h.

i...»m....5r “ween...

1.4 t . r .. t... $0.1!!!

:1 gr.alv\‘5.\t2:0‘a§§
x...$,:.§..s .

V!”

9""! Wii tun

[1tuptK

v .5}: I:
5... Kris?

... 4.5.... . 5...... ...
1 3L.

y(~1§y~ ‘(W‘L‘Y‘

(

q.

:15...
f. 3:!!!

Y?!

.3. . . 3.1a
Sisinl

u n...

J
.i 311 (£31. r. i-Ji.
.2... 5.5....- . fill-:35.
0;...

3.
7!:9';

P; . 1.... 3....

; ifiicrlttn .ys...v..l.

2...: 25.3.5
7... .2 .51..
.5 £3.53.

a...
. it) 3. (4...... h 1....
5.7 t I...

4. 1.... 4. r51... v
71.5..) 1....

. 3:, 7......

)\1.1. r 7 2:5... .. 1.11.42:
.1: o... . :2.

.5
55.3.1.9... f. 1....) J .. .T.)
2.. y)

. . .b) .b}!.v:1....>9
.029}: ..

{$.13}. .1.) , x.
v.15) :i...|.l.rrz .II
. 1. ta.....)\r....l..$...
r 1.55)}... 2.. .13.;
1.! 15‘ r. 15...... .
\. 3): .p . a . .139; 1.5:)?:
.i (I .0: \\:§ ...r£...):1\~.(!
nu §>.X.\)\.-{.!:ru | o

 

{KiritCanK 1)..
:7. {it .. Alrit (L.

...r!.....l..\.:x>5 1.35?!
):.l:.. i...) 5.). r1

.3..\T.\...;\f‘ it. 15...}. .
. .?r t. .33.)

 

THESIS

0-7639

MICHIGAN STATE UNIVERSITY LIBRARIES

lllllll l Il

31293 00881 1618

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

III II

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

thesis entitled

The effects of resin soluble binder
on mold filling with nonwoven preforms

presented by

Douglas James Backes

has been accepted towards fulfillment
of the requirements for

MS degree in ngneering

 

MSU ii an Affirmative Action/Equal Opportunity Inrlilution

 

PLACE IN RET

    
 
  

TO AVOID FINE

kout from your record.

URN BOX to remove this chec
5 return on or before date due.

 

 

THE EFFECT OF RESIN SOLUBLE BINDER ON MOLD
FILLING WITH NONWOVEN PREFORMS

By

Douglas James Backes

A THESIS

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

MASTER OF SCIENCE

Department of Chemical Engineering

1993

ABSTRACT

THE EFFECT OF RESIN SOLUBLE BINDER ON MOLD
FILLING WI'I'I-I NONWOVEN PREFORMS

By
Douglas James Backes

During resin transfer molding, a thermoplastic binder, used to hold a nonwoven pre-
form mat together, may affect the mold filling process and part strength if it is soluble in
the incoming resin. Resin (without catalyst added) was pumped through a flat, rectangular
mold with a nonwoven preform in it and was collected at the mold outlet The resin’s vis-
cosity was tested, from which the binder concentration in the resin and throughout the
entire mold was determined. Plaques were then produced and 3-point flex tests were per—
formed. Binder concentration increases with increasing preform fiber volume fraction and
mold temperature and decreasing operating pressure. The growth of fingers that occur dur-
ing mold filling will increases with higher operating pressure as a result of a faster pulse
rate. The flex strength and modulus scatters widely over the length of the plaque as a result

of inhomogeneities inside the mold.

 

‘To my parents and
’Dougfas Mums and tfie editorial anrcf

of t/ie ngrmg.

 

iii

 

 

ACKNOWLEDGEMENT

The author would like to thank Dr. Krishnamurthy Jayaraman for his technical guid-
ance and assistance during the course of this research.

He would also wishes to thank ling-Lei Chen for his assistance in performing many of
the RTM runs, without which this study would have been much more difficult. And also

Nancy Losure for providing much practical advice and technical knowledge.

iv

 

TABLE OF CONTENTS

List of Tables ..................................................................................................................... viii
List of Figures ..................................................................................................................... ix
1. Introduction ...................................................................................................................... 1
2. Background ...................................................................................................................... 3
2.1. RTM Process ........................................................................................................... 3
2.2. Prefonns .................................................................................................................. 4
2.3. Effects on Fiber/Resin Bonding .............................................................................. 5
2.4. Fingering ................................................................................................................. 5
3. Objectives ......................................................................................................................... 9
4. Materials and Equipment .............................................................................................. 10
4.1. Materials ................................................................................................................ 10
4.2. Equipment ............................................................................................................. 11
4.2.1. Mold ............................................................................................................. 11
4.2.2. Resin transfer machine ................................................................................. 15
4.2.3. Other equipment .......................................................................................... 18

5.Determination of Spatial Variations in Concentration and Viscosity
during Mold Filling ..................................................................................................... 19
5.1. Procedure .............................................................................................................. 19
5.1.1. Binder Concentration Vs. Resin Viscosity ................................................... 19

V

 

 

 

vi

5.1.2. Binder Solubility and Dissolution Rate ........................................................ 19
5.1.3. Binder Washout From the Mold ................................................................... 20
5.1.4 Concentration profile in the Mold ................................................................. 22
5.2. Results and Discussion .......................................................................................... 22
5.2.1. Binder concentration vs. resin viscosity ....................................................... 22
5.2.2. Binder solubility and dissolution rate ........................................................... 26
5.2.3 Binder washout tests ...................................................................................... 28
5.2.3.a. Fiber volume fraction of preform ........................................................ 30
5.2.3.b. Operating pressure .............................................................................. 34
5.2.3.c. Mold temperature ................................................................................ 40
5.2.3.d. Summary ............................................................................................. 49
5.2.4. Concentration profile in the mold ................................................................. 49

6. Effects of the concentration and viscosity variations on the mold filling process and

part strength .................................................................................................................. 60
6. 1. Procedure .............................................................................................................. 60
6.1.1. Binder effects on crosslinking ...................................................................... 60
6.1.2. Production of plaques ................................................................................... 60
6.1.3. Observation of fingering during mold filling ................................................ 62
6.1.4. Determination of flex strength and modulus of the plaque ........................... 62
6.2. Results and discussion ........................................................................................... 63
6.2.1. Binder effects on crosslinking ...................................................................... 63
6.2.2. Fingering during mold filling ....................................................................... 68
6.2.3. Flex strength and modulus of the plaque ...................................................... 78

7. Conclusions .................................................................................................................... 84

 

 

vii

8. Recommendations .......................................................................................................... 86
Appendix A: raw data ........................................................................................................ 87
Appendix B ...................................................................................................................... 102
B.1. Intrinsic viscosity ................................................................................................ 102
B2. Determination of mass transfer coefficient ......................................................... 105
Bibliography ..................................................................................................................... 107

 

 

LIST OF TABLES

Table 1: Binder concentration vs. viscosity at 23°C ........................................................... 87
Table 2: Binder dissolution rate .......................................................................................... 88
Table 3: Flushing runs ........................................................................................................ 90
Table 3.a: General Conditions ...................................................................................... 90
Table 3.b: Mold filling rate ........................................................................................... 90
Table 3.c: Viscosity and binder concentrations ............................................................ 91
Table 4: Concentration profile in mold ............................................................................... 95
Table 4.a: Viscosity and 23°C vs. viscosity at 28°C .................................................... 95
Table 4.b: Preforrns ...................................................................................................... 95
Table 4.c.: Viscosity and binder concentrations for flushing runs ................................ 99
Table 5: Growth of finger length and width ...................................................................... 100
Table 6: Flexural strength data ......................................................................................... 101
Table 7: Concentration calculations for intrinsic viscosity ............................................... 103
Table 8: Deviation of experimental viscosity from Huggins viscosity ............................. 103
v i i i

 

 

LIST OF FIGURES
Figure 1: Fingering in a Hele—Shaw cell ............................................................................... 7
Figure 2: Increase in Derakane viscosity with time ............................................................ 12
Figure 3: Mold for transfer runs ......................................................................................... 13
Figure 4: Resin transfer machine ........................................................................................ 16
Figure 5: Schematic diagram of the resin transfer machine ............................................... 17
Figure 6: Viscosity vs.%wt binder in resin at 23°C ........................................................... 23
Figure 7: Viscosity vs.%wt binder in resin at various temperature ................................... 26
Figure 8: Dissolution of binder in styrene .......................................................................... 26
Figure 9: Mass transfer coefficients for binder loss in styrene ........................................... 27
Figure 10: Pressure effects on resin viscosity for 1 and 3 ply preforms at 23°C ................ 29
Figure 11: Pressure effects on binder concentration for 1 and 3 ply preforms at 23°C ...... 30
Figure 12: Resin front vs. time for 1 and 3 ply preforms at 23°C ....................................... 32
Figure 13: Pressure effects on resin viscosity for 1 ply preforms at 23°C .......................... 33
Figure 14: Pressure effects on binder concentration for 1 ply preforms at 23°C ................ 34
Figure 15: Resin front vs. time for 1 ply preforms at 23°C ................................................ 36
Figure 16: Pressure effects on resin viscosity for 3 ply preforms at 23°C .......................... 37
Figure 17: Pressure effects on binder concentration for 3 ply preforms at 23°C ................ 38
Figure 18: Resin front vs. time for 3 ply preforms at 23°C ................................................ 39
Figure 19: Exit viscosity variations in washout runs at 40°C ............................................. 41
Figure 20: Exit concentration of binder in washout runs at 40°C ...................................... 42
Figure 21: Exit viscosity variation in washout runs at 60°C .............................................. 43
Figure 22: Exit concentration of binder in washout runs at‘60°C ...................................... 44
i X

 

 

Figure 23:
Figure 24:
Figure 25:
Figure 26:
Figure 27:
Figure 28:
Figure 29:
Figure 30:
Figure 31:
Figure 32:
Figure 33:
Figure 34:
Figure 35:
Figure 36:
Figure 37:
Figure 38:
Figure 39:
Figure 40:
Figure 41:
Figure 42:
Figure 43:
Figure 44:

X

Binder washout curves in flow through a 1 ply preform ................................... 47
Binder washout curves in flow through a 3 ply preform ................................... 48
Viscosity of flushed resin for different preform lengths at 20 psi, 28°C ........... 50
Viscosity of flushed resin for different preform lengths at 30 psi, 28°C ........... 51
Viscosity of resin in the mold at 20 and 30 psi, 28°C ....................................... 53
Viscosity profile with resin front at different lengths in the mold, 20 psi ......... 55
Viscosity profile with resin front at different lengths in the mold, 30 psi ......... 56
Concentration of binder in the mold at 20 and 30 psi, 28°C ............................. 57
Mass transfer coefficent along length of the mold ............................................ 59
Plaque cutting pattern for making flex test samples .......................................... 64
DSC run with 0% wt binder .............................................................................. 65
DSC run with 1% wt binder .............................................................................. 66
DSC run with 3% wt binder .............................................................................. 67
Fingering during mold filling at 20 psi, 28°C ................................................... 69
Fingering during mold filling at 30 psi, 28°C ................................................... 72
Finger length growth during mold filling .......................................................... 76
Finger width growth during mold filling .......................................................... 77
Flexural strength of an end loaded plaque ........................................................ 79
Flexural modulus of an end loaded plaque ....................................................... 80
Flexural strength of a center loaded plaque ...................................................... 82
Flexural modulus of a center loaded plaque ...................................................... 83
Specific viscosity/concentration vs. concentratoin ......................................... 104

 

1. INTRODUCTION

With the increasing importance of Resin Transfer Molding (RTM) with nonwoven
continuous fiber preforms in the production of automotive and other large size parts, it has
become imperative to fully understand all aspects of the mold filling process. One aspect
that has received little attention from researchers but can play a crucial role in the RTM
process is the preform binder. The binder, which is used to hold the preform together, can
affect the fiber/resin bonding, the flow of the resin inside the mold, and the crosslinking
reaction during curing. These aspects, if detrimentally affected, could severely weaken the
strength of the composite part produced.

This study will examine the case where the binder is soluble in the resin. From this
study, the dynamics of binder dissolution and its effects on the mold filling process and the
part’s strength will be better understood.

The objectives of this project are as follows:

1. Find how the binder concentration in resin affects the resin viscosity.
2. Determine the effects of resin flow rate, mold temperature and fiber
volume fraction of the preform on the dissolution of the binder and the
removal of the binder from the mold system.

3. Construct a viscosity profile and concentration profile of the resin
inside the mold during the filling process.

4. Determine the effect (if any) of binder dissolution on the crosslinking
reaction during the curing process.

5. Observe whether or not the viscosity profile in the mold that results

from binder dissolution will result in miscible fingering during the mold

 

 

filling process.
6.’Determine whether the binder dissolution affects the flexural strength
and/or modulus of the resulting plaque.
These results, it is hoped, will lead to further research into the influence of binder in
RTM and determine the precise effects these will have on the resulting parts. These results
may also lead to better modeling of the mold filling process by introducing another factor

for consideration, and therefore leading to more precise models being produced.

 

 

2. BACKGROUND

2.1 RTM PROCESS

RTM is a widely used process in the production of plastic composite parts. The basic
process involves the pumping of an uncrosslinked resin and a catalyst agent from separate
holding tanks into a static mixer. The resin/catalyst mixture is then transferred from the
mixer into the mold containing a fiber preform. After the mold is filled, it is heated to cur-
ing temperature and then to a postcuring temperature. When postcuring is completed, the
mold is cooled and the part is removed.

RTM has a variety of advantages over other processes including the ability to make
large and complex shaped parts, rapid production cycle time, low filling pressures, low
capital costs, greater reproducibility, better potential for automation, and the ability to
make two-sided finished parts(1,2). These advantages have made RTM increasingly
attractive for the production of high performance parts and have lead to a greater need to
understand how certain variables affect mold filling.

Preliminary testing by Pregelhof and Lien(3) showed that the RTM process was
affected by several factors. They found that increased pressure results in higher resin
velocity and a quicker fill time. It was also discovered that the preheating of the preform to
temperatures close to the molding temperature, the cycle time decreases to a minimum at
an optimum temperature, after which it starts increasing. Finally, if the resin is preheated,
the processing time will also decrease as a result of decreased resin viscosity. An extensive
study performed by Rudd, et al(4) verified the findings of Pregelhof and Lien. They dis—
covered that increased pressure and the preheating of the preform and the resin decreases

the impregnation time as well as the cycle time. They further found that the preheating of

 

the resin and the preform also will decrease the gel time by 20-40%.

2.2 PREFORMS

The research into making bigger and tougher parts has increased the number of types
of preform mat available. One of the early methods for making fiber preforms, directed
fiber preforming, has a variety of inherent disadvantages in the need for short fiber lengths
(under 2 inches) and a bulkiness of the resulting mat which leads to low glass content.
New types of mats have been developed including unidirectional, biderectional, tn’axial
stitched, etc., which have greater densities than fiber directed preforms. This results in the
ability to lay more plies inside the mold which will lead to greater flexibility in part
design. Studies have shown that parts made with long fiber preforms will have greater
flexural strength and modulus, creep resistance, fatigue endurance and impact strength by
25-100% than parts made with short fiber preforms(5). Unfortunately, these mats need to
be precisely placed on top of one another in the mold to give the final part the optimum
properties of flexibility and strength. This requires large amounts of time to cut, overlap,
tuck and stitch the fibers together and will result in decreased production and increased
labor costs(6).

An alternative to these directional mats are continuous strand nonwoven mats. Since
these mats are nonwoven, the plies do not need to be precisely aligned on top of one
another thus greatly reducing the production time involved. The main problem with these
mats is that in order to be able to handle them before the molding process begins, some
method is needed to hold the fibers together in the mat. The solution most manufacturers
have found is to coat the entire mat with a chemical binder that usually is a thermoplastic
powdered polymer. The binder is applied by spray coating it onto the mat after it has been
laid out on a conveyer belt. The binder is applied evenly on the mat in quantities of up to

10% of the total mat weight.

 

 

5

2.3. EFFECTS ON FIBER/RESIN BONDING

While the binder makes it possible to handle nonwoven mats, it can cause problems in
the filling and curing stages. If the binder is not removed from the fiber strands, it can
cause a weakening of the part by preventing complete bonding of the resin to the fiber mat.
When this happens, the part will be weakened. Owen, Middleton, Rudd and Revill (7),
have shown that plaque specimens made with a sizing soluble in the resin (which can be
dissolved off of the fiber surface) have a stiffness 15-50% greater than specimens with a
sizing insoluble in the resin.

A special case of interest, as mentioned above, occurs when the binder is soluble in the
resin. In this case, the binder will be dissolved by the resin as it is pumped into the mold
and can then be flushed out of the mold. Unfortunately, the resin will not completely
remove the binder unless an adequate amount of time is allowed for. This is impractical in
actual commercial processes since the production time will be slowed done and costs will
increase because a large amount of resin will be wasted by being pumped through the
mold solely to remove the binder. What usually occurs in industry is that the resin is trans-
ferred to the mold until it is just filled and then the inlet and outlet valves are closed for the
curing process to begin. The result of this will be a binder concentration profile in the
molded part with the lowest concentration occurring at the mold inlet and increasing con-
centration occurring as one moves further away from the inlet. Owen, et.al. (7) found that
in such situations, the part stiffness is 22% greater at the injection gate than it is at the edge

of the molding.

2.4. FINGERING

Another factor that can affect the mold filling process is uneven displacement of the
binder from the preform. If the dissolved binder causes an increase in the resin’s viscosity
(which is very likely since the binder is most likely a polymer) and the amount of increase

is dependent on the concentration of the binder, a viscosity profile will develop in the

 

resin. When resin first enters the mold, all the binder is present and the resin will dissolve
all binder it can at the molding conditions, leaving less binder that the succeeding resin
will be able to dissolve. This will result in the succeeding resin having a binder concentra-
tion less than that of the proceeding resin. A binder concentration profile will develop in
the resin and from this, a viscosity profile.

Situations where a less viscous fluid displaces a higher viscous fluid leads to the phe-
nomenon known as “fingering”. Fingering was first observed by the oil indusz (8) in an
immiscible system where water is injected into wells to displace more viscous petroleum.
Surface tension and capillary effects were present at the interface between the two liquids
which became unstable and long waves or “fingers” developed. This meant that the petro-
leum could not be evenly displaced out of the well without a large amount of water being
mixed in with it. There are two types of fingering that can occur, fingering during immisci-
ble displacement where the two liquids won’t mix and the fingers remain well defined and
fingering in miscible displacement where diffusion will occur across the fingers and
between the two phases. In the RTM process, since resin of low binder concentration is
displacing resin with high binder concentration, a viscosity profile will result and miscible
displacement and fingering may occur.

Some of the earliest work done on miscible displacement, with no interface surface
tension) in porous media was by Slobod and Thomas(9) who showed that increases in
flow rate will affect the shape of the fingering by forming more numerous and narrower
fingers (Figure 1). Paterson(10) went a step further by developing a method for determin-
ing the wavelength or width of maximum finger growth. In porous media obeying Darcy’s
law, Tan and Homsy(l 1) developed the quasi-steady-state-approximation (QSSA) for esti-
mating the growth rates and length scales of miscible fingers and for predicting the most
dangerous wavelength of unstable fingers. It must be understood that these results were
obtained with porous media comprised of spherical particles (such as sand) and their

applicability for porous beds compromised of long cylinders (such as a fiber preform) is

 

   

 

 

SLOW-RATE RUN (1.6 FT/D) — DARK AREA SHOF\S LOCATION OF INJECTED PHASE

Figure 1: Fingering in a Hele-Shaw cell

 

 

 

uncertain. They also found in their studies that fingers will grow faster at higher fluid
velocities. -

Hickemell and Yortsos (12) studied the linear stability of miscible displacement in
porous media with no diffusion or dispersion occurring. They developed a theoretical
model that showed this system is linearly unstable when the mobility profile contains seg-
ments of decreasing mobility. This means that if the fluid viscosity in this system is contin-
uously decreasing, then the interface will become unstable and fingering will occur.

Since a binder induced viscosity profile can occur in the RTM process, fingering may
also occur. If this happens, the binder concentration in the resin and the amount of binder
dissolved off the fiber will vary in the direction transverse to the resin flow as well as
along the length of the mold. This could result in the weakening of the part in both the
transverse and lengthwise directions.

The assertion that fingering can lead to a weakening of the part has been supported by
the results of Losure, et al.(13) with a reactive system. In this study, a vinyl ester resin
(Derakane, Dow corp.) was injected into a mold with a catalyst and accelerator added
(Benzoyl Peroxide and N,N-Dimethylaniline respectively). The resin was then allowed to
partially react (crosslink) in the mold and then fresh resin was transferred to the mold. Fin-
gering was observed as the more viscous, reacted resin was displaced by the less viscous,
unreacted fresh resin. After the molded plaque finished curing, it was cut into strips trans—
verse to the direction of the resin flow, and were then given a flexural strength test on an
Instron tensile testing machine. These studies showed that strips which had multiple stria-
tions as a result of fingering, had flexural strengths 36% less than those of strips with no
fingering occurring. This study shows that inhomogeneities such as fingering do cause a

weakening of the part when they are a result of differences in reaction time.

 

 

3. OBJECTIVES

The objectives of this work are as follows:

1. Determine the relationship between the binder concentration in the resin and the
resulting resin viscosity.

2. Evaluate the spatial viscosity profiles of binder in the resin at different instances of
time during mold filling. Determine how these profiles vary with the mold filling con-
ditions of fiber volume fraction, mold temperature, and inlet pressure (fill time).

3. Use the viscosity/concentration relationship to determine the spatial binder concen-
tration profiles for the mold filling runs.

4. Determine if these profiles cause inhomogeneities such as fingering in during the
mold filling process.

5. Determine the effects that the binder may have on the crosslinking reaction of the
resin. I

6. Test the flexural strength and modulus of plaques produced in this system. From the

results in 1-5, find the effects of binder concentration on part strength.

 

 

4. MATERIALS AND EQUIPMENT

4.1. MATERIALS

The resin under examination in this study was Derakane 411-C50 supplied by the
Dow Chemical Corporation. The Derakane resin is 50% by weight vinyl ester and 50% by
weight styrene monomer which acts as the crosslinking agent in this system. The Dera—
kane has the following physical properties (14):

Specific gravity: 1.020-1.060
Viscosity (23°C): 120 cps

Boiling point: 146°C

Vapor pressure: 7 mm HG at 20°C
Vapor density: 3.6 g / m2
Insoluble in water

The catalyst used in these experiments was Benzoyl Peroxide (BPO). One of the
brands of BPO that Dow recommends for use with Derakane is Cadox 40E manufactured
by the Akzo Chemical Company. Cadox 40B is an emulsion made up of 40% by weight
Benzoyl Peroxide (BPO), 40% diisobutyl pthalate, 7% silica, 3% surfactant, and 10%
water. The material added with the BPO serve as organic stabilizers and fillers to enable
the BPO to be in an emulsion form and, according to Dow, will have no affect on the cur-
ing process. Cadox 40E has a specific gravity of 1.152. The accelerator used was N,N-
Dimethylaniline (DMA) supplied by Aldrich Chemical.

For the production of plaques, the composition of the mixture entering the mold was
1.0% wt. catalyst, 0.05% wt. DMA and 99.5% wt. Derakane. It should be noted here that
if the dissolution of binder is to have any significance, the binder dissolution must take

place in a time frame much smaller than the resin gel time. According to literature sup-

10

 

11

plied by Dow, at temperatures of 80-90 °F(27-32°C), the gel time will be in the range of
20-40 minutes(15). Figure 2, (12) shows the viscosity increase during curing for a Dera-
kane system. As is shown in the graph, there is no significant increase in the viscosity until
10-12 minutes after the mixing of the catalyst and Derakane. In order to effectively study
the binder affects, the rate at which the binder dissolves (dissolution rate) into the resin
and mold fill time should be less than eight minutes to ensure that any changes in the resin
viscosity are the result of binder dissolution alone and not crosslinking. As will be shown
in Section 4.2, the dissolution time for the binder is five minutes at the lowest temperature
tested (23°C) and the slowest mold filling time is four minutes, indicating that for the con-
ditions tested, viscosity increases as the result of crosslinking will be negligible.

The mat used for the preform is Unifilo U-750 supplied by the Vetrotex/Certainteed
corporation. This is a continuous strand nonwoven E-glass (density of 2.54 g/cm2) mat
with an area density of 0.0450 g/cm2 and has a thickness of approximately 0.3175 cm. The
mat is composed of E-glass strands with a diameter 16+/-1.5 micrometers. The fibers were
grouped into tows of two types, 67% of the tows have 60 fiber filaments and 33% have
120 fiber filaments. The average diameter of the tows is 0.045 cm. These tows are coated
with a silane sizing that is insoluble in styrene.

The U-750 mat is spray coated with thermoplastic binder made of powdered polyester.
This binder comprises 8% of the total weight of the mat and has a medium solubility in
styrene. The manufacturer of the mat supplied a sample of the binder in its powdered form

for use in various experiments throughout this study.
4.2. EQUIPMENT

4.2.1 Mold.
The mold used in these experiments is shown in Figure 3. The bottom plate has dimen-

sions of 22.86 x 66.04 x 2.54 cm thick and is made of aluminum. At 3.81 cm from the

 

 

Pasca-

VISCOSITY.

10.00

i

010+—
0.00

    

 

2.09

Figuer

4.00

212.

12

DERAKANE 4} 1 “TH 1'31 BPO AVE .552 DMA

 

 

6.00 8.00 10.00 1200 14.00 16.00 18.00
TIME F'ROMMIX. min

Increase in Derakane viscosity with time

13

Figure 3: Mold for transfer runs

 

 

22.86
cm
Bottom plate
22.86
cm
Middle ring
Top plate
22.86
cm

<—————>
66.04 cm

 

 

14

mold’s front end is a 1.27 x 15.24 x 0.635 cm deep well with three 1.27 cm ports. This is
where the inlet valves for the resin are attached for most of the experiments. When the
resin first enters the mold, it fills up the well and succeeding resin pushes the resin into the
mold with an even flow front. At 4.445 cm from the edge of the well and every 6.35 cm
after that are holes that can be used as outlets for the resin for preforms of different
lengths. Valves can be attached to these holes and can be opened and closed as needed.
Holes that aren’t being used as either inlets or outlets are sealed shut with brass plugs. A
rectangular middle ring sets on top of the bottom plate. Its has a 22.86 x 66.04 cm outside
dimension, 19.05 x 62.23 cm inside dimension, and a thickness of 0.3175 cm. The top
plate is 22.86 x 66.04 x 2.54 cm thick and is made of plexiglass. Plexiglass is used for the
top plate because its transparency will allow observation of the resin flow during the fill-
ing process. At 2.75 cm in from the lengthwise edges of both the top and bottom plates are
a series of small screw holes. These are used to secure a 19.05 x 1.905 cm aluminum strip
that sets inside the middle ring. This strip, by placing it at different points, makes the mid-
dle ring shorter for producing plaques with smaller lengths than the maximum possible.
Long bolts are screwed up through the bottom plate through holes in the metal strip up
into the holes in the top plate. The strip is then held in place by the bolts during the mold-
ing process.

To hold the mold together, a series of c-clamps are placed around the outside edge of
the mold with metal strips between the clamp jaws and the top plate to prevent the plexi—
glass from cracking. To seal the mold at the juncture between the middle ring and the two
plates, 3 0.3175 cm thick and 1.27 cm wide silicon rubber gasket is placed on the inside of
the middle ring and the preform is placed inside the gasket. The plexiglass has a tendency
to swell after prolonged contact with styrene fumes. To counteract this, a thin sheet of vac-
uum plastic is placed between the top plate and the preform.

At 1 cm from the well of the bottom plate, a hole was drilled for the attachment of a

thermocouple. The thermocouple is a model CF-000-J-2-60-1 from the Omega company

 

 

15

and is screwed directly into the hole so that the its tip is flush with the inner mold surface
on the bottom plate. The thermocouple is connected to an Omega CN9llA temperature
controller. Connected in parallel to the controller are five 2.54 x 50.8 cm etched foil heater
strips. Four strips were placed length wise on the outside surface of the bottom plate on the
area between the screw holes and the center port holes. Since the strips were shorter than
the mold, a fifth strip was placed across the width of the mold, 6.25 cm in from the end of
the mold furthest from the inlet port. These strips were attached to the plate with a silicon

adhesive.

4.2.2. Resin transfer machine.

A Multiflow Model 5515 Resin Transfer Molding Machine manufacture by the Liquid
Control Corp was used for these studies. Figure 4 shows the complete machine while Fig-
ure 5 show a schematic of its operation. The RTM machine consists of two pneumatically
driven pumps, a 55 mm resin pump and a 10 mm catalyst pump. The pumps are operated
in unison by a driving bar attached to the pumps’ shafts and is powered with compressed
air supplied by an outside source. The resin and the catalyst are supplied to the pumps by
two holding tanks (the accelerator is premixed in with the resin). The resin and catalyst are
gravity driven to their respective pumps and are pumped out in separate lines to a mixing
head. This head has a static mixer attached to its outlet which mixes the catalyst and resin
thoroughly as they are pumped out to the mold inlet The RTM machine has no capability
for preheating the resin before it reached the mold, so all runs were performed with resin
entering the mold at room temperature.

The resin pump is stationary at the far end of the driving bar and has a volumetric out-
put/pump cycle of 181 cc. The catalyst pump is variable position and has a volumetric out—
put/pump cycle of 2.78-10.67 cc. The amount of catalyst the pump transfers is determined
by its location relative to that of the resin pump. The closer the catalyst pump is to the

resin pump, the higher the resin/catalyst ratio will be pumped. Both pumps are operated in

 

 

16

 

Figure 4: Resin transfer machine

 

 

17

Figure 5: Schematic diagram of the resin transfer machine

    
  
  
   
 

   
  

  
  
 

Resin Catalyst
tank tank
(fixed (movable
position) position)

   
   
 
  

Pivot
Driving bar point

Resin Catalyst
pump pump

      
 

Mixing head

___—_>

Resin and
catalyst flow
during
operation

to the mold

 

18

the range of 15-100 psi, lower than this will not give adequate pressure for pumping,

higher than this range can blow the'pump gaskets.

4.2.3. Other equipment

A Brookfield model LVF-DVIII viscometer was used to determine the viscosity of
various resin samples. A model SC—3l spindle with a diameter of 1.2 cm was used for
most of the testing. The temperature of the sample was controlled by a heater in which the
cup containing the sample was placed. A DuPont 910 Differential Scanning Calorimeter
attached to an IBM TA 2700 thermal analysis console was used for testing the crosslinking
reaction of the resin. For the observation and recording of finger development, a Nikon F3
camera with an automatic advance was used with Kodak 400 color film. For the flex test, a

UTS tensile testing machine was used.

J...

 

5. DETERMINATION OF SPATIAL VARIATIONS IN CONCENTRATION
AND VISCOSITY DURING MOLD FILLING

This section of the work deals with the determination of the viscosity and binder con-
centration profiles in the mold during the mold filling process (1—3, Section 3).
For these next two sections, the raw data is presented in tabulated form in Appendix A

for further study. Calculations are presented in Appendix B.

5.1. PROCEDURE

5.1.1. Binder concentration vs. resin viscosity

The resin viscosity was measured at varying concentration of binder in the resin. A
series of 20 ml samples of Derakane resin were placed in screw top vials (the vials were
always kept covered to prevent styrene evaporation). Varying amounts of binder in the
powdered form were then added to each resin sample. After the binder had dissolved com-
pletely, a 10ml specimen was taken from each sample and tested in the Brookfield viscom-

eter at a temperature of 23°C and a spindle speed of 100 rpm.

5.1.2. Binder solubility and dissolution rate

The binder dissolution rate and solubility in Derakane were determined as follows:
Preweighed styrene samples were heated to three different temperatures (23°C, 40°C,
60°C) and powdered binder was added until the point where it precipitated from the

binder. The styrene sample was reweighed and the solubility was found. The solubility of

19

 

20

the binder in Derakane was performed in a similar way for 23°C.

A preliminary study to determine the efi'ective mass transfer rate of the binder from the
fiber mat was conducted using pure styrene at the same three temperatures as before. Cir-
cular samples of the U-750 mat were cut out with diameters of approximately 8.5 cm and
each weighing from 2.0 to 2.3 grams. To prevent fiber loss in the samples during these
runs in which the mat lost its binder, the sample was placed between two preweighed wire
meshes of the same diameter. The sample was then placed on a metal wire hoop with a
handle and then submerged in a beaker containing 200 ml of styrene. Each sample was
submerged for a specific period of time and were then removed from the styrene using the
wire hoop. The mat, still between the wire meshes was dried for approximately 30 minutes
and then reweighed. The difference between the mats initial weight before submersion in
styrene and afterwards was the amount of binder dissolved in the styrene. The sample was
placed back into the styrene and the process was repeated until the mat’s weight stabilized,
indicating that all the binder was removed. This experiment was carried out at various sty-
rene temperatures with three trials performed for each temperature and the average of the

trials determined.

5.1.3. Binder washout from the preform

With the binder concentration vs. viscosity curve and dissolution rate determined, the
next phase was to determine the amount of binder being flushed out of the mold. A 56.52
x 15.68 x 0.318 cm preform (volume of 281 cc) made of varying numbers of U-750 plies
was placed inside the mold which was then clamped down. For each run, the fiber volume
fraction was found using the following equation:

Vf=W/(VmXDg)

Vf = fiber volume fraction
W = weight of the preform
Vm = volume of the mold

Dg = density of E-glass

 

 

21

In order to infer binder concentration from viscosity measurements, the crosslinking reac-
tion was suppressed. No DMA was added to the resin in the holding tank and the hose
leading form the catalyst pump was disconnected from the mixing head and placed in the
catalyst tank to allow the catalyst to circulate through the machine (the subsequent open-
ing in the mixing head was plugged with the catalyst hose end fitting). The RTM machine
was then operated at a predetermined pressure. The resin flowed, unheated, through a hose
connecting the static mixer to the inlet valve of the mold. On top of the plexiglass plate a
yardstick was placed running the length of the mold and the time was recorded when the
resin flow front reached various lengths in the mold. The resin flowed through the mold
and the preform and was collected at the mold’s outlet in a series of twenty vials. Each
sample collected had 22 ml of resin and the time was taken when each sample vial was
filled. Each vial was then capped to prevent styrene evaporation. Runs were done at sev-
eral different fiber volume fractions, operating pressures, and mold temperatures.

After the samples were collected, their viscosities were measured on the Brookfield
viscometer in a similar manner as for determining the concentration/viscosity curve. Two
10ml samples were taken form each vial and were tested on the Brookfield, one sample
was tested at the mold temperature, and the other at 23°C (if the mold temperature was
23°C, then only one sample was needed). The results of the 23°C tests were compared to
the concentration/viscosity curve to find the amount of binder being flushed out of the
mold.

A run was also performed with a pressure gage attached to a T-square joint placed in
the line between the static mixer and the mold inlet. This was to determine if any pressure
cycling occurred in the line during a run. The test was performed with a mold temperature
of 23°C and a machine operating pressure of 25 psi. The line pressure varied from 5 to 25
psi during cycling of 38 cycles/minute and a shot size of 8.78 ml. The reason for this vari-
ance in line pressure is the way the machine operates, it builds up pressure to the set oper-

ating pressure at which point it pumps a shot of resin. After the cycle, it needs to build up

 

 

22

pressure in the driving bar again. This causes a pulse behavior inside the mold during fill—
ing rather than a steady resin flow rate. This pulsing effect could affect any fingering that
may occur in the mold by changing the dyed/undyed resin interface stability. Any effects

of the pulsing should be seen in the fingering observation stage of this study.

5.1.4. Concentration profile in the mold

A primary objective of this work was to determine the varying concentration of binder
in the resin at different points inside the mold at a time when the mold has been just com-
pletely filled with resin. The direct approach of attaching several outlet valves along the
length of the mold and collect resin at these points would not yield the desired concentra-
tion profile. Extracting resin from the mold like that would alter the concentration profile
significantly, so it wouldn’t be representative of the actual concentration profile during the
mold filling process. Instead, the concentration profile had to be found indirectly. This was
done by doing a series of washout runs for different preform lengths. Resin samples were
collected in a similar manner to that for the washout curves. These samples were then
tested in the Brookfield viscometer and the concentration/viscosity curve was used to get
the concentration profiles for each run. The result of these runs was a series of washout
curves for different length preforms and from these curves, the viscosity and binder con-

centration curves were constructed.
5.2. RESULTS AND DISCUSSION

5.2.1. Binder concentration vs. resin viscosity

Figure 6 shows the relationship of binder concentration in the resin the resin’s viscos-
ity at 23°C. The runs were carried out with binder weight percents in the range of 0 to
8.5%. As can be seen in the graph, the viscosity increases from 125 cps to nearly 500 cps

as the weight percent increases from 0 to 8.5%. The viscosity varies nonlinearly with the

 

 

23

0.0

 

mm

NR

Damn a 5%.: E BEE .85 05 .w> $885 6 8:me

:68 E EUEQ EN
Wm v.0 we mm ”m we

rsoosm

O
O
N)
M

 

24

percent weight of binder and can be described by the polynomial:
u = 128.6 +23.7 xW+ 1.07 xW2+0.18 xw3

ti = resin/binder mixture viscosity
W =% weight of binder in resin

The density of the resin (1.05 g/cc) does not change significantly with binder concentra-
' tion, therefore the weight concentration (C) is related to the weight percentage by the
equation:
C = W x 1.05

Substituting this relationship into the polynomial above gives:

u = 128.6 + 22.6 xC +1.02 xc2 +0.17 x c3
The intrinsic viscosity of the Derakane/binder mixture was found to be 17.49 (Appendix
B). The mixture can be better explained using the Huggins equation:

p=tto+ttox[tt]xC+0.242xttox[u]2xC2+0.293x|.10x[u]3xC3

110 = pure resin viscosity (0% binder)
[ti] = intrinsic viscosity (17.49)

The Huggins equation is 5% accurate or better for the entire range of binder concentra-
tions that were studied here (0 to 8% wt. binder).

From Figure 6, it is obvious that if the binder is not dissolved evenly throughout the
resin, a viscosity profile will occur and fingering is a possibility. Also, the nonlinearity of
the curve indicates that during molding, at positions closer to the flow front, the viscosity
profile will be larger, which can lead to greater amounts of fingering. This can result in
larger inhomogeneities occurring near the flow front and any resulting part could be
weaker at positions further away from the inlet and closer to the outlet.

The specimens were also tested at 50 rpm to determine if the resin/binder mixture was
shear thinning. From Table l in Appendix A, the difference for the resin viscosity between
50 and 100 rpm was in the range of 0-2% with no clear trend being observed. The differ-

ence can be neglected and the resin/binder mixture is essentially Newtonian.

 

 

25

hem

 

mm

853—388 «not? 3 ER: 5 $33 4339 €885 a. 8%?—

cfimwu cfl Moccfin .pz R

NR. mam arm AWV @mw .naw w._

 

soosm

 

   

26

In section 5.2.3.c, different mold temperatures were examined for washout runs. For
runs at temperatures other than 23°C, collected resin samples were tested at two different
temperatures, the mold temperature at 23°C to find the binder concentration. Figure 7
shows the viscosity/concentration curve for 23°C, 40°C, and 60°C. As can be seen in the
graph, the viscosity goes from 125 to 500 cps for the resin temperature of 23°C, 48 to 142
cps for 40°C, and 21 to 76 cps for 60°C over a binder concentration range of 0-8% wt.
From Figure 7 , it can be seen that at higher resin temperatures, not only is the resin viscos-
ity lowered, but the change in viscosity as a result of the binder is suppressed. This means

at higher mold temperatures, the viscosity profile will be suppressed.

5.2.2. Binder solubility and dissolution rate

The solubility runs were performed at three different temperatures, 23°C, 40°C, and
60°C. The solubility was found to be 0.0814 g of binder/cc of styrene at 23°C, 0.206 for
40°C, and 0.385 for 60°C. The Derakane solubility check was performed at 23°C and
gave 11.86% wt. binder as the maximum concentration.

The dissolution rate of binder was studied at three different styrene temperatures,
23°C, 40°C, and 60°C. From Figure 8, it can be seen that as the styrene temperature is
increased, the rate of dissolution also increased. The vertical line represents a typical mold
filling time (1 minute) for the RTM system. At this time, only half of the binder was dis-
solved at 23°C, while 83% of the binder was dissolved at 40°C and 93% of the binder was
dissolved at 60°C. This would indicate that, during mold filling, as the resin temperature
increases, more binder will be dissolved and the viscosity profile will also increase. How-
ever, as the temperature increases, the resin’s viscosity will decrease, and this may negate
the effects of increased binder solubility.

To derive the equation for finding the mass transfer coefficients, starts with the mass

balance equation for the static case:

(l-Vr) (C - Co) = Vr (WBo ' WB)

 

27

O.

m

me

o

.v

ococbm 5 523 no E53035 “w 853m

$8.:ch 0E;
mm om mN ON m.

r

O.

r

00

 

 

 

28

C = binder concentration in resin
C0 = initial concentration of binder in resin
W}; = mass of binder/ preform volume (g/cc)
WBo = initial mass of binder/ preform volume (g/cc)

The rate of binder loss in the preform is represented by:
- d WE = k [b WB - ((l-Vf)/Vf) C)
d t

k = mass transfer coefficient (s'l)
b = solubility factor (=solubility/ WB) (=7.236 for 23°C, 18.311 for 40°C,
27.111 for 60°C)
t = time (s)

Rearrange the first equation, set Co=0, and combine with the second equation gives:

Ld—‘EB=k[bXWB'(WBo'WB)]
dt

The above equation can than be rearranged and integrated to give the following equation:
(- 1/b+1) 1n {[(b + 1) WE - wB0]/b WBo} =kt

The left side of this equation was labeled the X function for simplification. A plot of the X
function vs. time for the three different temperatures is shown in Figure 9. The slopes of
these lines equal the mass transfer coefficient, k, for each of the temperatures. From figure
8, the mass transfer coefficients are 0.00198 5'1 at 23°C, 0.00181 5'1 at 40°C and 0.00205
s'1 at 60°C. The mass transfer coefficient stays relatively constant over this range despite

the increase in dissolution rate shown in Figure 8.

5.2.3. Binder washout tests

For this set of experiments, three factors that affect mold filling behavior were varied
to study the flushing of binder out of the mold: fiber volume fraction, RTM operating pres—
sure (which determines the resin fiow rate), and mold temperature. All viscosity and
binder concentration graphs in this section are in terms of resin flushed, which is the vol-

ume of resin that was flushed out through the mold and been collected.

 

   

29

cm

 

e

Owe

0:25“ 5 83 825 8a 8:3qu8 hem—ab 332 um oSwE

Amocoommvmite
oer ONE cor om cm 0%

cm

 

 

uoqounjx

   

3O

5.2.3.a. Fiber volume fraction of the preform.

To find the dependence of preform’s fiber volume fraction on the binder washout, two
different types of preforms were used, the first made of one ply of Unifilo U750 mat, the
other with three plies. The one ply preforms had masses between 37 and 42 grams and
fiber volume fractions between 0.056 and 0.068. The three ply preforms had masses
between 100 and 125 grams and fiber volume fractions between 0.147 and 0.178.

Figurele and 11 show the viscosity and binder concentration curves, respectively for
one and three plies at a mold temperature of 23°C and two different operating pressures,
25 and 60 psi. For the 25 psi runs, the initial viscosity of the three ply (filling time, If = 226
s) test was twice that for the one ply test (tf = 84 s). The drop in viscosity was much
steeper for the three ply run, -1.0 cps/ml flushed compared to -0.067 cps/ml for the one ply
test during the flushing of the first 150 ml. In the 60 psi runs, the initial viscosity of the
three ply run (tf = 41 s) was 1.5 times greater than for the one ply run (tf = 28 s) and the
viscosity profile was -0.33 cps/ml at its peak for the three ply and -0.033 cps/ml for the
one ply. The reason for this is two fold, as the amount of plies of preform increases, the
amount of binder available for dissolution increases. The three ply preforms had three
times as much binder coated on them as the one ply preforms so the resin could dissolve
up to three times as much binder.

The other reason is that as more plies are used in the preform, the void fraction
decreases and the resistance to flow increases, which is Darcy’s law. Since this was a con-
stant pressure system, a slower resin flow rate and increased resin residence time occurred
inside the mold. This gave the resin more time to dissolve the binder as is shown in the
mold filling graph in Figure 12. As shown there, for an operating pressure of 25 psi, it took
84 seconds to completely fill the one ply preform while it took 226 seconds to fill the three
ply preform, nearly three times as long. For an operating pressure of 60 psi, the filling time
was 28 seconds for the one ply and 41 seconds for the three ply run, only 50% longer than

for the one ply. (If the flow rate was constant instead of constant pressure, the effects

 

   

 

31

 

Damn Ha 3:805 E m use a .8“ $883 fine so acute 2385 ”S oSwE

. AFCV @952:sz ®E3_o>
one. 00¢ com com 0mm 00m Om: ooh om

 

 

 

 

 

 

 

8 8 .36 m
66 8 .36 a
56 mm .36 m
as mm .36 _

00C!-

 

 

 

 

 

32

Comm 3 2:889 bu m 93 g .28 conga—8:8 BEE co 3800 Began n: ouswmm

050 096:: Em? eE:_o>
00¢ 00¢ 00m 00m 00m 00m 00e 00f 00

 

 

66 co .26 m
66 8 .36 _
66 mm .36 m
8 mm .36 _

OGD-

 

 

D
(rubraM 1093190) 0011011093003 raping

 

   

 

33

 

Damn 3 matches hm m can 0 no.“ we: 6..» E80 58m ”2 2:me

AEoveocgmfiEotEmmm
0v 0% mm 0m 0m 0m m: 0H

 

6Q 00 .30 m.
66 oo .3; _
8 mm .3; W.
Go mm .30 r

0.13-

1

 

   

 

34

would be opposite, it would take less time for the three ply to fill up since there is less void
volume to be filled by the resin.) In Figure 12, it can also be seen that the flow rate is linear
for the one ply runs and the three ply run at 60 psi, but decreases for the three ply, 25 psi
run. Low pressure is more affected by the preform resistance to flow but also the low pres-

sure results in higher resin viscosities which will also slow down the resin flow rate.

5.2.3.b. Operating pressure.

As well as showing the effects of fiber volume fraction on binder concentration in
flushing, FigureSIO and 11 also show the effects of the operating pressure on the viscosity
and binder profiles. As can be seen for both types of preforms, the amount of binder dis-
solved and the binder profile decreases as the operating pressure increases.

Figures 13 and 14 show the pressure effects on the viscosity and concentration profiles
for the one ply preforms at 23°C. Here, the viscosity and binder concentration profiles
were twice as large for the 25 psi run than for the 40 (tf = 25 s) and 60 psi runs. The viscos-
ity of the 25 psi runs was 17% larger than the 40 and 60 psi runs which varied by 8%
between each other. the concentration of the 25 psi run was nearly twice as large as that for
the 40 and 60 psi runs which varied by 40%. Both of these facts indicate that at higher
operating pressure, changes in the pressure have a smaller effect on the viscosity and
binder concentration of the resin. Figure 15 bears this out, the mold filling time for the 25
psi run was 84 seconds while it was 25 and 28 seconds for the 40 and 60 psi runs respec-
tively. The reason for this may be that the fiber volume fraction of a one ply preform is so
low that at higher operating pressures, the resistance to resin flow becomes negligible.

Figures 16 and 17 show the effects of pressure on the viscosity and binder concentra-
tion profiles for the three ply preforms at 23°C. Here a situation similar to that for the one
ply is seen but to a greater degree. The initial viscosity of the 25 psi run is 70% greater
than the 40 (tf = 575) and 60 psi runs which vary by only 2%. The viscosity profile for the

25 psi run is 1 cps/ml while it is 0.33 cps/ml for the 40 and 60 psi runs. A similar result

 

L t

 
      

, g.- . - '. '_ .I. u I '1 . I .1“
.‘l’: f In 1* r I .l ' l y g... 9"!“ E." ,v
. I I . -'.'.1 !d
I I. II |I- .. .
. I .‘L
J -._ f. I I I
I... . I I
l u. I
.ll- -.

”in (It;

 

 

35

Down 3 2:893 be ~ 80 $883 fine co $8.000 238i ”9 Emmi

ccevbmzmzcgtmog0omze:;o>
00¢ 00¢ 000 m_m 0mm 0mm 00_ mm? 00 me

fillflllJllJrlllellfllllqllqllqllqllJllT

 

 

_mao© a

_mao¢
_QON o

 

 

35m

Sdo)Kuso

 

 

36

00¢

 

00¢

Comm 3 Queen be H .8 count—50:8 .523 :o 808% 2385 ”E unswE

0C0 Dogma: :68 @E3_o>
00m 00m 00m 00m 0m: 00 e 00

 

 

0;

L”.

O.
(\I

L0.
N

O.
to

L“.
m

0.
fl.

(1060M z) 0011011093003 Japuig

 

 

37

00

 

mm

00

Damn 8 3:805 bu ~ 08 0E: .m> E80 58m ”2 Bami

AEUV mocofifi EC: 2QO
m¢ 0¢ mm 0m 0m 0m m:

0e

 

 

m

 

 

38

Comm 3 8:895 be m “8 .388? 580 so 88000 2885 ”3 Emmi

0:0 096:: :68 0o 6E3_o>
00¢ 00¢ 00m 0311 0km 0mm 00h mm? 00

 

m¢

 

 

39

00¢

 

00¢

Damn ”a 80895 in m .80 cougcoocoo 08:5 co acute 2385 ”S Semi

0:5 @063: :60: mEEo>
000 000 00m 00m 00 9 00 9 0

0

 

(\J \—

NU

fi—

2) 0011011093003 1913018

LO

(10619/0

(Q

 

   

 

40

also occurs for the binder concentration curves. For the three ply runs, a trend similar to
the one observed in the one ply runs is noted, that at higher operating pressure, changes in
the operating pressure have a smaller affect on the resin viscosity and the binder concen-
tration. Figure 18 shows the mold filling rate for the three ply runs. The mold filling time
for the 25 psi run was 226 seconds, for the 40 psi run it was 57 seconds, and for the 60 psi
run it was 41 seconds.

From the results shown, it can be seen that an increase in operating pressure of the
RTM machine will result in a decrease in the amount of binder dissolved in the resin.
Increasing the operating pressure results in an increase in the resin flow rate and hence, a
decrease in the resin’s residence time in the mold. This means that the resin will have less
time to dissolve the binder available in the preform, and a smaller binder concentration

and a more even profile will result.

5.2.3.c. Mold temperature.

The runs for determining the effects of the mold temperature were done at three differ-
ent temperatures, 23°C, 40°C, and 60°C. As stated in Section 5.2.1, the resin specimens
for the 40°C and 60°C flushing runs had their viscosities measured twice, at the mold tem-
perature (as shown in Figure 7) to get the actual viscosity profile and at 23°C so that the
binder concentration could be read from Figure 6.

Figure 19 shows the viscosity profile for the two different types of preforms during the
flushing runs performed at 40°C. The filling times for the three ply runs were 141 s for the
25 psi run and 54 s for the 40 psi run. The filling times for the one ply runs were 92 s for
the 25 psi run and 36 s for the 40 psi run.Here, the trends are similar to those in Figure 9,
with the three ply run having initial viscosities twice as large as the one ply run for 25 psi
operating pressure and 50% higher for the 40 psi run. In both cases, the viscosity profile
was also higher, at -0.58 cps/ml compared to -0.10 cps/ml for one ply for the 25 psi runs,
and -0.20 vs. -0.056 cps/ml for the 40 psi runs with less than 150 ml of resin flushed. As

   
  

..-="_i. 1!-.- -_-_|=.-';- :.-.u 1'11 .-'-'$"-m. i :"liiL-‘J’lifi 510') :ahnr-‘i

_.._ __ . ...a;,.,

 

 

41

Dona 3 8:82: 3: m :8 0:5 .m> “no...“ 58M ”M: oEwE

0¢

01:3 00:066 0:0: Emma

0¢

00

6‘

00

\V

0m

\\

0m 09

(V

\\

_mao@
_mgo¢
imamm

09

I11
\L'III. .

0

 

 

 

 

 

O.

 

“

0
0m

 

 

42

Uo0¢ an 83 Sofia? E acouatg €883 am ”2 Esmfi
A?:VU®:63:2:mmm
00¢ 00¢ 00m 000 00m 00m 009 009 00 0
0¢
00
00
0h
00

00

L0
CD

(Sd3>flpso3aA

 

  

8 or: .36 m
8 3 .2: _
8 mm .36 m
.8 mm .36 a

 

OD.-

 

 

   

 

43

would be expected, the viscosity at 40°C is much lower than at 23°C. Figure 20 shows the
binder concentration curves at 40°C derived from testing the resin samples at 23°C (Table
3.C Appendix A).

Figure 21 shows the viscosity profile for the 60°C runs with the two different types of
preforms. The filling times for the three ply runs were 150 s for the 25 psi run and 37 s for
the 40 psi run. The filling times for the one ply runs were 131 s for the 25 psi run and 34 s
for the 40 psi run. As in the other mold temperatures, the difference in the initial viscosity
and viscosity profile is much larger for the 25 psi runs than it is for the 40 psi runs. The
viscosity profiles for the 25 psi runs are -0.26 cps/m1 and -0.08 cps/ml for the three ply and
one ply runs respectively, and -0.096 and -0.02 for the 40 psi runs with less than 150 ml of
resin flushed. Figure 22 shows the binder concentration profiles at 60°C derived from test-
ing the resin samples at 23°C (Table 3.C Appendix A).

Figure 23 compares the binder concentration curves for one ply preforms and the three
temperatures. The steepest profile occurs at 60°C. This would be expected as shown in the
solubility curves in Figure 10. As the resins temperature is higher, it will dissolve more
resin quickly leaving less binder for successive resin to dissolve creating a steep profile.
Lower temperatures, as seen with the 40°C and 23°C runs, will result in more even disso—
lution and flushing of the binder.

Figure 24 compares the binder concentration curves for three ply preforms and the
three temperatures. As in Figure 23, the higher temperature results in the steeper concen-
tration profile. However the three runs are much more closely matched despite the temper-
ature differences particularly between 40°C and 60°C. They rarely vary by more than
one% binder concentration. The reason may be that besides although the higher tempera-
ture increases the rate of dissolution of the binder, it will also decrease the viscosity of the
resin and hence speed up the flow rate in the mold. The increased flow rate would reduce
the residence time of the resin in the mold, giving it less time to dissolve the binder. Still,

the dominate factor here clearly is the temperature. The 23°C run, although having the

          
 

it'l'r'i 1!; Mid: raw-01 affirm ti "1'“: .‘r' {rim-121v 3d:

 

 

44

UoO¢ “a 88 50:83 5 825 .8 8:83.88: Him ”0N 8:me

A:c066;mscitmmm
om¢ 00¢ 0mm com com com om_ oo_ on o

 

 

0
a
m M
a
pm.
mm“
d
9
¢ 0
a
W
m m.
U
Dr
Q 9
_mao¢.zam - 1
_mqo¢.26 : o
6: 0m .3: 0 D A
6: 0m .3: P o

 

 

 

45

Uooc a 82 80:83 E .283 .8 8:20:88: am JN 2&5

080 ©9630 :60m
00¢ 00¢ 000 000 00m 00m 009 009 00 0

 

 

  

8 9. .8 m
8 9s .8 _
8 mm .8 m
8 mm .8 8

 

OE].-

 

 

   

 

46

Omxw 00v

Uooo 3 EB :55?» 5 “8:5 mo cozabcoocoo :xm ”mm oEmE

QEV Umgmsz Emma
0mm com 0mm 00m Om: Om:

 

ma 3 .3a m
mg 9V in f
8 mm .3 m
9 mm *3 _

OE].-

Om

 

 

 

 

 

DU1q1ue3Jedw5!aA/x

 

47

@895 bu H a awsefi Bow 5 32:0 50:33 525 ”mN oSwE

A?cvbm;m3:r:w®LtumC::o>
Omw mofi owm mrm 0mm mmm owe mme om
fiJlflJJJIalJIH|14|q|JIIJI

mv

  

cor
me_
omr
mt?

Ooh

AMSOQS

LO
|\

/

<:
07
(sdo)

 

   

 

48

   

5.80:: bu m a 5:85 30: E 3350 50:33 505m ”Va 053.:

AmEoV bmgmzz Em? :o mE:_o>
00¢ 0mm 00m 0mm DON om; QQF

oomom .EENH: oooo
00m Fm ”SEN“: 009V
033 EEmH: 00mm

0

m

   

(\1

[\f‘)
JSDU!Q1USOJ9d1q5!e/\/\

q.

LO

49

largest residence time, has the most gradual profile because, as seen in Figure 8, the disso-
lution rate is so much lower than it is for either 40°C or 60°C. Decreased resin viscosity
should become more important at higher temperatures.

Finally, the higher mold temperatures decreased the resin’s viscosity, but also sup-
pressed the affects of binder concentration resulting in more gradual viscosity profiles. For
three ply, 25 psi runs (which have the highest resulting binder concentrations) the viscos-
ity profile dropped from -1.0 to -0.58 to -0.26 cps/ml as the mold temperature increased
from 23°C to 40°C to 60°C despite a relatively small increase in the binder concentration.
This means that fingering and inhomogeneities in the flushing of binder should be reduced

with increasing mold temperature.

5.2.3.d. Summary.

From Figures 9—24, it can be seen that all three variables, fiber volume fraction, operat-
ing pressure, and mold temperature, all have significant effects on the dissolution and
flushing of the binder. Although these are flushing curves, they do give a good indication
of what may happen inside the mold. A steep binder concentration transient in the washout
would indicate that the concentration profile inside the mold will be equally high. Since a
higher profile will lead to more fingering and inhomogeneities, this would indicate that as
either the fiber volume fraction or the mold temperature increases, or the operating pres-
sure decreases, the amount of fingering occurring during mold filling and the inhomogene-

ities in the finished pan will increase.

5.2.4. Concentration profile in the mold

For finding the viscosity and binder concentration profile, runs with eight different
preform lengths were performed with an operating pressures of 20 psi and 30 psi, a mold
temperature of 28°C, and the preforms all consisted of three plies of mat. Figure 25 shows

the results of the eight runs at 20 psi and Figure 26 shows the results for 30 psi (because of

 

     

't L- r.- , '=: .-.r!'T- 1.9. '4' :E rad; :wm! drum:

 

: I“

 

Uowm .qu cm 3 BER: E0905 “Enema: 8.: Emu: 8:3: no 3885 ”mm 85E":
9?; bmcmj: Emma ®E20>

one cow omm com 0mm com one on: cm, Q
00:

 

 

Om:

ow:

on:

50

  

misoosm

wao/

(8d

CON

0mm

H$l>+<><lClO

 

 

ovm

 

 

51

om

  

iv

Uowm :8 Om “a «59.2 €805 Eoeobmc 5.: 58: 00:3: «0 3:885 6a 233...:

Achbmgmscrtmmc@C::o>
00% com com 0mm com cm: 00: cm

 

 

 

 

 

><v$rl>+<><lDO

 

cor

of—

QM:

one,

AMSOQSV\

%
(ado)

QQ:

on:

ow:

 

   

 

52

the number of curves represented in this graph, only the smoothed data is shown, the
actual experimental data is shown in appendix). These particular lengths matched up with
where the outlets were located along the center of the mold. As would expected, longer
preforms will have more binder coated on them, and produce a greater resistance to resin
flow, therefore the binder concentration in the resin will increases and the concentration
profile will become steeper with increasing preform length.

To construct the viscosity and concentration profiles inside the mold, one starts with
the first point on the concentration curve of the longest preform. This is the binder concen-
tration at the end of the mold furthest from the inlet. From the second longest mold, one
would not take the first point on the concentration curve but rather the point that represents
when the entire mold would be filled. This is the where the amount of binder flushed from
the system is equal to the volume difference between the second longest and longest mold.
From the third longest preform, one takes the point from the concentration curve where
the amount of resin flushed through the mold is equal to the volume difference between
the third longest mold and the entire mold. This process is continued up to the shortest
mold which represents the point closest to the inlet. These points can then be put together
to get the concentration profile along the length of the mold from the inlet to the outlet.

Figure 27 shows the viscosity profiles of the resin at the time just when the mold is
completely filled for the two operating pressures. This graph was constructed from Figures
25 and 24 as was indicated above. For each curve, the first point in the 22.25 inch (56.52
cm) length mold run was taken as the viscosity at the outlet. The next point away from the
outlet was taken from the 19.25 inches (48.90 cm) mold run. The difference in length
between the two molds is 3 inches (7.62 cm) and the resulting volume difference (taking a
void fraction of 0.85%) is 32.3 cc. This is the amount of resin that would have to be
flushed through the 19.25 inch mold to equivocate the point at which the mold is filled.
The viscosity is read off Figures 25 and 26 and is the viscosity at 7.62 cm from the mold

outlet when filling is completed. This process was repeated until the viscosity profiles for

 

   

 

53

cm

 

m

m

on

Down .60 em 0:“ an as 22: 0:: E 58: m0 5:885 Km oEmE

me

AEQV 6:: 20E 80: @8ng

can

mm

om

mm

ON

m:

m:

 

 

of
c:
on _,
8 a
at
cm 3/
cm 2..
com
o _m
0mm
0mm

misoosm

 

   

 

54

the entire mold (Figure 27) was produced.

From Figure 27, it can be seen that the viscosity profile of the mold for an operating
pressure of 30 psi is not as sharp as the profile for an operating pressure of 20 psi. The vis-
cosity along the length of the mold increases form 115 to 220 cps for the 20 psi run, but
only 118 to 167 cps for the 30 psi run. The viscosity profile also increases as one moves
away from the mold inlet, for the 20 psi run, the rate of viscosity increase versus the dis-
tance from the inlet ranges from 0.394 cps/cm for distances of less than 10 cm to 3.35 cps/
cm at distances greater than 40 cm. For the 30 psi run, the rate varies less, ranging from
0.131 cps/cm for distances below 10 cm to 1.15 cps/cm for distances greater than 25 cm. If
fingering occurs during mold filling, it should be most pronounced for slower runs such as
the 20 psi run and will also be the most pronounced in the area of the mold furthest from
inlet where the viscosity profile is the steepest. However, a situation develops here that
hasn’t been examined closely before. Unlike the experiments of Slobod, Tan and Homsy,
etc., this is not a situation of a sharp, clearly defined boundary between two liquids of very
different viscosities, but rather a constantly increasing viscosity with no definite interface
present. Hickemell and Yortsos have shown theoretically that fingering will occur in such
a situation, but experimentation is still needed to show that fingering will indeed occur.

Another aspect of this experiment that has to be taken into account is that the viscosity
of the liquid flowing in the porous media is continually changing as a result of binder dis-
solution. This means that the fingering conditions are also in a continuous state of change
and must be documented. Figures 28 and 29 show (for 20 psi and 30 psi respectively) the
viscosity profiles in the mold when the resin flow front reached different points inside the
mold with the last curves being the same as those in Figure 27. Figures 28 and 29 were
constructed in the same way as for Figures 26 with the first point taken from a viscosity
curve in Figures 25 or 26 respectively where the preform length matched the flow front
location.

Figure 30 shows the binder concentration profile in the mold just at the point of filling

F.

 

   

 

55

cm

 

m

m

in on .208 05 E 2&5: “sauna: “a 20...: E02 ::3 2:80 3885 ”mm 0.59m

Arc

om

we

UV 6:: 908 E0: 8:055

ow

mm

on

mm

ON

or

m:

 

 

gexfifiet—mi

H+D+<><1DO

 

0:
cm:
on:

OOOOO
OOI\©LO<1'

 

CD
CD

<

com
0 EN
0mm
0mm

sdo) Kilsoosm

 

   

 

56

ow

mm

:00 an .208 05 E 0:095: “020%: 3 “not 502 5:3 2:80 3:885 ”0N 80$":

AcLov:®_c_U_orc Caoemm90c0:m_g

on me 0% mm on mm ON me Q: m

 

 

wwwwmwwm
049404CHQ<Q<ECV
% e D +—O)< D o

 
 

 

 

m__
cm:
mm:
om:
mmr
oer

OLOOLOOLO
NLOLOLOlflfi

 

LO
;\

(sd3>flysooaA

 

   

57

ow

 

mm om

Uowm :8 on 000 cm “a 20:. 05 E 009:: m0 00:00:00000 ”em 8:me

AEQV :07: EOE ED: 00:07.5
we 0.: mm om mm Om m_ of

 

 

”7. 0. L(7. C. 0 Q.
m m _ _ O O
/ 001011093003 Jepug

”J

O .

{V

 

 

58

for both operating pressures. For 20 psi, the binder concentration increased from 0.9% to
5.0% wt. from the mold inlet to the outlet, while for 30 psi the binder concentration
increased from 1.1% wt. to 3.1% wt. along the mold length. The binder concentration pro-
file is larger for the 20 psi runs, indicating that larger inhomogeneities will occur in the
resulting part that could weaken it to a greater extent than for 30 psi runs. It should also be
noted that for this profile as well as for most of the washout runs, the lowest concentration
was 1% wt. 0% wt. which may indicate that the powdered binder supplied by Vetrotex is
not exactly the same as the binder on the mat.

Figure 31 shows the bulk (resin) mass transfer coefficients (k) along the length of the
mold for the two operating pressures. As can be seen, in both cases the coefficient is very
large at the inlet of the mold and drops off rapidly. For the 20 psi run, k is 0.0030 cm/s at
the inlet and drops rapidly to 0.00075 cm/s at 18 cm from the inlet and remains steady up
to the outlet. For the 30 psi run, k is 0.0060 cm/s at the inlet, drops to 0.0011 cm/s at 18 cm
from the inlet, and after that drops gradually to 0.00080 cm/s at the outlet. This is what one
would expect since at the inlet, the resin’s binder concentration is 0% wt. and the binder
will be dissolved readily. At the outlet, the resin’s binder concentration approaches 3-5%
wt. and will lower the driving force for dissolution of binder and thus will slow it down.
The higher viscosity of the binder will also limit the dissolution of binder by hampering

the diffusion of the binder away form the fiber surface once it has been dissolved.

    

351133;! "2 "z." .zrnuzrmq

         

I ., . . _ u-n

 

59

om

208 05 “:0 Swan: 05 mac? 80208000 000050 was): ”8 oSmE

6:: DEE E0: 000806
me 00 mm om mm ON or or

 

 

||| Lil III I]!
L“. 0. LO. 0.
<1— fi' M I“)
3)

 

 

   

 

6. EFFECTS OF CONCENTRATION AND VISCOSITY VARIATIONS ON THE
MOLD FILLING PROCESS AND PART STRENGTH

This section examines the effects that the viscosity and binder concentration profiles
(studied in Section 5) have on the crosslinking reaction, mold filling process, and strength

of the resulting part (4-6, Section 3).

6.1. PROCEDURE

6.1.1. Binder effects on crosslinking

The next phase in this project was to determine the effects (if any) that the binder may
have on the crosslinking reaction once the mold was filled. First, two samples of resin
were weighed out, and the desired amount of accelerator was added to each sample. A cer-
tain amount of powdered binder was then added to one of the resin samples while the
other sample did not have any binder added. Then the desired amount of catalyst was
added to one of the resin samples and a small sample was placed in the DSC. The DSC
was set on the isothermal mode with the temperature set at the curing temperature needed
for making plaques later on. The graph was taken of the amount of heated needing to be
added or subtracted from the system to maintain the desired temperature. This process was
then repeated for the other resin sample and the two graphs were compared to see if the

binder did affect the reaction.

6.1.2. Production of plaques
The next two phases of this study involved the actual production of plaques. For these

parts of the experimentation, the first step was to determine the precise flow ratio of cata-

6O

 

   

 

61

lyst to resin. This was done by removing the catalyst hose from the tank and the resin hose
from the mixing head and placing them in separate preweighed sample jars. The RTM
machine was run for one cycle (single shot) and the samples of resin and catalyst were
weighed and compared. The catalyst pump was then moved either closer to or further
away form the resin pump as the need may be (the distance needed to move the pump was
determined from tables and equations provided in the RTM machine’s manual). This pro-
cess was repeated until the desired catalyst/resin ratio was attained. The resin and catalyst
hoses were then reattached to the mixing head.

The DMA accelerator was then added to the resin in the holding tank. The required
amount of DMA was first added, under a fume hood, to a beaker filled with about 700ml
of Derakane. A magnetic stirrer was used to thoroughly mix the DMA into the resin. The
covered beaker was then transported over to the RTM machine and the resin/DMA mix-
ture was poured into the resin tank containing the rest of the resin and the tank was
allowed to stand for several hours to make sure the DMA spread evenly in the tank. The
RTM machine was then run through 150 cycles to purge the resin in the lines that did not
contain any DMA.

The resin and catalyst where then pumped into the preform containing mold in a simi-
lar manner as for the flushing runs. The injection continued until resin first started coming
out of the outlet valve, at which point the machine was stopped and the inlet and outlet
valves were closed. Then the mold was cured at 50°C (122°F) for one hour and postcured
at 100°C (212°C) for another hour. While the plaque was curing, the static mixer and feed
line were purged to prevent curing inside the mixing head. After postcuring, the mold was
air cooled to room temperature and the plaque was removed (it is very important to coat
the bottom plate with mold release before placing the preform on top of it, otherwise it is

nearly impossible to remove the plaque without damaging it).

 

 

-u.
. ..'.

 

-.v_ a we

 

62

6.1.3. Observation of fingering during mold filling

For this step in the experimentation, plaques were made as stated above except here
dye was added to observe any fingering that may occur. Since this is a continuous process,
the dye had to be added to the resin in the line. This was done by placing in the line just
before the inlet valve a T-square with a septum attached on its top branch and a short static
mixer. The dye (Oil Red 0 biological stain from Aldrich Chemical Company) is added to
a sample of resin to about 0.6% weight and a syringe was then used to draw out a 10 ml
sample of the dyed resin. During a run, the syringe was inserted into the septum and the
dyed resin was injected into the line at the desired moment. The static mixer adjoining the
T-square then mixed the dyed resin with the feed resin thoroughly as it flowed into the
mold.

While the resin was filling the mold, the Nikon camera was mounted on a tripod over-
looking the mold so that the entire mold could be seen in through the camera lens. Pictures
of the filling process were taken every 2—4 seconds to record the development of any fin-
gers. A stopwatch and a ruler were placed on top of the mold so that length and time could
also be recorded by the camera. After the mold was filled, the plaque was cured as noted in
the previous section. Both end loaded and center loaded plaques were made in this experi-

ment.

6.1.4. Determination of flex strength and modulus of the plaque

During the runs with colored resin, a problem developed of air voids being produced
in the mold as a result of injection of the dyed resin. The resulting plaques were therefore
not used for determining the flex modulus and strength of the plaques. Instead, new
plaques were made without the dye but at the same conditions as for the dyed plaques.
Any fingering could be determined by comparing to the dyed plaques and the binder con-
centration profile can be determined by a comparison with the concentration profile pro-

duced in Section 5.

 

 

 

63

After the plaques were made, they were cut, using a diamond saw, into a series of
1.03-1.30 cm wide strips transversely to the resin flow in a pattern shown in Figure 32.
The unnumbered sections of the plaque had small bulges as the result of the plugged outlet
holes and were not tested since the bulges could alter the resulting flex modulus. Each
strip that was to be tested was numbered. The strips were then cut to a length of 6.35 cm
for three point flex testing.

The three point flex test, with the two bottom points as the supports and the top center
(cross-head) point bending the sample, was then performed on the numbered strips using
standard testing procedures (16). The support span was 5.12 cm and the cross-head speed
used was 1.3 mm/min. The flex test was carried out until each sample just reached the
breaking point. The flex modulus and strength of each sample could then be put together

to give a profile over the length of the plaque.

6.2. RESULTS AND DISCUSSION

6.2.1. Binder effects on crosslinking

The DSC runs were then performed to determine the affect of binder concentration on
the crosslinking reaction. The runs were carried out at 50°C, which was the curing temper-
ature used for making the plaques in subsequent experiments. The first run was carried out
for 150 minutes to make sure that the proper baseline had been established, the succeeding
runs were then performed for 30 minutes. The three graphs show the heat flow out of the
samples (since the reaction is exothermic) versus the time of reaction with 0 minutes rep-
resenting the starting time of the runs.

Figures 33, 34 and 35 show the isothermal runs for resin specimens with 0%, 1.04%
and 3.12% weight of binder respectively. The reaction occurs in the same time period,
from 2-6.5 minutes for the 0% run, 2-6.7 minutes for the 1.04% run, and 2-7.5 minutes for

the 3.12% run. The peak reaction occurred at 2.98 minutes for the 0% and 1.04% runs, and

 

   

. r' L-JII’.

 

80 «Wow

 

029:0» x0: mains: 08 50:00 miss 0035 “mm 80mm

 

   

 

65

ill .ii

.025 .3 so as» a: 80 ”mm 2&0.

. «Cueu afidh
comm uco no Ou v> HnLocow
0%— m« m.“ v.“ Na 0.“ m w v m o

Mold 393H

(MW)

 

gemmm mi“

mé

o.m

 

 

66

comm ucolan o«.v> Hmmeum
om mu m«

 
 

    

323 .E .2 53 s: 80 an 23E
ands. usah

v“ Na 0“ m m

  

    

0.0

N.o

v.0

w.o

m.o

0.“

m.«

V.“

w.«

m.“

o.m

Mold QEBH

(MW)

 

   

 

67

comm ucoman o«.v> ”Mmemw
om m« w«

   

523 .3 em a? 5: 0mm ”2 250E

Haas. mEdk

Va Na 0“ m w

   
  

    

:fiemm.m

o.o

m.o

v.0

w.o

m.o

o.“

N.“

v.«

m.«

m.“

o.m

Mold QBBH

(MW)

68

at 3.23 minutes for the 3.12% run. These results seem to indicate that the binder is acting
to slow down the crosslinking reaction. The heat of reaction was 7.862 J/g for the 0% run,
8.307 J/g for the 1.04% run and 5.238 J/g for the 3.12% run. The binder therefore may act
as an inhibitor of the crosslinking reaction or, because the binder increases the resin’s vis-
cosity, it may impede the movement of styrene monomers and impede the crosslinking
reaction. Because of this, the strength of the part will be weakened since less styrene

crosslinks will form.

6.2.2. Fingering during mold filling

Figure 36 on the next three pages shows six photographs taken at various times during
the mold filling run with an operating pressure of 20 psi and a mold temperature of 28°C.
Figure 37 on the three pages after that shows six photographs taken during mold filling at
30 psi and 28°C. The watch shows the time at which each photograph was taken and the
ruler shows the resin flow front and fingers position at each point. For both runs, 0:00 min-
utes is the time at which the stopwatch was started and the RTM machine began operating.
The edge of the entire mold was measured as 0 inches for these runs and two inches
marked the front edge of the preform. The large “2” written on the preform in the 20 psi
run and the “4” on the preform in the 30 psi run were just a markers used to keep track of
the runs.

For the 20 psi run (Figure 36), the resin first entered the mold at 34 seconds and the
dyed resin entered the mold at 54 seconds. Figure 36.A shows the mold at 58 seconds, just
after the dyed resin first entered the mold. The resin flow front was read off the yardstick
as 9 inches or 7 inches (17.78 cm) into the preform, beyond the flow front the preform was
untouched by the resin. At the front of the preform (left side of the photograph), the dyed
resin can just be seen (the darkest area in the mold). One major finger appears to forming
with a length of 3.6 cm and a width of 3.6 cm with several very thin fingers occuning

across the dyed resin front. These minor fingers may be the result of instability caused by

 

   

 

69

Figure 36. Fingering during mold filling at 20 psi, 28°C

 

 

 

 

 

 

 

 

 

Figure 36.B. 69 seconds

 

.Jtai :

   

 

7O

 

Figure 36.D. 100 seconds

 

 

71

 

Figure 36.F. 125 seconds

II“

 

   

 

72

Figure 37. Fingering during mold filling at 30 psi, 28°C

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 37.B.50 seconds

 

      

" ..-.-.- :1. 31271?“ i.-.'- - -: Maul.- gn'wgnifi

 

 

73

 

Figure 37D. 60 seconds

 

   

 

74

 

Figure 37.F. 72 seconds

 

   

75

the pulsing flow (as described in Section 4.2.2.) or by the miscible nature of the interface.

Figures-36B through 36.F shows the finger propagation during the mold filling in the
time period from 69 to 125 seconds. The finger length increases over this time period from
5.36 cm to 11.32 cm and the width increases from 4.8 cm to 8.9 cm. Figures 38 and 39
gives a more detailed account of the growth of the finger’s length and width respectively.
Also in this run, a second smaller finger also developed in the mold toward the mold edge
with the yardstick (bottom of the photographs). The smaller finger has a width of 3.6 cm
and the length increases from 4.4 cm to 4.8 cm until at 100 seconds, when this finger com~
bines with the larger one.

For the 30 psi run (Figure 37), the resin first entered the mold at 31 seconds and the
dyed resin entered the mold at 39 seconds. From Figures 37.A to 37.F, it can be seen that
the resulting finger grows at a much faster rate at 30 psi than at 20 psi. In the time period
from 42 to 72 seconds, the finger length increased from 5.9 cm to 22.2 cm and the width
increased from 7.9 to 11.9 cm. Figures 36 and 37 show the growth of the finger length and
width in more detail and in comparison with the growth rate at 20 psi. As before, the resin
has the same thin fingers occurring across the dyed resin front as before and the finger has
the same basic shape as in the 20 psi run.

As can be seen in the pictures in Figures 36 and 37, and Figures 38 and 39, the rate of
finger growth was nearly five times greater for the 30 psi run than for the 20 psi run even
though, as Figure 27 shows, the viscosity profile is twice as large for 20 psi than for 30 psi.
The reason is that what matters is not the increase in viscosity over a certain length of the
mold but rather the jump in viscosity at the point of interface. Because the change in vis-
cosity is continuous over the length of the mold, the jump in viscosity at any one point will
be very small regardless of the viscosity profile. Therefore the viscosity profile will not be
a large factor in determining the amount of fingering that will occur during mold filling.
The major influence on the amount of fingering will be increase in resin’s velocity. For the

20 psi run, the average resin velocity was 0.612 cm/s and for the 30 psi run it was 1.58 cm/

 

   

 

76

one

omr

cm?

95:: 20.: main: 5300» Swan: Swim ”mm 0.5mm”.

Amvate
of? 00: Ga cm on Om

 

 

 

 

 

 

 

   

mafia EoE mats: 530% 523 coma—E 5m 203m

388.;
9103030: 2: em 00 E 8 on 04 cm

 

<1.

77

L0

 

 

 

OO

 

(L03) 01609119601:

O

 

 

 

 

 

 

 

(\1

.0:

 

 

   

 

 

78

s, 2.5 times as fast as for the 20 psi run. This large difference in resin velocity will have a

larger effect on fingering than the small difference in their viscosity profiles.

6.2.3. Flex strength and modulus of the plaque

The plaques produced for these studies were made using three ply preforms and had a
mold temperature of 82°F (27°C) and an operating pressure of 20 psi as the mold filling
conditions. Figure 40 shows the flexural strength for a front end loaded plaque. The flex
strength has a wide variance from 21,000 to 31,000 psi and the regression curve shows
that the trend is linearly downward from 25,300 to 23,900 psi from the inlet to the outlet of
the mold. Comparing this to the binder concentration of the plaque (Figure 30) shows that
the flex strength decreases from 25,300 psi to 23,900 psi as the binder concentration
increases from 0.95 to 5.2%. Figure 41 shows the flex modulus of the same plaque. Here,
the modulus varies from 860,000 to 1,180,000 psi but the variance isn’t quite as large as in
Figure 40. The regression line shows a linear drop in the flexural modulus from 1,025,000
to 910,000 psi over the length of the plaque as the binder concentration increases from
0.95% to 5.2%. Because the drop in the flexural strength and modulus is slight (5%) and
the variance in the sample values is large (40-50%), a definite conclusion on the affect of
binder concentration on the physical properties of the part should not be made.

In her studies, Losure (13) observed, by using the same method used here, an average
variance in flex strength and modulus values for a nonwoven preform of 6.3%, much
smaller than seen in this study. Causes for this could include inhomogeneities in the pre-
mold or very small air voids in the plaque. Losure also observed a significant drop of
strength in sections of the plaque where fingering was present. This phenomena was not
seen here because, unlike in Losure’s studies where a partially cured resin was displaced
by an uncured resin, here there was no interface between two different liquids. Therefore,
the resin cured evenly throughout the plaque and there was no interface of two resins that

would have bonded relatively weakly.

 

   

 

mm

79

00020 00:00: 000 SN 00 5300.... :EBBE Nov 05mm":

AEOV :07: 908 80: 0000:05
om man 00 mm om mm cm of of m

C)

\f\f\f\f
Lr_JLr_Jl_LJLr_J
@0010

‘—

\f \f \1

IL] 11.1 LL.

N. Q 00.

N (\J r—
U

(18d) L116 9118 xalj

\T—
0

5i.
N

 

¢5@.m

 

 

 

 

80

mm

om

mv

262m 833 can 5 :6 3:62: 35on Av Bawfi

AEQV 6:: 908 80$ @8ng

ow

mm

\1

cm

mm

cm

'1

0

mm?

0
0

0

of

9

m

«.0

mmom

mm©©

mmofi

mmow

mea

d) snmpow ><S|_—f

8
@mo. 60
@mee

mime

 

   

 

81

Figures 42 and 43 show the flex strength and flex modulus, respectively, for a center
loaded plaque. Here, similar results to those of the end loaded plaques are shown. The
range of variance of the flexural strength and modulus varies wildly between 23,000 and
30,000 psi for the strength and from 860,000 to 1,110,000 psi for the modulus. The regres-
sion lines are also similar in shape showing a downward trend of 27,500 to 25,500 psi for
the flex strength and 1,070,000 to 1,020,000 psi for the flex modulus. Again, the variance
is too large to use these lines as conclusive evidence for any trend. However, the regres-
sion lines decreases form the front to the back end in a similar manner for both the end and
center loaded plaques which is something that would not be expected. What should occur
is that the center of the plaque should be strongest with the ends being the weakest Also,
it should be noticed that in comparison of the flex strengths of end and center loaded
plaques (Figures 40 and 42), the average strength of the center loaded plaque across its
length is approximately 2000 (10%) psi higher than that for the end loaded plaque. The
modulus of the center loaded plaque is 50,000 (5%) psi higher than that for the end loaded
plaque. This may be the result of more thorough flushing of the binder out of the center

loaded plaque.

 

   

 

82

mm

on

min

0:33 332 “850 a no fiwcobm :2:on ”NV Esmfi
AEQV $7: 908 :5: mocgmé
0v mm om mm ON m, F

9
a a
a
Q» ,
5

m:

\b

(O

m

he

 

 

   

 

83

mm

om

fr

mv

\‘

‘5

0:33 822 .850 a me 3:62: 35on av Emmi
A83 6:: 908 :5: mocgmfi
ON 0 f or

0%

mm Om

mm

mwom

mm©©

mmofi d
a

mmom

i» smnpotu x

mmoa

lSd

@mog/K

@mfe

mam;

 

   

7. CONCLUSIONS

The conclusions that can be drawn from this study are as follows:

1. Binder dissolution increases the resin viscosity and at higher binder concentrations,
the viscosity profile increases. This can cause a viscosity profile in a mold filling pro-
cess where the binder is not evenly dissolved.
2. Flushing runs have shown that binder dissolution is not evenly distributed over time
and that as more binder has been dissolved, there is less binder to be dissolved by suc—
ceeding resin, leading to a viscosity profile.
3. Increased temperature will increase the rate of binder dissolution in styrene (and
hence the resin).
4. The dissolution of binder is affected by three factors:
a. Increases in the preform density will cause a larger viscosity profile because
there is more binder that the resin can dissolve. The increased binder density also
increases the resistance to resin flow, slowing it down. This results in a larger resin
residence time in the mold giving the resin more time to dissolve the binder.
b. Higher operating pressure of the RTM machine will result in a decrease in the
binder profile. Once again, as the pressure increases, the resin flow rate increases
and the amount of time the resin is exposed to the binder in the preform decreases,
causing less binder to be dissolved.
c. Increased temperature increases the resin’s solubility which increases the disso-
lution of binder. However, the viscosity profile is lessened since the resin’s viscos-

ity decreases. The effects of binder on the viscosity is the thereby suppressed.

84

 

   

 

85

5. A viscosity and concentration profile also occurs along the length of the mold dur-
ing filling with the viscosity and binder concentration increasing from the mold inlet to
the outlet. When the mold is filled, the end closest to the inlet has been exposed to
resin longer and thus its binder was dissolved and flushed way more thoroughly. At the
outlet, the binder had the least resin exposure for the binder to dissolve and the resin
already has a high concentration of binder flushed from the front part of the mold. At
lower operating pressures, the viscosity and binder concentration of the resin increased
more rapidly along the length of the mold since the resin had a longer residence time.
6. Binder that is dissolved in the resin, decreases the total heat of reaction for the
crosslinking of the resin.This indicates that the binder impedes the crosslinking of the
resin and the resulting pan strength will decrease.

7. Miscible fingering occurs during the mold filling process. Essentially one large fin-
ger results that slowly grows as it travels across the length of the mold. The finger
grows faster with increased operating pressure even though the viscosity gradient is
more gradual at higher pressure.This is the result of the large difference in the resin
velocity for the higher operating pressure. This condition will take precedence over the
slight change in viscosity profile and will result in larger fingering at higher operating
pressure.

8. Three point flex tests performed on plaques were inconclusive.Although there was a
trend of decreasing flex strength and modulus at places in the mold with higher binder
concentration, the trend was too small and the scattering of the data too great to make
a definite conclusion on the effects of binder concentration.The scattering could be an

indication of inhomogeneities in the plaque resulting from the binder dissolution.

 

  
 

- . -_ -' --:"-,.151 an! '-_ -.-.u. 219-550 1.1:: 3 W11

 

' '.--‘-.' .mhua

4.
. J.lI-"I

 

8. RECOMMENDATIONS

1. Further study of the fingering phenomenon at higher temperatures to see if the
resulting suppressed viscosity profile will decrease the amount of fingering that occurs.
Also test at higher fiber volume fraction (up to 40%) to see if it occurs in conditions

related to industry.

2. Continue studying the strength of molded plaques with tensile strength as well as

flexural strength testing being performed. -

3. Study performs with higher fiber volume fractions (made with 4 plies or more). in
these runs care must be taken to ensure that channelling does not occur around the edges
of the preform. This will allow examination of whether these phoneme with binder will
also be apparent or if the much smaller void size will eventually impede binder dissolu-

tion.

86

 

   

 

APPENDIX A

 

   

 

RAW DATA

All data shown are accurate to three data points

Table 1: Binder concentration vs. viscosity at 23°C

First run Second run
% weight Viscosity Viscosity %weight Viscosity Viscosity
binder (50 rpm) (100 rpm) binder (50 rpm) (100 rpm)
cps cps cps cps

0.0462 121.5 122.0 0.0453 127.5 124.5
0.1022 126.0 126.5 0.1076 129.5 131.0
0.2166 132.0 132.0 0.2177 138.5 140.5
0.3147 135.0 135.5 0.3244 145.0 145.0
0.4165 140.0 139.0 0.4997 147.0 146.0
0.5044 141.0 141.0 0.4883 147.0 147.0
0.6173 142.0 142.0 0.6634 149.5 150.5
0.7995 144.0 143.0 0.7802 154.0 154.5
1.015 156.0 154.5 1.0282 155.5 156.5
1.193 154.5 152.5 1.1868 161.0 161.0
1.424 161.5 161.5 1.4374 166.5 167.5
1.642 166.5 167.5 1.6242 171.0 170.5
1.808 171.5 170.0 1.8210 180.0 180.0
1.990 177.0 175.0 1.9907 184.5 186.0
2.522 193.0 197.0 2.5111 199.5 200.5
3.004 209.5 210.0 3.0006 208.0 209.0
3.500 235.5 234.5 3.4930 235.5 237.5
4.016 253.0 255.0 3.9723 252.5 252.5
4.472 268.0 271.0 4.4893 277.0 274.5
5.008 291.0 296.0 4.9684 294.5 293.5

% weight Viscosity

binder (100 rpm)

cps
6.298 365.5
8.348 499.5

87

 

 

88

Table 2: Binder dissolution rate

Styrene temperature = 23°C

Tim Fiber weight % wt. drop % binder dissolved
(min) (grams)
Run 1 0 2.304 0 0
1:00 2.230 3.212 40.15
2:00 2.173 5.686 71.07
3:00 2.161 6.207 77.58
4:00 2.154 6.510 81.38
5:00 2.144 6.944 86.81
Run 2 0 2.339 0 0
0:30 2.315 1.03 12.83
1:30 2.237 4.361 54.51
2:00 2.208 5.601 70.01
3:00 2.189 6.413 80.16
4:00 2.181 6.755 84.44
5:00 2.174 7.054 88.18
Run 3 0 2.014 0 0
0230 1.987 1.341 16.76
1200 1.930 4.171 52.14
1:30 1.898 5.760 72.00
2:00 1.884 6.455 80.69
3:00 1.869 7.200 90.00
4:00 1.862 7.547 94.34
5:00 1.853 7.997 99.93
Run 4 0 2.390 0 0
0:30 2.368 0.92 11.51
1 :00 2.302 3.682 46.03
1:30 2.266 5.188 64.85
2:00 2.246 6.025 75.31
3200 2.238 6.360 79.50
4:00 2.219 7.155 89.44
5:00 2.212 7.498 93.10

 

   

Run 1

Run2

Run3

Runl

Run2

Run3

0

0:15
0: 30
0: 45
1.00
1:15

0:15
0: 30
0: 45
1 :00
1:15

0:15
0: 30
0:45
1:00
1: 15

0:15
0: 30
0: 45
1:00

0:15
0: 30
0: 45
1:00

0:15
0: 30
0:45
1.00
1:15

2.544
2.490
2.424
2.379
2.347
2.322

2.563
2.507
2.451
2.419
2.396
2.381

2.527
2.480
2.438
2.408
2.383
2.333

2.502
2.431
2.386
2.339
2.309

2.527
2.446
2.389
2.351
2.324

2.381
2.317
2.268
2.236
2.219
2.205

89

Table 2 (continued)

Styrene temperature =

0

2.126
4.724
6.513
7.763
8.000

0

2.197
4.362
5.638
6.535
7.105

0

1.840
3.514
4.710
5.687
7.710

Styrene temperature =

0

2.834
4.629
6.500
7.687

0

3.233
5.480
6.960
8.052

0

2.700
4.750
6.081
6.791
7.413

40°C

60°C

26.58
59.05
81.41
97.04
100.0

27.46
54.53
70.48
81.69
88.81

23.00
43.93
58.88
71.09
96.38

35.43
57.86
81.25
96.09

40.41
68.50
87.00
100.00

33.75
59.38
76.013
84.89
92.66

 

   

 

90

Table 3: Flushing Runs

Table 3.a: General conditions

Run Number Mold Operating Preform Fiber
number of plies temperature pressure weight volume
(°C) (psi) (g) fraction
1. 1 23 25 42.43 0.05966
2. 1 23 40 42.14 0.05925
3. 1 23 60 41.58 0.05846
4. 1 40 25 40.30 0.05666
5. 1 60 25 37.93 0.05333
6. 1 40 40 46.39 0.06844
7. 1 60 40 40.28 0.05944
8. 3 23 25 99.52 0.1399
9. 3 23 40 107.65 0.1514
10. 3 23 60 117.13 0.1647
11. 3 40 25 113.70 0.1599
12. 3 60 25 120.83 0.1699
13. 3 40 40 125.06 0.1845
14. 3 60 40 112.38 0.1658

Table 3.b: Mold filling rate

Mold Time (sec)

position run #1 run #2 run #3 run #4 run #5 run#6 run#7
cm

0 0 0 0 0 0 0 0
5.08 8 2 2 9 8 3 2
10.16 14 4 4 16 15 6 6
15.24 20 6 6 24 23 9 8
20.32 27 9 9 32 32 12 11
25.40 34 11 12 40 40 15 14
30.48 42 13 14 48 49 l9 17
35.56 49 16 17 56 59 22 20
40.64 57 18 20 65 72 25 23
45.72 65 20 23 73 89 28 27
50.80 74 23 25 81 113 32 ' 30

56.52 84 25 28 92 131 36 34

 

   

 

Mold
position
cm

5.08

10.16
15.24
20.32
25.40
30.48
35.56
40.64
45 .72
50.80
56.52

Resin
flushed
(CC)

22

66

88

110
132
154
176
198
220
242
264
286
308
330
352
374
396
418
440

run #8 run #9

0 0

9 2

19 6

28 10

38 14

52 18

68 23

87 28

110 34

143 41

181 47

226 57

run #1

viscosity % wt.
(23°C) binder
191.5 2.320
195.0 2.429
193.0 2.369
191.0 2.302
193.5 2.383
186.5 2.161
181.5 2.000
180.0 1.947
186.0 2.143
163.0 1.533
172.0 1.675
178.0 1.880
172.0 1.675
170.0 1.600
175.5 1.800
174.0 1.742
168.0 1.532
172.0 1.675
170.0 1.600

91

Table 3.b (continued)

Time (sec)
run #10 run #11 run #12
0 0 0
2 7 9
5 16 18
9 25 25
12 33 35
15 42 45
18 51 58
21 62 72
25 77 88
31 95 110
36 115 129
41 141 150

run#2

viscosity

(23°C)

150.0
152.0
152.5
153.0

% wt.
binder

0.869
0.950

run #13

13
16
21
25
29
35
40
46
54

Table 3.C: Viscosity and binder concentrations
viscosity in cps

run #3
viscosity
(23°C)

161.0
154.5

167.0
162.5

run #14

10
13
15
19
22
26
29
33
37

%Wt.

binder

1.280
1.037
1.500
1.450

 

   

 

Resin
flushed
(CC)

22

66

88

110
132
154
176
198
220
242
264
286
308
330
352
374
396
418
440

22

66

88

110
132
154
176
198
220
242
264
286
308
330
352
374
396
418
440

viscosity
(40°C)

77.0
75.0
71.5
71.5
69.5
68.0
61.0

92

Table 3.C (continued)

run #4
viscosity
(23°C)

206.0
200.0
198.0
191.0
170.0
179.0
176.0
171.0
170.0
154.0
166.0
165.5
162.5
150.0

% wt.
binder

2.865
2.585
2.524
2.302
1.600
1.913
1.810

viscosity
(60°C)

33.5
32.0
28.0
27.0
25.0
24.0

run #5
viscosity
(23°C)

243.0
225.5
206.0
181.5
178.0
168.0
164.0
161.0
155.5
154.5

93

Table 3.c (continued)

 

 

Resin run #8 run #9 run #10
flushed viscosity % wt. viscosity % wt. viscosity % wt
(cc) (23°C) binder (23°C) binder (23°C) binder
22 375.0 6.452 239.0 3.673 222.0 3.221
44 350.0 6.025 248.0 3.900 233.0 3.517
66 320.5 5.479 230.0 3.437 251.0 3.973
88 298.5 5.040 236.0 3.600 238.0 3.647
110 268.5 4.388 221.5 3.200 220.0 3.165
132 241.5 3.737 215.0 3.025 218.0 3.110
154 221.5 3.200 208.0 2.823 209.0 2.852
176 208.5 2.837 200.0 2.584 195.0 2.429
198 197.0 2.500 188.0 2.207 203.0 2.675
220 197.0 2.500 180.5 1.962 191.5 2.320
242 185.0 2.1 11 ---------------------------
264 182.0 2.013 176.5 1.827 179.0 1.913
286 174.0 1.750 -------------- ------- -------
308 161.0 1.280 168.0 1.532 170.0 1.600

330 163.0 . 1.353 -------------- .......

 

 

 

94

Table 3.c (continued)

Resin run #11 run #12_____
flushed viscosity viscosity % wt. viscosity viscosity % wt.
(cc) (40°C) (23°C) binder (60°C) (23°C) binder
22 142.0 406.0 6.950 70.0 410.5 7.010
44 141.0 400.0 6.860 76.0 461.5 7.750
66 124.0 343.0 5.900 65.0 383.5 6.596
88 107.0 299.5 5.063 60.0 346.0 5.955
110 94.0 25 3.0 4.031 46.0 308.0 5.240
132 86.0 225.0 3.316 44.0 248.5 3.915
154 79.0 193.0 2.369 34.0 201.5 2.630
176 77.0 180.0 1.947 32.0 188.0 2.210
198 67.5 168.0 1.533 30.0 173.0 1.710
220 ----- -------------- 28.0 162.0 1.318
242 60.5 152.0 0.950 -------------------
264 ----- -------------- 26.0 157.0 1.133
286 62.5 142.0 0.550 -------------------
308 ----- ------------- 27.0 152.0 0.950
330 61.5 139.0 0.430 -------------------
352 ----- - ------------ 26.0 147.0 0.750
374 60.0 137.0 0.350 ------------------
396 ----- .. ------------ 26.5 147 .0 0.750
418 59.5 135.0 0.270 ------------------
440 ----- ------- ------- 25 .0 145 .0 0.700
run #13 run #14
22 94.0 270.5 3.410 34.0 320.0 5.471
44 98.5 304.5 4.434 36.0 298.5 5 .063
66 105.0 304.0 5.154 30.0 260.0 4.190
88 98.5 282.5 4.700 27.0 240.5 3.713
110 92.0 255.0 4.071 25.0 224.0 3.275
132 87.5 239.0 3.674 24.0 214.0 2.997
154 84.0 221.5 3.208 22.0 200.0 2.5 85
176 77.5 212.0 2.940 20.5 184.0 2.080
198 --------------------------------------
220 72.0 189.5 2.258 17.0 173.0 1.710
242 --------------------------------------
264 63.5 179.0 1.913 16.0 167.0 1.500
286 --------------------------------------
308 59.0 165.0 1.425 16.0 162.0 1.318
330 ----- -------------- ----- --------------
352 61.5 160.0 1.244 ----- --------------
374 ------------------- 17.0 158.0 1.170
396 63.0 155.0 1.078 ----- --------------
418 ----- ---------------------------------

440 61.0 158.0 1.170 14.0 149.0 0.830

 

 

Table 4: Concentration profile in mold

95

Table 4.a: Viscosity at 23°C vs. viscosity at 28°C

viscosity
(28°C)
241.5
232.0
209.5
174.0
165.0
157.0

viscosity

(23°C)
326.5
331.0
270.0
224.5
214.5
206.5

viscosity
(28°C)

152.0
142.5
134.0
122.0
116.0

199.0
184.5
174.5
161.5
152.0

viscosity
(23°C)

V23 = -49.25 + 2.54 x v28 - 9.598 x 10'3 x v282 + 2.348 x 10'5 x v283

run#

OOQQ'JIDUJNn—a

20 psi
preform
length (cm)
3.18
10.80
18.42
26.04
33.66
41.28
48.90
56.52

Table 4.b. Preforms

preform
weight (g)
6.52
22.83
35.85
57.09
63.71
82.68
94.67
123.25

9

10
11
12
13
14
15
16

run#

30 psi
preform
length (cm)

3.18

10.80
18.42
26.04
33.66
41.28
48.90
56.52

preform
weight (g)
5.71
22.93
36.74
52.52
71.86
81.56
106.38
116.67

 

 

 

96

Table 4.c: Viscosity and binder concentrations for flushing runs
viscosity in cps

Resin run #1 run #2
flushed viscosity viscosity % wt. viscosity viscosity % wt.
(cc) (28°C) (23°C) binder (28°C) (23°C) binder
22 124.0 163.0 1.353 134.0 175.5 1.793
44 120.5 158.5 1.188 131.5 172.0 1.672
66 116.5 153.5 1.001 128.0 168.0 1.532
88 ------- ------ ------- 126.5 166.0 1.461
110 117.0 154.0 1.020 125.0 164.0 1.389
132 ------ ------- ----- - 122.0 160.5 1.262
154 -------------------- ------- --------------
176 116.5 153.5 1.001 118.5 156.0 1.095
198 -----------------------------------------
220 --------------------- ------- --------------
242 ------- ------------- 118.0 155.5 1.076
264 115.0 151.5 0.925 ------- ------- --—----
286 ---------------------------- ------- -------
308 1 15 .0 151.5 0.925 116.0 153.0 0.982
330 ------- -------------- —------ ------ ------
352 ----------------------------------------
374 113.0 149.0 0.828 115.0 151.5 0.925
396 ------- - ---------------------------------
418 ----------------------------------- -------
440 113.0 149.0 0.828 115.0 151.5 0.925
run #3 run #4
22 146.5 190.5 2.288 159.0 206.5 2.778
44 142.5 185.5 2.127 164.0 212.5 2.953
66 140.0 182.5 2.028 158.5 205.5 2.749
88 137.0 179.0 1.912 151.0 196.5 2.476
110 129.5 169.5 1.585 142.0 185.0 2.111
132 127.0 166.5 1.479 137.0 179.0 1.912
154 123.5 162.5 1.335 133.0 174.0 1.742
176 123.0 161.5 1.299 125.0 164.0 1.389
198 ------------------------------------------
220 119.5 157.5 1.151 121.5 160.0 1.243
242 ------------------------------------------
264 119.0 156.5 1.114 119.0 156.5 1.114
286 --------------------- ------- --------------
308 116.5 153.5 1.001 118.0 155.5 1.076
330 ------------------------------------------
354 ------------------------------------------
374 115.5 152.5 0.963 116.0 153.0 0.982
396 ------------------------------------------
418 ----------------------------------- -------

440 115.0 151.5 0.925 117.0 154.0 1.020

 

 

 

97

Table 4.c (continued)

Resin run #5 ' run #6

flushed viscosity viscosity % wt. viscosity viscosity % wt.
(cc) (28°C) (23°C) binder (28°C) (23°C) binder
22 164.0 212.5 2.953 170.0 220.5 3.179
44 164.5 213.5 3.041 177.0 230.0 3.437
66 158.0 205.0 2.734 179.0 232.5 3.504
88 152.5 198.0 2.522 162.0 210.0 2.881
110 148.5 193.0 2.367 158.0 205.0 2.734
132 143.0 186.5 2.159 152.0 197.5 2.507
154 136.5 178.5 1.895 140.5 183.5 2.061
176 131.5 172.0 1.672 141.0 184.0 2.079
198 -------------- ------- ------- --------------
220 127.5 167.0 1.497 131.0 171.5 1.655
242 -----------------------------------------
264 125.0 164.0 1.389 123.5 162.5 1.335
286 ------------------------------------------
308 119.5 157.5 1.151 123.0 161.5 1.299
330 ------------------------------------------
352 ------- ------- --------------------- -------
374 119.5 157.5 1.151 118.0 155.5 1.076
396 ------------------------------------------
418 ------------------------------------------
440 117.0 154.0 1.020, 122.0 160.5 1.262

run #7 run #8

22 202.0 265.5 4.319 206.0 272.0 4.468
44 202.0 265.5 4.319 241.5 335.0 5.753
66 194.0 253.5 4.035 232.0 316.5 5.402
88 179.0 232.5 3.504 209.5 277.5 4.591
110 169.0 219.0 3.137 190.0 248.0 3.900
132 158.0 205.0 2.734 ---------------------
154 150.0 195.0 2.429 166.5 216.0 3.053
176 143.0 186.5 2.159 154.0 200.0 2.584
198 ------- ----- ----------------------------
220 133.0 174.0 1.742 140.0 182.5 2.028
242 -------------- ------- ---------------------
264 126.5 166.0 1.461 132.0 173.0 1.707
286 -------------- -—----- ---------------------
308 121.5 160.0 1.243 127.0 166.5 1.479
330 -------------- ----___ ---------------------
352 ------------------------------------------
374 119.0 156.5 1.114 121.0 159.0 1.207
396 ------------------------------------------
418 ----------------------------------- -—-----
440 118.5 156.0 1.095 122.0 160.5 1.262

|P

 

      
 

JJI 141187

.111. uum-:15.-

 

 

98

Table 4.c (continued)

Resin run #9 run #10

flushed viscosity viscosity % wt. viscosity viscosity % wt.
(cc) (28°C) (23°C) binder (28°C) (23°C) binder
22 122.0 160.5 1.262 134.0 175.5 1.793
44 120.0 158.0 1.170 131.0 171.5 1.655
66 120.5 158.5 1.188 129.5 169.5 1.585
88 --------------------- 126.0 165.5 1.450
110 120.0 158.0 1.170 125.5 165.0 1.425
132 ------- ---------------------------------
154 ------------- ------- 118.5 156.0 1.095
176 118.5 156.0 1.095 ---------------------
198 -------------------- 121.0 159.0 1.207
220 ---------------------------------- -------
242 119.0 156.5 1.114 120.0 158.0 1.170
264 ------------------------------------------
286 ------------------------------------------
308 118.0 155.5 1.076 116.5 153.5 1.001
330 ---------------------------- ------- -------
352 ------- - ------------------- ------ -------
374 116.0 153.0 0.982 118.0 155.5 1.076
396 ------------------------------------------
418 ------------------------------------------
440 117.5 155.0 1.078 117.5 155.0 1.078
run #11 run #12
22 136.0 177.5 1.861 142.5 185.5 2.123
44 135.5 177.0 1.850 146.5 190.5 2.288
66 134.0 175.5 1.793 139.5 182.0 2.013
88 131.0 171.5 1.655 136.5 178.5 1.895
110 128.0 168.0 1.532 133.0 174.0 1.742
132 --------------------- 132.5 173.5 1.725
154 124.5 163.5 1.371 130.0 170.5 1.620
176 --------------------- 128.5 168.5 1.545
198 121.5 160.0 1.243 ---------------------
220 --------------------- 125.5 165.0 1.425
242 119.0 156.5 1.114 ---------------------
264 --------------------- 122.5 161.0 1.280
286 --------------------- ------- --------------
308 121.5 160.0 1.243 120.5 158.5 1.188
330 ------------------------------------------
352 ------------------------------------------
374 116.5 153.5 1.001 119.0 156.5 1.114
396 ------------------------------------------
418 --------------------- ------- --------------

440 114.0 150.5 0.880 —————————————————————

 

   

 

Resin

flushed viscosity
(cc) (280C)
22 152.0
44 157.0
66 150.5
88 146.5
110 141.5
132 137.5
154 131.0
176 128.5
198 -------
220 125.5
242 -------
264 123.5
286 -------
308 121.0
330 -------
352 -----
374 119.5
396 -------
418 -------
440 117 5
22 155.5
44 161.0
66 161.5
88 156.5
110 152.5
132 152.0
154 ' 150.5
176 143.5
198 -------
220 146.0
242 -------
264 137.5
286 -------
308 131.5
330 -------
352 -------
374 128.0
396 -------
418 -------
440 122.5

run #13

99

viscosity % wt.

(230C)

197.5
204.0
195.5
190.5
184.5
179.5
171.5
168.5

binder

2.507
2.704
2.445
2.288
2.094
1.929
1.655
1.550

Table 4.c (continued)

viscosity
(280C)

154.0
152.5
150.0
149.0
146.5
143.0
141.0
136.5

run #14
viscosity

(230C)

200.0
198.0
195.0
194.0
190.5
186.5
184.0
178.5

 

   

..'.a '

 

100

Table 5: Growth of finger length and width

Time
(seconds)

0
41
54
58
62
66
69
73
77
80
84
88
96
100
104
109
113
117
121
125

31
39
42
46
50
52
55
57
60
62
65
68
72

20 psi

length (cm width (cm) Flow front
) (cm from inlet)

RTM machine and stopwatch started

resin entered the mold

dyed resin entered the mold

3.6 3.6 17.8
4.4 3.6 21.4
4.9 4.0 23.7
5.4 4.8 26.3
6.0 5.4 28.3
6.5 6.0 30.1
6.8 6.4 32.1
7.1 6.8 34.1
7.5 7.1 36.0
8.3 8.3 40.0
8.7 8.5 41.6
9.1 8.5 43.2
9.5 8.7 44.8
9.9 8.7 46.8
10.3 8.9 48.1
10.7 8.9 49.5
11.3 8.9 50.7

30 psi

RTM machine and stopwatch started
resin entered the mold
dyed resin entered the mold

6.0 7.9 21.3
11.9 11.9 27.5
13.7 11.9 32.6
14.7 11.9 35.4
15.7 11.9 38.2
16.7 11.9 41.4
17.9 11.9 44.7
18.9 11.9 47.5
19.8 11.9 50.1
21.0 11.9 54.1

22.2 11.9 56.5

101

Table 6: Flexural strength data

End Loaded Center loaded___
Sample Flex Flex Flex Flex
# (from strength modulus strength modulus
Figure 7) (x104 psi) (x106 psi) (x104 psi) (x106 psi)
1 2.1514 0.93085 2.5754 1.05864
2 2.5134 0.93745 2.6153 1.00854
3 3.0451 1.17286 2.6045 1.03003
4 2.4438 1.02546 2.9123 1.07707
5 2.4364 0.97632 3.0540 1.11007
6 2.4213 0.97881 2.4815 0.94512
7 2.6157 0.97427 2.8062 1.04301
8 2.3141 0.94609 2.7440 1.03068
9 2.2783 0.93055 2.7788 1.08825
10 2.8807 1.07411 2.6167 1.07419
11 2.5784 1.00662 2.3404 1.04552
12 2.6557 1.03164 2.3985 0.96294
13 2.1478 0.93467 2.7841 1.05410
14 2.2903 1.00771 2.7119 1.01014
15 2.3521 0.84858 2.5668 0.98812
16 2.2111 0.89246 2.8186 1.07831
17 2.1035 0.89331 2.3636 0.97362
18 2.5086 0.93629 2.5483 0.92483
19 2.5717 0.95103 2.9228 1.01439
20 2.2711 0.91843 2.2173 0.86142
21 2.1921 0.92485 2.3845 0.92262
22 2.3363 0.91915 2.2966 0.88422
23 2.2882 0.88998 2.5371 1.02417
24 2.4741 0.90876 2.6998 1.03694
25 2.1742 0.86279 2.8053 1.05265
26 2.7080 0.98252 2.4263 0.99392

 

 

 

 

   

 

B.1 Intrinsic viscosity

The intrinsic viscosity of the Derakane] binder mixture was found using the following
method outlined by Sperling (17):
First find the ratio of the mixture’s viscosity to the solvent viscosity, this is the relative vis-
cosity:
“rel = l81 / 110

“rel = relative viscosity
1!. = solution viscosity
1,10 = Solvent (Derakane) viscosity

Then find the specific viscosity which is just the relative viscosity minus one:

lisp = “rel ' 1
The specific viscosity is then divided by the weight concentration of the binder, which is
extrapolated to zero, yielding the intrinsic viscosity:

[lisp / C] C=0 = [11]

C: weight concentration of binder
[1,1] = intrinsic viscosity

The specific viscosity can then be related to the intrinsic viscosity via the Huggins equa-
tion:
rap/C = [111 - k’ [1112 C - m’ 1103 c2
Where k’ and m’ are constants. The Huggins equation can then be rearranged in terms of
the solution viscosity:
11 = 11., + C 110111] + 161101112 c2 + m’ 1101113 c3

For this equation, k’ was calculated at low viscosity to be 0.242 and m’ was 0.293. The
viscosity values calculated from the Huggins equation was within 5% of the experimental
viscosity up to 8% wt. binder. Table 7 shows the calculations to derive the intrinsic viscos-
ity, Figure 44 shows the resulting graph and Table 8 shows the deviation of the actual vis-

cosity from the viscosity using the Huggins equation.

102

 

 

_,.

 

r. ' "15:1 11‘1“!

 

weight%
binder

0

0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0
7.5
8.0

103

Table 7: Concentration calculations for intrinsic viscosity

concentration viscosity

(g/CC)

0
0.005250
0.01050
0.01575
0.02100
0.02625
0.03150
0.03675
0.04200
0.04725
0.05250
0.05775
0.06300
0.06825
0.07350
0.07875
0.08400

(CPS)

128.6
140.7
153.5
167.1
181.6
197.3
214.1
232.4
252.1
273.4
296.5
321.6
348.7
377.9
409.5
443.6
480.2

relative
viscosity

1

1.0941
1.1936
1.2994
1.4126
1.5342
1.6653
1.8069
1.9603
2.1263
2.3061
2.5008
2.7114
2.9390
3.1847
3.4495
3.7346

- specific

viscosity

0

0.0941
0.1936
0.2994
0.4126
0.5342
0.6653
0.8069
0.9603
1.1263
1.3061
1.5008
1.7114
1.9390
2.1847
2.4495
2.7346

From Figure 38, the regression equation is

W, /c = 17.49 + 76.51 x c + 1222.92 x c2
C = 0, Vist = 17.49 = intrinsic viscosity

at

specific
viscosity/conc.

Table 8: Deviation of experimental viscosity from Huggins viscosity

weight%
binder
0

0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
5.5
6.0
6.5
7.0

7.5
8.0

experimental calculated

viscosity
(CPS)
128.6
140.7
153.5
167.1
181.6
197.3
214.1
232.4
252.1
273.4
296.5
321.6
348.7
377.9
409.5
443.6
480.2

viscosity
(CPS)
128.6
140.7
153.5
167.2
181.9
197.8
215.2
234.1
254.8
277.4
3021
329.1
358.5
390.5
425.4
463.2
504.2

% deviation

0

0.060
0.165
0.253
0.514
0.731
1.071
1.463
1.889
2.332
2.810
3.334
3.883
4.418
4.998

   
    

{OI

Jr? 1:: 10: monarch:

'I. .T'I'I

 

 

104

04.0 00.0

 

:oumbcoocoo .m> :ounwbcoocookzmoomg ocmuomm H34 Emmi

Aoo\0v cozobcmucoo
000 K00 00.0 00.0 40.0 n110.0 N00

00 000
he

1km

0m

B.2 Determination of the mass transfer coefficient

The mass transfer coefficient in rhw bulk phase, 1:, is determined from the flushing
runs using the following equations used by Geankoplis(18) for calculating the mass trans-

fer rate in a packed bed:

NB A = A k gilt-.1185;
1n [(Ci'C1)/(C1'C2)]

NB A=V(C2'C1)
NB = the mass flux of the binder
A = the total surface area of the preform
Ci = the binder concentration at the interface of the resin and fibers
C1, C2 = the binder concentration at points 1 &2 in the mold
V = volumetric flow rate of resin

combining and solving for k gives:

k = meg-£11.11 [Eiilmifizn
A (C, - C1) - (Ci - C2)

To find A, assume that the preform is made of one continuous cylinder so therefore:
A = 2 1: x r x 1

r = the radius of the cylinder (fiber)
1 = the length of the cylinder (fiber)

The volume of the preform is the volume of the mold (281 cc) multiplied by the fiber vol-

ume fraction (V f). Setting this equal to the volume of the cylinder gives:

u r2 x 1 = 281 x vf
Taking the last equation in terms of l and substituting into the surface area equation gives:
A = 562 x Vf/ r

This equation can then be substituted into the k equation to calculate the experimental k
values. The Ci at the interface is taken as the saturation concentration (11.86%). The k val-

ues were founds over the entire length of the mold and C1 is the binder concentration at

105

 

_ . ,tmin'nr::'.=‘.1

    
 

 

 

 

 

BIBLIOGRAPHY

 

BIBLIOGRAPHY

1.Hayward, 1.8., and B. Harris, “The Effect of Process Variables on the Quality of RTM
Mouldings”, SAMPE Journal, vol. 26, N0. 3:39-46, (1990).

2.Hoover-Siegle, Laurie, “Settling into its Niche: Resin Transfer Molding”, flasjies
World, April 1982, pp. 40-4.

3.Pregelhof, RC. and RM. Lien, “Getting the Best from Transfer Molding”, Megem
Plastics, Nov. 1987, pp. 106-110.

4. Rudd, C.D., M.J. Owen, and V. Middleton, “Effects of Process Variables on Cycle Time
During Resin Transfer Moulding for High Volume Manufacture”, Materials Science and

Technelogy, July 1990, p.656-65.

5. Travis, J.E., D.A. Cianelli, and CR. Gore, “The Long and the Short of Fiber Reinforced
Thermoplastics”, Maehine Design, Feb. 12, 1987, pp. 193-8.

6.Hom, S.W., “Advances in Binders for RTM and SRIM Fiber Preforming”, Advanced
m ' M '1' D lmn n 111 nf Pr in,
Detroit, Sept. 30- Oct. 3, 1991, pp.41-7.

7.0wen, M.J., V. Middleton, C.D. Rudd, and ID. Revill, “Fibre Wet-out in High Resin
Transfer Moulding”, nt rf i 1 Ph n m n in m it Material nferen
University of Sheffield, Sept. 5—7, 1989, pp. 208-19.

8. Stalkup, F.I., Wm, Society of Petroleum Engineers, 1983.

9. Slobod, R.L., and RA. Thomas, “Effect of Transverse Diffusion on Fingering in Misci-
ble-Phase Displacement”, Seeiety Qf Petroleum Engineers leumal, Mar. 1963, p. 9-13.

10.Paterson, L., “Fingering with Miscible Fluids in a Hele Shaw Cell”, Bhysieal Fluids,
Jan. 1985, pp. 26-30.

11.Tan, GT, and GM. Homsy, “Stability of Miscible Displacements in Porous Media:
Rectilinear Flow”, Physieal Fluids, Nov. 1986, p. 3549-56.

12. Hickemell, El, and Y.C. Yortsos, “Linear Stability of Miscible Displacement Pro-
cesses in Porous Media in the Absence of Dispersion”, Studies in Applied Mathematics,

107

108

1986, vol. 74, pp. 93-115.

13. Losure, N.S., C.A. Petty, and K. Jayaraman, “Effects of Reactive Filling Time on
Inhomogeneities in Liquid Molding”, presented at the Polymer Processing Regional Meet-
ing, Knoxville, TN, October 1992.

14. Dow Corporation, W, 1992.
15- DOW Corporation. BMW 1991.

16. “Standard Test Methods for Flexural Properties of Unreinforced and Reinforced Plas-

tics and Electrical Insulating Materials”, W3, vol
08.01, sec. D790-l, Pp. 269-77.

17. Sperling, L.H., Intredeetien te Physieel Pelyr_ner Sejenee, 1986, John Wiley & Sons,
New York.

18. Geankoplis, C.J., Transpegt Preeesses and Unit Qperatiens, 1983, Allyn and Bacon,

Boston.

19. Smith, J .M., Chemiegl Engineering Kigeties, 1981, McGraw-Hill, New York.

IF

   

 

 

 

 

 

 

 

 

 

1|11|11111|1111||111|111111I111|1|E1§11|E111|11|111131111111111

 

312938.88