i

.-. "fl
9

 

*1

w: u»
gn'ofi

‘
LP?

.9
‘

39??

, .3»-

; ., -
~.:§~?.p.v

~.,-.
‘i

Mi?

.4
V ‘ ‘( .a w
5.7 ‘A-sz- .
V -;V‘;'A-g-1L—
$1.3..u“ ‘

-i

312:

“r
I.

‘fié

2.3

< « mi. I
v rev-r.
«-

 

..i a,,..

V , . . . 1‘ . . 5 1-“.-
“J'ur r 'v . u - . ’ , ‘ ‘ {1‘5"
a» Av _ A 7 ‘31 3.: z? .
‘ r - 1‘" ’3

‘45: .. . "
#33:"?!

£1.

 

1 , J
‘ g 1, ”M .e"
13:34:58. 31““: V > i=2.“
‘ may: {r I; 5““.
‘H I. '1'... ".‘ESC‘EI 1*
m}: ,. ~32».
mwfiu w
..,5,-,~.:...v,e...Ԥ .3
4.743,. 0'1".“ "2;; N... a. a".— W
53., :3 :9 . ,r.-. sung-,-
2 ‘_ ‘ - aggné‘mzdmw
. '. -. “v“ 93m

{JR

~...<4-...—
-.<
-

§
.' ,4
‘9?!“
\ :t “ -
‘Jl’tLKjW’ ‘ - ‘ ‘3‘ " \x' ‘1 'x'a A u A "1'.
"‘«1 " ”xx.“ . ' ‘" " ‘2-‘43-"9'7'3Ev'353'
' 4 ' ‘ " .7

J
x
‘-
’,.
m, 'i
. V‘L ,
9: . .XA 9
“way;- I -
J

1}:

 

was

ll|lll|lll||lllllllllllllllllllllllllll‘Illllllllll

3 1293 01020 1493

This is to certify that the '
thesis entitled. '-
A Water Based Manufacturing Method
For Two Phase Matrix Composite Materials
presented by

James Hector Fernandes

has been accepted towards fulfillment
of the requirements for

Masters degree in Chemical Engineering

Major professor

 

 

Date 7/12/94

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

 

 

LIBRARY

M'Chlgan State
University

 

 

 

PLACE N RETURN BOX to man this chock" from your mood.
To AVOID FINES Mum on or More data duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An AffinnJativo Action/EM Opportunity Insulator!

W1

 

A WATER BASED MANUFACTURING METHOD
FOR TWO PHASE MATRIX COMPOSITE MATERIALS

By

JAMES HECTOR FERNANDES

A THESIS

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

MASTER OF SCIENCE

Department of Chemical Engineering
Composite Materials and Structures Center

1 994

ABSTRACT

A WATER BASED MANUFACTURING METHOD
FOR TWO PHASE MATRIX COMPOSITE MATERIALS

By

J ames Hector Fernandes

A novel processing technique for the manufacture of continuous fibers with a two
phase matrix from a well-mixed aqueous colloidal dispersion bath was developed. In this
study, carbon fibers were spread using acoustic energy and then coated with the two
particles, i.e. an epoxy resin and an elastomer, by passing through a bath of the mixture.
The prepreg tape was then consolidated into a uni-directional multi-layered part, under
vacuum in an autoclave and under a specific consolidation cycle. Several composite parts
were consolidated by preparing prepreg tape with different ratios of epoxy resin particles
to elastomeric particles, by changing the concentration of the elastomer to the epoxy resin
in the dispersion bath. The parts were subjected to mechanical testing and morphological

study, and the results indicated two distinct phases in the composites.

- I Tomy‘nate.Fathen-,.Mummy, . °
. Sweetie; Sana& Julie

iii

ACKNOWLEDGEMENTS

My journey down the meandering, and at times tortuous, path of graduate
research work could not have been completed without the selfless assistance of co-
workers and associates. In the foremost stands Dr. L. T. Drzal, my graduate advisor,
under whose didactic tutelage and continuing support I was able to complete all of my
research work with complete freedom, in a timely and advantageous manner. I would
like to extend my gratitude to Mr. M. J. Rich and Mr. Brian Rook of the CMSC lab for
all the help that they gave in training me on different lab equipments. I would like to

thank Mr. Terry Casey and Mr. Mike of the machine shops for their timely cooperation
and guidance in the mechanical set-up of my process. I would like to thank Julie David
for all the assistance she gave me in taking SEM pictures and in the presentation for my
oral defense. Besides this I would like to take this opportunity to thank all of my co-
workers for the congenial atmosphere that they created around me to make my stay at

MSU pleasurable and rewarding.

iv

TABLE OF CONTENTS

List of Tables

List of Figures

Introduction

Process Description
2.1 Introduction

2.1.] Slurry Processing
2.1.2 Aqueous Foam Processing

2.2 Proposed Process

2.2.1 General Description
2.2.2 Fiber Motion System
2.2.3 Spreader Unit

2.2.4 Dispersion bath
2.2.5 Heating section

2. 3 Process Automation

2.3.1 Fiber Motion
2.3.2 Oven Temperature

2.4 Supplementary Equipment

2.4.1 Prepreg Tape Winder

2.4.2 Autoclave Consolidation

2.4.3 Volume/Void Fraction Analysis
2.4.4 Mechanical Testing

ix

13
15
15

l7

17
19

l9

19
21
21
23

vi
3. Material Properties
3.1 Introduction
3.2 Fiber Reinforcement
3. 2. I Magnamite° Carbon Fiber Type 1M7
3.3 Matrix
3. 3.1 Colloids

3.3.1.1 Electrical Double Layer Stabilization
3.3.1.2 Steric Stabilization

3. 3. 2 Material Selection
3.3.2.1 Waterborne Epoxy Dispersion
3. 3. 3 Epi-RezO W60-3515 Epoxy Resin Dispersion
3.3.3.1 Performance Highlights/Features
3.3.3.2 Uses
3.3.3.3 Typical Properties
3.3.3.4 Curing Agents
3. 3. 4 Carboxylated Styrene Butadiene Rubber (CSBR) Latex
4. Sample Preparation
4.1 Introduction
4.2 Dispersion Bath Preparation

4.2.1 Matrix Dispersion Preparation
4.2.2 Dispersion Dilution

4. 3 Prepreg Tape Preparation
4.3.1 Processing

4.3.2 Mechanism for Coating
4.3.3 Heating

25

25

25

26

26

26
29

29
29
31
31
32
32
32
34
39
39
39

39
40

43
43

45
49

vii

4.4 Composite Manufacturing Process 51
4. 4.1 Prepreg Tape Winding 51

4.4.2 Autoclave Molding 55

4.5 Sample Cutting 60
4.5.1 Sample Preparation for Volume/Void Fraction Analysis 60
4.5.1.1 Casting of Samples in Plastic Mold 60

4.5.1.2 Polishing of Sample Surface 64

4.3 Summary 65
5. Process Modelling 66
5.1 Introduction 66
5.2 Process Model Development 66
5. 2. 1 Process Data 68

5. 2.2 Development of Equation 69

5.2.3 Experimental Evaluation of Heat Transfer Coefiicient 70

5. 2. 4 Moisture Removal in Convection Heater 76

5.2.5 Moisture Removal by IR Lamps 78

6. Results and Discussions

6.1 Introduction 80
6.2 Volume/Void Fraction Analysis 81
6.3 Mechanical Testing 82
6.3.1 Three-Point F lexure Test 82
6. 3. 2 Short Beam Shear Test 86
6.3.3 Fracture Energy Analysis 88
6. 3. 4 Discussion of Mechanical Testing Results 90
6.4 Dynamic Mechanical Analysis (DMA) 92
6.5 Morphological Study 97

6.6 Summary 103

7.

viii

Conclusions and Future Work

7.1 Conclusions

7.2 Future Work
Appendices

A Method for Winder Speed Evaluation
Autoclave Program # 7
Program to Evaluate Temperature Profile in Slab due Convection
Flexure Test Data

ENF Testing Method

mmunw

Compliance Calculation Data Sheet
G EN F Test Data

Bibliography

104

104

105

107

109

110

112

115

120

121

124

Table 1.1
Table 3.1
Table 3.2
Table 4.1
Table 4.2

Table 5.1
Table 6.1
Table 6.2

LIST OF TABLES

Comparison of matrices

Typical physical properties of 1M7 12k carbon fiber tow

Typical properties of CSBR

Sample dimension for 3-point flexural testing method for ASTM D790
Sample dimension for short beam shear testing method as per ASTM
D2344

Composition of prepreg tape prior to heating

Volume/ void fraction analysis

Flexural properties of composites with varying CSBR concentration

ix

Figure 2.1
Figure 2.2
Figure 2.3
Figure 2.4
Figure 2.5
Figure 2.6
Figure 2.7
Figure 2.8
Figure 2.9
Figure 2.10
Figure 2.11
Figure 2.12
Figure 2.13
Figure 2.14
Figure 2.15
Figure 3.1
Figure 3.2
Figure 3.3
Figure 3.4
Figure 3.5
Figure 3.6
Figure 3.7
Figure 3.8

LIST OF FIGURES

Production of prepreg tape using polymer slurry
Process schematics for foam towpregging
Schematic of aqueous colloidal dispersion prepregging process
DC gearrnotor

Pinch rollers

Take-up winder

Spreader unit

Dispersion bath

Heater

Circuit diagram of variable speed controller
Calibration graph for speed controller

Circuit diagram for convection oven temperature controller
Schematic for frame

Autoclave

Equipment for volume fraction/ void analysis
Model for electrical double layer stabilization
Model for steric stabilization

Theoretical his A epoxy molecule

DSC analysis of Epi-Rez W60-3515

TGA analysis of Epi-Rez W60-3515

DSC analysis of CSBR

TGA analysis of CSBR

Fractograph of F85/ graphite composite sample

Figure 4.1

Figure 4.2
Figure 4.3
Figure 4.4
Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
Figure 4.9
Figure 4.10
Figure 4.11
Figure 4.12

Figure 4.13

Figure 5.1
Figure 5.2

Figure 5.3

Figure 6.1
Figure 6.2
Figure 6.3
Figure 6.4
Figure 6.5
Figure 6.6

Figure 6.7
Figure 6.8
Figure 6.9
Figure 6.10

x1
Coating of 12k 1M7 carbon fibers with EpiRez W60-3515 at different
dilution levels
Schematic illustration of the various degrees of wetting
Coating versus time
SEM photographs of prepreg tape
TGA analysis for prepreg tape
Photographic representation of winding system
Vacuum bag molding schematic
Vacuum bagging arrangement in autoclave
Consolidation cycle
Diamond saw
Sample dimension for three-point flexure and short beam shear test
Sample dimensions for ENF testing
Polished sample for volume/ void fractions analysis. Composite part
embedded in plastic mold
Heating arrangement in convection oven
Experimental set-up for evaluation of heat transfer coefficient
Experimental evaluation of average heat transfer coefficient:
temperature versus time
Micrograph of polished composite surface at magnification of 65X
Arrangement for three-point flexure testing
Flexural strength of composite with varying CSBR concentration
Flexural modulus of composite with varying CSBR concentration
Arrangement for short beam shear strength analysis
Short beam shear strength data for varying CSBR concentrations in
composite samples
Fracture energy of composite parts with varying CSBR concentration
DMA analysis of 0% CSBR composite part
DMA analysis of 5% CSBR composite part
DMA analysis of 10% CSBR composite part

Figure 6.11
Figure 6.12
Figure 6.13
Figure 6.14
Figure 6.15
Figure 6.16
Figure 6.17
Figure 6.18
Figure 6.19
Figure 6.20
Figure 6.21
Figure 6.22
Figure D.1
Figure D.2
Figure D.3
Figure D.4
Figure G.1
Figure G.2
Figure 6.3
Figure G.4
Figure G.5

xn
DMA analysis of 15% CSBR composite part
DMA analysis of 20% CSBR composite part
ENF fracture surface for 0% CSBR composite sample at 5,400X
ENF fracture surface for 0% CSBR composite sample at 10,000X
ENF fracture surface for 5 % CSBR composite sample at 5,400X
ENF fracture surface for 5% CSBR composite sample at 10,000X
ENF fracture surface for 10% CSBR composite sample at 5,400X
ENF fracture surface for 10% CSBR composite sample at 10,000X
ENF fracture surface for 15 % CSBR composite sample at 5,400X
ENF fracture surface for 15 % CSBR composite sample at 10,000X
EN F fracture surface for 20% CSBR composite sample at 5,400X
EN F fracture surface for 20% CSBR composite sample at 10,000X
Three-point flexural test data for 5 % CSBR composite samples
Three-point flexural test data for 10% CSBR composite samples
Three-point flexural test data for 15% CSBR composite samples
Three-point flexural test data for 20% CSBR composite samples
ENF data for 0% CSBR composite samples
ENF data for 5% CSBR composite samples
EN F data for 10% CSBR composite samples
ENF data for 15% CSBR composite samples
ENF data for 20% CSBR composite samples

Chapter 1

 

INTRODUCTION

 

A composite material may be defined as a substance comprised of two or more
identifiable constituents. The individual constituents in a composite material are
combined in such a way that the final composite part has superior properties than the
individual components. Composites are both natural and man-made. These materials
can be categorized into the following groups:

a) fibrous composites consist of either short, long or continuous fibers in a matrix.

b) laminated composites consist of layers of different materials.

c) particulate composites consist of particles of one or more materials enveloped in
a matrix.

In this dissertation we will confine our study to fibrous composites. Some of the
more popular fiber types are carbon, glass, and Kevlar (aramid) fibers. Fiber diameters
vary from 5 to 15 microns. The matrices associated with such composites are usually
of two types:

a) Thermoplastics: They are one- or two-dimensional in molecular structure and
show a discrete melting point. In these compounds the process of softening and

regaining rigidity upon cooling is a reversible onem e.g. polyethylene, polystyrene,

2
polypropylenes, polyamides, and nylon.

b) Thermosets: They form a well-bonded three-dimensional molecular structure upon
curing. Once cross-linked or hardened, these polymers will decompose rather than
melt. e.g. epoxy, polyester, phenolic polyimide resins.

Table 1.1 compares the properties of thermoset and thermoplastic matrices for
composites.

Table 1.1 Comparison of matricesm

 

PROPERTY THERMOSET THERMOPLASTIC

 

Weight +
Material Cost +
Processing Cost-Reduction

Potential
Simplicity of Chemistry +
Melt Flow +
Prepreg Tack and Drape ' +
Long Prepreg Shelf Life +
Low Processing Temperature
Low Processing Pressure
Low Processing Cycle
Low Cure Shrinkage
Quality Control Data Base
Mechanical Property Data Base
Ability to Translate Fiber Properties
Solvent Resistance
Corrosion Resistance
Resilience
Toughness +
Lack of Time-Dependence
Interfacial Adhesion
Low Thermal Expansion

++++++ ++
+ ++ +++

++

+

 

3

Unidirectional prepreg tapes are generally the precursors for fibrous composite
materials. Prepregs are fibers preimpregnated with a resin system (thermoplastic or
thermosetting). The finished composite part is formed by consolidating unidirectional
prepreg tapes that have been stacked in a predetermined sequence of orientation e. g. 0°,
45°, 90°. Consolidation cycles for thermoplastic or thermosetting prepregs are different.
Thermosets undergo a chemical reaction and so have to be cured by the application of
heat and/ or pressure during processes such as vacuum bagging, autoclave molding, hot
press molding, etc. Thermoplastics are usually in their final polymerized form and so
their consolidation involves heating under pressure to achieve fusing between adjacent
prepreg tapes. Hergenrother and Johnstonm proposed the following requirements for an
ideal polymeric system to be used as a composite matrix :

a) Easy prepreg preparation
b) Good shelf life
c) Acceptable tack (stickiness) and drape (shape relaxation)
d) Acceptable quality control procedures
e) Low processing temperature and pressure
f) N o volatile evolution
g) Dense void-free matrix
h) Good mechanical performance over the desired temperature range, time and
environmental conditions
i) Acceptable repairability and cost effectiveness
No currently available matrix conforms to all these properties and hence the

choice of a matrix usually represents a compromise“).

4

The primary objective of this study is to investigate the manufacture of
unidirectional prepreg tape on a continuous basis from aqueous colloidal dispersions of
multiple resins in order to control the microstructure and morphology of the prepreg and
composite. Such a process would offer complete control on the final coating of epoxy
resin on the fiber surface and at the same time provide a tape of uniform dimensions and
a high degree of drapability, so that it can be formed into any desired part. These tapes
would then be consolidated within a prescribed curing cycle to give a void-free composite
material. The combination of high-strength fibers with the adhesive properties of epoxy

resin yields physical properties unsurpassed by other similar materials of construction.

It should be noted that epoxies are inherently brittle materials, only slightly
tougher than inorganic glasses. Considering the fracture energy as a measure of
toughness of materials, it can be seen than most thermosetting polymers have fracture
energies ranging from 80 to 200 J/mz, whereas certain thermoplastics have fracture
energies of 1kJ/m2. [5] An important aspect of this thesis, is to demonstrate that the same
process can deposit two particles, an epoxy and an elastomer, from an aqueous colloidal
dispersion. Prepreg tapes prepared from epoxy resin coating with elastomeric inclusions
in the matrix, can be consolidated and optimized for toughness by controlling the size,

density and nature of these elastomeric inclusions.

Chapter 2

 

PROCESS DESCRIPTION

 

2.1 INTRODUCTION

Since prepreg tape is the prerequisite for most continuous unidirectional fibrous
composite parts, it is exigent for developing a process for the manufacture of prepreg
tape on a continuous basis. While several processes have been developed for the
manufacture of prepreg tape, this thesis only encompasses coating from aqueous
dispersions of resins, therefore only those processes that make use of slurries will be
discussed.

2.1.1 Slurry Processing

In this slurry process, spread fibers are drawn through a liquid suspension of the
polymer powder. The liquid carrier is removed by subsequent drying in the heating
zones and the prepreg tape is wound onto a take-up winder. O’Connor“ developed a
patented process for prepreg tape manufacture using a polymer slurry. Major drawbacks
of the process were determination of right concentration of slurry, maintaining this
concentration and removal of all liquid so that voids are not formed during the
consolidation step. The minimum void content in the slurry processed tape were 2-4%.

Dyksterhouse et alm made use of a polymer gel to impregnate the fibers. See Figure 2.1

 

 

 

 

 

 

° 4.
. rdlu'sm
‘ . ... .o
. f .Hth 4.le duh!“ . 55
l ’ '—z—*-z' ' 19%
. I —— .
3 . 7 -. Paw .7 s1
4 ( - 27
Mule 53.7327' W: L]

Figure 2.1 Production of Prepreg Tape using Polymer Slurry

Slurry

Mixing Held

 

Ill

 

 

 

 

C9)
Tow Spreader

 

 

Figure 2.2 Process Schematics for Foam Towpregging

7

for details on the process description using slurry processing techniques.
2.1.2 Aqueous Foam Processing

Chary et all“ developed a process wherein an aqueous foam is used as the carrier
medium to deposit polymer onto the fiber tow. This technique has the advantages of
controlled deposition of polymer onto the tow with considerable ease in process control
and a reduction in drying times as compared to solution/ slurry coating. The major
problem with this method is that agglomeration of molten polymer takes place, due to
high temperatures in the heating section being required for removal of surfactant, which
leads to a stiff prepreg tape. Figure 2.2 is a schematic representation of this process.

2.2 PROPOSED PROCESS

In conceiving the proposed process, whereby an aqueous colloidal dispersion of
a resin is the coating medium, all the currently available prepregging techniques were
evaluated and the salient features of each process were adapted into the final process.
Figure 2.3 shows the unit operations involved in the process.
2.2. 1 General Description

A tow of continuous fiber is unwound from the feed spool by means of a DC
motor. The fiber tow is then directed through a guide slot and passed through the first
of the four sets of pinch rollers. The four sets of pinch rollers are driven by a single DC
motor and are interconnected with chain links and sprockets. The fiber tow then passes
over a spreader unit, where the tow is separated into individual filaments. The second
set of pinch rollers helps maintain the spread and facilitates its passage through the
dispersion bath. In the dispersion bath, the separated fiber filaments are coated with the

dispersed solid particles. Once the fibers are coated, all the moisture is removed in the

ramp-

   

       
 

/
[I
/ O

#1

Hum

  

 

/

mm
/
/

 

3pm

Figure 2.3 Schematic of Aqueous Colloidal Dispersion Prepregging Process

9

heating section. The heating section comprises of a convection oven and two infra-red
lamps. Most of the heat energy is expended in evaporating the moisture. The final two
sets of pinch rollers facilitate the passage of the coated fibers through the bath and the
heating section. The dry prepreg tape is then directed through another guide slot and is
then wound on a take up drum, such that the winding is both in the rotary and traverse
as given below:
2.2.2 Fiber Motion System
The fiber motion system performs the following functions:

a. Unwind the unsized carbon tow from the feed spool.

b. Facilitate the motion of the fibers through the various units in the process line.

c. Provide rotary and lateral motion of the coated fibers on a take-up spool, so that

the prepreg tape is wound over the entire surface of the spool.

The fiber motion system comprises two DC motors, an AC take-up winder, pinch
rollers, guides and chain links. The two motors are 90VDC permanent magnet, ball
bearing, continuous operation (10 hr/day, 40°C max), 1/ 10 HP type with gears for speed
reduction. See Figure 2.4 for details. Both these motors are controlled by a variable
speed controller. Incoming AC voltage is converted to an adjustable full wave rectified
DC voltage to operate the DC motor. (Details of the circuitry will be given in the
section on process automation).

The processing line is provided with four sets of pinch rollers. Dimensions and
construction of rollers are shown in Figure 2.5. Two sets of pinch rollers are installed
before and after the spreader unit, one after the dispersion bath and the other after the

heating section. Each set of rollers is provided with a bleeder cloth wiper at the top and '

 

 

 

 

Figure 2.4 DC Gearmotor

 

 

 

 

 

 

 

ELEVATION
win-a —>ml+ mm
M o
:5 T T i »
I; E] “3.3:. :I 13’: a
A”; ‘ 3‘ ’l
PINCH ROLLERS

Supportflotfortopmlla *l 5/8'l(_
\

\

Nu : i
MVATION 1.5.

 

 

 

 

 

45' 1 T
% p I j i 2.5' *l 2' K
I / T

HOUSING FOR PIN CH ROLLERS
Suppmholefubottom roller withfixedbeuing

 

 

 

Figure 2.5 Pinch Rollers

11

the bottom. The function of these wipers are to gently remove any stray carbon fiber
thus preventing entanglement of the tow. The bottom roller of each set of pinch roller
is connected to the next set of rollers by means of a chain link and sprocket mounted on
the protruding end of the roller. Fibers are unwound from the feed spool, as supplied
by vendor, by utilizing the first of the two DC motors. Speed of this motor is adjusted
so as to match it as closely as possible to the second motor. The second or drive motor
controls the motion of all the four sets of pinch rollers. The speed of each roller is
matched with the drive motor by the interconnected link chain. The unwound fiber then
passes through a guide and taken up by the first of the four pinch rollers. The
subsequent rollers assist the motion of the fibers through the spreader, dispersion bath
and heating units. The line speed is adjusted via the speed controller located on the drive
motor.

The prepreg tape is wound on an empty cardboard spool. This spool is mounted
on a commercial take-up winder, Leesona #959. The take-up winder has been
specifically retrofitted to match the process line speeds. It is provided with a three-phase
220V, AC, 0.5 HP motor. To the motor is linked a cam shaft and an expanding flange
tube holder. (See Figure 2.6). An empty cardboard spool is placed on the flanged tube
holder. Air or nitrogen is filled into a diaphragm located within the tube holder. The
flanges expand and form a tight fit with the spool. Prepreg tape is passed through the
guide and fixed to the spool. Speed of the motor is adjusted by a torque controller.
It adjusts the tension between the final pinch roller and the take-up winder such that a
winding similar to that of the feed spool is attained. The motor provides rotary winding,

whereas the cam shaft provides for transverse winding.

12

3-Phase, 220V, AC Motor

     

Belt m ’1/” Camecttr ft! Air Supply

Cam For Traverse Motion

Figure 2.6 Take-Up Winder

13
2.2.3 Spreader Unit

The spreader unit is a patented device developed at the Michigan State
University”). A schematic representation of this unit is as shown in Figure 2.7. It is
necessary to spread the fibers to achieve uniform and maximum coating of polymer
particles from colloidal bath. The spreader unit consists of a 10" dia woofer speaker
mounted in a wooden housing with dimensions as shown above. The speaker has a
frequency response to 4,500 Hz and an input power of 20 W nominal and 50 W
maximum. The speaker is driven by a frequency generator and a power amplifier. A
BK Precision 2MHz digital display function generator is used to generate a signal of
specified frequency. This signal is amplified by a 5W Optimus STA-20 AM/FM stereo
receiver and supplied to the speaker. The speaker vibrates at a constant predetermined
frequency and amplitude. The acoustic energy thus delivered assists in the spreading of
the narrow incoming fiber tow to any required width. For a particular fiber tow, there
is a narrow band of operation at which the fibers absorb the most energy and give an
optimum spread. The narrow fiber tow is passed through the 3/8" dia case hardened
polished steel shafts arranged as shown in figure. The shafts are held in place by means
of an aluminum block fitted with precision bearings. The function of the shafts is to hold
the tow in its spread form as it is being conveyed forward. The speaker provides the
acoustic energy for spreading the fibers. The fibers are maintained in a spread manner
by ensuring zero tension between the ends of the spreader. This is achieved by the two

by two pinch rollers at both ends of the speaker.

14

3/8' Die Case Hardened
Polished Shafts Boating block

Narrow fiber-tow /\ smlifiborm
- , ___1 L_ ‘

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 2.7 Spreader Unit

15
2.2.4 Dispersion Bath

A schematic representation of the dispersion bath is as shown in Figure 2.8. The
dispersion bath comprises of a 6.5"x 6.5" x 12" plexiglass box in which is placed a 1.5"
dia aluminum roller mounted on sealed bearings. The colloidal dispersion of a polymeric
resin is diluted with deionized water to the desired level of pickup. Spread fibers pass
through this bath and the suspended solid particles are coated onto the fiber surface. The
roller within the bath facilitates the passage of the fibers. The volume of the dispersion
and the line speed determine the residence time of fibers within the bath. With the
passage of time, the concentration of the solid particles in the bath would decrease as it
is being coated onto the fiber surface. However, the concentration of the bath is
maintained by periodic replenishment of bath with fresh colloidal dispersion.

2.2.5 Heating Section

The main purpose of the heating section is to remove moisture from the fiber and
sinter the polymeric particles to the fiber surface. The heating section consists of a
convection type laboratory oven and 375 W infra-red lamps. The oven is a Grieve,
120V, LW 201C convection type oven with a maximum operating temperature of 200°C.
The oven has been retrofitted to meet these material and process requirements. (See
Figure 2.9 for details). The temperature in the convection oven is controlled by a
dedicated controller details of which will be covered in the section on process
automation. This oven is not capable of providing all the energy required to remove all
the moisture from the fiber surface at the given line speeds. Therefore, heating is
supplemented by two infra-red lamps installed downstream of the oven. The first of the

two lamps is a 37 5 watt infra-red lamp with intensity control range from 15% to 80%.

 

 

l I l
x i

\\ k 6.5" ,I

 

 

 

 

 

 

 

 

 

JI-

K

 

 

 

Side View Tip poi Draining
Top View
Figure 2.8 Dispersion Bath
Side View Elma“

H 141/8' ->i #20. —>|

 

 

 

 

 

 

 

Figure 2.9 Heater

 

 

17
The second lamp is also 375 watts, but is housed in an ordinary lamp holder. It is

possible to adjust the height of the bulb from the prepreg tape, thus varying the intensity
of heating. Both lamps can supply a maximum of 120°C to the fiber surface. Partial

sintering of the solid particles is also achieved in this section.

2.3 PROCESS AUTOMATION

Process automation is achieved in the different sections of the processing line by
different controllers. Each will be described separately.
2.3.1 Fiber Motion

As described earlier, the motion of the fibers through the processing unit is
controlled by means of a drive motor, pinch rollers, chain links and sprockets. The
motors are controlled by a variable speed controller, DART Controls 125K Series. This
controller is a dedicated speed controller and adjusts the current to the motor according
to the desired setting. (See Figure 2.10 for details on the circuit diagram). The
controller was calibrated against the dial settings to ascertain the line speed. For this
purpose the process line was run for one minute time periods with carbon fiber tow to
calibrate the line speed. For e.g. if for a dial setting of l, the resulting fiber length was
60 cm; then the line speed at 1 would be lcm/sec. At least three sets of samples were
taken for each dial setting. (See Figure 2.11 for calibration chart). The speed of the
motor linked to the feed roller is adjusted manually such that there is very little tension
or a slight slack between the feed roll and the first pinch roller. The take-up winder
is controlled by a 2100 Series Fincor DC motor speed controller. This controller has
been set for maintaining a constant tension between the last pinch roller and the take-up

spool. The initial speed has to be dialed in and after this the torque controller takes over.

18

 

 

OI ............................ m C
a: a: 125 SEHES SCI-GMAT“: mamas.- .......... rm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

III .

IN ..

It! . no OH I12 ...... I“

.20 .. ”K I!) ...... autism
R20 ..... 0.7! I" no:

In ..... 91K '13 ...... man

02 ....... mount an 3" flAVfl

03 ....... m m mom

 

 

 

 

 

Irma; 0403:!

soon. CHIS: .. d5

 

 

.l
“.1
-ISA ...OINC‘ ...CM 16V

'1' -.mos ...za-nev

d! -.mcs 63M 16V

 

Figure 2. 10 Circuit Diagram of Variable Speed Controller

 

 

 

N
01

 

 

 

/

 

.9

 

Line Speed (Inches/sec)
... 01 N

 

 

/

0 1.7

00'!

 

 

 

 

2 2.3 2.6 3
Dial Setting

 

Figure 2. ll Calibration Graph for Speed Controller

 

 

19

2.3.2 Oven Temperature

The temperature of the convection oven is controlled by a dedicated controller.
See Figure 2.12 for details of the circuit diagram. The oven temperature is measured
by a J—type thermocouple, which is converted to a millivolt signal and fed to a CN 300
Omega Temperature Controller. This temperature controller operates on an ON-OFF
principle. As long as the temperature of the oven is below the set point, the controller
keeps the heating circuit on; but as soon as the temperature exceeds the set point the
controller switches off the heating circuit and does not restart it until the measured
temperature goes below the set point. Temperature control of between i5°C was

obtained by by-passing the oven’s original thermostat.

2.4 SUPPLEMENTARY EQUIPMENT

Additional equipment is required for consolidation of the prepreg tape into final
composite parts. Besides this it is also necessary to identify equipment for
characterization of prepreg tape, prepregging material and composite parts. Given below
are brief descriptions of the various ancillary equipment required for this:

2.4.1 Prepreg Tape Winder

Prepreg tape produced from the process has to be wound onto a frame so that a
multi-lamina composite part can be consolidated from it. A sketch with dimensions of
the winding frame is given in Figure 2.13. The winding frame is fixed to a rotary
winder. This winder has been specifically designed to wind prepreg tape or fiber tow.
(See Figure 2.13 for details). The winder has an AC motor, which rotates the frame
about its axis. The motor is also provided with a counter, which counts the number of

revolution. The fiber guide is attached to a second motor. This motor moves the fiber

20

 

 

 

 

 

"fig rag-7g Oven Tweentrol
o o o

 

Q
a .

L
“HT!

 

 

 

fiv

 

 

 

 

 

 

 

 

 

Figure 2. 12 Circuit Diagram for Convection Oven Temperature Controller

 

 

 

 

 

 

 

 

 

Figure 2. 13 Schematic for Frame

 

 

21

tow in a traverse motion. Both the motors are operated simultaneously. The speeds of
these motors are based upon the width of the fiber tow. Detailed calculations for
determining the speeds of the motors are provided in the following chapters. The
prepreg tape is taken from the product spool and passed through the tow guide. The tape
is fixed to the rotary frame. When both the motors are operated, the tape undergoes
rotary and traverse motion, so that it is wound over the entire width and length of the
frame.
2.4.2 Autoclave Consolidation

The multi-layered prepreg tape part must be consolidated into the final composite.
Consolidation involves the curing of the resin under controlled conditions, i.e. according
to a specific heating, pressurization cycle. This is usually done in an autoclave under
vacuum. A United McGill autoclave is used to carry out the curing of the epoxy resin.
The consolidation cycle has been programmed into the control loop of the autoclave.
(See Figure 2.14 for details). The finally layed-up part is placed in an autoclave vacuum
bag and then placed within the autoclave. In the autoclave, nitrogen is used as the
pressurizing and heating gas. After the part has undergone the consolidation cycle it is
removed from the autoclave. Details of the consolidation cycle and autoclave molding
process are provided in later chapters.
2.4.3 Volume/Void Fraction Analysis

Consolidated composite parts have to be void free and have a certain volume
fraction of fibers and material to ensure optimum mechanical properties. Volume
fraction and voids within the composite part are determined by the volume fraction

analyzer as shown in Figure 2.15. Samples are cut from the final plaque on a diamond

22

 

 

Pressurized Nitrogen from Cylinders

It"

To Vent
run 19'. mm - _____________ _| .c. I . I

on [swoon

A sum 4 L
"M To Vacuum
01 mm
M

W

D E: ‘2:

van mn- '

 

 

 

 

 

 

v—In‘

iguana ....
.1 1, MM ""

.-. '4'

To Drain

 

 

31..

To Drain

 

Figure 2.14 Autoclave

 

 

 

 

Figure 2.15 Equipment for Volume Fraction/Void Analysis

 

 

23

saw. These samples are then mounted in an acrylic mold and polished in successive
steps using a 80 grit belt sander, Struers Abrahim Automated Polisher (240, 320, 1000,
2400 and 4000 grit), LECO GP20 Grinder/Polisher (5 microns and 1 micron) and a
Buelher Vibromet (0.05 micron). Details of polishing and cleaning are provided in later
chapters. The polished surface of the sample is placed on an Olympus microscope,
which is viewed with a video camera. Images from the video camera are sent to a video
monitor and to an Amiga personal computer in order to digitize random images of the
sample surface. These images are then manipulated with brightness and contrast, so that
the voids appear as dark regions. These dark regions are then analyzed as a void fraction
of the total surface area by using the ON VFA (Optical Numeric Volume Fraction
Analyzer). This analysis method uses the grey levels from a histogram to determine the
void fraction. The same arrangement is also used to determine volume fraction of fibers
except the digitized images need not be altered for brightness and contrast.
2.4.4 Mechanical Testing

Mechanical testing for the composite parts consists of short beam shear, 3-point
flexure, and Mode 11 fracture analysis. For all three tests the United SFM Testing
Machine is used . It is composed of two interconnected yet separate components :

a) The Loadframe : The actual testing of a specimen takes place here.
b) The Console : It contains the computer, which processes and stores data, and
controls the loadframe.

The power source is a variable speed DC motor. Testing speed is controlled by

two feedback loops. A DC tachometer provides instant feedback of motor speed to

stabilize the servo loop and provide coarse control. The microcomputer senses and

24

provides precise control of position and speed, strain, or load and load rate as required.
Every testing methodology has a separate testing fixture, which is adaptable to the

loadframe. Prior to the test, the fixture is attached to the loadframe along with an

appropriate load cell.

Chapter 3

 

MATERIAL PROPERTIES

 

3.1 INTRODUCTION

In this chapter, the different materials used for manufacture of water-based
composites will be discussed. An in-depth presentation of the material properties and
material selection procedure will be made. It is exigent to understand the various
physical properties like particle size, fiber diameter, thermal properties, viscosity etc. for
optimal operations of the prepregging, consolidation and characterization processes.

3.2 FIBER REINFORCEMENT

Various types of fiber reinforcements are available in the market today. Selection
can be based upon processing, final properties of composite and compatibility with
matrix. Commonly used fibers are glass, carbon, and Kevlar fibers. For our process,
it is essential for the fibers to be free of sizing or have one applied at very low
concentrations so as to be spreadable. Unsized carbon fibers meet this criterion most
adequately. For this research Hercules Magnamite' Type 1M7 12k tow carbon fibers
were used.

3.2.1 Magnamite" Carbon Fiber Type 1M7

Magnamite' Type 1M7 12k tow carbon fiber is a continuous, high performance,

25

26
intermediate modulus, PAN -based fiber having 12,000 filament-count tows. The fibers

are provided in an unsized state, so as to facilitate spreading. Presence of any sizing on
the fiber surface will inhibit spreading by acoustic waves. This type of fiber is ideal for
prepregging and filament-winding processes. It has a high tensile strength, shear strength
and modulus, so can be used in stiffness-critical and strength-critical applications.
Typical properties of the fiber are as given in Table 3.1

Table 3. 1 Typical Physical Pr0perties of IM7 12k Carbon Fiber Tow

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WRTIES U.S. Units S.I. Units
ensile Strength 800,000 psi 5,518 MPa
Tensile Modulus 44 x 10° psi 303 GPa
Ultimate Elongation 1.86 % 1.86%
ensity 0.0643 lbs/in’ 1.78 gm/cm’
Typical Epoxy Composite Properties @ R.T.I
Tensile Strength‘ 630,000 psi 4,34I MPa
Tensile ModulusT 40.0 x 10":psi 276 GPa
Ultimate Elongafion 1.50% 1.50%
Compression Strength 265,000 1,826 GPa
Compression Modulus 23 x 10'5 psi 1,633 MPa
Flexural Strength’ 237,000 psi 1,633 MPa
exural Modulus3 24.1 x 10F psi 166 GPa
SBS Strength’ 18,700 psi 129 MPa
Waracteristics
Filament Diameter 0.197 mils 5 u
Filament Shape Round Round
Twist None None
Tow Cross-Sectional Area 390 x 10“ in2 0.25 mm2

 

 

I. All values are based on 3501-6 resin
2. Tensile values an: normalized to 100% V,
3. Values are normalized to 62% V,

27
3.3 MATRIX
In this thesis, emphasis is being placed on prepregging using water based epoxy
resin systems. Prior to outlining the properties and features of the selected materials it
is essential to have a brief knowledge of colloidal dispersions, which are the basis for
water dispersed epoxy resins.
3.3.1 Colloids’ml
By definition colloids are liquids having dispersions of particles in the size range
of 0.01-1pm. The particles remain in suspension by two modes of stabilization, namely:
1. Electrical Double Layer Stabilization
2. Steric Stabilization
3.3.1.1 Electrical Double Layer Stabilization
Electrostatic repulsion arises from the presence of ionized moieties at the particle
surface. In the region around the surface a charge cloud is formed, which is also known
as the electrical double layer (EDL) and the charge cloud consists of evenly distributed
opposite charges, known as counterions. See Figure 3.1 for a schematic representation
of the model. The Stern Layer is a small space separating the ionic atmosphere around
a surface, the actual double layer, from the "Steric" wall of the charged plane just
adjacent to the surface. The overlap of the counterion clouds leads to repulsion between

ions of the same charge, and hence to a repulsion of the particles.

 

 

 

 

 

 

Figure 3. 1 Model for Electrical Double Layer Stabilization

 

 

i
i
i

qr---------

 

i
i

l
I
1
1

 

Figure 3.2 Model for Steric Stabilization

 

 

29
3.3. 1.2 Steric Stabilization

Steric stabilization can be used in all liquid media. The basic requirements for
steric stabilization are that the particle surface be covered by a contiguous layer of
solvated chain molecules anchored to the surface either chemically or strongly physically
absorbed. Steric stabilization agents are soluble macromolecules or non-ionic surfactants.
A schematic representation of the steric stabilization model is give in Figure 3.2. When
h < 2L, the layers interfere with each other and so have either to interpenetrate or indent,
or undergo a combination of the two. This leads to a change in the local osmotic
pressure and free energy. This may be attractive or repulsive in nature. When h< L,
there is an additional contribution to the steric potential due to loss of configurational
freedom experienced by anchored chains. This force is purely repulsive in nature and
hence gives rise to steric stabilization.

3.3.2 Material Selection

Since this thesis encompasses prepregging via aqueous based resins, only those
materials will be discussed that can be obtained as a water-based dispersion.
3.3.2.1 Waterbome Epoxy Dispersions

A variety of waterborne epoxy dispersions are currently available. They will
adhere to a wide variety of substrates and when properly cured will develop high strength
and resistance properties. Non-ionic reactive surfactants such as polyoxyethylene,
polyols etc. , are used to effect dispersion of epoxy resins in water. These dispersions
are easily blended with other waterborne chemistries to produce hybrid systems. thne-
Poulenc has developed a wide variety of such waterborne epoxy dispersions. The term

epoxy or oxirane refers to a chemical group consisting of an oxygen atom bonded with

30

two carbon atoms which are also bonded in some way‘“). The most widely used
family of epoxy resins are known as "Bis A epoxies" based on the condensation reaction

of bisphenol A and epichlorohydrin. A theoretical Bis A epoxy molecule is as shown in

 

Figure 3.3.
BISPHENOL A BASED EPOXY
FLEXIBILITY
I
1““ /°\
04! cucnp© ocmcucuzo ©f©ocuzeu en,
en,
CROSSLINKING
SITE
HYDROLYSIS ADHES'O"
RESISTANT

Figure 3.3 Theoretical Bis A Epoxy molecule

These resins come in different molecular weight series. As molecular weight
increases, the epoxy equivalent weight and the number of hydroxyl groups available for
reaction increases. Also polyfunctional epoxies are also available in which there are
more than two epoxy groups per molecule. Advantages of these systems are as follows:

0 Excellent volatile release.
0 Adhesion to a wide range of substrates.

0 Can be formulated to be formaldehyde free.

31

0 Excellent electrical properties.

0 Compatible with other resin types.

0 Two phase for longer pot life.

0 Shear stable.

0 Can be used with conventional epoxy curatives.
0 Uniform particle size.

0 Low water sensitivity.

Since a variety of waterborne epoxy resins are available, selection for the most
suitable dispersions system was made based on comparing its utility for prepregging in
terms of ease of coating, comparatively low viscosity, high mechanical properties,
simplistic curing cycle and compatibility with elastomer latexes. Based upon this
selection criterion Epi-Rez° W60-3515 epoxy resin dispersion was chosen.

3.3.3 Epi-Rez' W60-3515 Epoxy Resin Dispersion

Epi-Rez W60-3515“” is a non-ionic, aqueous dispersion of a bisphenol A epoxy
resin with an equivalent weight of 220-260. It has a moderate viscosity and is
mechanically stable. N 0 organic solvents are present. It is completely water reducible,
providing wide latitude for viscosity reduction.
3.3.3.1 Performance Highlights/ Features

0 High epoxy functionality

High solids

Completely water reducible

Compatible with most latex resins and some waterborne phenolic resins

Good water resistance

32

0 Good chemical resistance

3.3.3.2 Uses

0 Industrial textile binders or finishes

0 Adhesives

Fiberglass reinforced plastics

Coating and electrical dip coatings and varnishes
0 Performance modifiers for other waterborne resins

3.3.3.3 Typical Properties

0 Viscosity at 25°C, cps 11,000
0 Non-volatile, % 63
0 Volatile portion Water
0 Pounds/Gallon 9.2

0 Particle size, average, microns < 3
0 pH 3 .7
0 Weight per Epoxide, on solids 250
When applied at room temperature, Epi-Rez W60—3515 coalesces to form a clear,
continuous, tacky film after evaporation of water. A DSC and T GA analysis of the Epi-
Rez W60-3515 dispersion are as shown in Figures 3.4 and 3.5, respectively.
3.3.3.4 Curing agents
Dicyandiamide is used as a curing agent with W60-3515 to provide extended pot
life formulations which cure at 300-350°F to form highly crosslinked systems suitable for
service at high temperature. Dicyandiamide is a high functionality (4-5) curing agent

having a molecular weight of 84.08 and a molecular formula HN--C--(NH2)--NH--C--N.

33

80.010: EPI-FEZ HBO-3315 SOLUTION D S C F110: I: EPIREZ

Size: 13.5000 no Operetor: JAMES FERNANOES
Method: 10°C/MIN To 350°C Run Dete: :o-uur-ea 00:02
Connent: UNCURED

 

 

 

 

 

 

 

 

     
 

 

 

 

0"
1
E
a -2.
x
O
F.
u.
8 d
e
i
«i
435--.. r ya * reo— “‘
Teepereture (’C) Oenerel V4.10 DuPont 2200
O O 0
Figure 3.4 DSC Analysrs of Epr-Rez W60-3515
Sample: Epoxv RESIN VARNISH T G A File‘ C ERTGA 001
Size 21 7540 mg Operator JAMES FERNANDES
Method 10°C/HIN TO 650°C Gun Date: 22-520—93 05 07
Comment' 10 DES C/MIN T0 600 DEG C
100
80‘
l3l.25°C
59.15: 325 55'C
58 6:!
504 T
E
U
C
0
GI
‘ 40~
20-' 456 ES’C
a 240x
1
0 m I ' * fl fir I a I I
0 too 200 300 400 500 900 700
I M V5 1A DuPont 2200

Temperature PC)

Figure 3.5 TGA Analysis of Epi-Rez W60-3515

34

It is available as odorless, white rhombic leaflets or plates or monoclinic prisms. It has
a melting point of 412-414°F, specific gravity of 1.404 @14°C and solubility in water
2.26 %@ 13°C. 2-methyl irnidazole catalyst is used as the dicy accelerator. It has a
molecular formula of C4H6N2, melting point of 142°C to 143°C and is available as a
white crystalline solid. The formulation with the above mentioned curing agent and

accelerator is ideal for prepregs. A typical formulation for the epoxy varnish is as given:

 

 

 

 

 

Material Parts by Weight
Epi-Rez W60-3515 100
Dicyandianfide 3.50
2-methyl imidazole 0.10
Deionized Water 13

 

 

 

 

3.3.4 Carboxylated Styrene Butadiene Rubber (CSBR) Latex

Selection of an appropriate elastomer as the second coating particle is based upon
its compatibility, stability and cured state properties when added to the epoxy resin
dispersions. Research work done at Rhone-Poulenc indicated that carboxylated styrene-
butadiene rubber (CSBR) is fairly compatible with epoxy resin W60—3515. CSBR is
available as a latex from Gencorp Polymer Products as Gen-Flo' 3000 series latexes.
Gen-FloO 3000 is a carboxy-modified, formaldehyde free, styrene/butadiene latex
designed for non-woven and saturation application. It has a chemical formula :
[—(C4H6)x-(C8H3)y-(R-COOH),-]n. The recommended uses of this product are toweling,
laminated fabrics and bedding fabrics. Typical properties of this material are as given

in Table 3.2. DSC/TGA analysis of CSBR latex are as shown in Figures 3.6 and 3.7.

35

Table 3.2 Typical Properties of CSBR

 

 

 

 

 

 

 

 

 

 

 

 

 

 

‘—PHYSICKITPROPERT1ES CHEMICAL]
MECHANICAL PROPERTIES
m by weight 50.7 Mechanical Stability Excellent
Wet Weight/ Gallon 8.3 Multivalent Ion Stability Excellent
Dry Weight/ Gallon 4.2 Shelf Life Excellent
We Gravity 1.0 Foam Low
pH 9.0 Odor Low
Brookfield Viscosity 150 Emulsifier Anionic
Surface Tension (dyn/cm) 46
Glass Transition Temp,°C -37
Particle Size 0.18-0.19 u

 

 

36

STYRENE DUTADIENE RUBBER SOL
10.2000 Io
SO'C/HIN T0 250°C

Bennie:
sue:
Method:

DSC

fleet Flee We)
I
.-
O
l

-1.5<

 

File: I: see
JAMES FERNANDES
10:09

Ooeretor:
nun Dete: ao-uur«es

 

 

 

 

  
     

-g_oa 11e.e1°c
4
-!.5+—-—- w w— -- f 1— ---—.-.--—---i
25 75 12!
Teloereture (’CI Ienerel V4.10 DuPont 8800
O 0
Figure 3.6 DSC Analysrs of CSBR
Sample‘ sen SOLUTION ’r'CS [A File c:san-rcs 001
Size 30 0480 mg Operator: FERNANOES
Method ib’C/wIN to 600°C nun Oate‘ 13-Sea-93 14'10
Comment: FIRST nun N2 GAS PUQGE up to GOO'C
100
J 32 ei'c
99 52x
60—
137 so’c .
60- 54 291 3‘0 63 C
5 5: 61:
U 4
c
0
It
2

 
 
 

 

 

 

20«
l 481 zs'c
o 95241
0 Dr V ' Y 46 550 do 700
c ‘00 20° 30° ° '81 v5 1A 0090"! 2200

temperature I'Cl

Figure 3.7 TGA Analysis of CSBR

37

Epoxies are inherently brittle materials, only slightly tougher than inorganic glass.
In order to mitigate the brittleness of epoxy materials, a soft rubbery inclusion is
dispersed in the brittle epoxy matrix to create a multi-phase system. During crack
propagation, the crack front requires a lot more energy to pass through the soft rubbery
inclusions, than it would take in passing through the brittle epoxy matrix. Thus, as the
concentration of the rubbery inclusions increases, the fracture energy for the material
also increases and makes it a tougher part. Increase in fracture energy reaches a
maximum with the increase of elastomeric particles, and then begin to decrease. This
decrease is closely related to the resin morphology as it transfers from a "particulate” to
a "blend" morphology. Other mechanical properties like flexural strength, flexural
modulus and short beam shear strength all go down with the increase in the concentration
of the elastomeric additives. This is all due to the increase in the quantity of the "softer”
second phase”). Typical morphological studies of the fracture surfaces of rubber-
modified epoxy matrixes indicate gross plastic deformation in the matrix and a highly
porous structure in the resin. These holes results from the cavitation of elastomer
particles in response to the dilational stress field, as the crack propagates through the
material. Figure 3.8 shows the typical morphology of the fracture surface of a rubber

modified epoxy/carbon composite“.

38

 

Figure 3.8 Fractograph of F85/Graphite Composite Sample

Chapter 4

 

SAMPLE PREPARATION

 

4.1 INTRODUCTION

In this chapter, the complete set of experimental details for the process for
manufacturing of multi-component water based composite materials will be discussed.
The composite parts have varying concentrations of a second elastomer phase in the
epoxy matrix. These multi-phase composite materials were mechanically and
morphologically characterized.

4.2 DISPERSION BATH PREPARATION

Prior to manufacturing prepreg tape by using the continuous prepregger, a the
epoxy resin dispersion has to be mixed with the catalyst and accelerator and then diluted
with deionized water to give the right concentration of the dispersed solids.

4.2.1 Matrix Dispersion Preparation

Epi-Rez W60-3515 is supplied as a non-ionic, aqueous dispersion of a bisphenol
”A" epoxy resin having approximately 63 % by weight of solids. This colloidal solution
does not have any catalyst to promote cross-linking reaction during the consolidation
cycle of the prepreg tape. Therefore, catalyst and accelerator has to be added to it. The
catalyst used is dicyandiamide and the catalyst accelerator is 2-methyl imidazole. The

39

40

procedure is given below:

Materials

Parts by weight, as supplied:

Epi-Rez W60—3515 : 100.0
Dicyandiamide : 3.5

2-methyl imidazole : 0.1
Deionized or distilled water : 13.0

Mixing Procedure

Heat the deionized or distilled water to 70-80 °C.

Add the dicyandiamide (dicy) to the warm water and agitate until completely
dissolved. Additional heating will be required to maintain the required temperature
of 70-80°C.

When solution is complete, add the 2-methyl imidazole and dissolve.

Add the warm dicy solution to the epoxy dispersion (W 60—35 15) and agitate slowly
at room temperature until well mixed (30-60 min). If the dicy solution cools and
begins to crystallize before it is added to the dispersion, it should be reheated to
70-80 °C.

An induction period of at least one hour is required prior to use.

4.2.2 Dispersion Dilution

It has been observed that a good composite part having a Vf of 60-70% is

obtained when the coating on the prepreg tape is around 25-30% by weight of epoxy

matrix. Such levels of coating cannot be achieved with the existing concentration of the

solution. To attain lower coating levels, the as-supplied epoxy resin dispersion has to

be diluted with deionized water. The level of dilution was determined by a simple

experimental procedure, which is described as below:

41
0 Take 6-6.5" length of carbon fiber tow.

0 Spread the to about an inch in width and weigh.

0 Fill a flat pan (6" x 4" x 2") with the epoxy dispersion.

0 Different levels of dilution are maintained in the pan by the addition of deionized
water. Dilution is carried out per volume.

0 Dip tape lengths in the dispersion for around 15 seconds.

0 Remove water by drying in an oven and measure the coating pick-up by weighing.

A graphical representation of the results is as given in Figure 4.1. From the
graph it is evident that at a volumetric dilution level of 6: 1, a coating of around 30% by
weight will be obtained. When the process was run, this dilution level did not give the
desired level of coating. The main reason was that the residence time of the fibers in the
bath was around 6-7 seconds in comparison to 15 seconds taken during the dipping step.
Eventually it was found that a dilution level of 1:1 gave the optimum coating level of 25-
30% by weight of epoxy resin in the prepreg tape.

The prepregging process was run with the diluted epoxy resin dispersion. In
subsequent runs, CSBR latex was added to the diluted dispersion so that a coating
medium containing 5, 10, 15 and 20% of CSBR by weight were obtained and prepreg
tape were produced with each of these dispersion mixtures. Addition of CSBR latex to
the dispersion to give the desired concentration is based upon the following formula:

Vss PasWas Wsu
WEMWSL Ps1.

 

VSL =

where VSL is the volume of CSBR latex added to epoxy resin dispersion

VES is the volume of epoxy resin dispersion

42

 

 

100
90%
so:

70 \\

60E \

50;
40: 3K
30: X
20: \.
10: 4
02.... ...H !
0 5 10 15 20

Dilution Level (ERzDIW)

 

 

 

 

 

 

 

 

 

Coating in Weight %

 

 

 

 

 

Figure 4.1 Coating of 12k IM7 Carbon Fibers with EpiRez W60-3515 at different
Dilution Levels

43

pES is the density of epoxy resin dispersion

WES is the weight fraction of solid epoxy in dispersion

WSM is the weight fraction of CSBR particles in epoxy resin/CSBR mixture
WEM is the weight fraction of epoxy resin particles in epoxy resin/CSBR mixture
WSL is the weight fraction of CSBR particles in CSBR latex

pSL is the density of CSBR latex

4.3 PREPREG TAPE PREPARATION

The process is operated with the matrix dispersion as the coating medium to give
prepreg tape with epoxy resin coated at a level of 25-30% by weight and CSBR coated
at a concentration of 0, 5, 10, 15 and 20% by weight.

4.3.1 Processing

A spool of Hercules' IM7 12k unsized carbon fiber is located on the feed spool
frame. The fiber tow is unwound, threaded through the first guide slot and then passed
over a pinch roller. Refer to Figure 2.3 for details. The tow is then alternately passed
through the guide shafts on the spreader. From the spreader the tow passes through
another pinch roller. It is then passed under the support roller in the dispersion bath,
through a set of pinch rollers and through the heating zone. From the heating section
the tow passes though a final guide slot and is then taken up by a spool located on the
rotating arm of the automatic torque controlled winder. Once the threading of the fiber
tow is done through the continuous prepregging unit, the diluted epoxy resin dispersion,
with CSBR if required, is added to the dispersion bath. The convection oven is switched
on and the temperature is maintained at 190°C. The infra-red lamps are switched on and

the temperature at the fiber surface due to these lamps are about 120°C . The feed spool

44

unwinding motor and the main motor are switched on. The setting of the main motor
is around 1.75 on the dial, which corresponds to around 1 inch/sec in terms of line
speed. The speed of the feed spool motor is adjusted manually, so that there is a slight
slack between the spool and the first pinch roller. The take-up winder is started and its
speed is adjusted manually to match that of the main motor. The torque controller
adjusts the speed of the winder so as to maintain a constant torque between the last
pinch roller and the winder. Winding of the prepreg tape is done in such a way that the
spool is rotating along its axis, whereas the prepreg tape is traversing the entire length
of the spool. In this way the geometric configuration of the wound spool is similar to
that of the feed spool. The speed of all the rollers are synchronized by inter-linking them
with chains.

Once the fibers are set in motion, the spreader is switched on. The underlying
principle of the spreader is that when the fibers pass over the vibrating speaker, acoustic
energy assist the fibers in separating into individual filaments. Normal operating
conditions of the speaker are 38112 and 11.2-11.6V to spread the fiber tow by 3"-4". It
is also important to maintain zero tension in the fiber tow through the spreader to achieve
the desired effect. This is done by using two pinch rollers at either end of the speaker
to control fiber tow tension.

Spread fibers then pass through the dispersion bath. Suspended epoxy resin
particles and/or CSBR particles along with water adhere to the surface of the carbon

fibers. The possible mechanism responsible for coating is as follows:

45
4.3.2 Mechanism for Coating

Coating of the carbon fibers can take place by any of the following factors:
0 Binding by physical interactions.
0 Mechanical binding, including anchor effect.
0 Binding by chemical interactions.
Whenever a liquid is brought into contact with a solid material, the liquid is said
to undergo a wetting or non-wetting phenomenon. Consider a drop of liquid on a solid

surface as shown in the model Figure 4.2.

         

(a) (b) (c)

0'11. Wash. "dbl Compute Non-Wotan.

Figure 4.2 Schematic Illustration of the Various Degrees of Wetting

As seen in the model, when a drop of liquid is placed on a solid surface, the
liquid will either spread across the surface to form a thin film or it will spread to a
limited extent but remain as a discrete drop on the surface”. The final condition of the

applied liquid on the surface is taken as an indication of the wettability of the surface by

46

the liquid or the wetting ability of the liquid on the surface. The quantitative measure
of the wetting process is taken to be the contact angle, 0, which the drop makes with the
solid. In the case of a liquid that forms a uniform film, i.e. where 0 = 0°, the solid is
said to be completely wetted by the liquid. If a contact angle is formed that lies between
0° < 0 < 89° constitutes partial wetting and anything greater than 90° forms a non-wetting
system. The contact angle forms the basic framework for the thermodynamics of

wetting. This is represented as Young’s equation.

st ' Ysr. = ywcose
where 'st is the interfacial tension of the solid in contact with the liquid.

‘th is the interfacial tension of the liquid in contact with vapor.

'st is the interfacial tension of the solid in contact with vapor.

There are three types of wetting phenomena of importance: adhesional, spreading,
and immersional wetting. Adhesional Wetting refers to the situation in which a solid,
previously in contact with a vapor, is brought into contact with a liquid phase. During
the process, a specific area of solid-vapor interface, A, is replaced with an equal area of
solid liquid interface. The free energy change of the process is given by the work of

adhesion:

We = YSV+YLV—YSL
where Wll is the thermodynamic work of adhesion.
From this equation it can be seen that any decrease in the solid-liquid interfacial
energy will produce an increase in the work of adhesion, while an increase in solid-vapor

or liquid-vapor interfacial energies will decrease the energy gain from the process.

47

Spreading Wetting applies to the situation in which a liquid and the solid are
already in contact and the liquid spreads to displace a second fluid, usually air. Here the

spreading coefficient Susm is defined as

Susm : st ' YLV " Ysr

If this term is positive then the spreading liquid will spontaneously displace the
second fluid and spread completely over the surface. If it is negative, then spontaneous
spreading will not occur.

Immersional Wetting results when a solid substrate not previously in contact with
a liquid is completely immersed in it, displacing all of the second fluid with which it was

in contact prior to immersion. The free energy change in this case is given as:

-A G = Achy-y”)
where AG is the free energy change.

A is the total surface area of the solid

From these relationships it is quite clear that the interfacial energies between a
solid and any contacting liquid and the interfacial energies between the solid and the
second fluid (usually air), control the manner in which the liquid will spread and adhere
to the solid surface.

This process involves the irnmersional wetting of carbon fibers when they pass
through the dispersion bath. From the above discussions it can be seen that larger the
work of adhesion better is the adhesiveness. To examine the effect of wettability on
adhesion strength, Zosel et aim] showed that when the surface free energy of the

adherent is lower than that of the solid adherend (i.e. wetting is not complete), but

48
approaching that of the adherend, the adhesive strength is large and adhesion is favored.

In general the formulation should have surface tension less than or equal to the surface
free energy of the solid adherend to wet its surface. The surface tension for the given
epoxy resin formulation is around 40 dynes/cm and that for CSBR is 46 dynes/cm,
whereas the surface free energy for unsized IM7 carbon fibers is around 45-50 dynes/cm.
In this case there is a positive thermodynamic driving force, and so the epoxy resin and
the CSBR latex partially wet the fiber surface giving good adhesion.

Another factor governing the adhesion of the coating mixture onto the fiber
surface, is the penetration of the fluid into the cavities, which are characteristic of the
inherent surface roughness of the fibers”: “‘1, while passing through the bath. Capillary
driving forces via penetration through surface features are governed by the capillary

penetration pressure AP, which can be described by the Zisman equation“ as shown:

AP: 1.1-ii. i+i
2 2b r1 r2

where AP is the capillary penetration pressure

71 is the liquid surface tension of a homologous series of liquids

r1 & r2 are the principal radii of curvature of the pore interstices

b is a so-called solid-liquid interaction constant

Also in the process the fibers are in a spread condition, causing a larger exposure
of surface area to the epoxy/ rubber particles. Without spreading the fiber tow a coating
level of 23 % by weight was obtained, whereas with spreading a coating level of 29% by

weight was obtained with the same dilution of 1.4:1 by volume of epoxy resin

49

formulation in deionized water. This shows that spreading is also important to coating.

In many cases a sizing is applied to the carbon surface, so as to protect it from
damage. This provides sites on the carbon fibers for chemical interaction between the
sizing and the epoxy matrix to give a complex interface structure, which may result in
a strong and in many cases irreversible bonding. For this study, use of unsized fibers
have been made, therefore bonding by chemical interactions is not taken into account.

As the solid epoxy resin and/or CSBR particles are coated onto the fiber surface
the concentration of the particles within the bath decreases. This results in a decrease
in the amount of particles coated. This can be seen graphically by plotting the quantity
coated onto the bare fiber surface versus time (Figure 4.3). From the best fit line, it can
be seen that the coating concentration decreases from 21% by weight to around 18.5 %
by weight in about 16 minutes. In order to maintain a constant coating level it is
observed that over a period of about five minutes the decrease in coating level is not
substantial. So 5cm3 of fresh solution was added to the system every ten minutes to keep
the coating level steady.
4.3.3 Heating

The coated prepreg tape has a significant amount of moisture in it. Moisture is
removed in two stages within the heating section. The first section consists of a
convection oven in which the prepreg tape is heated to 190°C for a residence time of 18
seconds at a line speed of linch/second. Here most of the heat is used up in the removal
of water. At these process speeds, the oven is undersized and heating must be
supplemented by use of an additional pair of infra-red lamps located in the second section

of the heating unit. The IR lamps heat the fibers to 120°C for about 10 seconds. In this

Coating in Weight %

N
01

50

 

llll

N
O

 

 

 

 

lllJ

 

 

 

 

 

 

..x
01

llll

_a
O

 

J

l l

01

 

ill]

0

 

 

 

 

 

 

 

 

 

 

O 2 4 6 810121416
Time in Minutes

Figure 4.3 Coating versus Time

51

section the moisture is removed and the epoxy resin particles melt onto the fiber surface,
completely wetting the surface and forming an even coating on every individual fiber.
Figure 4.4 contains SEM photographs of the coated prepreg tape after it has passed
through the heating section. As is evident from the picture, the coating is very even
and there is complete wetout of the fiber surfaces. Prepreg tape thus produced is very
drapable and tacky. Despite this additional heating, moisture is not completely removed
from the prepreg tape. Figure 4.5 is a TGA analysis of the prepreg tape. From the
analysis of data for different samples, it can be seen that approximately 004-0. 12% by
weight water is still present in the prepreg tape. It is mandatory to remove this
moisture prior to curing the epoxy resin, otherwise the moisture would vaporize creating
excessive voids in the final part. Removal of this moisture will be touched upon later.

4.4 CONIPOSITE MANUFACTURING PROCESS

This prepreg tape is now ready for consolidation into a multi-laminate composite.
Several steps are involved in the composite part manufacturing operation.
4.4.1 Prepreg Tape Winding

Prepreg tape is unwound from the spool onto a specially fabricated aluminum
frame. The dimensions and details of the frame and the winder are given in Figure 2.13.
A 1.18" x 8" non-porous teflon tape is fixed to one end of the frame so that it traverses
the entire width. The purpose of this strip is to act as the initial crack propagator in the
ENF samples. Figure 4.6 is a photographic representation of the winding system.
Winding onto the frame is done in the rotary direction and in the traverse direction.
Adjustments to the speed of the two motors are made in such a way that there is no gap

between adjacent strips of prepreg tape traversing the length of the frame. As mentioned

52

IOKU X508

 

Figure 4.4 SEM Photographs of Prepreg Tape

53

 

 

 

 

 

Sample: COATED C FIBERS ER/SBR 5% ‘r‘[3 Z: File' C CFTGA.002
Size: 18 8440 mg Operator: JAMES FERNANDES
Method: 2 DEG C/MIN T0 650°C Run Date: 23-580-93 06:14
Comment: 10 DEG C/MIN T0 650 DEG C
105
101.56°C o
‘ 99,931 273.44 C
99.50%
100m
95~
g .1
g 90—
CD
8
I d
85-
432.81°C
. 80.10%
80—
75 V ' ' I . 3'0 - r0 ' 560 600 . 700
4
0 100 200 ' 0 O TGA V5 1A DuPont 2200

Temperature (°C)

Figure 4.5 TGA Analysis for Prepreg Tape

54

 

Figure 4.6 Photographic Representation of Winding System

55
previously, the prepreg tape is very tacky. Care has to be taken while unwinding the

tape from the spool, so that the prepreg tape layers do not stick together, entangle and
break. A low motor speed allows for flexibility in operations and any hard-to-handle
tacky part of the tape can be manually guided onto the frame. From calculations a
traverse speed setting of 32 and a rotary speed setting of 20 rpm were selected.
Appendix A contains details of these calculations. At the completion of this procedure
a prepreg plaque of 6" x 8" in dimensions is obtained. The plaque is then heated in the
oven at 65°C for about an hour to remove all the residual moisture. On an average, about
1 gm of water is lost from each 63—70 gm plaque. The plaque is now ready for
consolidation in the autoclave.
4.4.2 Autoclave Molding
The part is placed in a vacuum bag as per arrangement shown in Figure 4.7.

Details of vacuum bagging are as given.

0 Two clean flat steel plates of dimensions greater than that of the plaque are covered

with release ply. Bleeder cloth is added on top and wrinkles are removed.
0 Cut two pieces of porous teflon having dimensions equal to that of the plaque.
Place one teflon sheet on the bleeder cloth of one steel plate.

0 Place the unconsolidated prepreg tape on the teflon sheet.

0 Build a cork-dam around all four edges of the part with 1/8" thick cork-dam tape.

0 Place the second teflon sheet on the part and tape the ends.

0 Place the second plate on top of the part.

0 Make a hole in the vacuum bag having dimensions equal to that of the vacuum

adaptor.

56

 

To Vacuum ‘__

 

 

 

Vacuum

 

 

 

 

 

 

 

Figure 4.7 Vacuum Bag Molding Schematic

57

0 Place the entire arrangement along with the adaptor within the vacuum bag.

Seal the edges of the bagging with tacky tape, so that it is completely airtight.

Place this part in the autoclave as per the arrangement shown in Figure 4.8.

The part is then subjected to a consolidation cycle graphically represented in Figure 4.9

and which is as given below:

Pull vacuum on the part.

Pressurize the part to 150# @20#/min.

Heat the part to 212°F @20°F/ min for about ten minutes.

Heat the part to 340°F @20°F/min.

Maintain part at 340°F and 150# for about an hour.

Fifteen minutes into the cycle remove vacuum from the part.

After an hour cool the part to 100°F @10°F/min.

At about 150°F, reduce the pressure to atmospheric pressure 20#/min.
Remove the part from the autoclave once the curing cycle has ended.

This cycle is controlled through a program in the controls of the autoclave and

is given in Appendix B. The part is then removed from the autoclave and can now be

subjected to mechanical testing and micro-graphical analysis.

 

Figure 4.8 Vacuum Bagging Arrangement in Autoclave‘

59

_ 350

 

 

 

 

E300 ”u:
a 3250 s
Q" - V
E E—ZOOE
a @150 g
8 _2_ 0
a: 2100 2‘

2‘50 i3

10

O

2

 

Remove vacuum

Figure 4.9 Consolidation Cycle

4.5 SAMPLE CUTTING

Composite parts were prepared from solutions containing 0, 5, 10, 15 and 20%
by weight of CSBR particles in the epoxy resin dispersion. These parts were then cut
into test coupons in accordance with ASTM standards. All the samples were cut on a
diamond saw. See Figure 4.10. For 3-point flexural tests, dimensions of the samples
conformed to that given in Table 4.1. Typical dimensions for the test coupons are given
in Figure 4.11““). For short beam shear the sample dimensions conformed to that given
in Table 4.2. See Figure 4.11 for typical dimensional details. For End Notch Flexural
(ENF) testing the dimensions conformed to those given in Figure 4.12. Samples for
DMA, fiber volume fraction and void fraction do not have any standard specifications
and were cut from residual parts. Both volume fraction and void fraction can be
analyzed from the same sample but they have to be cast in a plastic mold.

4.5. 1 Sample Preparation for Volume/ Void Fraction Analysis
4.5.1.1 Casting of Samples in Plastic Mold
0 Cut small rectangular specimen of the composite part.
0 Place the specimen in the mounting cup, so that the fibers are oriented in a
longitudinal manner.
0 Prepare the mounting material, by mixing well 1 part by volume of hardener with
2 parts by volume of epoxy resin powder for 30-60 seconds. Place the mounting
material in vacuum bell jar to remove air particles. Care must be taken that the
mounting material does not harden prior to pouring it into the cup.
0 Carefully pour the material into the cup, so that the composite sample does not

move .

61

 

Figure 4.10 Diamond Saw“

 

 

 

 

[ l [:33
P/ZL uz .L LIZ jP/z F—d
If —I fl w
3-Ptnex‘rea mums.
wa- L- LTD? L'WDW
"' l' "015'
I - WDQI'M' h " "DEW

“I." I'll?

 

 

 

Figure 4.11 Sample Dimensions for Three-Point Flexure and Short Beam Shear
Testl‘

62

Title-4.1 mmmmammmmm-

rumlrsmw

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

wmmmun
Lid-1001 Lid-8201 L/d-wbl Ila-0001

m that Md Rad Md
Span-1mm that: ...—.... 9.1:"... m
M‘mm.)mwwmwmmwmww Mela)
MM Lam. Spam m Law. Sell. Wrea- tam Mm (Pin Lam 81mm m
mm M) MN. MN NM NM)- mlh) In.) 0'40. mm M) GIN

mh-V may "lily In!!!

IraI‘ IIaI‘ IraI‘ aaI‘
010/11) 25(1) 50(2) 1000? 050.02) 50(1) 25(1) 1.30.05) ”(2%) ”(1%) 2.00.13) ”(30 4.11%) 4.001!
1006.) 25(1) 00(2) 25(1) 0.00.03) [1(3) 50(2) 2.00.11) ”(3%) math) 4.30.17) 124(4'14 UM 0.40.31)
2407712511) 000%) 4001/.) 1.00.04)100(4) 0(3) 41010) 120047.) um.) 0.4025) 105017.) 143041) 142050
3201.) 25(1) I113) 50(2) 1.30.05) 1:00) 110(4) 5.1021) 100(7) 130(5) 0.4033) 2470?.) 1000171) 1000.70
41m.) 131%) 100(4) 110(3) 7.00.00) mav.) 150(6) 01032) 24007.) 101017.)I:.7050) 37204-7.) 200(1qu 20.40.12)
0.404) 130/1) 100(5) 100(4) 2.00.11)250(10) 210(0) 10.00.43) 3:003) 250(10) 17.00.01) 4950“) 31(15) 37.011.“
0.006) 130/.) math) 150(0) 4.10.16) 300(15) 300(12) 10.30.64) 400m) ”(15) 25.40.10) 7440417..) 5720216) 500024)
11701) 1304) 25000 210(0) 53021) “(10%) 410110) 21.01005) 040125) 51000 34.01134) 0010!) 702m 702“
10.190) “(W HHS) 110(12) 0.10.12) 740123 01000 32.5020) 0400?) mm 5001200 1480“) 110(6) 114 (4..
25.40) 25(1) 4.509%) 41004) 10.00.43) 040m 010(32)4a.4(1.71)1240(40) 11mm 07.00.07) 1041(70) 1524(0)» 152 an

 

 

 

‘mwnumammmaaor W001 RMIIIMITOWIMRIMBMMI‘DMOW

crewman-unto.maaauwuaeuwmmmmanwmauuum
d

M
'mwmmswwroaramawmmmmumm

mun Immune-hm“ ...-ems 1mg

Nnrr l—Ranfmnu may be in the form ofeinde mien.
strands. yarn. 1M woven cloth. m cloth. etc.

Nan Z—The ratio-in. are vacuum woven can to
thickness and spacers-en km to finches-Ia ralioa: M.
then are many specification in cause-nee with dill’en-m mute-d
soodm thick-Iona and man to thicken. rat-ca which take
Mm.mmwwymfytumm
nocturnal stricture. and support can so thickne- new «ad and
the specification with M it complies» Specific-m with
aluminum-satiation)” I upIoOIo I mean

identified.

Reinforces-cues Span/Thicket.“

Glass fibers

Graphite titan

Carbon fiber:

Stat Wore

Boron Manama

H10) 0011qu was:
made (am
ac.)

Abbubu

90000-4

Length]
“new

‘Smawhflmmmbvnuammfie

Fig.1.
'Toulknataforllasoau'aeeee.

63

 

 

 

 

 

 

25
' ‘Nmmrdlm
1-2 mm
0 l
. I L
LO 1
I 35
I
I
I

L 50 5O

 

 

Figure 4.12 Sample Dimensions for ENF Testing“

64

0 Cure the material at room temperature for about 60-90 minutes.
0 Remove the mounted sample from the cup.
4.5.1.2 Polishing of Sample Surface

0 The base of the cast samples are sanded down to a flat smooth surface using an 80
grit belt sander. Also one of the edges of each sample is bevelled with this sander.

0 The samples are placed in a rotary sample holder, such that the bevelled end is
exposed.

0 The holder is fixed to the rotary arm of a Struers Abrahim Automated Polisher.
Polishing of the sample surfaces are successively done using 320, 600, 1000, 2400
and 4000 grit paper at 150 and 300 rpm, respectively, for about 2 minutes at each
cycle. Each polishing cycle should remove scratches from the previous cycle.

0 The samples are then polished in a LECO GP20 polisher using 5 microns and 1
micron aluminum oxide dispersions.

0 These samples are then mounted in heavy steel holders and polished for about 12
hours in a Buelher Vibromet, which uses 0.05 micron dispersion for polishing the
sample surface.

0 Samples are then placed in an ultrasonic bath and cleaned using deionized water.
At the end of the entire process the sample surface get a mirror like finish. See

Figure 4.13 for a photograph of a typical sample.

65

manner-mosh
\ I" ,. mefllfiamag
Ion-face
(—
Mined-thee

~—— Mew-dd

 

 

Figure 4. l3 Polished Sample for Volume/Void Fraction Analysis. Composite Part
Embedded in Plastic Mold.

4.6 SUMMARY

The process was run with coating medium containing 0, 5, 10, 15 and 20% by
weight CSBR in epoxy dispersion. The prepreg tape produced at the end of each run
was then consolidated into composite plaques, as per procedure outlined in the previous
paragraphs. These plaques were then cut into samples for void/ volume fraction analysis,
mechanical testing and morphological studies. These tests would establish the validity
of the process to coat two particles from an aqueous dispersion mixture, with the final
composite having two distinct phases and the second phase being well dispersed in the

matrix.

Chapter 5

 

PROCESS MODELLING

 

5.1 INTRODUCTION

The water-based continuous prepregger is dependent upon a number of physical
parameters. These parameters can be incorporated into process models to define the
characteristics and dynamics of the unit operations involved. The entire process can be
defined in terms of discrete operations : fiber motion, fiber spreading, coating and
heating. The heating, aerosolization for coating and spreading phenomenon have been
discussed in detail by Iyerm and Padaki‘m for dry powder prepregging processes. The
aim of this chapter is to develop a process model for the heating unit and discuss the
removal of moisture from the coated tape in light of heating bottleneck in the convection

oven.
5.2 PROCESS MODEL DEVELOPNIENT

Prepreg tape coated with epoxy resin and/or CSBR particles and a thin film of
water enter the heating section. The purpose of the heating unit is to remove moisture
from the prepreg tape. The tape passes as a very thin flat sheet through the convection
oven, where the temperature of the fiber tow is raised to the boiling point of water by

both convection and radiation. (Figure 5.1). The presence of water is very detrimental

66

67

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1
Insulated aluminum walls
a
Y
7 Side View in y-Direetion
_ See ”A" .i
5 9A; 5
./7 ‘ 18" ’
Slot for Fiber Motion E Direction of Motion
).( h

T

0.358 mm Carbon Fiber Tape

_Y_

Thin Water Film will
Epoxy Ruin

 

"All

Front View in x-Direction

Figure 5.1 Heating Arrangement in Convection Oven

68

in the consolidation step, since water vaporizes at the high consolidation temperatures
causing excessive void formation in the final composite part.

The composition of the prepreg tape based upon a material balance carried out on
the tape at the exit of the dispersion bath and prior to entering the heater is given in

Table 5.1:
Table 5. 1 Composition of Prepreg Tape Prior to Heating

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
  
   

 

 

Material Weight Percent Volume Percent
Carbon Fiber 42.42 30.14
Epoxy Resin 17.42 19.14

Water 40.16 50.72
5.2.1 Process Data

Line speed 1 inch/sec
Coating of solid, by weight 29.0%
Dilution of solution in bath 1:1 by volume
Weight of bare fiber 0.3583 gm/32"
Density of epoxy resin solids 1.137 gm/cm3
Density of carbon fibers 1.783 gm/cm3
Oven Temperature 190°C
Emissivity of IR lamp 0.39
Emissivity of water 0.96
Emissivity of aluminum 0.18
Thickness of dry coated tape 0.252 m I
Width of dry coated tape 2.30 mmJ
Length of tape travel in oven 18" I
Thermal conductivity of air 0.02195 Btu-ft/hr' -ft2-°F I
Thermal conductivity of water 0.353 Btu-ft/hr-ft’-°F
Thermal conductivity of resin 0.35 W/m-“K
Thermal conductiv=ity of fibers 4.5 Btu-ft/hr-ft2-°F

 

 

 

69
5.2.2 Development of Equation

Consider the prepreg tape, shown in Figure 5.1, moving as a thin flat sheet
through the convection oven at constant line speed. Individual fiber strands are all
moving at the same speed. The temperature is constant in the y—direction. This tape is
being heated with hot air over the entire surface of the fiber. It is also assumed that the
bulk temperature of air, Tb, is constant. At any time, 1, from the start of the heating
operation i.e. when the prepreg tape enters the oven, the quantity of heat, dQ, transferred
in a short time, dt, depends upon the surface area of the tape, the difference in
temperature between the air and the surface of the tape and the heat transfer coefficient

of air, 11,, such that” :

d
.3? = hcA(Tb-T) (5.1)

where T is the surface temperature of tape after time dt.
A is the surface area of tape after time (it.

The heat of conduction in the slab is given as:

do = mcpdT (5.2)

Equating equations 5.1 and 5.2 results in:

(IQ = mcpdT =h‘A (Tb- Dd: (5.3)
As explained earlier, part of the heat transfer in the oven is convective in nature.
This can be either forced or natural. As the tape is moving within the oven some forced
convective currents are also produced, which can result in heating of the tape. If the

forced-convection effects are very large, the influence of natural—convection currents may

70

be negligible, and similarly, when the natural-convection forces are very strong, the
forced-convection effects may be negligible. The ratio of the Grashof‘s number to the
square of the Reynold’s number establishes the dominance of one mode of heating over
the other”. If the ratio is much larger than one, then natural convection is dominant
and vice versa. In this case, the ratio was much greater than one indicating that the
mode of heat transfer in the oven is predominantly natural convection.

In equation 5.3, the value for heat transfer coefficient is difficult to determine as
it depends upon the units employed, physical properties of the fluids, dimensions of the
apparatus, velocity of the fluid past the surface and often even on the temperature
difference, AT “9‘. For an accurate estimation of heat transfer by convection it is
essential to employ point or local surface coefficients. However, for most engineering
applications, it is quite safe to use average heat transfer coefficient values based upon the
average of the local heat transfer coefficients over the entire surface area. For this
model the average heat transfer coefficient will be used. Even when average heat
transfer coefficients are used, its absolute value can fluctuate tremendously. For air, in
case of heating or cooling, the average heat transfer coefficient can range from 1.2 to 57
Watt/mz-K. Thus the importance of a fairly accurate estimate of the average heat
transfer coefficient is essential.

5.2.3 Experimental Evaluation of Heat Transfer Coefficient

The average heat transfer coefficient can be estimated experimentally by heating
an aluminum slab in the oven from a certain inlet temperature to the set temperature of
the oven. The experimental set up is as shown in Figure 5.2. Two aluminum slabs are

taken having the dimensions L x w x h. A k-type thermocouple is placed between the

71

k - Type Therm ocouplc

 

Flow of Air from Convection Oven

Figure 5.2 Experimental Set-Up for Evaluation of Heat Transfer Coefficient

72

two slabs, such that the sensing tip is in the center of the slab, far away from the edges.
The convection oven is heated to a certain set temperature, such that the bulk temperature
of air is Tb. The slabs are held tightly together with insulating tape stuck around the
edges. It is assumed that heat transfer from the sides of the slabs is negligible, as the
slabs are fairly thin. There is continuity of heat flow at the faces of the slabs that are
in contact with the flowing air. It is also assumed that temperature remains constant in

the yz—plane. The energy equation reduces to :

6T_ azr
—-a

__ (5.4)
a: 3x2
where a is the thermal diffusivity of the material
The following initial and boundary conditions hold true:
Initial Condition t = 0 0 s x s h T = 'I‘i
Boundary Condition 1 Fort > 0 x = 0 fl]. = 0
(37: =0
Boundary Condition2 Fort > 0 x = j; h 4% = hc(T-Tb)
=1]!

where k is the thermal conductivity of the aluminum slab.

hc is the heat transfer coefficient.

Equation 5 .4 can be converted into dimensionless form by substituting the
dimensionless variables 5 = x/h, r = ta/hz and 9 = (T), - T)/(Tb - Ti), such that the

equation now becomes:

73

66 826 (5.5)

and the initial and boundary conditions now become:

 

LC 1:0 0523] 9:1

B.C.1 Forr>0 2:0 532 =
65...,

BC. 2 Forr > o z: 21:1 § = -Bi9
aEE-tl

 

By separation of variables and only considering the transient part of the equation,

the solution of Equation 5.5 with the given conditions is as given below:

 

 

 

T -T ‘° 2
b = 2 Bnexp - A" at cos[ Ex] (5")
Tb-Ti n=1 . ’12 h

where 131] are the Eigen Coefficients and has the form 25in xn/(x,+ sin 2M2)

Kn are the Eigen Values and solutions lie in Antan X, = Bi

x is the Slab Location

A program to solve equation 5.6 using Fortran is given in Appendix C. The
program solves the given equation for an inputed value of the heat transfer coefficient
and generates the temperature at x = 0 versus time. The eigen values of the above

equation are calculated in a sub—routine in the program using the Newton Raphson

74

Iteration method and the values are converged to a difference of less than 1 x 10'”.

The oven is heated to a preset temperature and allowed to stabilize for some time
at this temperature. The two slabs, conforming to the arrangement as shown in Figure
5.2, are placed in the oven and the temperature at the center of the slab ,i.e. at x = 0,
is noted after short intervals of time (after every 5 seconds). This process is continued
until the temperature at the center is near that of the oven set point. Data is plotted for
the experimental results of temperature versus time. (Figure 5.3). The program
generates theoretical values of center line temperature versus time at different values of
heat transfer coefficient, with all the other physical parameters like thermal diffusivity
and thermal conductivity of aluminum, initial temperature, bulk temperature and half
thickness are constant. A heat transfer coefficient of 25 Watts/m2-°K was found to give
the best fit with experimental data, with the conductivity and thermal diffusivity of
aluminum as 201 Watts/m-°K and 8.5 x 10‘5 mzlsec, respectively, half thickness as
0.00635 m, bulk temperature as 129°C and initial temperature as 22°C. (Figure 5.3).

It was discussed earlier that in the convection oven, some heating also takes place
due to radiation. So it is important to evaluate the heat transfer coefficient due to
radiation and see if there is an overall impact on prepreg tape heating due to radiation.
The heat transfer coefficient due to radiation in a grey body is evaluated by the given
formula“:

11. = 71.2 F r (5.7)

where h, is the radiative heat transfer coefficient

.9'1_2is the gray-body shape factor

FT is the temperature factor

75

I40

l20

coefficient = Watts/sq

80

60

40

Temperature at Center Line in Degrees C

20

0 100 200 300 400 500

Time in seconds

Figure 5.3 Experimental Evaluation of Average Heat Transfer Coefficient:
Temperature versus Time

 

600

76
The value of h, based upon Equation 5.7 is only 6% of the heat transfer

coefficient due to convection, therefore heating due to radiation is very minimal and so
can be neglected.
5.2.4 Moisture Removal in Convection Heater

Prior to using Equation 5.3 for the evaluation of time taken to raise the prepreg
tape to 100°C and amount of water removed after that, it is essential to ascertain the
mode that limits heat transfer to the tape. For this the Biot number has to be evaluated.
The Biot number” measures the significance of internal (heat conduction within the
solid) to external (heat convection to the surface of the solid) heat transfer resistances and

is given by :

Bi = — (5.8)

where h is the heat transfer coefficient which includes contribution from convection
and radiation.
ks is the thermal conductivity of the solid.
L is the significant length dimension obtained by dividing the volume of the
body by its surface area.
If the Bi < 0.1 then the resistance to heat transfer is due to external heat transfer
i.e. convection. The thermal conductivity of the solid for the prepreg tape is a
combination of the conductivities of water, resin and carbon fibers. It is assumed that
the resistance to heat flow in the water/resin system is parallel in nature, whereas that
of the fibers is in series with the two. The thermal conductivity of the water/resin

system was evaluated using a parallel resistance analysis and this came out to be 0.534

77
Watts/m-°K. Relationships for conductivity through two-phase materials have been

discussed as Generalized Rayleigh-Maxwell Relationships‘zo]. In the case of parallel

fibers and transverse heat transfer, the Rayleigh form of the relationship is as follows:

k
2(k—f-1]Vf
k.=l+k "'k (5.9)
—f+1 - —f-1 V
km km f

where RC is the thermal conductivity of the composite.

kf is the thermal conductivity of the fiber.

km is the thermal conductivity of the matrix (water/resin system in this case).

Equation 5.9 was used to evaluate the thermal conductivity of the prepreg tape
and this was 2.03 Watts/m-°K. Based upon the evaluated values of heat transfer
coefficient and thermal conductivity, the Biot number was calculated to be 0.0019 which
is much lesser than 0.1. This implies that external heat transfer resistances are rate
controlling and conduction across the thickness of the tape is not significant. In this case
a lumped model for heat transfer gives a fairly accurate representation of the physical
situation existing within the oven. Therefore, integrating Equation 5.3 from zero time,
when the temperature of the prepreg tape is T), to any time t, when the temperature of
the prepreg tape is T, gives the following equation:

, = flm(fl] (5.10)
hcA Tb-T

where m is the mass of the material in lb/sec.

cp is the specific heat capacity in Btu/lb-°F.

78

A is the surface area of the material in ft’lsec.

From Equation 5.10, the time taken for the tape to reach 100°C is around 13.5
seconds. After the tape has reached this temperature, all of the heat transferred is used
up in evaporating water from the tape surface. For the remaining time approximately
5-6 % by weight moisture is removed from the convection oven. Hence, the convection
oven does not have the capacity to remove the moisture content from the prepreg tape.
This phenomenon was also observed physically, since the tape produced with this
configuration had a lot of moisture on it.

5.2.5 Moisture Removal by IR Lamps

Since the convection oven is too small in size to remove water from the prepreg
tape completely, heating is supplemented by two IR lamps. Each lamp is provided with
a 375-watt IR bulb. The mathematical formula used to evaluate the heat transfer due to

radiation for each of these bulbs is as given in Equation 5.11m].

FAAlo

“(i)+(-l-)—l‘”'°
e1 62

where Q is the heat flow or net heat exchange

 

51,152 are the emissivities of surfaces 1 & 2, respectively

F A is the geometric factor

Al is the surface area of the emitting surface

T),T2 are the absolute temperatures of surfaces 1 & 2, respectively
a is the Stefan-Boltzmann constant

Based upon Equation 5.11, each of these lamps is capable of supplying a

79
maximum quantity of 73.35 Btu/hr heat to the prepreg tape, when the lamp bulb is about

2.5" above the surface of the tape. The first of the two lamps, which is located 2.5"
above the surface of the tape, removes around 95 % by weight moisture. The second
lamp is adjusted in such a way that some moisture is removed, but the resin is not
effected. It was physically observed that if the second lamp was placed very near the
surface of the tape the resin would begin to smoke and smell, indicating some reaction
or burning of the resin. It was therefore decided to place the second lamp well above
the tape, so as to remove maximum moisture and at the same time maintain the integrity
of the matrix material. As will be discussed in the later chapter, the final prepreg tape
has some moisture, which is removed by heating in the convection oven at a lower

temperature prior to consolidation of laid up part.

Chapter 6

 

RESULTS AND DISCUSSIONS

 

6.1 INTRODUCTION

The primary objective of this study was to develop a continuous prepregging
process for the manufacture of two-phased composite materials using a two mixture
aqueous dispersion as the coating medium. Chapters prior to this deal with the
development of the process, consolidation of the prepreg tape into composite parts and
preparing samples for characterization of the parts to see if the two coated particles exist
as distinct phases in the composite part. As mentioned earlier, the composite part is
manufactured of two materials in the matrix, i.e. the epoxy resin and styrene/butadiene
rubber (CSBR). Epoxies are inherently brittle materials and so have very low fracture
energies. Addition of CSBR particles to create a multi-phase system, may result in a soft
rubbery inclusion dispersed in the brittle epoxy matrix, which if properly bonded to the
epoxy, increases its fracture energy. Forthcoming topics will deal with the different
experimental techniques used to identify the two phases, evaluate the effect of CSBR on
the epoxy matrix and discuss these results in the light of earlier work done around

rubber-toughened epoxy matrixes.

80

81
6.2 VOLUME/VOID FRACTION ANALYSIS

The stiffness and strength parameters of fiber composites are determined by the
internal packing geometry of the fibers and the constitutive behavior of the fiber and the
matrix”. The geometry of the composite part is generally defined by a transverse plane,
in which the fibers appear as circles embedded within the matrix. The volume fraction
of the constituents of a composite may be determined by chemical matrix digestion or by
photomicrographic techniques. Emphasis will be placed on the latter of the two
techniques, as it is the preferred method of analysis in this study. This technique also
can be used to determine the void content within the part. It is imperative that an
adequate number of samples are taken randomly throughout the composite part and
enough surface should be examined under the microscope to make a definitive analysis
of the voids and fiber content. For a good composite part it is necessary for it to have
a void content of less than ~ 1% to carry out mechanical testing on it. Also for the
mechanical analysis to be consistent from part to part, the fiber volume fraction for most
plaques must be fairly consistent. Details for preparation of samples for volume/void
fraction analysis are given in section 4.5.1 and the equipment set-up/operations is given
in section 2.4.3.

Thirty images of the polished surface for each sample were scanned by the
digitizer and the images were then analyzed by the Optical Numeric Volume Fraction
Analyzer (ONVFA). Results of this analysis are given in Table 6.1. It can be seen that
the void fractions of all the composite parts are well within the prescribed limit of less
than ~ 1%. A typical micrographic image at a magnification of 65 is as shown in Figure

6.1.

82
Table 6. 1 Volume/Void Fraction Analysis

 

 

 

 

 

 

 

 

 

 

 

CSBR Content Void Fraction Vol Fraction

(%) (%) (%)
0 0.41 65.57
5 0.70 65.48
10 0.64 67.92
15 1.08 63.56
f 20 0.52 _ 61.57

6.3 MECHANICAL TESTING

Mechanical testing for the composite parts consists of three-point flexure, short
beam shear and Mode H fracture analysis. All these tests are carried out on the United
SFM Testing Machine. Details of this equipment are given in Chapter 2.4.4.

6.3. 1 Three-Point Flexure Test

The primary objective of the three-point flexure test as shown in Figure 6.2 is to
determine the stress-strain response of the laminated composite part in bending”. In this
test, the central load, applied by the central loading pin, is monitored versus strain,
which is measured by a deflectometer located on the tension side of the beam (bottom
surface). Commonly, a unidirectional laminate with the fiber direction along the beam
axis is loaded in bending to determine the flexural modulus in the fiber direction, the
flexural strength in the fiber direction and the ultimate flexural stress and strain. The
flexural modulus and strength is a combination of the tensile and compressive properties

of the material.

83

~1r a ' .r.‘_-Nl""" 1 swat-03' 1‘
0“ ‘. " 4”,. “...-r ..-' 0P1“ ."
n “4 av «.fi. .:.:‘ 33-52:. 0.-

\
‘7‘ :P‘r‘f I la . “(’0‘.’ ‘r
. r- ( ‘g“~"£’ 4“.
. :5 ‘50" :j,
'3- .o. 7 . :3 361:3
£73.33de
‘G'SJ'I'T'a

 

Figure 6.1 Micrograph of Polished Composite Surface at Magnification of 65X

 

Caller-inhale
“HMO“ mm“.
Swath _Swltl‘h
W

 

 

Figure 6.2 Arrangement for Three-Point Flexure Testing

 

34

All the testing was done as per ASTM D790—86. Five sample coupons were

tested from each composite plaque containing 0, 5, 10, 15 and 20% by weight CSBR

particles in the epoxy matrix. Consolidated results of this test for the different samples

are given in Table 6.2. A graphical representation of the flexural strength and flexural

modulus with varying concentration of CSBR particles are given in Figures 6.3 and 6.4,

respectively. Data output in graphical form for load versus extension as generated by

the UTS software is given in Appendix D.

Table 6.2 Flexural Properties of Composites with varying CSBR Concentration

 

 

 

 

 

 

 

 

 

 

 

CSBR Flexural Standard Flexural Standard
Content Strength Deviation Modulus Deviation
(% by wt) (MPa) (MPa) (GPa) (GPa)
0 1,210 87 124 6
5 1,175 20 122 3
10 965 31 118 3
15 848 103 104 21
20 788 1 13 85 19

 

85

1400

1200 L—e. I

 

 

 

 

 

 

 

 

 

 

 

 

 

’3
E. 10003 \l\;\
E 8003 _, 4;
m 600i 1'
E 400:
2003
0 5 10 15 20

CSBR Content (by weight %)

Figure 6.3 Flexural Strength of Composite with
varying CSBR Concentration

140
120;
100i
sof
605
40f
20:
0‘.... .... .... ....

0 5 10 15 20

 

 

 

 

 

 

 

Flannel Modulus (GPa)

 

 

 

 

 

 

 

CSBR Content (by weight %)

Figure 6.4 Flexural Modulus of Composite with
varying CSBR Concentration

86
6.3.2 Short Beam Shear Test

Short beam shear testing methodology is described in ASTM D2344-84. The
testing equipment is similar to that used in three-point flexure test, the only difference
being the loading nose and the distance between supports. (See Figure 6.5 for details).
This test determines the inter-ply strength, with failure occurring in horizontal shear.

The apparent shear strength for each sample is calculated by the following formula:

P
s” = 075—2 (6.1)
bd

where SH is the short beam shear strength in N/m2 or psi.

PB is the breaking load in N or lbf.

b is the width of specimen in m or in.

d is the thickness of specimen in m or in.

Samples cut as per guidelines set in section 4.5, were tested for short beam shear
strength. The load at break for each of these samples was noted and the shear strength

was then calculated as per equation 6.1. These results have been tabulated in Figure 6.6.

87

 

 

 

 

 

Figure 6.5 Arrangement for Short Beam Shear Strength Analysis

 

100
ix 1

80 j' ‘\

soi \

40f

 

 

 

 

20f

 

 

 

 

 

 

OE... --..
0 5 10 15 20

ShortBeam Shear Strength (MPa)

CSBR Content (by weight %)

Figure 6.6 Short Beam Shear Strength for
varying CSBR Concentrations in
Composite Samples

 

88
6.3.3 Fracture Energy Analysis

The failure mode of epoxy resins is generally characterized by flaw growth and
progressive crack propagation. Fracture mechanics deals with the presence of flaws and
crack propagation, fracture analysis of the tested coupon is generally used to understand
the failure mechanism in epoxy matrixes. Fracture occurs when sufficient energy is
released from the applied stress field to generate new fracture surfaces at the instant of
crack propagation. This strain energy release rate provides a measure of the energy
required to extend a crack over a unit area and is termed the fracture energy (Id/m2) [5].
Several tests exist for the evaluation of the delamination failure mode. Most of them are
limited to unidirectional laminates where the crack propagates between the plies along
the fiber direction. In this study, the End-Notched Flexure (ENF) test was chosen to
determine the fracture energy of the material. Details of the ENF test are given in
Appendix E. From the elastic beam theory an expression to evaluate the fracture energy
in pure Mode II loading is given asml:

2 2
G” = 9P Ca (6.2)
‘ 2w(2L3 + 3a 2)

 

where GIIc is the fracture energy at critical load during pure Mode 11 loading in kJ/mz.
Pc is the critical load in N.
C is the compliance of the material in mm/N.
a is the crack length in mm.
w is the specimen width in mm.

L is the span between the central loading pin and the outer support pins in mm.
The compliance may be calculated from the following formula based on the beam

theorym‘:

89

= 2L3+3a3
8Ewh 3

C

where E is the flexural modulus in the axial direction of the beam in GPa.

h is the semi thickness of the beam in mm.

It is, however, recommended that the compliance be measure experimentally.
Test specimen for EN F testing are so prepared that an initial crack having length equal
to "a" , such that a z L/2. At this condition the crack growth is unstable even under fixed
grip conditions. (Details of this derivation are given in Appendix E). A three-point
flexure fixture with a total span of 100 mm is used to test the ENF specimen.
Consequently, the initial crack length is 25 mm for the crack growth to be unstable.
Prior to testing, the crack should be carefully wedged open and extended about 2 mm
beyond the insert to achieve a natural starter crack.

ENF samples for 0, 5, 10, 15 and 20% by weight CSBR in epoxy matrix were
tested as per guidelines of Appendix E. Compliance was evaluated by linearly regressing
twenty random data points of each test sample, as shown by a sample spreadsheet in

Appendix F. The equation“ used for compliance calculation was:

 

b = (nEXY - EXEY)
nEX2 — (2)02
where b is the material compliance in mm/N.
X is the load on the test specimen in N.
Y is the extension in mm.

The spreadsheet also incorporates equation 6.2 for calculation of the material

fracture energy. Appendix G gives the graphical output of the ENF test data plotted as

90

load versus extension. (See Figure 6.7 for a graphical representation of the fracture
energy as a function of increase in CSBR concentration in epoxy matrix).
6.3.4 Discussion of Mechanical Testing Results

From Figures 6.3, 6.4, 6.5, and 6.7 it can be seen that the flexural strength,
modulus, the short beam shear strength and the fracture toughness all decrease with the
increase in the CSBR concentration in the epoxy matrix. These results are to be expected
as the increase in the volumetric amount of the " soft" rubbery inclusions proportionally
lowers the mechanical properties of the matrix like flexural strength, modulus and short
beam shear strength. Work done by Ting et. al. ’12:]. [2“ show similar results for a
graphite epoxy system modified by a CTBN elastomer except that the fracture toughness
increases with increasing elastomer concentration. As the concentration increases, the
softer rubber particles present in the second phase begin to dominate the properties of
the epoxy matrix by a simple rule of mixtures phenomenon. As mentioned earlier, epoxy
materials in itself are very brittle and are highly susceptible to rapid crack propagation
upon application of load. On the other hand rubbers are very tough and resilient
materials and are capable of withstanding large loads before complete failure occurs.
When these rubber particles are dispersed within the epoxy matrix, they act as crack
arresters in case of a crack propagation. The softer rubber particle requires a lot more
energy for the crack to propagate through it, therefore as the concentration of the rubber
particles increases in the epoxy matrix the fracture energy should increase, i.e. the
amount of energy required for the crack to propagate throughout the epoxy matrix. As
seen in Figure 6.7, the fracture energy of the composite samples decrease with the

increase in CSBR concentration. This behavior is quite contrary to what should have

91

 

 

 

 

 

 

 

 

 

 

 

 

 

1400,
1200f 1’
<3? 2.
E 1000 (\-
a a.
as 800; ..\
33 : \ 0
L5 600: .. \-
a 400: "
a _
is? 3
LL. 200_
03....
o 5 10 15 20

CSBR Content (by weight %)

Figure 6.7 Fracture Energy of Composite Parts with varyings CSBR
Concentration

92

been expected from the above discussion.

An important aspect to consider for the propagation of the crack through a rubber-
toughened epoxy matrix is the interfacial adhesion between the epoxy matrix and the
rubber particles. Work done by Kinloch et. al.9514261, Miilhaupt et. al.“, Qian et. 31,1231
and other relevant literature place a strong emphasis on the interfacial bonding of the
rubber particle and the epoxy matrix for improvement in the toughening characteristics.
In toughened resins, as the crack propagates through the matrix a dilational stress field
is formed causing cavitation of rubber particles”. The cavitation is initiated by
interfacial debonding around the particles. In case of a strong interfacial bonding, the
epoxy matrix will get tougher. A possible explanation for a decrease in the fracture
energy of CSBR toughened epoxy/carbon composites with the increase in CSBR
concentration, is the weak interfacial bonding between CSBR particles and epoxy matrix.
This study is beyond the scope of the current work being done as the primary thrust of
this thesis is to develop a continuous prepregging process for coating two particles from
a water-based mixture of the two particles and consolidating them to form two phase
composites.

6.4 DYNAMIC MECHANICAL ANALYSIS (DMA)

Dynamic mechanical instruments measure the deformation of a material in
response to vibrational forcesm]. This equipment measures the dynamic modulii over a
wide range of temperatures. These dynamic parameters are used to determine the glass
transition region, relaxation spectra, degree of crystallinity, molecular orientation,
crosslinking, phase separation, structural or morphological changes resulting from

processing, and chemical compositions. For this study the DMA will be used to see if

93

two materials, i.e. the epoxy and the rubber particles, exist as two distinct phases within
the composite matrix. While running the DMA over a range of temperature, it is seen
that there is a distinct peak in the loss modulus of the material at its glass transition
temperature. In case of the presence of a second phase in the material, there will be
another peak corresponding to the glass transition temperature of the second material“.

The samples for the DMA are cut as long rectangular strips, with the fibers
aligned along or traverse to the axis of the samples. For consistency in results the fiber
orientation of all the samples should be same, as the absolute values of the modulli are
very sensitive to the fiber orientation. Since sufficient material was not available from
each individual plaque, the fibers are not aligned in the same direction for all the
samples. But in this study, as only the presence of two phases is of significant
importance, consistency in the alignment of the fibers is not important.

Figure 6.8 through Figure 6.12 show the results of the DMA analysis. Analysis
results for the 10, 15 and 20% by weight CSBR samples show two distinct peaks
occurring at or around the glass transition temperatures of the two materials. This
clearly indicates that the CSBR particles exist as a distinct phase in the epoxy matrix.
DMA analysis for the 0 and 5% by weight CSBR samples show only one peak. This is
expected for the 0% part, but not for the 5 % part. One possible explanation for this
could be that the instrument is not able to detect CSBR particles at such low
concentration levels. Phase separation occurs during the reaction because the increase
in the molecular weight of the epoxy resin diminishes the entropy of the blendm]. The
chemical reaction drives the change in free energy, so this entire process is kind of a

"chemical quench ". The morphology formed (diameters, number and volume fraction

94

Sample 1 or 5811 IN tooxv/cmaou cow

 

      
 

 

 

 

File ‘ C OflAOSBR 1
S129 ‘ 27 4300 n B 1600 I 2 3500 an [3 h4 Z\ Operator JAMES FERNANOES
Method - -7o'c to 225°C nun Oata. zs-Jan-Qa 10 37
Conaenr
60 4
Amplitude In-o) -0.20 aa
115 9c ’
50-4
- 3
A ‘0'“ \ b o;
O \ Q
i i 9
. 4 I a
U \ ~2 u
— so~ \‘ :
\ I
I I
\ , I
... J \,,ae"’" 'N--\ r h.
x/
20" 'ft )— 1
r‘l/
4 /
i / >
10- "W“.
T Y I I 1 I o
-100 -50 o 50 100 150 200 an H 2c
Teloeraturt 1°C)

Figure 6.8 DMA Analysis of 0% CSBR Composite Part

 

 

 

 

 

Sanola f 51 SSE IN EPOXY/CARBON coap File ' C DIASSBR 2
5120 35.5000 1- 6 2000 n 2 7400 not DM A Ooerator' JAHES FERNANUES
Method : -70'C To 225°C Run Oata‘ 25-Jen-94 12 35
Conant
10 800
AnoirtqulD-Dl-O.2O an
a
b
fig 4‘ 101.7'1:
I
I
" L-600
4 I
I
I
.. 6~ '. . -
O ) I
n.
2 '1 3
. ‘ I u
I
u, 1 b 400 w
4- 1 T
' 1
‘ 1
| 4
_. , 1 > ..
1
i
l
2- ‘1
1 F 200
4 ‘.
‘\
0—1
I Y Y Y 1 O
~100 -50 0 so 100 150 200 an v4 21:
Temperature 1°C)

Figure 6.9 DMA Analysis of 5% CSBR Composite Part

95

Sample ' 10: 584? IN Even/cm
. SON COHP
5 file .
4:00 . 3509303012254300 n 2 4800 an D M A Operator, Sagzgigzgznloes
Consent FlNGtfi TIGHYENEO CLAIPS Run Dale. tQ-Jan-Qd 08 ‘0
10

   

 

 

no) nude Ie-o) men an 500

 

  

 

 

 

 

 

 

\
‘1‘ 400
1
‘1
- 6 ‘. i
O I
9, ' -
7' " .— 300 o
.. .| g
.. ‘ “ . W
I -—
I 1
‘. l
‘ )- 200 I
1 .1.
‘i
2 -35.19'c ,
7» "I
.'— ‘\ I
‘ \\ l b
\y ,I 100
\_‘ ’1
O ----- L
«100 -52, 5 .
50 100 15:0 ' -o
2
’I-OePaturet'C) 00 m V‘ 2:

Figure 6.10 DMA Analysis of 10% CSBR Composite Part

Sewiv ' 15! 569 CWVOSIIE

 

    
 

 

 

 

Hie ‘ C 00141558312 1
S119 ' 45 1800 l 6.4300 :1 A 5500 In DM A Operator” JAMES FERNANOES
Method ' -7o'c to east Run Date io-Fao-Ql “‘07
Comment.
50 AaolituueIo-ol-o 10 a-
~ 4
1
’1 “79°C
I
' 1
40-4 ,' I r
, 1
1 '.
1 1 )-
4 1 , 3
' 1
' 1
1' ' 3'
3 ' ’
30‘ 1
a 1 g
\— . I
u 4 |‘ _ 2 w
| 'T
I 7
‘1 :
I mm x r IL
— I
1
+ -22.22°C ‘ )- ,
I 1
I
W]
10~ ,
. r0
9.0.. ' -go a go .60 so ' 260 an .. 2c

tewaratura I‘C)

Figure 6.11 DMA Analysis of 15% CSBR Composite Part

96

 

    

 

 

55-012 1 20! $89 COHPGSIIE File (2‘ 044420589 4
$171.- ' 33 1400 n 5 9800 - 2 6800 a- DM A Operator“ JANES FEflNmOES
IPthOO : -70'C Io 225°C nun Oate' 104215-94 08' 34
Cement
8
8° 4”“!qu ID-Dl .0 20 n-
116 3°C _ 6
1
60-1 .
i
I
1 1
‘1
.. J .3
o ' o
§ ‘1. _ . e
. 1 w
‘” ‘0‘ '. _.
‘" 1‘ . i
1 l
‘ l
.- I _'
‘1 ~ 2
I
20 -14 12°C
"‘ r
_/ .
.-’A.\~ /,,.
.1 ‘\.-‘_'-¢’
‘ hf” (.0
9100 $0 15 570 160 150 260 an "ac
rowerature 1°C)

 

Figure 6.12 DMA Analysis of 20% CSBR Composite Part

97

of the dispersed spherical particles) is determined by the competing effects between
phase separation and polymerization rates. Shift in the glass transition temperature could
be attributed to the possibility to chemical reactions at interface during the sample’s high
temperature cure cycle.

6.5 MORPHOLOGICAL STUDY

Crack surfaces obtained after EN F testing, were studied under the SEM to
identify the morphology of the rubber-toughened epoxy matrix. Figures 6.13 through
6.22 show the photomicrographs of the morphology of the cracked surfaces of the ENF
composite samples with varying concentration of CSBR particles taken at a magnification
of 5,400 and 10,000. Figure 6.13 and 6.14 shows the fracture surface of 0%CSBR
composite part. Features of the surface are typical of a pure epoxy matrix failure.
There are well defined ridges and planes. Figure 6.15 and Figure 6.16 shows the
fracture surface of 5% CSBR composite part. The features are very similar to that of
the 0% CSBR part, except that there are small cavities evenly dispersed all over the
surface. These cavities are an identifying feature of a rubber toughened epoxy
composite. This porous structure results from the cavitation of rubber particles in
response to the dilational stress field, as a crack propagates through the material”.
Figures 6.17 through 6.22 show the fracture surface of 10, 15, and 20% CSBR
composite parts. The porous structure in all these cases is more pronounced and the
cavities are dispersed all over the epoxy matrix indicating a larger concentration of CSBR
particles in the epoxy matrix. Also the morphology has a spongy texture indicating gross
plastic deformation. The morphology of all the fracture surfaces are typical to that cited

in literature.

98

is,

151w x5400 0001 1.0—u 05094

 

Figure 6.14 ENF Fracture Surface for 0% CSBR Composite Sample at 10,000X

99

 

Figure 6.15 ENF Fracture Surface for 5% CSBR Composite Sample at 5,400X

 

ISKU X10060 @052 1.0U CE094

Figure 6.16 ENF Fracture Surface for 5% CSBR Composite Sample at 10,000X

 

101w x5400 0103 1E1 05094

Figure 6.17 ENF Fracture Surface for 10% CSBR Composite Sample at 5,400X

 

10KU X10000 0102 1.0U CEU94

Figure 6.18 ENF Fracture Surface for 10% CSBR Composite Sample at 10,000X

 

Figure 6.20 ENF Fracture Surface for 15% CSBR Composite Sample at 10,000X

4K

15KU X5480 0203 1.8U CE094

R -

15KU X10000 0202 1.8U CE094

 

Figure 6.22 ENF Fracture Surface for 20% CSBR Composite Sample at 10,000X

103

6.6 SUMMARY

Results of mechanical testing indicate that as the concentration of the "soft"
rubber inclusions in the epoxy matrix increases the mechanical properties like flexural
strength, flexural modulus and short beam shear strength, decrease. Including a rubber
phase should increase the fracture energy of the composite material, but in this study the
fracture energy decrease with increasing concentration of CSBR particles. A possible
explanation could be a weak interfacial adhesion between the epoxy matrix and the
rubber particles. DMA analysis show two distinct phases in the final composite parts.
Morphological studies show an even distribution of the CSBR particles in the epoxy

matrix and the features are similar to other rubber-toughened epoxy matrix systems.

Chapter 7

 

CONCLUSIONS AND FUTURE WORK

 

The primary objective of this study was to develop a continuous prepregging

process for coating fibers with at least two polymers suspended in an aqueous mixture.

Based upon this study a number of conclusions can be drawn and a variety of

recommendations can be made for future work.

7.1 CONCLUSIONS

1.

The developed water-based prepregging unit was successfully used to manufacture
quality prepreg tape coated with two different polymer particles from a colloidal
aqueous mixture of each. SEM study of the tape indicated an even coating spread
over the entire surface of the carbon fibers. This process is independent of
solvents. Small colloidal particles improve wetting of fibers, prevent agglomeration
and coating is more even. Water, as a dispersing medium, is environmentally
friendly. The prepreg tapes produced were fairly tacky, but despite this fact they
wind off quite easily from the spool during composite part lay-up.

The prepreg tapes were consolidated to give void-free composite parts having
varying concentrations of CSBR particles dispersed all through the matrix. These
multiple phase matrix composites were fabricated independent of the compatibility
of the matrix components.

As the concentration of the CSBR particles was increased in the epoxy matrix,

mechanical properties like flexural strength, flexural modulus and short beam shear

104

105

strength all went down. This is because increasing the concentration of the "softer"
CSBR particles causes the mechanical properties of these particles to dominate the
properties of the epoxy/carbon composite. These results are consistent with other
research work done in the area of rubber-toughened epoxy matrixes.

A secondary objective was to study the improvement in the fracture behavior of the
epoxy matrix by the increase in concentration of the elastomer. In this study it was
observed that the fracture energy decreased with the increase in concentration of
the rubber particles. A possible explanation could be the poor interfacial adhesion
between the rubber particles and the epoxy matrix, causing the crack to propagate
through the interface with relative ease.

Morphological study of the fractured surface indicated distinct cavitation in the
epoxy matrix due to propagation of the crack. These results are consistent with
that seen in literature. The rubber particles are well dispersed through the entire
volume of the matrix. Also a lot of plastic deformation was noticed in the fracture

surfaces that had a higher concentration of the rubber particles.

7.2 FUTURE WORK

The ease with which two particles can be coated from an aqueous dispersion onto

the fiber surface independent of the compatibility of the two particles in suspension, as

long as they remain in suspension, leaves several avenues open for modifications in the

current set-up, developing appropriate mathematical models and application to other

processes. These are as given below:

Mechanisms of coating should be studied in more depth and a model should be

developed to understand the coating phenomenon.

106

The process should be run with suspensions of modified rubber particles and the
effect of the interfacial adhesion should be studied on the epoxy matrix.

The convection oven should be replaced with an IR heater. This will also help if
a scale-up and high speed version of the process is to be developed.

Coating on fibers is currently done in a dispersion bath. Other means of coating
should also be looked into. Some prospective coating methods would be spraying
with micron spray-heads and humidification of suspension.

The fiber motion system will have to be improved by incorporating proper
controllers and motors at each pinch roller. It is imperative to synchronize all of
the motors to produce motion in the fibers such that there is no tension in them.
Zero tension helps during the spreading process.

The current study was carried out by using an aqueous dispersion of epoxy resins.
The process should also be run using aqueous dispersions of other thermoplastic
and thermoset resins. It could also be envisaged that more than two materials could
be used during the coating process.

The current study was carried out by using 12k tows. Another possibility would
be to use tows having a much larger quantity of fiber filaments and maybe other
fiber materials like glass, Kevlar etc..

Look into the possibility of impregnating chopped fiber preforms with aqueously
dispersed resins.

Develop a high speed version of the current set up. For this it would be imperative
to make modification in the fiber motion, heating and coating set-ups. These have

been discussed above.

APPENDIX A

 

METHOD FOR WINDER SPEED EVALUATION

Given below is a table of the speed versus the dial setting for traverse motor.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DIAL SPEED
SETTING s
(in/min)
32R 2.44
32F 2.35
34R 2.89
34F 2.87
36R 3 .48
36F 3.43
38 3.98
40 4.36
42 4.83
44 5.33
46 5.65
48 6.12
50 6.58
L___.52 7.01

 

 

 

 

Where F denotes forward motTon of motor.
R denotes reverse motion of motor

107

108
Width of prepreg tape w = 1/8"

Width of composite plaque W = 6"

Number of prepreg strips in entire width N = W/w = 48

Therefore, the rotary winding motor will have to rotate 48 times to fill the entire width
of the frame with prepreg tape.

Choose a conservative rotary speed of w = 20 rev/min.

Time taken to make up one layer t = N/w = 2.4 min.

Traverse speed s = W/t = 6/2.4 = 2.5

From table dial setting for traverse motor is around 32.

Therefore, rotary motor is set at 20 and traverse motor is set at 32.

APPENDIX B

 

This program ramps pressure at 20pai/m1n to 150pai,

340

10.0
4
5

100
5
11

0.00
700

0.00

10.00
0.00

0.00

60.00
0.00

0.00
0.00

so '00
0.00
30.00

Autoclave

For James

1600

0.00

0.00

OH

.00

.00
.00

00

0.00
30.04

Program 0 7
Pernandee
CH2

SOAKl O
PVEVT 2
PVEVT 6
RAMP? 20.0
TMEVT 9
GSOAK 150'
TMEVT B
TMEVT 12
RAMPT 20.0
BOAR 0

0.00
330
400

0.01
1.00

0.01
0.05

1.00

holds, and

ramps temperature at 20°P/min to 212°F, holds for 10 minutes, ramps

temperature at 20°F/min to 340°F,

holds for 1 hour.

ramp-

temperature at 10°P/m1n to 100°F, holds for 30 minutes, then ramp-
pressure 20pei/min to Opai.

Brien Rook
Sept 13, 1993
Updated Jun 14, 1993

109

APPENDIX C

 

00000

15

Program to Evaluate Temperature
Profile in Slab Due Conduction

PROGRAM FOR ONE DIMENSIONAL HEAT CONDUCTION THROUGH
AN INFINITE SLAB CONVECTIVELY HEATED FROM BOTH
SURFACES

WRITTEN BY SANJAY PADAKI

real lambda(50),bn(50),bi,k,h
character*80 infile,outfile

common lambda,bn,bi

write(*,*)’INPUT FILE NAME :’
read(*,2)infile

format(a80)

write(*,*)’OUTPUT FILENAME :’
read(*,2)outfi1e
open(unit=1,file=infile,status=’old')
open(unit=2,file=outfile,status=’new’)
read(1,*)k,h,alpha,b,y,t0,tinf
write(2,*)’THERMAL CONDUCTIVITY :',k

write(2,*)’TRANSFER COEFFICIENT :’,h
write(2,*)’ALPHA :',alpha

write(2,*)’SLAB HALF THICKNESS :’,b
write(2,*)’LOCATION OF INTEREST :’,y
write(2,*)’INITIAL AND FINAL TEMPERATURES :',t0,tinf
bi=h*b/k

GENERATE EIGEN VALUES and COEFFICIENTS
call eigen

do 16 j=1,1ooo

t=(j-1)*S.

sum=0.0

do 15 n=1,30
suml=bn(n)*exp(-1.*lambda(n)**2.*a1pha*t/b**2.)
sum2=cos(y*lambda(n)/b)
sum=sum+sum1*sum2

continue

temp=tinf+(tO-tinf)*sum
write(*,*)t,temp

110

16
17

10

111

write(2,*)t,temp
tempdiff=abs(temp-tinf)

if (tempdiff.lt.1.e-1) then
goto 17

endif

continue

stop

end

SUBROUTINE FOR EIGEN VALUE GENERATION
subroutine eigen

common lambda,bn,bi

real 1ambda(50),bn(50),bi
pi=acos(0.0)*2.

do 10 i=1,30

x0=lambda(i—1)+pi

if (i.eq.1.0) then

x0=pi/4.

endif

f0=0.

fprime0=0.

x1=0.

if (bi.ge.(pi/Z.)) then
f0=(xO/bi)-(1./tan(x0))
fprimeO=(1./(sin(x0)*sin(x0)))+(1./bi)
goto 6

endif

f0=x0*tan(xO)-bi
fprimeO=(x0/(cos(x0)*cos(x0)))+tan(x0)
x1=xO-(fO/fprime0)

diff=abs(x1—x0)

if (diff.1t.1.e-8) then

goto 7

endif

x0=xl

goto 5

lambda(i)=x1

bnum=2.*sin(lambda(i))
bden=lambda(i)+sin(lambda(i))*cos(lambda(i))
bn(i)=bnum/bden

continue

return

end

APPENDIX D

 

FLEXURE TEST DATA

mat-I can have“!
”all Min

3

I"!!!
up”
um"
um

 

Figure D.1 Three-Point Flexural Test Data for 5% CSBR Composite Samples

112

113

”I!
Iuwmllllmit'
Mil-[Mum
-.
WI!
fill, I-lll-~Illl
HI“ l-II-—Il‘l
-_ MIWH--Dll
M“ -‘--b04
Ml ---”)...
-.
-.

 

 

fl.
Ill.
4
1..
1'.
m I
0.. I.“ ... 8.. A." 4.. l. I.‘ 1.. 0..

Figure D.2 Three-Point Flexural Test Data for 10% CSBR
Composite Samples

 

 

I“.
IMI-umllumnv
minimal-“WV
WE
m‘.-_-—M0|
‘1'. .1.-~llll
.v'. _--~IIOI
- ~‘s_--~Illl
MI. t.---IIOI
ID.
7.-
.. g
-
.
-
u
I.
ma
0.. L- I.. l- l... 4.. L. I.‘ 7.. I.“

Figure D.3 Three-Point Flexural Test Data for 15% CSBR
Composite Samples

114

no.3
[' mun em. may-um
man my. mum
as.
J an: m

"I"
III"
ll!!!

 

 

Figure D.4 Three-Point Flexural Test Data for 20% CSBR
Composite Samples

APPENDIX E

 

ENF TESTING METHOD

13.2 ANALYSIS OF THE ENF SPECIMEN

The End Notched Flexure (ENF) specimen and the test principle are shown in Fig.
l3-7. The purpose of this specimen is to determine the critical strain energy release
rate in pure Mode ll loading [6]. This specimen has been found to produce shear
loading at the crack tip without introducing excessive friction between the crack

surfaces.
From elastic beam theory an expression for the strain energy release rate can
be derived [6].
G _ 9P3Ca’
" 2w(2L’ + 3.3) “3’

wherePistheappliedload,Cisthecompliance,aisthecracltlength,wisthe
specimenwidth. and L is the span between the central loading pin and the outer
support pins (see Fig. 13-7).

The compliance may-be calculated from the following formula. based on
beam theory [6].

21.’ + 3a’

C = W “"

 

Figure 13-7. ENF specimen.

115

116

where E is the flexural modulus in the axial direction of the beam, and h is the
semi thickness of the beam. It is, however, recommended that the compliance be
measured experimentally.

In the derivation of eqs. (l3) and (I4), deformations due to the action of
interlaminar shear stress and the influence of friction between the crack surfaces
were neglected. These effects may influence the results in some cases as shown by
Carlsson et al. [7] and Gillespie et al. [8]. The localized high shear stress and strain
in the vicinity of the crack front especially appear to influence the strain energy
release rate computation [8]. However, as a conservative estimate, the beam theory
expression, eq. ( l3) appears satisfactory.

13.2. 1 Stability of Crack Growth

The stability of crack growth may be judged from the sign of dGulda analogous to
the treatment for the DCB specimen in Sec. 13.1.] (see also [7]).
For fixed load conditions, eqs. (13) and (14) give

dG" 9a?2
712' = SEw2h3 “5)

This quantity is positive, hence the crack growth is unstable.
For fixed grip conditions, eqs. ([3) and (14) give

191' _ 982a _ 9a3 ([6)
da 8Ew2h3C (2]..3 + 3a3)

For stable crack growth dGnlda has to be less than or equal to zero. This gives
3
a 2 U\/_ a 0.7L (17)

Consequently, for the commonly used a z L/2, the crack growth is unstable even
under fixed grip conditions.

13.2.2 ENF Testing

A three-point flexure fixture with a total span of 100 mm (Fig. l3-8) should be
mounted in a properly aligned and calibrated test machine (Fig. l3-9). The crack
should be carefully wedged open and extended about 2 mm beyond the insert in
order to achieve a natural starter crack. The specimen shall then be placed in the
loading fixture so that the initial crack length is about 25 mm (see Fig. 138). The

117

25

 

l-2

 

 

 

 

 

 

I I *'|

Figure 13-8. ENF specimen geometry and loading. All dimensions
are in mm.

 

ml”. Nut-m

118

crosshead rate should be approximately 5 mm/min and the displacement under the
central loading pin may be recorded by an extensometer or LVDT. If correction is
made for the machine compliance, the beam compliance may be evaluated from
the crosshead travel. In either case, a real time plot of the load versus displacement
should be made on an x-y recorder. it has been observed that the crack generally
propagates to the central loading point in an unstable manner (see Sec. 13.2.1).
This means that only one value of One is obtained for each specimen.

13.2.3 ENF Data Reduca'on

Figure l3- IO shows a typical load-deflection curve for an ENF specimen. The actual
dimensions of the specimen were: L = 50.8 mm, a = 27.9 mm, 2h = 3.5 mm,
and w = 25.3 mm. The critical load was 762 N, and the specimen compliance
was 2.3 x IO“3 mm/N. These data in eq. (13) give: Guc = 0.56 kJ/mz.

The compliance calculated fromeq.(l4)withE = l4OGPais:C = 2.2 x [0'3
mm/N . The small difference between the measured and calculated compliances may
be due to scatter in the material properties or influence of interlaminar shear de-
formation [7].

The initial crack length, a, can best be evaluated by opening the specimen
in two parts after completion of the fracture test, and measuring the distance between

 

800 l l I

ENF

 

    

L = 50.8 mm

 

 

 

200 o = 27.9 mm ..
O
1 12.
o l l l
0 LC 2.0 3.0 4.0

Displacement,mm

Figure 1340. Typical load-deflection curve for an ENF specimen
made from AS4I3SOl-6 graphite/epoxy.

119

+lwl+/P

shear ply
(flop)

strep ‘

... W» L

-

Figure 13-11. CLS specimen.

the end of the precrack and the imprint of the outer loading pin on the specimen
surface. ln cases where the compliance is evaluated from the crosshead travel, it
may be important to correct for the machine compliance.

73.3 ANALYSIS OF ”IE CLS SPECIMEN

The Cracked Lap Shear (CLS) specimen was developed by Wilkins et al. [2] as a
Mode ll specimen for composites (see Fig. lB-l I). This specimen was originally
designed for shear dominated fracture investigations of adhesive joints [3]. How-
ever, this specimen does not produce a pure Mode ll loading at the crack tip [2].
The unbalanced configuration of the CLS specimen results in peel stress (Mode l).
Consequently. the CLS specimen is a mixed-mode specimen.

The compliance of the specimen, obtained from simple strength of materials
analysis with the notations given in Fig. l3-ll is

_ L 3(hj — hz)
C " wEh. + wEh.h2 “8)

From this equation and eq. (2) the strain energy release rate is

_____ P2011" l12)

2W25h1h2 (19)

APPENDIX F

 

COMPLIANCE CALCULATION DATA SHEET

Width lmrnl -

Halt Span L (mml -
Crack Length e lmml -
Thickness (turn) I
Helf'l’hleltneeelml -

Break Leer! P -

No X 11611
1 5.0366
2 10.3785
3 17.2466
4 24.7253
5 31.5934
6 47.819
7 59.8291
8 65.1709
9 74.949

10 81.1966
11 91.2698
12 99.51 16
13 107.448
14 1 1 3.553
15 1 22.1
1 6 130.9524
1 7 133.852
18 136.752
1 9 140.873
20 150.183

Compliance

Error

Gllc -

ENFOS3 .XLS

Linear Regression of EM Data for Complenee Caleulerlone

Y 1%)

0.663919
1.0302
1.6102
2.2263
2.7472
3.9911
4.9603

5.36
6.1966
6.6239
7.4557
6.1349
6.7301

9.264

9.69
10.636
10.6622
11.1 57
11.536
1 2.546

C -

f-

Y (in)

0.006639
0.010302
0.0 1 6102
0.022263
0.027472
0.03991 1
0.049603
0.0536
0.061966
0.066239
0.074557
0.061349
0.067301
0.09264
0.0969
0.10636
0.106622
0.1 1 1 57
0.1 1536
0.12546

0.00456
0.999742

0.692962

01‘ 8811
Sam 93

101-

X lNl

22.4036
46.16564
76.71633
109.9631
140.5336
21 1.6166
266.1316
269.6932
333.3661
361.1767
405.9663
442.6475
477.9502
505.1065
543.1252
562.5025
595.4005
606.3002
626.6313

666.044

7313.907

mm/N

ltJ/rn‘2

120

672.1166 N

Y (mm)

0.166635
0.261671
0.406991
0.565966
0.697769
1.013739
1.25991 6
1.36652
1.573936
1.662471
1 .693746
2.066265
2.217445
2.353056
2.51206
2.702052
2.764079
2.633676
2.930652
3.166664

34.45956

X'Y

3.776075
12.0602
31.37627
62.24913
96.06269
214.7291
335.3036
396.1449
524.7317
607.6726
766.6357
914.6269
1059.626
1 166.544
1364.363
1573.952
1 645.734
1723.649
1636.436
2126.645

16491.14

X'2

501.9304
2131.267
5665.395
1 2096.26
19749.74
44667.22
70626.1 4
64036.07
1 1 1 147.7
130450.1
164624.9
1 96936.6
226436.4
255132.5
294965
339309.1
354501.7
3700292
392666.6
446262.6

3523799

Y‘2

0.026436
0.066472
0.167273
0.320343
0.466909
1 .027666
1 .567369
1 .667377
2.477276
2.630707
3.566261
4.269449
4.917064
5.536673
6.310445
7.301065
7.640132
6.030665
6.566721
10.15495

77.19772

APPENDIX G

 

ENF TE T DATA

seo._e
emanates-mimics
nnmrruemm
m2; mm
ares-teenager
acumen-tees
"66--3006
“H "antennae
Inna-nee

 

 

‘ a A a A a .a
v v r v v

I.“ 4.” 6.66 6.“ 16.66 12.60 14.66 26.“ 16.06 ”.66

Figure G.l ENF Data for 0%CSBR Composite Samples

121

122

luau-u Reta mien-etc!
lweteel team w

"1":
66666
III"8
11!!!

/"

/.......

2.66 4.” 6.” 6.” 16.66 16.66 14.66 16.66 16.“ ..66

Figure G.2 ENF Data for 5% CSBR Composite Samples

 

mu “ 011m“!
M1001 7.8”. W

are

+ g a A A 4
v f v

6.“ 4.“ 6.“ 6.” 16.” 16.“ 14.66 16.“ 16.66 ’ .66

Fiugre G.3 ENF Data for 10% CSBR Composite Samples

123

 

A A

 

‘
—V

6.“ 4.” 6.” 6.. ... 16.“ 14.. 16.. 16.“ ..ee

Figure G.4 ENF Data for 15% CSBR Composite Samples

sen
mute-um
”6.67.61”
are events
eons-teem.“
eels-union
.. arcane-uses
‘ censure-nee
avenue-uses
are.
“a

 

#-
v

6.“ 4.. 6.66 6.. 16.66 16.“ 14.66 16.“ 16.66 ..66

Figure G.5 ENF Data for 20% CSBR Composite Samples

 

BIBLIOGRAPHY

 

[1]

[2]

[3]

[4]

[5]

[6]

[7]

[8]

[9]

Weeton, J.W. , Engineer’s Guide to Composite Materials, American Society for
Metals.

Carlsson, L.A. , Thermoplastic Composite Materials, Elsevier Science Publishing
Company, 1991.

Taylor, N. , & McAinsh,J . , "Thermoplastic Compositions" , U. S. Patent
3785916, 1974.

Iyer, S.R., & Drzal, L.T., "Continuous Processing of Unidirectional Prepreg",
PhD. Dissertation, Michigan State University, 1990.

May, C. A. , Epoxy Resins: Chemistry and Technology, Dekker, 2nd Edition,
Revised and Expanded.

O’Connor, Phillips, "Reinforced Plastic ", U.S. Patent No 4,680,224, July 14,
1987.

Dyksterhouse R. , & Dyksterhouse, J. A. , ”Production of Improved
Preimpregnated Material Comprising a Particulate Thermoplastic Polymer
Suitable for use in the Formation of Substantially Void-Free Fiber-Reinforced
Composite Article", U.S. Patent 4,894,105, January 16, 1990.

Chary, R.R. & Hirt, D.E. , "Coating Carbon Fibers with Thermoplastic Polymers
using Aqueous Foam”, ANTEC '92, pg 1181-1183, 1992.

Iyer, S.R., "Method and System for Spreading a Tow of Fibers", U. S. Patent
No. 5,042,111, 1991.

124

[10]

[11]

[12]

[13]

[14]
[15]

[16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

125

Myers, D., "Surfaces, Interfaces, and Colloids Principles and Applications" VCH
Publishers, Inc., 1991.

Briggs, Kathy L. , ”Waterborne Epoxies for High Performance Nonwovens",
Rhone-Poulenc Technical Literature, Performance Resins and Coatings Division,
9808 Bluegrass Parkway, Louisville, Kentucky 40299.

Zosel, A. , "Adhesion and Tack of Polymers; Influence of Mechanical Properties
and Surface Tensions", Colloid Polym. Sci, 263, 541 (1985).

Kaelble, D. H., in Treatise on Adhesion and Adhesives (R. L. Patrick, ed.),
Marcel Dekker, New York, 1967, p. 169.

Huntsberger, J. R., Chem. Eng. News, Nov 2, 1964, p. 82.
Zisman, W. A., Ind. Eng. Chem, 55(10): 19 (1963).

Carlsson, L. A., & Pipes, B. R., "Experimental Characterization of Advanced
Composite Materials", Prentice-Hall Inc., 1987.

Padaki, 8., "Powder Processing of Composite Prepreg Tape : Particle Size
Effects", Master’s Thesis, Michigan State University, 1992.

Kreith, F. , "Principles of Heat Transfer", International Textbook Company, 2nd
Edition, 1966.

McAdams, W.H. , "Heat Transmission" , McGraw-Hill Book Company Inc. , 3rd
Edition, 1954.

Charrier, J .-M. , "Basic Aspects of Structure-Property Relationships for
Composites", Polym. Eng. Sc., 15:731-746 (1975).

Kern, D. Q., "Process Heat Transfer", McGraw-Hill Book Company, 1950.
Russel, A. J. and Street, K. N., ”Moisture and Temperature Effects on the
Mixed-Mode Delamination of Unidirectional Graphite/ Epoxy " , Delamination and
Debonding of Materials, ASTM STP 876 (1985), p. 349.

Ting, R. Y., & Moulton, R. J., SAMPE Tech. Ser. 122265 (1980).

Moulton, R. J., & Ting, R. Y., Paper presented at the Int. Conf. on Composite
Structure, Paisley, Scotland, Sept. 1981.

Kinloch, A. J .,& Young, R. J. (eds.), "Fracture Behavior of Polymers" Elsevier
Applied Science, London 1983.

[26]

[27]

[28]

[29]

[30]

[31]

[32]

126

Kinloch, A. J ., in "Structural Adhesives", Kinloch, A. J. (ed), Elsevier Applied
Science PubL, Essex 1986, p. 127.

Miilhaupt, R. , & Buchholz, U. , "Compatibilized Liquid Rubbers as Epoxy
Toughening Agents" , F reiburger Materialforschungmentrum (FMF) and Institut
fiir Makromolekulare Chemie ", Stefan-Meier-Str. 31, D-79104 Freiburg e. Br.,
Federal Republic of Germany.

Qian, J. Y., Pearson, R. A., Dimonie, V. L., and El-Aasser, M. S., "Epoxy
Polymers Toughened with Novel Latex Particles", J. Mater. Sci, 26 3828
(1991), 24 2571 (1989), 21 2462 (1986).

Bascom, W. D., Cottington, R. L., Jones, R. L., Peyser, P., J. Appl. Polymer
Sci, 19:2545 (1975).

Murayama, T. , "Dynamic Mechanical Analysis of Polymeric Material", Elsevier
Scientific Publishing Company, Vol 1 1978.

Zeng, S. , Holsington, M. , & Seferis, J. C. , "Particulate Interlayer Toughening
of Dicyanate Matrix Composites" , Polymeric Composites, December 1993, Vol.
14, No. 6.

Moschiar, S. M., Riccardi, C. C., Williams, J. J. R., Verchere, D., Sautereau,
H., & Pascault, L., J. Appl. Polym. Sci, 42, 717, (1991).