRODUCTION

The experimental program of this study focussed on developing an understanding
of how a commercial sizing interacts with the matrix in the glass ﬁber—matrix interphase
and its effect on composite properties. Measurements of the glass transition temperature
and tensile properties of bulk samples of epoxy/ sizing system blends were made to
evaluate the range of properties of expected interphases. The mechanical response of the
ﬁber—matrix interphase (interfacial shear strength and failure mode) was evaluated using
the embedded single ﬁber method and the Interfacial Testing System. Composite
mechanical prOperties were measured (short beam shear, 0° and 90° ﬂexure, 0° and 90°

tension) and fractography conducted on unidirectional laminates.

2.2 MATERIALS

2.2.1 SYSTEM CONSTITUENTS. A typical epoxy matrix, a commercial sizing system
and E-glass reinforcing ﬁbers were selected as the materials to be used in this study. A
diglycidyl ether of bisphenol A based resin, D.E.R. 383° (Dow Chemical Company), was
selected as the epoxy resin. 1,2-Diaminocyclohexane (DACH) (Aldrich Chemical

Company) was used as the hardener at 15.6 parts per hundred of resin by weight. E-

10

glass tows (1600 ﬁlaments/ tow, 13 am nominal ﬁlament diameter) were supplied by the
Pittsburgh Plate Glass Company (PPG). The ﬁbers were supplied with two surface
treatments, the ﬁrst being deionized water only and the second a commercial epoxy-
compatible sizing system. The water—treated ﬁbers are referred to as "bare" and the
ﬁbers coated with the size are referred to as "sized". The sizing system was also
supplied in bulk form for the preparation of epoxy/ size blend samples which were used

for mechanical property measurements.

2.2.2 COMPOSITE FABRICATION. Unidirectional laminates were prepared for single
ﬁber microindentation measurements with the Interfacial Testing System (ITS) and for
short beam shear, ﬂexure and tensile mechanical testing. A ﬁlament winding prepregger
was used to prepare glass/epoxy prepreg tapes. The glass ﬁber tow was supplied on a
cardboard core and stored in a polyethylene bag to prevent drying out of the tow. The
core was mounted on a frame and axle to allow the unwinding of the tow from the
outside of the package. The tow passed through a drying oven held at 110 °C with a
residence time of about 5 seconds. The tow then entered the resin pot where it travelled
between 3 polished stainless steel pins that helped separate the ﬁlaments and aid
impregnation of the resin into the tow. The tow exited the resin pot through a slit die
of dimensions 0.20 x 1.8 mm. The impregnated tow was guided by a roller and laid
onto a rotating drum. The drum was covered with a layer of release ﬁlm. The resin
pot, guide pins and rollers were held at room temperature with the process requiring
about 2 hours to produce a tape 11 inches wide and 6 feet long. The tapes were cut into

12 inch lengths and stacked to produce unidirectional laminates of 12 or 18 plies. The

ll

uncured laminates were placed between 1/2 inch thick aluminum caul plates that were
covered with 2 plies of bleeder cloth and 1 ply of porous Teﬂon' (DuPont) coated glass
release fabric. Self-adhesive cork dams surrounding the prepreg prevented its escape
from between the caul plates during processing.

The temperature cycle used to cure the laminates was 16 hours at room
temperature, 2 hours at 80°C and 2 hours at 175°C with 5°C/min ramps between soaks.
Directly after lay-up, the caul plate-laminate assembly was placed in a Carver Laboratory
Press and sufﬁcient load applied to produce 0.7 MPa (100 psig) in the laminate. This
load was maintained for 30 minutes in order to bleed-out excess resin. The assembly
was then placed in a vacuum bag and loaded into an autoclave for the elevated
temperature portion of the cure cycle. A pressure of 0.9 MPa (125 psig) was applied to
outside of the vacuum bag assembly for the duration of the autoclaving. Vacuum of -
0.09 MPa (-28 in Hg) was applied for the ﬁrst 30 minutes of the 80°C soak period and

vented to atmosphere for the remainder of the cure cycle.

2.2.3 LAMINATE CHARACTERIZATION. Laminate quality was assessed after
fabrication though the use of Optic Numeric Volume Fraction Analysis18 (ONVFA), a
technique which allows quantitative assessment of ﬁber volume fraction using optical
methods. The procedure requires the preparation of a composite cross section for
observation with a reﬂected light microscope, digitizing a video image of the specimen
and then counting the ﬁbers in a deﬁned area of the cross section. One other variable,
the area per ﬁber, is obtained from manufacturer’s lot data or determined from the ﬁber

tow directly. The ﬁber volume fraction is then determined from ( 1):

12

Area per Fiber (1)
Area per Image

Volume Fraction Fibers = Fibers per Image *
For this study, three randomly selected sections about 2 cm x 2 cm were removed from
the laminate of interest and embedded in polyester metallographic mounts with the long
axis of the ﬁbers perpendicular to the mount surface. The specimens were polished using
standard metallographic methods to a 1 pm ﬁnish on a rotating lap polisher. This
produced a surface with well deﬁned ﬁber ends. However, as shown in Figure 6, the

contrast between ﬁber and matrix was poor.

 

Figure 6. Polished Laminate Cross Section

ONVFA was originally developed for carbon ﬁber-reinforced epoxy composites which

exhibited a high contrast between ﬁber and matrix after polishing. To increase the

13

contrast between ﬁber and matrix in the glass ﬁber-reinforced epoxy composites, the
polished sections were etched for 5 minutes with CHROMERGE' (VWR Scientiﬁc),
rinsed with deionized water for 10 minutes in an ultrasonic bath and dried with
compressed air. As shown in Figure 7, this treatment retained the proﬁle of the ﬁber
ends while the loss of matrix has increased the contrast to allow accurate counting of the

ﬁber ends.

   
 

   
 
 

0..

a“? “553;“ '0': W’rﬁh

Figure 7. Polished and Etched Laminate Cross Section

 
   

 

The specimen was mounted on a metal slide and adjusted to give uniform focus across
its cross section. A reﬂected~light image was fed to a CCD camera and captured with
a frame-grabber for analysis. The counting parameters were adjusted to give about the
same number of double counted and missed ﬁber ends per digitized image. Fifteen

images from each cross section were digitized and the ﬁbers counted. Measured sections

14

of bare ﬁber tow were dried at 100 °C for 2 hours and weighed to determine the average
area per ﬁber of 125.6 i 1.5 umz/ ﬁber. The void volume fraction was determined using
the measured ﬁber volume fraction and the density of the laminate measured

gravimetrically.

Table 3. Laminate Characteristics

 

 

 

 

 

 

 

Laminate Number Fiber Fiber Void Test
of Type Volume Volume Specimens
Plies Fraction Fraction
(%) (%)
042390—1 18 Bare 63.8 i 0.45 + Short Beam
1.9 0.39 Shear
0° Tensile
040990-2 l2 Bare 65.6 i 0.63 + 90° Flexure
1.5 0.28 90° Tensile
050290-2 l2 Bare 63.2 i 0.62 :1: 0° Flexure
1.9 0.37
041390-2 18 Sized 65.1 + 0.65 + Short Beam
2.2 0.45 Shear
0° Tensile
042090-1 12 Sized 63.6 i 0.67 + 0° Flexure
1.6 0.30
0409901 12 Sized 63.6 -_I— 0.48 + 90° Flexure
2.0 0.39 90° Tensile

 

 

 

 

 

 

 

 

2.3 FIBER TENSILE STRENGTH

Fiber tensile strengths were measured on a test ﬁxture speciﬁcally designed for
reinforcing ﬁbers. The apparatus consists of a sample preparation jig and a loading
frame mounted on the stage of an Olympus BH-2 microscope.

The jig holds two

mounting blocks to provide a reproducible gage length between the blocks. A groove

 

15

machined in the t0p of each block keeps the ﬁber axis parallel with the loading axis. The
ﬁber is secured to the blocks with 1,5-Diphenylcarbohydrazide (DPCH) (EM Science),
which is melted with a soldering iron applied to the underside of the mounting block.
The ﬁber is checked for alignment and to ensure that a meniscus of DPCH did not form
along the gage length of ﬁber.

The mounting blocks are secured to a carrier for transfer to the loading frame.
The blocks rest on rail-mounted supports. One block is attached to a load cell, the other
to a linear motion stage. Care is taken during the mounting and loading procedures to
not apply tension to the ﬁber. The carrier is removed and any slack is removed from
the ﬁber. The diameter of the ﬁber is measured with a calibrated ﬁlar eyepiece at three
points along its diameter. The output of the load cell is fed to a digital voltmeter and
chart recorder to record the load at failure. Breaks that occur at the mounting blocks are
rejected. Prior to testing, sections of ﬁber tow were dried at 80 °C for 10 minutes. Ten
individual ﬁber strengths were obtained for each ﬁber type at gage lengths of 25.4, 10.2

and .65 mm.

2.4 BLEND MECHANICAL PROPERTIES

2.4.1 DYNAMIC MECHANICAL ANALYSIS. A DuPont model 983 DMA interfaced
with a 9900 model controller was used to perform dynamic mechanical testing on the
epoxy/ sizing system blends. Stoichiometric mixtures of D.E.R. 383 and DACH were
ﬁrst prepared, then sufﬁcient sizing system was added to produce the desired

concentration on a weight percent basis. The mixture was degassed and poured into

 

16
silicone RTV—664' (General Electric Company) molds that contained 4 cavities 3.2 x 12.5

x 60 mm. The specimens were cured at room temperature for 16 hours in a desiccator
then placed in a forced convection oven and ramped to 80°C at 5°C/ min and held at that
temperature for 2 hours. The samples were allowed to cool to room temperature,
removed from the mold and replaced in the oven on a metal sheet for a 2 hour postcure
at 175°C. The free surface of the specimens was ground on a Struers Abramin polisher
using 320 grit SiC paper and water to produce parallel surfaces.

The DMA was operated in the ﬁxed frequency mode at 1 Hz and a peak to peak
amplitude of 0.6 mm. A DuPont Liquid Nitrogen Cooling Apparatus (LNCA—II) was
used to achieve subambient temperatures. The temperature proﬁle used was: a ramp to -
20°C, soak for 5 minutes then ramp at 5°C/ min to 220°C. Three specimens were tested

for each blend.

2.4.2 TENSILE TESTING. Tensile properties of the blends were measured according
to ASTM standard D638-8419 using Type I specimens which were cast in a manner
identical to that used to prepare DMA test specimens. The specimens were polished to
a 1 pm ﬁnish on all surfaces and stored in a desiccator until tested. Five specimens of

each blend were tested on a United Testing System SFM-20.

2.5 INTERFACIAL CHARACTERIZATION

2.5.1 EMBEDDED SINGLE FIBER TEST. Two methods, the embedded single ﬁber
test and the Interfacial Testing System (ITS) were used to examine changes in ﬁber-

matrix adhesion and failure behavior. The embedded single ﬁber test uses a model

 

l7
composite system consisting of a single ﬁber embedded in a tensile coupon of the matrix

of interest.”21

The coupon is placed in a jig mounted on a microscope stage that
permits simultaneous tensile loading of the coupon and direct high magniﬁcation
observation of the failure process of the ﬁber. Kelly22 ﬁrst analyzed the failure of the
ﬁber which occurs through the transfer of tensile loads on the matrix to the ﬁber through
shear forces at the ﬁber-matrix interface. As the tensile loads in the matrix increase, the
ﬁber fractures at points where ﬂaws have reduced the ﬁber’s tensile strength. As this
population of ﬂaws is reduced, the load required to fracture the ﬁber exceeds the stress
that can be supported by the interface and the ﬁber breakage stops. Once the ﬁber has
stopped fracturing, it is said to have reached the critical transfer length (critical length),
1,, which is a measure of the interfacial shear stress. Shear lag analysis of the specimen,

assuming ideal ﬁber properties and uniform distribution of loads yields the relationship

(2) for the interfacial shear stress, 7':

t z (3:) .. (ft) (2)
2 z,

where df = ﬁber diameter and of = ﬁber strength at a gage length equal to the critical
length. In practice, the measured fragment lengths are not a single value, but rather a
distribution of values between lc/2 and 1,. This distribution is principally due to the
random population of flaws along the length of the ﬁber. To account for the
inhomogeneities in ﬁber strengths, the measured ﬁber lengths are best ﬁt to a two-
parameter Weibull23 distribution. This distribution is used to characterize the strength

of brittle materials such as glass ﬁbers. The modiﬁed expression (3) for the interfacial

18

shear strength, 1:

 

”I °f w
2*[3 a

now contains a and B, the Weibull shape and scale parameters and the gamma function,
1‘. During the loading process the behavior at the ﬁber-matrix interface is observed.
Transmitted light photomicrographs can reveal the presence of matrix or interfacial
cracks at ﬁber breaks and other phenomena that evolve as the coupon is loaded. Epoxy
matrices are stress birefringent, allowing qualitative evaluation of the stress ﬁelds at the
ﬁber breaks through the use of crossed polarizing ﬁlters.

The single ﬁber embedded shear strength test specimen is a single ﬁber
encapsulated in an epoxy tensile coupon. A silicone RTV—664' (General Electric
Company) mold with eight cavities is used for the fabrication of the coupons. The
cavities are ASTM standard tensile coupons, 60 mm long with a gage section 25.4 mm
long, 4 mm wide and 1.5 mm thick. Sprue slots are molded into the end of each cavity.
A single ﬁlament is removed from a ﬁber tow and secured in the Sprue slots with hot-
melt adhesive. Care is taken to ensure that no tension is applied to the ﬁlaments during
separation or while they are secured in the slots. The components of the matrix are
thoroughly mixed and degassed in a vacuum oven for 5 minutes at room temperature to
remove air entrained during the mixing process. The molds with the single ﬁbers are
also treated in the vacuum oven. The mold is inﬁltrated by ﬁlling one end of a cavity
and allowing the resin to flow to the other end, thereby preventing the entrapment of air

bubbles in the coupon. The mold is then transferred to a desiccator for a 16 hour cure

19

at room temperature followed by a 5°C/ min ramp to 80°C for a 2 hour soak. After the
molds have cooled, the test coupons are removed, placed on an aluminum sheet, and
ramped to 175°C at 5°C/ min for a 2 hour postcure. The coupons are allowed to cool at
5°C/ min to room temperature.

The free surface of the coupons are polished to a 1 pm ﬁnish using a Struers
Abramin polisher to allow clear microscopic observation of the fracture process.
Specimens are mounted in a tensile testing apparatus that permits direct observation of
the fracture process via transmitted light. The specimen is strained in increments
indicated on a dial gauge, at each increment the strain and number of ﬁber breaks is
recorded. After the number of breaks has reached a plateau, the ﬁber fragment lengths

and ﬁber diameter are measured with a ﬁlar eyepiece.

2.5.2 INTERFACIAL TESTING SYSTEM. The interfacial testing system (ITS)
(Schares Instrument Corporation, Houston, TX) measures the interfacial shear strength

of ﬁbers in an actual composite,”25

calculated from the load at debond of a single
ﬁber. Figure 8 is a schematic of the process of the test, considering only the test
specimen and microindenter. The ﬁber of interest is loaded by the microindenter until
failure at the ﬁber—matrix interface is detected, either by a load drop or the appearance
of an interfacial crack. The result is calculated using the ﬁnite element modelling results
of Tsiang26 and Grande.27 The model was axi-symmetric, consisting of a cylinder
representing the ﬁber surrounded by a constant thickness of matrix material. Beyond the

layer of matrix, the material properties were assumed to be those of the composite. To

model continuous ﬁbers, their lengths were chosen to be many times their diameter. The

 

20

Diamond

— Tipped
Microindenter

 

Composite
Specimen

 

 

 

 

 

 

 

 

 

 

x\\\\\\\\\\\\\\\\
s\\\\\\\\\\\\\\\‘
s\\\\\\\\\\\\\\\\8
W

 

(3 Fiber
Matrix

Figure 8. ITS Microindenter and Composite Specimen

load was modelled as an applied pressure to the surface of the ﬁber and the results did
not vary appreciably with loading areas ranging from 10 to 60% of the ﬁber cross
section. The maximum shear stress occurred 0.25 to 0.55 ﬁber diameters below the
surface of the specimen. The maximum shear stress was considered a reasonable choice
as the failure criteria as the other interfacial stresses were smaller and compressive in
nature.

The ITS, shown in Figure 9, consists of a Mitutoyo Metallographic FS100
microscope with the stage modiﬁed to house Klinger linear motion stages for x, y and
z translation with 1 pm x, y resolution and 0.04 pm resolution in the z direction.
Mounted on the stages are a Sartorius L610 weighing mechanism and a specimen holder.

Attached to the objective lens is a collar which holds the diamond—tipped indenter, the

21

proﬁle of which is a 90° cone with an included tip radius of 10 um. Stage controllers
and the balance readout are interfaced with a Zenith 386 20MHz microcomputer for
control through the software. Observation and selection of ﬁber ends for testing is done

via a video camera and monitor.

 

Figure 9. The Interfacial Testing System

The specimen used for ITS testing can be any section of composite, provided that
after mounting and polishing, ﬁbers normal to the polished surface can be found. A
section approximately 1 cm x 1 cm was cut from the laminates previously prepared. The
sections were embedded in 2.54 cm diameter phenolic ring mounts with epoxy-based
room temperature cured mounting media, oriented so that the ﬁber ends were normal to
the face of mount to be polished. The initial steps of the polishing were accomplished

on a Struers Abramin polisher and the ﬁnal step on a Beuhler Vibromet I vibrating lap

22

polisher. The polishing used protocol is listed in Table 4.

Table 4. ITS Specimen Polishing Protocol

 

 

 

 

 

Abrasive Lubricant Duration I
320 grit SiC Water 0:30 min
1000 mesh SiC Water 2:00 min
2400 mesh SiC Water 3:00 min
4000 mesh SiC Water 4:00 min
1pm diamond Oil 30:00 min

 

 

 

 

 

After polishing, the specimens were stored in a desiccator until testing. Once mounted
in the specimen holder, the specimen was scanned for ﬁbers that were spaced a minimum
of 2 pm and a maximum of one-half of their diameter from the nearest neighbor. Matrix
conditions near the ﬁber of interest also were considered. Fibers near voids, cracks, or
surface damage or those ﬁbers that showed a pre-existing interfacial crack were rejected.
Once a suitable ﬁber end was located, its diameter and distance to nearest neighbor were
entered for data reduction purposes. The ﬁber tensile modulus and matrix shear modulus
are also required for the analysis. The ﬁber’s coordinates are recorded directly from the
stage controllers to the computer. The Operator begins the test from the keyboard. The
x,y stages move the ﬁber end to a position directly under the tip of the indenter, the 2
stage then moves the sample surface to within 4 pm of the tip. The 2 stage approach is
slowed down to 0.04 pm/ step at a rate of 6 steps/ second. The balance readout is
monitored, at a load of 2 grams the loading is stopped and the ﬁber end returned to the

ﬁeld of view of the camera. The location of the indent is noted and corrections are made

 

 

23

if necessary to center the point of contact. Loading alternating with observation of the
ﬁber-matrix interface is then continued from 4 grams in approximately 1 gram
increments. Debond is determined to have occurred when an interfacial crack is visible
for 90 to 120 degrees on the ﬁber perimeter. The load at which this occurs is used to

calculate the interfacial shear strength.

2.6 COMPOSITE MECHANICAL PROPERTIES

2.6.1 SHORT BEAM SHEAR TESTING. Short beam shear tests (AST M D2344-84)
were conducted on 18 ply unidirectional laminates. Specimens were cut from laminates
using a water-cooled diamond saw. The support-span-to-thickness ratio used was 5 to
1 with a corresponding specimen-length-to-thickness ratio of 7 to l. The specimen width
was ﬁxed at 6.4 mm. The specimen was three point loaded, the two end supports were
set to the span previously determined and the load was applied to the midpoint via a
loading nose. The crosshead speed was 1.3 mm/min. The load at break was recorded
along with the failure mode, i.e. shear or tensile failure. All composite mechanical

testing was conducted on a United Testing System SFM-20 test ﬁxture.

2.6.2 FLEXURE TESTING. Flexure testing (ASTM D790-76) was performed on 12
ply unidirectional laminates with the ﬁbers oriented at 0° and 90° to the support span.
Specimens were 25 mm wide, nominally 100 mm long and tested at a span-to-thickness
ratio of 32 to l. The specimens were three point loaded and the rate of strain in the
outer ﬁbers was kept constant at 0.01 (mm/ mm)/ min. The deﬂection was measured with

a deﬂectometer for determination of stress versus strain curves and the modulus.

 

24
2.6.3 TENSILE TESTING. Tensile testing (ASTM D3039) of 00 and 900 specimens

was performed on 18 and 12 ply unidirectional laminates, respectively. 0° tensile
specimens were 12.5 mm wide, 225 mm long with a 150 mm gage section. Crossply
epoxy/glass tab material was attached to the ends of the specimens with D.E.R.
383/DACH and cured at 40 °C for 4 hours. 90° tensile specimens were 25.4 mm wide,
165 mm long with a 90 mm gage section and tabbed in the same manner as the 0°

specimens. Testing was done at a constant crosshead strain rate of 0.5 %/ min.

2.7 FRACTOGRAPHY

Scanning electron microscopy (SEM) provided high magniﬁcation
photomicrographs of the composite failure surfaces. The failed composite specimens
required trimming, mounting and coating before they were suitable for examination. The
short beam shear specimens were cut through the thickness along the line where the load
was applied which allowed a section to come free that displayed the failure surface. The
section was mounted on a stub with conductive sliver paste and sputter coated with gold
to minimize charging effects from the electron beam. Flexure and tensile specimens
were trimmed to approximately 25mm x 25mm, mounted on stubs and gold coated.
Prior to coating, delaminated material was removed, when necessary, to reveal the

failure surface on the tensile side of ﬂexure specimens.

CHAPTER 3

RESULTS and DISCUSSION

3.1 INTRODUCTION

The interphase of a glass ﬁber-reinforCed epoxy composite was characterized and
its effect on composite mechanical properties was measured. Mechanical properties of
epoxy/ sizing system blends were measured to provide some indication of the properties
of the interphase. Both ﬁber—matrix interphase sensitive mechanical tests (short beam
shear, 90° ﬂexure, 90o tension) and interphase insensitive tests (0° ﬂexure, 0° tension)
were conducted on high volume composite samples fabricated from the same materials
to determine if the interphase and its properties altered the composite mechanical

properties and in what manner.

3.2 EFFECT OF THE SIZING ON FIBER TENSILE STRENGTH
Both the test gage length and the presence of the commercial sizing system had
an effect on the ﬁber tensile strength. The dependence of tensile strength on gage length

2829.30.31.32

of E-glass ﬁbers is known, however, the effect of a sizing system on
strength is not so well characterized.”32 The effect of gage length on tensile strength for
brittle ﬁbers (e. g. glass, boron, graphite) is to increase the strength with decreasing gage

length. The rate of increase is not uniform, decreasing below a certain gage length. In

25

 

26

glass ﬁbers this transition occurs at a gage length of about 1 to 2 cm.28 This is thought
to occur as the result of the presence of two types of ﬂaws, internal and surface, along
the ﬁber. Internal ﬂaws are produced during manufacture of the ﬁbers and limit the
ultimate strength of the ﬁbers. Internal ﬂaws include: bubbles, inclusions and other
defects. Surface ﬂaws are created after manufacture as the result of damage during
handling and processing. The failure of a ﬁber will occur at the "weakest link" present
in the gage length. The " weakest link" is the ﬂaw of greatest severity with internal ﬂaws
usually considered less severe than surface ﬂaws.3O The probability of capturing a
surface ﬂaw in the gage length is dependent upon the spacing between the ﬂaws. A large
gage length will include more surface ﬂaws and reduce the average strength of the ﬁber
at that gage length. Hence the observed increase in the average strength as the gage
length decreases. Figure 10 is a plot of the mean ﬁber tensile strength versus gage
length determined for the bare and sized E-glass ﬁbers used in this study.

A two-sided Student’s t analysis indicates that the differences between the ﬁber
types and gage lengths is signiﬁcant at the 99% level of conﬁdence. The effect of the
sizing system probably occurs through protection of the ﬁber surface. The sizing is
applied immediately after manufacture, before the ﬁbers come in contact with one
another. The bare ﬁbers were treated with deionized water only. The increase in mean
ﬁber strength is about 12% at all gage lengths for the sized ﬁbers. This increase is
similar to that found for E-glass ﬁbers coated with a silane coupling agent alone (12%
versus bare ﬁbers) or a model sizing agent (polyvinylacetate) alone (7% versus bare
ﬁbers). The silane coupling agent in combination with the polyvinylacetate produced an

increase intermediate of either agent used by itself (8% versus bare ﬁbers).3o

27

 

 

 

 

3.00 T I I I I I I I T T I I I I I
T __ ..... B .
2.75 _ . are Fibers -

’3 | '—- Sized Fibers

0.. 2‘.‘

w l \\\\\\

E 2.50 i— i .\'\,\.‘ T. — _.

En f T" ‘-L.\ l

2 2.25 ‘ ......................... l ' 1 a

+3 . ............ _ .

U2 - ""o. i i

Q) -

z 2.00 ~ .- I

”7 :

i: : - _

CD *

9‘ :
1.75 - ; a
1.50 1 l I l i L i i l l l l 1

5 2 3 4 5 6 3
100 101

Gage Length (mm)

Figure 10. E—Glass Fiber Strength Versus Gage Length
3.3 BLEND PHYSICAL PROPERTIES

3.3.1 GLASS TRANSITION TEMPERATURE. Dynamic Mechanical Analysis
provides information on the thermomechanical properties of the epoxy/sizing system
blends as a function of temperature. The Tg of a polymer is qualitatively the point at
which it undergoes a transition from the glassy state to the rubbery state. For purposes
of comparison, the Tg is taken to occur at the inﬂection point in a plot of the storage
modulus (E’) versus temperature. Figure 11 is a plot of Tg determined in this manner
versus the amount of epoxy-compatible size added to the stoichiometric mixture of

D.E.R. 383/DACH on a weight percent basis.

28

 

 

 

5O - l l l l I l l l l
0 5 10 15 2O 25
Percent PPG Sizing (Wt)

 

Figure 11. Tg Versus Sizing System Concentration in D.E.R. 383/DACH

It is clear that the polymer created by the addition of the commercial size exhibits a
monotonic decrease in the Tg. This indicates that the silane coupling agent and other
ingredients in the sizing system are acting to reduce the crosslink density of these
mixtures.33 The silane coupling agent present in commercial sizing systems usually
contains a monofunctional chemical group which is reactive with the matrix constituents

and can terminate network growth.

3.3.2 MECHANICAL PROPERTIES. The tensile properties of the blends: Young’s
modulus, Poisson’s ratio and tensile strength were determined via mechanical testing.
The storage and 0% tangent Young’s modulus determined via DMA or by standard
ASTM tensile coupon and extensometer methods, respectively, were relatively the same.

Toughness, a measure of the material’s ability to absorb energy before fracturing was

29

calculated by integrating the area under the stress versus strain curve.34 Table 5 lists
the values of Young’s modulus (E), storage modulus at 25 °C (E’), tensile strength (a),

crosshead failure strain (6), toughness and Poisson’s ratio (v) for the blends.

Table 5. Mechanical Properties of D.E.R. 383/DACH/Sizing Blends

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sizing Young’s Storage Tensile Failure Toughness Poisson
Concen- Modulus Modulus Strength Strain (MPa) Ratio
tration E E’ o e (v)
(wt %) (GPa) (GPa) (MPa) (%)
0 3.09 3.46 72.9 4.11 1.58 0.41
$0.17 $0.02 $5.0 $0.40 $0.34 $0.03
5 3.38 3.30 71.8 3.39 1.19 0.45
$0.04 $0.02 $0.7 $0.05 $0.20 $0.01
10 3.69 3.39 73.4 3.25 1.16 0.45
$0.06 $0.04 $5.2 $0.35 $0.22 $0.01
15 3.71 3.62 72.3 3.39 1.25 0.43
$0.10 $0.06 $7.3 $0.29 $0.28 $0.01
20 4.01 3.76 78.0 3.33 1.19 0.44
$0.07 $0.03 $4.8 $0.61 $0.21 $0.01
25 3.97 3.82 81.6 3.39 1.36 0.45
$0.12 $0.09 $3.5 $0.28 $0.12 $0.02

 

 

Considering the neat D.E.R. 383/DACH as representative of bulk matrix properties, it
is apparent that adding the sizing system has signiﬁcant effects on the properties of the
matrix in the interphase region. For comparison, Figure 12 presents the Young’s
modulus, tensile strength and toughness versus weight percent sizing added to the
stoichiometric D.E.R. 383/DACH.

The Young’s modulus and tensile strength increase while the toughness decreases

with increasing sizing content. Since all samples were prepared in an identical manner,

30

 

4.1

'9-1

3.7 r

l’i

lil

3.3-:

2.9- l L L l L 1 J_ l
90

 

 

lfl
O

70

l
1 .rﬁ

Tensile Strength (MPa)Young's Modulus (GPa)

 

 

2.1
LBb
1.5‘ 7‘ T

*— i 7 . "ii

.

 

 

Toughness (MPa)

0.9 JL ' t T L l l l i
O 5 10 15 20 25
Percent PPG Sizing (Wt)

 

Figure 12. Blend Physical Properties Versus Sizing Concentration

the results must reﬂect changes in the material properties although the fracture properties
are certainly sensitive to ﬂaws. This may be a similar phenomenon35 to that reported
previously in which difunctional additives with differing reactivities when added to an
epoxy matrix produced increases in tensile strength. Taking the 25% sizing blend as a
model for the interphase near the ﬁber surface, this material would have a lower Tg than
the bulk by about 70°C (176°C vs. 100°C). The Young’s modulus and tensile strength
of the interphase would be increased by 28 and 12% respectively versus the bulk matrix.
The failure strain and toughness would decrease by 18 and 14% respectively. Overall,

the interphase created by the commercial sizing system and the matrix would be

 

 

31

characterized as a stiffer, stronger, lower toughness interphase.

These results suggest that the interphase formed when the sized ﬁber is brought
into contact with the matrix during formation of the composite can have signiﬁcantly
different mechanical properties than the bulk matrix. In an actual composite, the
solubility of the sizing system and matrix components would determine the extent of the
interaction and the resultant composition. There would undoubtedly be a gradient of
properties where the interphase material near the ﬁber surface would have a high
concentration of sizing and low bulk matrix. As the distance from the ﬁber surface
increases, the sizing content decreases until the bulk matrix properties are reached at
some distance away from the ﬁber surface. Currently there is no acceptable method for
quantitatively determining this gradient and distribution in a real interphase, the values
obtained here serve to provide some limits that "bracket" the expected material behavior

of the ﬁber-matrix interphase.

3.4 EFFECT OF THE SIZING ON INTERFACIAL BEHAVIOR

3.4.1 FIBER-MATRIX FAILURE LOCUS. Embedded single ﬁber and ITS tests
provide insight into how modiﬁcation of the interphase by the sizing system affect ﬁber—
matrix failure modes and adhesion. Figure 13 is a polarized transmitted light
photomicrograph of a typical ﬁber break in a D.E.R. 383/DACH/Bare E-glass embedded
single ﬁber coupon. The specimen was strained in a direction parallel to the ﬁber axis
until the ﬁber started to fracture in the coupon. The photomicrograph shows the region

around the break taken immediately after the ﬁber fractured. A small matrix

32

 

Figure 13. Fiber Break at Low Strain, D.E.R. 383/DACH/Bare E—Glass, X67

crack extends outward from the ﬁber into the matrix. The light regions extending
approximately 3 ﬁber diameters into the matrix are photoelastic stress birefringence
patterns. The stress birefringence patterns qualitatively correspond to a decreasing stress
gradient away from the ﬁber. As the coupon was strained further, the nature of the
ﬁber—matrix failure changed. Figure 14 is a photomicrograph of a ﬁber break in
polarized light. The matrix crack has not extended appreciably, however there are
changes at the ﬁber-matrix interface. An interfacial crack has appeared at the ﬁber break
and extends down both sides of the break for about 3 ﬁber diameters. The breadth of
the birefringence patterns has decreased to less than 2 ﬁber diameters and the
photoelastic area extends down the length of the fragment away from the break,

indicating a more uniform stress state along the ﬁber. The ﬁber—matrix failure mode

 

33

changes from matrix cracking to ﬁber-matrix interfacial debonding at the ﬁber break.
The arresting of the matrix crack by the bulk matrix away from the ﬁber surface has

shifted the failure locus to the ﬁber-matrix interface.

 

Figure 14. Fiber Break at High Strain, D.E.R. 383/DACH/Bare E—Glass, X67

This is not the behavior observed in D.E.R. 383/DACH and other epoxy
systems36 when the epoxy-compatible sizing agent is present in the ﬁber-matrix
interphase. Figure 15 is a photomicrograph of a typical ﬁber break in this system using
polarized transmitted light. The initial matrix crack is larger than that observed in the
bare E—glass system, in this case extending on both sides of the ﬁber and forming a
penny shaped crack if viewed end—on. There is no debonding at the ﬁber-matrix
interface. The birefringence pattern includes two bright regions at the tip of the crack,

indicating a high level of stress. At this level of strain there are very few ﬁber breaks

 

34

 

Figure 15. Fiber Break at Low Strain, D.E.R. 383/DACH/Sized E—Glass, X67

in the coupon, three on average. When the specimens were strained past the point where
these initial cracks appeared, the coupon failed due to continued matrix crack propagation
perpendicular to the ﬁber axis. These results indicate that modiﬁcation of the interphase
by the sizing system changes the failure mode. Instead of additional strain energy being
absorbed by interfacial debonding, as in the bare E—glass system, the matrix crack
propagates and results in the failure of the coupon. Recalling the epoxy/sizing system
blend data, the interphase could be reasonably expected to be lower in toughness, leading
to larger matrix cracks under the same stress concentration at the broken ﬁber end. Once

these cracks exceed critical ﬂaw size the test coupons fail catastrophically.

 

35
3.4.2 FIBER-MATRIX ADHESION. Embedded single ﬁber and ITS testing were used

to measure the level of ﬁber-matrix adhesion in model and actual composites. Embedded
single ﬁber results were obtained for bare E—glass ﬁbers only, failure of the test coupons
containing sized ﬁbers occurred after the very ﬁrst breaks in the coupons. The ﬁber
tensile strength results were extrapolated to the critical length, 1., to determine the
interfacial shear strength. Table 6 lists the values measured for the distribution of

fragment lengths and the interfacial shear strength calculated for a ﬁber tensile strength

of 2.34 GPa at 10.

Table 6. Embedded Single Fiber Results for D.E.R. 383/DACH/Bare E—Glass

 

 

Number of Weibull oz Weibull 6 1c Interfacial
Fragments (um) Shear
Strength
(MPa)
294 3.23 28.2 376 54.7

 

 

 

 

 

 

 

This value of interfacial shear strength is similar to that measured in an epoxy/polyamide
system with bare E—glass ﬁbers of 58.1 MPa.32

The Interfacial Testing System (ITS) provides comparative data on the interfacial
shear strengths of the bare and sized E-glass ﬁbers in actual composites. A value of 76
GPa37 was used for the tensile modulus of E-glass ﬁbers and the shear modulus of
D.E.R. 383/DACH was measured as 1.10 GPa. Table 7 lists the mean interfacial shear

strength, standard deviation and number of ﬁber ends tested for the two ﬁber types.

 

36
Table 7. ITS Results for D.E.R. 383/DACH/E—Glass Composites

 

 

 

Interfacial Standard Number of
Shear Strength Deviation Fibers
(MPa) (MPa) Tested
Bare 44.6 3.0 30
ll Sized 60.1 3.3 28 ll

 

 

 

 

The interfacial shear strength (IFSS) increased 35% for the sized ﬁbers versus the bare.
A two-sided Student’s t-test indicates that the difference is signiﬁcant at the 99% level
of conﬁdence. These results indicate that the application of the sizing and the consequent
formation of the interphase increased the level of ﬁber-matrix adhesion. It could
reasonably be expected that under the multi-axial state of stress at the ﬁber-matrix
interphase, stronger ﬁber-matrix adhesion would reduce the tendency to grow an
interfacial crack thereby placing more energy into driving the matrix crack.

The increase in interfacial shear strength is consistent with theoretical mechanical
models of ﬁber-matrix adhesion. Rosen”, Cox”, Whitney and Drzal40 have shown
that the square root of the shear modulus of the matrix appears explicitly in any model
of the interfacial shear strength. It has been demonstrated experimentally“"42 that the
ﬁber—matrix interfacial shear strength has a dependence on both the product of the strain-
to-failure of the matrix times the square root of the shear modulus as well as on the
difference between the test temperature and Tg when the interfacial chemistry is held

constant.

 

 

37
3.5 EFFECT OF THE SIZING ON COMPOSITE PROPERTIES

3.5.1 SHORT BEAM SHEAR. Short beam tests provided information on the apparent
interlaminar shear strength (ILSS) of laminates made with bare and sized E—glass ﬁbers.
In all cases the specimens failed in shear at or near the midplane, allowing comparisons
to be made between ﬁber types. The ILSS, standard deviation and number of specimens

tested are given in Table 8..

Table 8. SBS Results for D.E.R 383/DACH/E—Glass Laminates

 

 

 

 

 

 

 

Fiber Apparent Standard Number of
Type Interlaminar Deviation Specimens
Shear Strength (MPa) Tested
(MPa)
Bare 71 .3 1 .6 9 l
Sized 80.2 2.2 12 ll

 

The increase in the apparent ILSS for the sized versus bare ﬁbers is 12.4%,
approximately 50% of the increase observed in the interfacial shear strength as measured
by ITS testing. The apparent ILSS is certainly sensitive to the macroscopic properties
of the laminates, but changes in ﬁber-matrix adhesion and interphase properties are
expected to inﬂuence the results. Recalling that the ﬁber and void volume fractions are
similar for the SBS specimens, direct comparison between ﬁber types can be made. The
apparent ILSS values of unidirectional composites have been found to rank with ﬁber-
matrix bond strengths,”43 however, changes in the apparent ILSS were not found to be
proportional with changes in ﬁber-matrix adhesion. The short beam shear test will be

sensitive to macroscopic strength limiting defects such as voids and ﬁber misalignments

 

 

38

that do not affect ITS results. Recalling the lower toughness of the model interphase
blends, the increase in adhesion at the interface may be partially offset by a decrease in
the toughness of the interphase. An interphase lower in toughness would provide a path

of lesser resistance for the propagation of defect-initiated intralaminar failure.

3.5.2 90° FLEXURE. The 90° ﬂexural properties were determined using the procedure
described above. Table 9 lists the properties measured for laminates with bare and sized
E-glass ﬁbers. The ﬂexural modulus was not affected by the ﬁber type. The amount
of sizing applied to the ﬁbers is too small to increase the modulus of the bulk matrix

away from the ﬁber and affect the result between ﬁber types.

Table 9. 900 Flexural Properties of D.E.R. 383/DACH/E—Glass Laminates

 

 

 

 

 

 

 

 

Fiber 90° Flexural 90o Flexural 90° Flexural Number of
Type Strength Strain at Modulus Specimens
(MPa) Failure (GPa) Tested
(%)
I Bare 75.6 i 1.8 0.48 -_I— 0.01 15.7 i 0.26 5 I
ll Sized 102 i 5.2 0.69 i 0.05 15.6 i 0.23 5 ll

 

In the 900 direction, the ﬁber-matrix interphase is a controlling factor in the ﬂexural
strength. Failure of the test specimens occurred on the tensile surface, forming a hinge
crack that did not extend though the specimen. Here a 35 % increase in flexural strength
is measured in direct proportion to the increase in interfacial shear strength measured
with the ITS tests. This indicates an increase in the transverse tensile strength of the
ﬁber-matrix interphase. Apparently the lower toughness of the interphase and its

tendency to create matrix cracks and not debond under shear loading does not affect the

 

 

39

interphase when it is loaded in transverse tension. An increase in the strain to failure of
the sized versus bare ﬁber laminates is not paralleled by the results of the model
interphase. Increased ﬁber-matrix adhesion would limit the formation of ﬁber—matrix
interfacial cracks that would lead to premature failure of the specimens and thereby

increase the strain to failure.

3.5.3 0° FLEXURE. Flexural properties in the 0° direction are largely determined by
the ﬁber properties; however, the matrix or interphase may control the failure mode or
path which may in turn affect the strength of the composite. The 0° ﬂexural properties

of the composites made with bare and sized E—glass are listed in Table 10.

Table 10. 0° Flexural Properties of D.E.R. 383/DACH/E-Glass Laminates

 

 

 

 

0° Flexural 0° Flexural 0° Flexural Number of
Fiber Strength Strain at Modulus Specimens
Type (MPa) Failure (GPa) Tested
(%)
Bare 1010 -l_- 60 3.0 i 0.2 39.5 :1; 1 3 6
ll Sized 1260 i 50 3.5 i 0.2 44 2 i 3 3 6 ll

 

 

 

 

 

Laminates with either bare or sized E—glass ﬁbers failed ﬁrst on the tensile surface
of the test specimen but the nature of the event differed. Figures 16 and 17 are
photographs of the tensile surfaces and edges of failed test coupons that were fabricated
with bare E-glass. Figure 16 illustrates the breakage of ﬁber bundles which then
propagated delaminations along the tensile surface of the specimen. This process

occurred over the last 10 to 15 seconds of the test, and failure occurred gradually rather

than catastrophically. The lack of penetration of the failure zone into the interior of the

 

4O

 

 

I l"II '|"'1] III I'il I3I' I""'|4'I”'IIII ISIIIITITI'EII
I “IIIIIIIII'IIIIIIIIIIIIIllIflIIIIlIIﬂIIIIIIIIIIIIIIIIIIIIIIIIIIII|IIII|IIII|IIIIIIIII IIIIIIIlIIIIIIIIlIIIIlIIIlIIIm IIIIlII III!

I

 

 

 

Figure 16. 0° Flexure Specimens, Tensile Side, D.E.R. 383/DACH/Bare E-Glass

 
 

i-E. .

8.! ~ V“
a‘ifw Q’s-o. -‘ WA ,_
II '
in
il E

Figure 17. 0° Flexure Specimens, Edge View, D.E.R. 383/DACH/Bare E—Glass

 
    
 

‘.d I. -*

  

41

specimens, Figure 17, indicates that strain energy was being absorbed by delamination
rather than the growth of matrix cracks into the specimen.

The failure behavior of the laminates made with sized E—glass ﬁbers was different.
Figures 18 and 19 are photographs of the tensile surfaces and edges of failed specimens.
The specimens failed catastrophically through the thickness, and the delaminations shown
in Figure 18 occurred during the failure, not before. Regions of tensile failure (ﬁber
pull-out) and compressive failure (crushing) can be seen in Figure 19. The change in
failure at the single ﬁber level as shown by the embedded single ﬁber test is mirrored

by the failure behavior at the macroscopic level.

 

  

Il'l"'|"'|'1I'l'|"ZI'I'I'IQI'I'III‘III'“VIE”.

iii A ,

     

 

Figure 18. 0° Flexure Specimens, Tensile Side, D.E.R. 383/DACH/Sized E—Glass

 

 

Figure 19. 0° Flexure Specimens, Edge View, D.E.R. 383/DACH/Sized E—Glass

The increase in the 0° flexural strength is 25% for the sized versus bare ﬁber
laminates. This can be partially attributed to the measured 12% increase in the tensile
strength of sized versus bare ﬁbers. Reduction of the 0° ﬂexural strength of the sized
ﬁber to account for the increase in ﬁber strength yields a value of 1,121 MPa. The
difference between the bare and sized ﬁber laminate 0° ﬂexural strength is reduced to
11%. The contribution of the ﬁber tensile strength difference may be lost during
production of the laminates, so the increase in 0° flexural strength due to the sizing
system is likely between 11 and 25%. The change in failure mode from delamination
to through—the-thickness failure indicates that the arresting of delaminations is an
important modiﬁcation afforded by the use of a sizing system. The increase of the strain

at failure of sized versus bare ﬁber laminates is the result of the increase in strength.

 

43

3.5.4 90° TENSILE. Differences in the 90° tensile properties of laminates made with
bare and sized E—glass ﬁbers were noted, however, quantifying these differences was
made difﬁcult by specimen failure at or in the grips. Table 11 lists the 90° tensile

strength and crosshead strain at failure and modulus.

Table 11. 90° Tensile Properties of D.E.R. 383/DACH/E—Glass Laminates

 

 

 

 

 

 

 

 

Fiber 90° Tensile 90° Tensile Tangent Number of
Type Strength Strain at Modulus of Specimens
(MPa) Failure Elasticity Tested
(%) (GPa)
I Bare 23.5 i 2.2 1.1 i 0.2 1.92 i .10 6 I
II Sized 37.0 i: 4.8 2.1 i 0.2 2.06 :1; .18 6 II

 

If the stress concentrations due to the tabs and grips is taken to be constant between all
specimens, then these values can still be used for comparisons between ﬁber types. The
increase in 90° tensile strength corresponds to the change in the level of ﬁber-matrix
adhesion measured via ITS testing. The complex state of stress introduced by the grips
makes the difference in strength impossible to explain in terms of loading at the single
ﬁber level. No effect on the modulus of the laminates was noted, the amount of sizing
system introduced into the interphase is too small to effect the bulk matrix properties.
No differences in the macroscopic failure mode was noted between the bare and sized

ﬁber specimens.

3.5 .5 0° TENSILE. Specimen failure at or in the grips also occurred for the 0° tensile
specimens. Table 12 lists the properties determined for the bare and sized ﬁber

laminates.

44

Table 12. 0° Tensile Properties of D.E.R. 383/DACH/E-Glass Laminates

 

 

 

 

Fiber 0° Tensile 0° Tensile Tangent Number of
Type Strength Strain at Modulus of Specimens
(MPa) Failure (%) Elasticity Tested
(GPa)
Bare 580 i 33 6.2 i 0.3 11.5 i 0.5 4 I
II Sized 944 -_I- 28 9.2 i 0.4 11.8 i 0.1 5 II

 

 

 

 

 

If the stress concentration due to the tabs and grips is taken to be constant between all
specimens, then these values can still be used for comparison between ﬁber types. The
properties in the 0° direction are expected to be ﬁber dominated, as is the case for the
modulus. However, the increase in the strength is greater than the single ﬁber tensile
strength increase for bare versus sized ﬁbers of 12%. The complex stress state in the
grips may introduce ﬁber-matrix interfacial cracks that reduce the bare ﬁber laminate
No change in the macroscopic failure behavior between ﬁber types was

strength .

observed. The increase in failure strain is a result of the increase in laminate strength.

3.6 EFFECT OF THE SIZING SYSTEM ON COMPOSITE FRACTOGRAPHY

3.6.1 SHORT BEAM SHEAR. Differences in the microscopic failure surfaces of bare
and sized ﬁber short beam shear specimens were noted. A scanning electron microscope
(SEM) was used to examine the failure surfaces of specimens prepared as discussed in
Chapter 2. Figure 20 is a SEM photomicrograph at X1,000 of a SBS coupon made with
bare ﬁbers. The surfaces of the ﬁbers are smooth and appear free of matrix. The lack

of matrix adhering to the ﬁbers is interpreted as an indication of a weak

45

y
I S
I

[SKU X1,BBB

 

Figure 20. SBS Failure Surface, D.E.R. 383/DACH/Bare E—Glass

ﬁber-matrix interface.44 The macroscopic failure of the specimen was interlaminar,
conﬁned to a plane between plies. SEM examination revealed that the failure on the
microscopic level was also contained to a plane and no broken ﬁbers were observed.
This indicates that the failure occurred in pure shear, corroborated by experiments
designed to load the interface in pure shear or mixed modes45 in glass/epoxy
composites. The macroscopic failure of the SBS specimens with sized ﬁbers was the
same as the bare ﬁber specimens. Figure 21 is a SEM photomicrograph of the failure
surface of a sized ﬁber SBS specimen. There is a layer of matrix adhering to the ﬁbers
which is a qualitative sign of good ﬁber-matrix adhesion.46 The planar microscopic
failure surface and the hackled appearance of the matrix are considered indications of

shear failure in the matrix.45 The resence of the sizin has increased the
P g

 

 

46

lg.

15KU XlaBBB

 

Figure 21. SBS Failure Surface, D.E.R. 383/DACH/Sized E-Glass

ﬁber—matrix adhesion as characterized by microscopic examination of the failure surfaces

and correlates with the increase in ﬁber—matrix interfacial shear strengths.

3.6.2 0° FLEXURE. The macroscopic failure of 0° ﬂexure specimens made with bare
ﬁbers was characterized by delaminations on the tensile surfaces of the coupons. Figure
22 is a SEM photomicrograph of a representative microscopic failure surface of a bare
ﬁber specimen. The surfaces of the ﬁbers are devoid of matrix, indicating poor adhesion
and failure at the ﬁber—matrix interface. The fact that the delaminations occurred in
overlying planes combined with the appearance of the matrix indicate a mixed failure
mode of shear + tension.45 The macroscopic failure of 0° ﬂexure specimens made with

sized ﬁbers was through the thickness,

 

 

47

 

 

Figure 22. 0° Flexure Failure Surface, D.E.R. 383/DACH/Bare E—Glass

regions of ﬁber pull-out and crushing could be observed. Figure 23 is a X35 SEM
photomicrograph of a section through the thickness of a failed specimen. The neutral
axis runs from A to B, above this line the specimen failed in tension (ﬁber pull—out),
below this line compression. Figure 24 is a X1,000 view of ﬁber surfaces on the
compressive side. The appearance of the matrix adhering to the ﬁbers is characteristic
of mixed mode shear + compressive failure at the surface.45 Fibers on the tensile side
of the neutral axis are shown in Figure 25. The ﬁbers are covered with a layer of
matrix, indicating increased adhesion between the ﬁber and matrix and a shift in the
failure locus to the interphase. The pull-outs combined with the appearance of the matrix
are characteristic of mixed mode shear + tensile failure at the surface.45 Inspection of

the failure surfaces of bare and sized ﬁber 0° ﬂexural specimens has revealed qualitative

 

48

III
. / 1'
.13" °‘

f 3,.ﬁ/
51‘. (r.-. ’\
15KU X1,806 109m 519125

 

Figure 24. 0° Flexure Failure Surface, Compressive Side, D.E.R. 383/DACH/Sized E-
Glass

 

49

"C

SKU X1:BBB

 

Figure 25. 0° Flexure Failure Surface, Tensile Side, D.E.R. 383/DACH/Sized E-Glass

increases of the adhesion between ﬁber and matrix of sized ﬁber specimens. The failure
locus shifted from the ﬁber—matrix interface to the interphase for both mixed mode
loading conditions of shear + tension and shear + compression for the sized ﬁber

specimens.

3.6.3 90° TENSILE. The failure of the 90° tensile specimens was inﬂuenced by the
stress state introduced when the failure occurred at or in the grips. The complex stress
state was revealed in SEM photomicrographs of both the bare and sized specimens.
Figures 26 and 27 are SEM photomicrographs at X1 ,000 of 90° tensile specimens made
with bare and sized ﬁbers, respectively. The specimen made with bare ﬁbers has very

little matrix adhering to the ﬁbers, again indicating poor ﬁber—matrix adhesion. The

 

X1,898

\\\M\\\\\\\\K \
\W‘ ”V"? . . "I/ '

 

Figure 27. 90° Tensile Failure Surface, D.E.R. 383/DACH/Sized E-Glass

 

51

failure surface is not conﬁned to a plane and the matrix in this region has widely spaced
hackles. This indicates that this surface failed both in tension, as expected from the
loading, and shear, possibly introduced by the grips. The specimen made with sized
ﬁbers has matrix adhering to the ﬁbers, indicating an increase in adhesion and a shift in
the failure from the interface to the interphase. The non-planar failure surface and

hackled matrix again indicate a mixed mode of failure of tension + shear.

3.6.4 0° TENSILE. The failure of the 0° tensile specimens also occurred at or near the
grips. Figures 28 and 29 are SEM photomicrographs at X1,000 of the failure surfaces
of specimens made with bare and sized ﬁbers, respectively. The bare ﬁber specimen has
clean ﬁber pull-outs, again indicative of poor ﬁber—matrix adhesion. The failure path
traveled along the ﬁber-matrix interface, similar to the behavior observed in the
embedded single ﬁber test. Sized ﬁber specimens also reﬂected the behavior observed
in the embedded single ﬁber test; the failure surface is relatively smooth with blocks of
ﬁber and matrix at different levels. If the large matrix cracks observed in the single
ﬁber test occurred at ﬁber breaks in the tensile test coupon, the failure surface would
tend to be conﬁned to a plane perpendicular to the ﬁber’s long axis. The level of
adhesion and failure behavior at the ﬁber-matrix interface have again inﬂuenced the

failure at the microscopic level.

 

52

.,' Il""»~"1\

IOKU X1,BOO 10H":

 

Figure 28. 0° Tensile Failure Surface, D.E.R. 383/DACH/Bare E—Glass

3:715?

IBKU X1,BBO 10”!!! 12?115

 

Figure 29. 0° Tensile Failure Surface, D.E.R. 383/DACH/Sized E-Glass

 

CHAPTER 4

CONCLUSIONS

It has been demonstrated that a sizing system’s interaction with the bulk matrix
in a glass ﬁber-reinforced epoxy matrix composite has effects at three levels: molecular,
micrOSCOpic and macroscopic.

At the molecular level, modelling of the interphase via bulk measurements of
epoxy/sizing system blends has shown that the sizing system can signiﬁcantly alter the
physical properties of the interphase. An epoxy blend containing 25 % sizing system used
to model the interphase exhibited increases in the Young’s modulus and tensile strength
of 28% and 12% respectively. The toughness of the model interphase was found to
decrease by 14%. The Tg of the blend decreased by 72 °C versus the bulk epoxy.

At the microscopic level, embedded single ﬁber and ITS testing revealed effects
of the sizing system on level of ﬁber—matrix adhesion and failure mode. The presence
of the sizing system increased the level of ﬁber—matrix adhesion by 35% and shifted the
locus of failure from interfacial debonding to matrix cracking. The change in adhesion
can be attributed to the increased modulus of the sizing system-matrix interphase and
chemical bonding through the silane coupling agent. The increased level of adhesion and
decreased toughness of the interphase combine to shift the ﬁber-matrix failure locus.

At the macroscopic scale of composite test Specimens, the changes in the

53

 

54

interphase primarily affected the failure behavior of ﬁber dominated tests (0° ﬂexure, 0°
tension) and affected both the failure behavior and mechanical properties of interphase
sensitive composite tests (short beam shear, 90° ﬂexure, 90° tension). All interphase
sensitive tests measured increases in the strength values and no changes in the moduli.
Macroscopic failure of the specimens was identical between bare and sized ﬁber
specimens. Interphase insensitive tests showed effects on the physical properties due to
the presence of the sizing system; but the properties were still largely ﬁber dominated.
The macros00pic failure of 0° tensile specimens did not change with ﬁber type, however,
the 0° ﬂexure specimens changed from delamination to through-the—thickness failure.

SEM examination of failure surfaces indicated qualitative increases in ﬁber-matrix
adhesion in all stress states, and a shift in the failure path from interfacial to the

interphase for the sized ﬁber specimens.

 

 

LIST OF REFERENCES

 

 

10.

11.

12.

13.
14.

15.

LIST OF REFERENCES

E. P. Plueddemann, Silane Coupling Agents, 2nd ed., (New York: Plenum Press,
1990), 86.

Ibid., 116.
Ibid., 55.
Ibid., 35.

C. H. Chiang, H. Ishida and J. L. Koenig, J. Colloid Interface Sci. 74
(1980): 396.

K. P. Hoh, H. Ishida and J. L. Koenig, Polymer Compgsites 11, no. 2
(1990): 121.

H.-J. Kang, W. Meesiri and F. D. Blum, Materials Science and
Engineering A126 (1990): 265.

D. J. Pawson and F. R. Jones, in: Controlled Interphases in Composite Materials,
ed. H. Ishida (Amsterdam: Elsevier, 1990), 407.

Plueddemann, 101.

H. Ishida and J. L. Koenig, Journal of Polymer Science: Polymer Physics
Edition 17 (1979): 615.

 

H. Ishida, in: Molecular Characterization of Composite Interfaces vol. 27,
Polymer Science and Technology, ed. H. Ishida and G. Kumar (New York:
Plenum Press), 25.

Dow Corning Corporation, A Guide to Dow Corning Silane Coupling Agents
(1985), 15.

Plueddemann, 87.
Ibid., 147.
Ibid., chapter 5.

55

 

 

15.
16.

17.

18.
19.

20.

21.

22.
23.

24.

25.

26.

27.

28.

29.

30.

31.
32.

56
Ibid. , chapter 5.

L. T. Drzal, SAMPE 19 (1983): 7.

L. T. Drzal, in: Advances in Polymer Science II vol. 75, ed. K. Dusek
(Spring-Verlag: 1985)

M. C. Waterbury and L. T. Drzal, J. Reinf. Plast. & Composites 8 (1989): 627.

American Society for Testing and Materials, Annual Book of ASTM Standards,
vol. 8.01 (Philadelphia: ASTM, 1988), 167.

L. T. Drzal, M. J. Rich, J. D. Camping and W. J. Park, Proceedings of the
35th Annual Technical Conference. Reinforced Plastics/ Composites Institute, (The
Society of the Plastics Industry, Inc.: 1980), 1.

L. T. Drzal and P. Herrera-Franco, in: The Engineered Materials Handbook:
Adhesives and Sealants vol. 3 (ASM Intl., 1990), 391.

A. Kelly, Proc. Royal Soc. A. 319 (1970): 95.
A. C. Cohen, Jr., Technometrics (7): 579.

D. L. Caldwell, D. A. Babbington and C. F. Johnson, in: Interfacial Phenomena
in Composite Materials ’89 ed. F. R. Jones (Butterworth, 1989), 44.

D. L. Caldwell and F. M. Cortez, Modern Plastics (September 1988): 132.

T.‘-H. Tsiang, Damage Development in Fiber Composites Due to Bearing,
ScD Dissertation, M.I.T. (1983).

D. H. Grande, Microdebonding Test for Measuring Shear Strength of
Fiber/Matrix Interface in Composite Materials, M.S. Thesis, M.I.T. (1983).

R. A. Larder and C. W. Beadle, J. of Composite Materials 9 (July 1975): 241.

C. Zweben, W. S. Smith and M. W. Wardle, Composite Materials: Testing
and Design (Fifth Conference) ASTM STP 674, ed. S. W. Tsai (American
Society for Testing and Materials, 1979), 228.

W. A. Fraser, R. H. Ancker, A. T. DiBenedetto and B. Elbirli, Polymer
Composites 4, no.4 (October 1983): 238.

K. K. Phani, J. Appl. Phys. 62, no.2 (15 July 1987): 719.

A. T. DiBenedetto and P. J. Lex, Polymer Engineering and Science 29, no. 8
(April 1989): 543.

 

 

33.

34.

35.
36.

37.

38.

39.
40.

41.
42.
43.
44.
45.

46.

57

E. A. Turi, Thermal Characterization of Polymeric Materials (Academic
Press Inc., 1981), 515.

R. W. Hertzberg, Deformation and Fracture Mechanics of Engineering
Materials (New York: John Wiley & Sons Inc., 1976), 27.

A. Garton and G. S. Haldankar, J. Adhesion 29 (1989): 13.

A. N. Netravali, P. Schwartz and S. L. Phoenix, Polymer Composites 10 no.6
(1989): 385.

D. Hull, An Introduction to Composite Materials (Cambidge: Cambridge
University Press, 1987), 14.

B. Rosen, in: Fiber Composite Materials chap. 3 (American Society for Metals,
1964), 37.

H. L. Cox, Brit. J. Appl. Phys. 3 (March 1952): 72.

J. M. Whitney and L. T. Drzal, in: Toughened Composites STP 937
(Philadelphia: American Society for Testing Materials, 1987), 179.

V. Rao and L. T. Drzal, Polymer Composites (in press) (1991).

L. T. Drzal, Mater. Sci. Eng. Trans ASME A126 (1990): 289.

Hull, 48.

Ibid., 54.

A. Voloshin and L. Arcan, J. of Composite Materials (July 1979): 240.

Hull, 55 .

 

 

 

NICHIGQN STRTE UNIV L

I II

III III III
412EBL09093

IIIIII'
471

 

 

...... ......
.-.I...,