The Reinforcement Mechanism of Fiber-Glass
Reinforced Plastics Under Wet Conditions:
A Review
H . ISHIDAand J . L. KOENIG
Department of Macromolecular Science
Case Western Reserve Unizjersity
Cleveland, Ohio 44106
Reinforcement mechanisms of fiber-glass reinforced plastics
(FRP) under wet conditions are reviewed with emphasis on
molecular structures of glassimatrix interfaces. Included are
studies on glass surface, the glassicoupling agent interface,
silane coupling agents on glass surfaces a s well as in solution,
the coupling agentiniatrix interface, extending to the interphase of particulate-filled composites, and matrix resin.
For a better understanding of wet strength of FRP, the structures under dry conditions are extensively reviewed. The
chemical bonding theory still dominates other reinforcement
theories. The importance of other fictors such as orientation of
silane coupling agents and the restriction of matrix polymer
conformations are also considered. Based on recent development i n spectroscopy, molecular level research of the glass/
matrix interfaces has been initiated in the past decade, yet only
a few spectroscopic investigations on the function of water
have appeared. It is concluded that the correlation between
spectroscopic investigations and mechanical properties of a
FRP is indispensable.
INTRODUCTION
ince fiber-glass reinforced plastics (FRP) appeared
industrially many years ago, fabrication techniques
have been developed yielding high performance materials. The principal limitation for industrial usage is sensitivity to moisture. To elucidate the role of water in
reinforcement, a number of macroscopic studies have
been attempted. A majority have utilized mechanical
tests comparing dry and wet properties of FRP. Generally, studies on the molecular level have been unsuccessful, although many techniques, such as electron
microscopy, cllipsometry, X-ray photoelectron spectroscopy, auger electron spectroscopy, mass spectroscopy, nuclear magnetic resonance spectroscopy, infrared and laser Raman spectroscopy, thin layer
chromatography, contact-angle measurements, sorption and desorption measurements, radioisotope techniques and dielectric constant measurcrnents have h e n
employed. The sensitivity and selectivity of thesc techniques have lieen problematic. There are many theories
elucidating the behavior of FRP based on the results of
the above techniques. Usually, conclusions were derived from indirect evidence, and such conclusions were
often the cause of much confiision. It is the purpose of
this review, therefore, to suinmarizc the widesuread
work from the molecular hasis in order to evaluate thc
reinforcement mechanism of FRP undcr wct conditions.
S
128
Most FRP research has focused on m e of five
categories depending upon the materials and/or int erfaces that w w e heing studied :
1. Glass (surface)
2 . Glass/coupling agent
interface
Glassiina t I-is i 11t e r face
3. Coupling agent
4. Coupling ageiitimatris
resin interface
5. Matrices
Those regions are depicted in Fig. 1 and the types of
honding possible in Fig. 2 a s schcmatic diagrams. It is
generally recognized that the glass/matrix interface is
the determining f k t o r ofthe reinforcement mechanism,
especially wet strength retention. The strength and
modulus changes of the resin are relatively iinimportant
in the water degradation process (I), though little research has appeared on the effect of water on t h t matrix
interphase. Hence, the bulk properties of the glass
fibers and the matrices will lie briefly rcvicwed with the
main interest on the coupling agent, the glnssi coupling
agent intcrfacc, and the coupling agentiinatrix rcsin inte rface.
I
THE ROLE OF GLASS FIBERS
The most widely used glass fiber is “E-glass fiher”
which contains S5 percent SiOz as the main component
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, NO.2
The Reinforcement Mechanism of Fiber-Glass Rein-forced Plastics Under W e t Conditions: A Review
Table 1. A Comparison Between Auger and ESCA Analysis of
the Components of €-Glass Fiber
A.
Elements
J
Si
Al
Mg
Ca
B
F
0
6.
I GLASS I
u
I
I
C.
:
V
I
; i G W S - S I L A N E IMERFACE
SILANE
1
Fig. 1 . Schematic diagranis of glussjiber reinforced plastics
showing a composite structure.
A . H-BOND
B . CHEMICAL
BOND
R
I
-0-si-oI
R
0
.**
H
\
I
\
*
-
H
0.I
si
-0-
si-oI
0
I
si
-
F I B E R GLASS
Fig.2. Siinplijied molecular mode1.s of the gla.w/coupling ageitt
interfuce.
with the remainder being oxides of other metals such as
A, Mg, Ca, etc. The filjer surfaces have distinct properties from the bulk. After contact with water, the surface
oxides are hydrated and form metal hydroxide groups
which are considered to be adsorption sites of coupling
agent molecules. The compositions of metal oxides near
the surfice are much different from the hulk (2, 164). A
comparison of surface compositions obtained from auger
electron spectroscopy (AES) and X-ray photoelectron
spectroscopy for chemical analysis (ESCA) was made as
shown in Table 1 . Although these two techniques do not
necessarily give the same results due possibly to the
different sampling depth and varying sensitivities to
different elements, specific feature of the fiber surface
compositions is well depicted. These nonsilica components exist as microheterogeneities with dimensions es-
Percent atomic composition
Fiber surface
Bulk
Auger
ESCA
18.6
6.1
2.2
6.3
4.1
0.4
61.8
24.1
8.4
0.7
1.8
3.0
1.8
61.1
29.6
5.0
1.8
3.4
0.0
0.4
59.9
timated to be from 15 to 200 8, (3).The environment of
the glass fiber affects these nonsilica portions. Diffusion
of Ca to the surface at various temperatures and leaching
of A1 by acid as a function of p H were studied by ESCA
(4). The effect of heat cleaning was also studied by
Adams e t al. (5).The strength loss after heat cleaning of
glass fabrics was attributed to exchange of sodium ions,
resulting in weakened tensile strength at the fiber surface.
Although the detailed function of these nonsilicate
components on the adsorption of the coupling agent is
not yet clear, some attempts have been made utilizing a
titration m e t h o d . A chromium complex agent
chemisorbs onto every negatively charged site on the
glass surface in the p H range 3-6. Therefore, Volan@
adsorbs more on the strongly acidic boro- and aluminosilicate sites than on the silanol sites (6). Johanson and
others (7) reported fewer acid sites in porous silica than
E-glass by a factor of one-fifth for (n-C4H1))4NOH
and by
one-half for (CH3),NOH as titrants. This effect may be
due to the limited availability of the silica surface. Furthermore, the incomplete coverage may be due to the
large molecular size of the titrants, which imply that a
large coupling agent molecule may not cover all the
available surface on active sites. The same result was
obtained by Eakins (8). Hence, the smaller water
molecule may attack the glass surface at sites incompletely covered by the silane coupling agent molecules,
though there are some indications that the coupling
agent layer does not act as a water repellent film.
Water molecules cause a hydrolysis reaction of the
siloxane linkages at the glass surface and produce silicic
acid hydroxyls. Berstein and Nikitin extensively studied
this phenomenon using infrared spectroscopy (9-11).
Attenuated total internal reflection spectroscopy (ATR)
with a specially designed ATR plate made of silica glass
was utilized, and spectra obtained at various moisture
treatments are shown in Fig. 3 . Polished glass shows
only a weak broad band around 3200-3400 cm-‘ due to
the liquid water condensed on the defects and capillaries
at the surface. Water boiling for one hour indicates
appearance of some weak feature at 3650-3550 cm-*,
which is strongly enhanced b y the water vapor treatment at 200°C. This band is due to the hydrogen bonded
surface hydroxyls and water interacting with surface
hydroxyls. In addition, a relatively sharp peak is seen at
3710 c n - l which is assigned to relatively “free” hydroxyls. Etching by H F solution eliminates surface hy-
POLYMER ENGlNEERlNG AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
129
H . Ishida a n d ] . L. Koenig
B
A
100
80
60
E
4 0
3900 3 0 0
3100
3900 3500
3100 cm-1
Fig. 3. Znfrared spectra of citreous silicu obtained by ATR
method. a: polished .sample, b: water hoiling,for 1 h , c: water
r;apor treatment at 200”C,d : etched by H F solution, und e: heut
treatment at 500°C.
droxyls while heat treatment at 500°C drives off the
adsorbed water molecules as can be seen in Fig. 3 . They
clearly demonstrated that vitreous silica, considered to
be a water resistant glass, can be attacked by water at
moderate temperatures (100-200°C). There are other
reports which assert that water seriously degrades glass
fibers (12-14).
It is important to consider the hydroscopic nature of
the glass surface since collection of water molecules at
the interface enhances hydrolysis. Studies have attempted to determine the amount of water adsorbed by
various kinds of glass as a function of relative humidity
(15-19). E-glass fibers adsorb more water per unit area at
100°C than porous silica at 30°C (20).
It is evident that surface silanols arc a major factor in
determining molecular adsorption onto a glass surface.
The surface silanols exhibit various structures with different environments. Koenig and others (52) have demonstrated that the structural differences of the surface
silanol can be detected by utilizing laser Raman spectroscopy. An example of Raman spectra of glass surface
at different conditions is shown in Fig. 4 . There are
several lines around 1050-950 cmpl region due to the
SiO stretching mode of the surface silanols. Glass microspheres display two lines at 980cm-’ for a wet sample
and at 1005 cm-’ for a dry sample. These two frequencies were related to the silanols hydrogen bonded to the
1
1400
1200
I000
800 cm” 600
Fig. 4 . Rumurr .spectra of gluss rrricro.c.pheres under dry uric1 rcct
conditions.
130
adsorbed water molecules. There is an additional line at
992 cm-’ for the glass microspheres. This line is responsible for the silanols hydrogen bonded to the adjacent
silanols.
The catalytic effect of the glass surface on chemical
reactions is well known (21), and must be considered
during the reaction and curing of composite systems.
Fratzsher observed that the higher the concentration of
fumed silica in the resin, the slower was the rise in
temperature due to cobalt adsorption by the glass surface (22). Infrared spectroscopy demonstrated the preferential adsorption of an amine catalyst in epoxy resin by
glass fibers. In conjunction with this phenomenon, it is
interesting to consider the extent of the influence of
filler surface. We shall discuss this problem in the cobpling agentlmatrix interface section.
THE GLASS/COUPLING AGENT INTERFACE
The existence of covalent bonds at the glass/coupling
agent interface has been questioned for over thirty
years. Its existence has been disclaimed (23-27), but a
great majority of researchers assume its existence (6-9,
28-58). This work has already been reviewed elsewhere
(59-63). We will review the results as related to water
susceptibility. Although a few reviews have appeared in
the literature on the function of water, these discussions
were mainly based on the mechanical properties in contrast with our emphasis on the molecular vffects.
The predominant thinking is that the weakest portion
of the FRP system is the glass/coupling agent interface
(35,43, 64). This interface is especially weak when water
is present. In the dry state, mechanical rupture may
occur either at the SiO bond (the glass coupling agent or
the coupling agent), the S i c bond (the coupling agent)
or the C C bond (the coupling agent, the coupling
agent/matrix interface or the matrix itsel0 depending
upon the performance of the FRP. Scission occurs predominantly at the S i c bond when induced thermally.
This is understandable, since the bond energy ofthe SiO
group is 89.3 kcal/mole compared with 57.6 kcal/mole
for the S i c bond (6,5).When water is present, however,
the situation is quite different. The activation energy of
hydrolysis of siloxane was reported to be 22.8 kcal/mole
when reacted with a KOH solution and 5 kcal/mole with
benzoic acid (66). Similar results were obtained by
Matellock (67), who reported a value of 6 kcal/mole for
benzoic acid and 23.6 kcal/mole for water.
Thus, water hydrolyzes the siloxane bond relatively
easily once water reaches the interface, but a question
arises whether water can penetrate to the glass surface.
A water repellent film may be formed (56). Furthermore, good wet strength retention of silane coupling
agent-treated F R P has been attributed to the hydrophobicity of the siloxane bonds formed by the coupling agent (68, 69). This is contrary to the statement by
Bascom: “. . . silanes did not significantly improve resin
wettability nor did they impart an unusually hydrophobic character to the fiber. However, silane finishes
are effective against molecular water penetration by
diffusion along the glass resin interface” (70). Also, the
results obtained by Laird and Nelson (36) revealed that
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
The Reiiiforcenient Mechanism of Fiber-Glass Reinforced Plastics Under W e t Conditions: A Recieu
coupling agents are mainly effective for bond permanerice and not for mechanical strength or wetting.
Water has been reported at the glasshesin interface
but the mechanism of water penetration into FRP is
questioned. McNeil and others (71) propose three
mechanisms: transport ( i ) through resin (nearly
pressure independent up to 10,000 psi), (ii) through
flaws at the surface of FRP, and (iii) along the fiberhesin
interface. The rate of penetration along the fiber is generally 100-400 times faster than through the resin. Dielectric measurements using glass laminates (no edge
exposed) showed that the penetration through the resin
is very slow (72). The penetrating water is believed to
cause a hydrolysis reaction of the siloxanes and a debonding at the glasshesin intt'r fiace.
Plueddemann (41, 42, 73) suggests that the reformable nature of silanol is understood as a dynamic equilibrium of the coupling agent competing with small penetrant molecules, e.g., water molecules. This mechanism
is confirmed by both mechanical property studies (1,74)
and a laser Raman study (49). Scola (1) demonstrated
reversibility of the mechanical strength of FRP after
boiling water (or methanol) followed by evacuation at
elevated temperature. Koenig and Shih (49), utilizing
laser Raman spectroscopy, observed a reversibility in
the intensity of the SiOSi symmetric stretching line
which appears at 788 cm-' for the honded coupling
agent. The system used was E-glass fibers treated with 2
wt percent vinyltriethoxysilane aqueous solution and
dried in air at an elevated temperature. After boiling the
fibers in water for two h, this line shifted to 783 cm-'
which agrees with t h e position of vinylsilane
homopolymer. Similar results were obtained when the
samples were placed in a humidity chamber at 38°C with
relative humfdity of 100 percent for four months. Subsequent heat treatment at 110°C for three days reproduced the original line at 788 cm-' indicating reversibility of the siloxane bond with hydrolysis.
Proper silane treatment of glass fibers improves the
dry and wet strength as seen in Table 2. However, the
optimum number of chemical bonds at the interface
needed to give a lasting strength is not yet known.
Monolayer coverage by the coupling agent usually does
not give optimum performance. Flexural and compressive strength of E-glass fiber laminates under dry and
wet conditions were measured by Johanson et d . (7)
with respect to the concentration of a silane treating
solution. Their data is plotted in Fig. 5. It is clear from
Table 2. Polyester Laminate Strength Data with Various Silane
Treatment*
Flexural
strength, psi
DrY
Wet
112 (heat clean)
Ethyltrichlorosilane
Vinyltrichlorosilane
Propyltrichlorosilane
Allyltrichlorosilane
From reference 117
56,000
34,500
72,000
34,500
57,800
34,800
26.00
59,000
26,600
58,400
Retention,
percent
62
77
82
77
101
A
B
.--
0
-
0
Compressive
0
Dry
0 Wet
0
0
1,OO
2.00
3,OO
Concentration
4.00
(wt
5.00
%I
Fig. 5 . Flexural und compressilje strength of gluss-jiber reinforced po/!jestcr with respect to the concentrations of silune
freuting solution ( u n aqueous .solution of y-rnethucryloxy~irop,yltrirnetlioxysilune).
Fig. 5 that the strength of the laminates increased drastically within a very small concentration range from zero
to 0.01 wt percent. According to their study using a
radioisotope labelled silane coupling agent, the lowest
concentration of the silane solution studied (0.01 wt
percent) was sufficient to yield a monolayer of the silane
monomer on the E-glass fiber surface. Nonetheless,
monolayer silane coverage did not yield the highest
strength but a maximum appeared at higher concentrations for both flexural and compressive strength. This
phenomenon may be attributed to an imperfect availability of the coupling agent due to the surface geometry
or other foreign materials. It is interesting to notice that
the optimum concentration for the two types of measurement is dfierent. Chamberlain et al. present a different view (80). A Grignard reagent having the same
organofunctionality as the silane coupling agent was
chemically reacted with the fluorinated glass surface.
The mechanical performance of epoxy laminates was
compared with either a modification of 5 percent of the
glass surface by a Grignard reagent or a silane coupling
agent applied from aqueous solution. A higher extent of
Grignard modification lead to inferior mechanical performance. This fact suggests that a small number of
covalent bonds are sufficient to give good mechanical
performance. A similar result was obtained by Ahagon
(81)using polybutadiene rubber bonded to glass. Glass
plates treated with 1wt percent of either ethylsilane or a
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
131
H . lshida a n d ] . L. Koenig
50:50 mixture of vinyl functional silane/ethyl functional
silane were coated with crosslinked polybutadiene. The
peel strength of the silane-mixture-treated sample
under a methanol environment gave nearly the same
results as the dry condition. The ethyl functional silanetreated sample gave scattered values ranging from the
dry strength value to the theoretically predicted lower
limit of Van der Waals adhesion under the presence of
methanol. The scattered values were explained by a
small amount of chemical reaction by hydrogen extraction between the ethyl group and polybutadiene. This
result implies that even a small number of covalent
bonds prevents liquids from deteriorating the joint
under wet conditions (see Figs. 6 and 7).
It is desirable to measure quantitatively the number
of chemical bonds on the glass fiber surface in order to
elucidate the role of covalent bonds in the strength of
FRP. At the present time, an appropriate technique is
not available. Nonetheless, the recent advances in infrared and laser Raman spectroscopy provide a potential
for these measurements. Choosing a proper model system, it is now possible to study the interfaces quantitatively. Ishida and Koenig (78) applied Fourier transform
infrared spectroscopy (FT-IR) to the study of the coupling agenvhigh surface area silica interface. Full range
infrared spectrum was disclosed by digitally subtracting
the absorbance contribution of the silica as shown in Fig.
8. A peak at 893 cm-' was attributed to the residual
SiOH groups of the coupling agent employed. The
number of the residual SiOH group per chemical repeat
4
20°C
0-20"C
0-40"C
0
n
E
0
\
0
D*
0'
If:. -'-
2O0C
0-200C
0-4 O°C
Q
3 n
-
E
---_
---_
*_---
Dry
0
J
- I4
I
I
-I 0
-I 2
I
I
-8
L o g Ro, ( c r n l r n i n )
Fig. 7 . W e t and dry peel strength of crosslinked polybutudiene
rubber coated on a glass plate. T h e glass surface was treated
w i t h only an ethyl functional silane.
unit was calculated from the measured intensity of the
893 cm-' band. This value for the fully dried silica
sample was almost zero within the experimental error.
Since the amount of the coupling agent was .measured to
be less than a monolayer and it was predominantly randomly distributed, they concluded that covalent bonds
exist at the coupling agent/silica interface. Qualitative
evidence which is shown in Fig. 9 also siupports this
conclusion. A silane oligomer treated porous silica
showed a disappearance of the surface SiOH groups (970
Y
3
0.;;
----Dry
09'
3
f'
h
+
?/'
,
,
\
M
Q)
,d
,
-02
1. CAD-0-S1L
J
,*
0
1042
J
0
C. K)(YVINYLSILOXANI
ON
CM-0-SIL
I
-16
-14
-12
-10 -8
Log Ra, ( c m i m i n )
-6
Fig. 6. Wet and dry peel strength of crosslinked polybutadiene
rubber coated on a glass plutc at various peeling rates and
temperatures. T h e g1a.s~.surface was treated with a t h y 1 und
ethyl functional silane mixture.
132
1400
600
Fig. 8. Fourier transform infrared spectrcr of high surface areo
1800
1000
silica ~Cab-O-Sil)lcinylfiInctioncllsilane s!ystcm. A: Cab-0-Si!
treated tcith 1 % b!y w e i g h t ciqueous s o l u t i o n of cin!yltriethoxysilone. B: Cab-0-Sil heut cleaned ut I 5 0 - C f o r srwrnl
hours, and C : Polycinylsiloxnne on gl11,s.s .yurfuce rcitli the contrihution of the glass excluded.
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
The Reinforcement Mechanism of Fiber-Glass Reinforced Plastics Under W e t Conditions: A Review
ON
A.
CAB-0-SIL
B€#)R€
H U l lR€AlMENl
j
'
~.
.... .........- . ....
c.
DIR€RENCC
~
0-A
~
D
893
1800
1400
lo00
6OOcrn-1
Fig. 9. Fourier transform infrured spectra of high surface areu
silica (Cab-0-Sil~lpolyoinylsiloxaneo l i g o m e r s y s t e m . A:
Cab-0-Sil coated w i t h polyuinylsiloxane oligomer before heat
treatment, B: Sample A wus heated at 150°C f o r 3 0 m i n , und C:
Difference spectrum before and after the heat treutment.
cm-l) and residual SiOH groups (893 cm-l) of the coupling agent after a thermal treatment. In addition, new
bands appeared in the SiOSi group frequency region
(1170 and 1080 cm-l). These new bands did not agree
with either the coupling agent or the silica in frequency
and, therefore, they were assigned to the bands due to
the SiOSi linkages at the coupling agent/silica interface.
Although laser Raman spectroscopy was successfully
applied to the glass fiber/coupling agent system, the
Raman effect of the inorganic part o f the coupling agent
is undesirably weak; however, the weakness of interference by the bulk glass is advantageous. Therefore,
a monolayer of the coupling agent is very difficult to
study. On the other hand, resonance Raman spectroscopy may be used because of its unusual enhancement of
the peak intensity. Ultraviolet resonance Raman spectroscopy is of particular interest since many of the chemical species exhibit electronic transition in the UV region, which is a requirement of the resonance Raman
spectroscopy, while a few synthetic polymers have electronic transitions in the visible region. The specific absorptivity of silanes was studied by Schreiber (84) and
the SIC group was reported to he resonance Raman
active in the UV region studied. The SiOSi group, the
important group for the structural analysis of the coupling agent layer, was unfortunately resonance Raman
inactive in the region studied.
X-ray photoelectron spectroscopy (ESCA) is capable
of studying sinall amounts of surface species (85). This
method is, however, insensitive to the intramolecular
effects beyond nearest neighbor and is of limited utility
for this area.
FRP strength (86). The coupling agent should be
applied either as triols or possibly silanol rich small
oligomers. Condensed siloxanes were reported to be no
more useful as coupling agents (87). Silanetriols readily
polymerize in aqueous solution (in which the triols are
isolated). E v e n freshly hydrolyzed vinyltrimethoxysilane contains 82 percent monomer, 15 percent dimer, and 3 percent trimer. The composition
changes to 34 percent monomer, 23 percent dimer, 30
percent
trimer
and~ 13 percent
l
s
~ tetramer when the precipitation is initiated (88). The measurement of silanol
groups terminated by trimethylsylil groups was achieved
by liquid-gas chromatography.
On the other hand, silanetriols can be readily studied
by laser Raman spectroscopy (89). Vinylsilanetriol and
methacryloxypropylsilanetriol give rise to the SiO symmetric stretching mode at 672 and 710 cm-', respectively (as 2 wt percent aqueous solutions). The hydrolysis reactions of vinyltriethoxysilane and y-methacryloxypropyltrimethoxysilane were followed using the
above lines as shown in Fig. 10. The concentration of the
triols rapidly reached a maximum value. Then it started
decreasing slowly with time. A similar result was obtained by gas chromatography (90). Figures 11 and 12
2
I
VS-!I,O
-I
CY
CY-'
1410 CM-'
710cu-'
THE ROLE OF SILANE COUPLING AGENTS
Silane coupling agents are usually applied to the glass
surface from aqueous solution. The concentration employed depends upon the nature of the coupling agent.
Ordinarily 0.025 to 2 percent hy weight solution of
coupling agcnt deposit inorc than a monolayer. Most of
the silanes are used as ti-ialkoxysilanes. Differences in
the alkoxy group did not lead to marked changes in the
0
30
100
2 00
a00
4 00
TIME,wNs.
Fig. 10. Norkulized Ramnn intensities of the SiO symmetric
stretching mode of silnnetriols derioed f r o m uinyltriethoxysilune ( u p p e r curces) und y-methucryloxytrimethoxysilane
(lower c u r u s ) . Dotted lines represent the intensity change due
to the hydro1ysi.s and condensation of t h e silanes. Solid lines
.show the intertsity change of the organic fiinctional groups.
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
133
H . lshida a n d ] . L. Koenig
a) initial Mixture
t
w
“1
Aicohol and wotcr
/
0
C
-
t
0
0
- -
Time
b) Portly Reacted Mixture
‘I
Hydrolyzed, monomeric
c
C
Q,
0
5
I *
W
Time
Fig. 11. Gus chromatogram.s of a: u silunelulcoho1lu;uter mixture hefore hydrolysis und 11: the mixture ufter hydrolysis.
95
ucid
90 -
h’
b
Q,
c
h
80Q,
L
*
60-
Pny‘
cr
0
instability of the silanetriols. Utilizing the combination
of laser Raman and Fourier transform infrared spectroscopy, Ishida reported first complete assignments of vibrational modes of vinylsilanetriol in aqueous solution
(83), and phenylsilanetriol in aqueous solution and crystalline phenylsilanetriol (83). This combination is particularly advantageous since, in the case of vinyl functional silane, only the trio1 showed an intense Raman
line at 672 cm-I and intense infrared doublet at 927 and
848 cm-’. When the vinyl silanetriol was polymerized,
no lines with strong intensity were observed in the
Raman spectrum while a strong single band appeared at
893 cm-’ in the infrared spectrum. As aresult. structural
changes of the adsorbed silanetriols could be followed
(78, 91).
Unlike short-chain alkyl substituted silane coupling
agents, the silanes with the polar groups attached to the
y-carbon have distinct properties. Amino functional
silanes, especially, decompose almost instantly in water
and form very clear solutions over long periods of time
even at high concentrations. However, such solutions
do not contain triols. Raman spectroscopy revealed that
even freshly prepared 10 percent hy weight y-aminopropyltriethyoxysilane aqueous solution s,howed no
noticeable SiO symmetric stretching mode indicating
almost no triols (119).It probably forms low molecular
weight cyclic structures with the amine groups protruding. In order to explain this stability, Plueddemann (88)
proposed low molecular weight siloxanes either as having a six membered ring in which the nitrogen atom
forms a zwitterion with one of the unreacted silanols, or
a five membered ring in which the nitrogen atom is
interacting with the silicone atom (see Fig. 1 3 ) .
When a large amount of silane containing a long-chain
organic group was deposited on the glass surface, a waxy
appearance resulted even after curing (92). Wright and
Semlyen (93) found that longer-chain organosilanes
tended to yield more cyclic oligomer than thosr having
short chains. Perhaps, the wax-likc appearance is caused
either by cyclic oligomer formation or h y cyclic oligomer
deposition on the surface. Our infrared data indicate
that E-glass fibers treated with 0.5 wt percrmt y-methacryloxypropyltrimethoxysilane aqueous solution and
dried at rooin temperature for about a wet.k still contained unreacted silanols but the amount was very small
(91).
Time, hours
show gas chromatogram and kinetic curves, respectively. Since the triols were very unstable, especially at
the temperature employed, the peak due to unhydrolyzed silane was used to trace the extent of hydrolysis.
The silane hydrolysis rate was shown to be first order in
silane concentration. A s cited above, vibrational spectroscopy is very useful in studying silanetriols and the
condensation products. In spite of its usefulness, few
attempts to assign vilmtional modes of organosilanetriols had been made until recently due primarily to the
134
There is a correlation between the strength, FRP and
the pEi of the silane treating solution. This effect is
especially large when an ionic silane is used. Electrokinetic effects at the surfice were suggested to explain this phenomenon (94). Wet strength of FRP with
respect to the pH of the silane treating solution is shown
in Fig. 1 4 . The elcctrokinetic effect, rather than simple
-0
\
/
\
H2C-CY2
CH,
S i
\goN<
H
2
or
0,
-0-Si-
o/
/
H2
,C-CH2
I
\+
N-CH2
R2
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, VoI. 18, No. 2
Tlw Re i 11f o rc e in e 11t 2lec h
iTiz 5 iri
of F i 1x3 1-42
r i 5 5 Rcz T if0 I-ccd Pla 5 t i
5
UtI rle t W<>
t C o t i dz t 1 o ii s A R L'Lze i~
Table 3. Contact Angle Measurements of Silane Treated Glass Plates
Wetting liquid
Water
Formarnide
Thiodiglycol
Methyleneiodide
u-Chloronapthalene
WBromonap h t helene
Surface
tension,
CH2=CHSi(OEt); adsorbed from
Cyclohexane
MEK*'
Propanol
dynelcm
H,,cl,
6,<
0 ,ti\
H,,,
72.8
58.2
54.0
50.8
42.9
44.6
86
78
72
54
38
40
0
40
50
20
34
5
24
40
25
0
0
0
34
10
@id\
or,,
43
27
32
38
20
37
0
0
31
0
NH,(CH,),Si(OEt)l
adsorbed from
Cyclohexane
Water
@"I,
Or, I
O'Kl\
53
20
36
39
0
0
0
30
38
10
28
0
0
0
19
26
14
14
14
10
Ht,,
' Et = CHICH,.
''
MEK
=
Methyl ethyl ketone
electrostatic force Iietwc.en silaiic-modified plustics and
mineral siirf;lcrs, ilia!. be important in controlling tlic
orientation of t h e coupling agent. The optiinum 1111 of
the treating solution ilia!' be predicted for siiiiplc oxide
siirfiices from t h e kiiown isoelrctric point of the srii-face.
The o p t i m u m conditions for g l a s s filwrs intist lw
e\amiiied e~I"rii~ieiit"lly duc' to t h e coiiiplesity of tile
composition and t h e thermal history (93).
A kno\\.leclgt. of tlic oriclntation o f t h e silane iiiolccules
is indispeiisAle when we,t sti-tmgtli is considered. -11tliougli detailed study on tlie relation of thc siloxamc
linkagrs lwt\\reeii acljacent silanrl inolcciilrs to \vet
strength h;is not I ~ e e nm&.
it ma!- I)c, one of the ke!,
fkt o r s in the \\Y>t s t r e i i gt li re te i i t ion ,
Bascoin's approwli (114) to thc sti-uctural :inal!.sis o f
silanes tising many techniques iiicludiiig contact aiiglc
iiiwsiireiiieiits. iiifriirrd spectroscopy. scanning e l ( v
troll iiiicroscopy, m c l c~llilxoiiietr!~
addcd iisc&l kiiow+
edge to this area, The contact u i g k iiie:i\rireiiieiits
h \ f x d ;i large hystc'resis m c l soiiie of the data are slio\\.ii
in Tnh/v 3 , The ad\.aiiciiig angles ;ire iii~iclilarger than
t h c, r w etli i i g iiii g1e s , TIi i s ol )ser\.ati on is significant 1)ccatis? it inc1ic;itt.s that sol\,c)nts tmil!. poncti.atc~silane
coiipliiig iigcnt 1ayt.r~.Thcrcfore, he eoiic~~idc.d
that a i i
open striicttirc exists in tlit, \ icinit!- o f tlitl glass srirf:icc~.
NONIONIC SILANE
CONTROL
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, VoI. 18, No. 2
135
H . Ishido m i d ] . L. Koenig
L).
9 0-
e 0N
70-
0
;c o -
gX
50-
oio
0
oio
oio
oio
050
SOLUTION COW~CNTRRITION. IT
-
1000
u)
%
0
c
c
>
g
c
X
060
070
SIUNE
VTMS I E-GLASS F IOER
800
r
.-
a
600
_.I
E
u)
c
3
0
L
400
200
I0
o
a2
a4
46
0.8
1.0
1.2
1.4
14
CONCENTRATION
-4-J-
1.8
20
WTX
4000 3000 2000 1800 1600
c m -I
These ol)ser\xtions qqxtreritly suggest the esistencr. of
tnultilayers, Miiltilayer adsorption w i s quantitativvly
detnonstrated b y iising radioisotope 1al)cllcd silanc. coupling agents I,y Johanson and others (7).Figitre 17 reproduces their results. Under ii proper assumption of
the area occiipicd h y thc coupling agent molcciile, o n e
can ciilciilate monolayer cqiiivalents, thougli calculated
value has a ccrtain uncc,rtainty I,ecausc~c,sact structure
of the adsorlwd niolecrilc is not yet kiiowii. Figirre 17
re\-ealed that the iimount of the coupling agents depends
on tlw concc,ntrations of the silane treating solutions.
Their restilt was also recorifirmed by FT-IH (91) a s
show11 in Fig. I N . This o1)scrvation is \ ery important
liecause we are iio\v dealing with the interphase its well
;IS the interfiice. LVater attack to the interface, might as
well estcnd to thc interphase. It should he pointr.d out
that the term “intt,rfke” has l w c n rised synonymously to
the term “interpliitsc”.”~ 2 ~ h ethe
n interfacial prolwrties
are under considcration, this might introduce s o m r ’ con136
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol 18, No. 2
The Reinforceivienf Jifechariisiri .f Fiber-Glas, Reiriforced Plustics Under Wet Conditions: A Review
IMIVINYLSILOYANI
ON €-GLASS F l 1 l
A.
I W I ~w r s ~ n u l ~ l w
WlID FOR 24 HIS A1 1.1.
IVACUAlID PQIOW.
A 1 110' C
B
.
Y
?
6 -
-
v,
3
I-GLASS F l l i n
z
v,
4 -
a
C.A-a,
lox
2 -
1800
1400
1000
600
3
c
Cm.1
. , , ~ , . ~ . , . , , , . / . . , , , , , , , / , , . , , , . .
100
200
300
4M)
SO0
600
0
Fig. 19. Fourier t r a m f o r m infrared spectra of E-glass f i b e r /
coupling ugent system. A : E-glossfiher treuted with 1 percent
by a e i g h t ciiiyltriethos!/silNneund dried f o r 24 h at room tentperuture w i f h .siiliseyuent cwucucition f o r 30 miit u t 100°C, B:
E-gla.us filler heut clcuned (it .500"C f o r 24 la, C : Polyninylsiloxcinc, on the E-glass fiher with the contribution o f t h e g1u.n
esclrided.
6
b
BOILING WATER EXTRACTION TIME. MIN
Fig. 2 1 . Rudiocictioe uniinopropyl functionul silune remuining
silica.put (geometric surfuce ureu 3 cm') during hoiling inuter
extructioti which .folloreed preliminury rittses in henzene und
cold w u ter
011
R o d ~ o o c f t r ~on
t y surfore
(rou*ii!m~?
1
70.030
,
(:oun-rmr
835.
~
,
6C.000
50.000
4 0.090
3C.393
20,000
/
2
0
Fig. 22. A m o u n t of rcidiouctice nmino))rop!/lfiinctionu/ . d u n e
reinciitting on poli.shed P!yrex surfuce after extrriction with zcuf e r .(i:Estrciction ut25"C cindb:Extrclctio?t continuedclt1OO"C.
lic molecule formation. This result is
results derived fi-om contact angle
liic~Lsllrcll~cnt
I,y Hasco111 (1 14).
It should l)c notcd that iu spite of tlie largc. amouiit of
tlie coupling agent on the siirfiice, not all reactive sites
a r e co\xwd. \\'lien the coupliiig agent \vas extracted by
I)oiliiig \vattlr, tlie rate of dcsorption was the largest
initially I)rit decreased i y i d l y . A monolayer eqiiivaleiit
i . c . n i u i i i r d on the siir1;ice (40, 60). According to
Sc1ir;ider. c k s o i y t i o n regions ;wise from physisorl)ed,
cl~t~iiiisoi.l)(,(l.
aird cliemicall!. reacted coupling agent,
rv3p(.cti\ el!.. Sterinaii and Bradley also reportcd the
tlifficiilt~.of coniplctely rc,nio\-iiig the coiipling a p i t
f r o i i i tlit) gla5s s r i r f a c . ~ (76). E y a i i i p l e s of t h c s e
i-atlioi\otopo stiic1it.s are slio\\w i n Fig:.$.21 and 22.
Sci-c~rthelrw.t h e w are s o m e indications that the rc.action Iwt\~-ceii
tliv coupling agent :inti the glass surf:~cc~
w e sIo\vcr than I)rt\vec,n the coupling agent itself. I,c.c.
( 7 7 )proposed ;I pol>mc.rization nicclranism of thcx c o ~ i -
pling agent on the glass surface shown in Fig. 23. Infrared evidence favors his postulation though definitive
information could not be obtained (78).
A radioisotope study by Schrader and Block (95),
using a low-specific-"ctivity aminopropylsilane (APS),
determined the amount of the silane on the glass surface
after the fiilui-e of Pyrex/coupling agent/epoxy resin
adhesive joint under wet conditions. Throughout the
entire range of surface coceragc (from one to six
monomers per 100 .Ae), halfof the silane remained on the
glass surface. Their result is shown in Fig. 2 4 . They
speculated that the hydrolytic failure occurred in the
coupling agcnt structure hecause otherwise a vertically
zigzag tearing mechanism would result in wide vai-iations from sample to sample. Based on this result, and
tliv result that half the coupling agent remains after joint
fiilure, they posttilated a particular dimer adsorption
mechanism. It should be noted, however, that the
thickness ofthe coupling agent is negligible compared to
the joint a r c x The zigzag traring could occur on the
molecular level which would statistically average the
resulting amount of the coupling agent on the glass
surf:icr. If all the organic groups of tlie coupling agent
: ~ r cnot rcxacted with the resin, tliis statistical averaging
corild also occur for a sample having less than a
Ill onola!~el-.
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
137
c
I
R'O-St-OR'
HO-:t-OH
oti
h
TRIALKOXYSILINES
CHIMI-
uyp74h
STABLE SILAWETRIOLS
?
'/,
S, ,/GLASS
,,,, SURFACE
,,,,
/ /
YIDRMIhDWDLD
I
9
TOTAL RATING
HO-51-OH
lDVPlYG
LVAPWIAlIOh I
I
0
I80
I60
0
a
3-
-
glass f i l ) c ~ s The
.
sprtctra arc shown in Fig. 26 indicating
t h e pi-efei-entinl adsorption of tht. aininc c~atalvst.The
r e s t r a i n d layer tlirory assumes that tlic siirface of the
glass fiber affects the curing or morpliolog) of the resin
surrounding t h e fibers and prodnces thc layvr having an
interinediate modulus I)etwecn the glass and resin. The
cocfficient of friction theory claims that the important
factor of reinforcement niechanisni is the friction bet\veen the fiber a r d matrix rathcr than t h e chr~niical
Imiding I)et\veen them. The prcf;wwtial adsorption
tlicory and tlic rcstrained layer theory arcs essentially
seccoirflicting thcorirs. As mentioncd in thc prt'\'IOUS
'
tion, ho\vever, t h e r r is a report which assclrts that thc
glass fil)crs iiihi1)it curing of the resin ( 2 3 ) .On the contrary. Plueddcinann (%) and .ilksue ct trl. ( 3 8 )showed
that sonic coupling agents iiiiprov? t h r ciiring of tlw
r&n. Both results niay occnr depending upon th(3
glassicoupliiig agentiresin system. Participation o f ii
coupling agent to thr curing oftlit. resin is &wonstrated
I)\- FT-IR. A sni;i11 E-glass mat treated n i t h inyltriW a v r n u m b e r ( cm")
300020001500
70 0
1100 900
10 0
a
6 0
2 0
too
6 0
2 0
138
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
The Reinforcement Mechanism of Fiber-Glass Reinforced Plastics Under Wet Conditions: A Review
methoxysilane was impregnated in a polyester resin.
Spectra of the composite and the polyester resin sanipled nearby the composite are shown in Fig.23. Difference spectrum shows higher extent of curing in the
composite system.
Chemical bonding between the coupling agent and
the matrix seems well established (81, 97, 104-111).
Such bonding is necessary to obtain a high performance
of FRP (97). Bjorksten and Yaeger (30) proposed a
chemical reaction between vinylsilane and polyester
resin through the vinyl group. Erickson et al. (31), on
the other hand, proposed either transesterification or
esterification with a terminal C=O group or OH group.
However, the evidence was difficult to obtain for the
vinylsilane and styrene (34). Vanderbilt reported that
the vinyl group of the CH,==CH--Si=
structure polymerized with difficultv (112). The C=C group of
y-inethacryloxypropyltrimethoxysilanehad higher reactivity to styrene or acrylate than the above vinylsilane
and copolymerized completely (10.5). While polyester
resins, phenols. and thiols copolymerized by a chain
transfer mechanism, epoxy and peroxy groups copolymerize by means of a bond scission mechanism, and
vinyl and acryl groups copolymerize in the usual manner
(107).
Spectroscopic evidence as shown in Fig. 28 was obtained by laser Raman spectroscopy on chemical reaction between methacryl silane and styrene monomer
(113). The styrene monomer was polymerized on the
methacryl silane-treated glass fibers. The styrene spectrum of this sample was identical to t h a t of
homo pol y in er , But the carbony1 stretching frequency at
1718 c n - ' of the unreacted silane shifted to 1702 c1n-l
after the polyinerization. These results indicate that
st!.rcne w a s homopolymerized though grafted to the
silane at the end of the chairi. Furthermore, neither
methacry1 silane nor vinyl silane react with polyester
resin if the resin coritains no styrene monomer. On the
contrary, thcx mcthacryl functional silanc was reported
B . POLYESTER
0.11%
A.
3n JL
Jr
PS-GF
B.
r'
n
1702
C.
1718
PS-ME S-GF
I\
D.
I600
I600
1400
CM-'
uf the g1as.s filierlmutrix interfuce. A:
monomer, B : Untreuted E-glass-filier coiited with polystyrene, C : E-glass ,fiber trecited with y-nietkucryloxypropyltriniethoxysilane. Styrene monomer i(;mpolymerized on thc
g/(i.r.s surfcice, und D: E-g/n.r.c.fiber trecitecl with y-metlauer!y/oi~ypr~~~J!/ltrimethox!/silnnc.
Fig. 28. Rumuri spctrci
St!yrene
to copolymerize with styrene while the vinyl functional
silane tended to terminate the styrene polymer chain,
i . e . , graft ( 8 7 , 118). Furthermore, Plueddemann
suggested the possibility of chemical reaction between
metliacryl functional silane and unsaturated polyester
resin without styrene monomer (118).
Chemical reaction between silane coupling agents
and a matrix was also demonstrated hy FT-IR (119).
High surface area porous silica was treated with 1 wt
percent vinyltrimethoxysilane aqueous solutions. Then
the silica was mixed with styrene monomer and 1 m7t
percent methylethylketoneperoxide ( M E K P ) and
heated for 3 h in air at 75°C. After being washed with
sufficient amount of carbon disulfide, the infrared spectrum was compared with the silica sample physisorbed
by polystyrene with sulxequent carbon disulfide wash.
Difference spectrum shown in F i g . 29 shows t h e
presence of polystyrene on the silica surface (e. g., 699
cm-l) and disappearance of the vinyl group (e.g., 1411
cm ) .
Unlike thermosetting plastics, thermoplastics usually
have no reactivc groups. Nonetheless, silane treatment
of glass fibers improves mechanical performance.
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
139
H . lshidu and J . L. Koenig
0)
CAB-0-Sll TREATED WITH
1 W l % VTMS M. SOL.
A.
0
A
J
c
A
0
A
t
t
.-
0.096
8 0
I-
E
u)
c
0
L
I-
$
1 698
1452
c.
I
A-B
800
16 X
Fig. 29. Fourier trunsform infrnred apectru o f h i g h surfuce ureu
~
silicci (Cab-O-Sil)lpolystyrene q s t e m . A: Cab-0-Sil w a treated
with 1 percent b!y weight oinyltl-imethosysilune and mixed with
styrene monomer. The mixture was polymerized at 75°C w i t h 1
percent by weight methylethylketoneperoxide (MEKPJ. Then
the mixture was dissolved and washed w i t h u sufficient umount
c$ curbon disulfide, B : Same (1s A except f o r poly,styrene Z G U S
physisorhed und uushed b y curbon disulfide, and C : Dqference
spectrum.
According to Plueddemann (108), the requirements for
silanes to function properly as coupling agents for fiberglass reinforced thermoplastics (FRTP) are: (i) the
silanes must be thermally stable at the maximum temperature encountered in molding (no decomposition),
and (ii) the silanes must be reactive with the resins
under molding conditions. A chain transfer mechanism
between polystyrene and alkylsilane was proposed by
Sterman and Marsden (109). The temperature dependence of the strength of FRTP molded at 200°F (93.3"C)
showed that all silanes yield lower strengths than heatcleaned glass. All silanes were better than the control at
500°F (260°C) molding. which implied that a chemical
reaction had taken place. Atkin and others (111)employing infrared spectroscopy give indirect evidence of a
chemical reaction between polyethylene and the hydrolyzate of vinyltriethyoxysilane. A mixture was ground
in a ball mill and extruded at 180°C. The polyethylene
was then extracted from the mixture with boiling xylol
until the control specimen was completely dissolved.
The infrared spectrum shown in Fig. 30 revealed
polyethylene residue in the silane indicating that chemical bonds exist.
Aside from the chemical reaction at the coupling
agent/matrix interface, there is a special effect of solid
particles. It is known that solid inclusions affect the
morphology of matrix, whether crystalline, amorphous
or crosslinked polymers. A number of investigations
have been made of the supermolecular structure of the
matrix at the interphase. On this subject matter, readers
may refer to excellent books by Manson and Sperling
(120) with emphasis on mechanical properties and h y
Lipatov (121) concerned with the molecular structure of
the interphase. Although most of the researchers
studied particulate filled polymers under dry conditions, the results are still useful for the hydrothermal
degradation of FRP. Since the degradation is a result of
140
700 cm-'
Fig. 30. Infrared spectru of polyoinylsiloxanelpol~~ethylene
SYStem. u: Polyethylene, b: Polycinylsiloxune, and c: Polyethylene
and polyvinylsiloxane mixture wus ground by Elull mill. The
mixture wus extruded a t 1NO"Cfollowed b y boiling xylol extraction.
water penetration, the morphology of the interphase
may be of great importance.
Thermal transition temperatures of particulate filled
polymers were studied by many researchers (122-149).
Their results are summarized in Tables 4a and 4 b , including the types of matrix and filler, filler contents,
glass transition shift (AT,) caused by fillers and techniques used for the T , measurements. The tables reveal
specificity of the AT, on filler/matrix system, yet in
general the T , increases upon introduction of fillers.
Since the T , change is the consequence of filledmatrix
interaction, the higher the total surface area of the filler
in the matrix, the larger the ATg. However, this relationship is not Straightforward, especially when the
filler size approaches the size of the matrix molecules.
Particle size, shape, nature of surface and filler content
affect the size and extent of agglomerate formation,
which causes heterogeneous filler distribution. Therefore, the interphases overlap. Surface modification is
one of the standard techniques to change the solid
surface/matrix interaction. It is interesting to note that
the surface treatment of the high surface area silica by a
silane coupling agent causes no shift in TSofpolystyrene
while the same filler content showed AT, = +20"C
when the surface is untreated. This may be due to the
interaction between the benzene ring and the surface
silanols. However, the activation enthalpy of segmental
motion is rather insensitive to the filler content whereby
the entropy shows marked change (12). Thus Lipatov
postulated that the nature of the interface is independent of the chemical nature of the surface unless there is
a specific interaction hetween the filler surface and the
matrix.
The increase of the T,, for the filled polymer suggests
that the motion ofthe backbone chain is restrictedby the
solid surface. This restriction decreases the number of
possible conformations. The higher sensitivity of the
cntropy term to the filler contents stated previously
supports this statement (121). Consequently, restricted
polymer chains form a loosely packed layer near the
filler surhce. It suggests that the side chain motion in
the loosely packed layer is more flexible, though the
main chain motion is still restricted. This is in agreement
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
The Reinforcement Mechanism of Fiber-Glass Reinforced Plustics Under W e t Conditions: A Recieu:
Table 4a. Effect of Fillers to the Glass Transition Temperature (T,) of Epoxy Resin
Literature
no.
Matrix
Filler
content
Filler
129
EPOXY
TiO, (0.2 p m )
124
128
EPOXY
EPOXY
TiO, (0.2 p m)
AI,O,, (0.05pm)
AI,O, (0.3pm)
144
EPOXY
136
EPOXY
Glass beads
(soda lime:
44v)
GIass beads
(soda lime)
5-10 p m
10-20 prn
30-40 prn
75-90 prn
146
Quartz
(0.4 d i g )
EPOXY
(diglycidylether
type)
(bisphenol A
type)
T,, "C
10 wt%
20 wt%
30 wt%
40 wt%
12 vol%
1.0 vol%
1.7 ~ 0 1 %
3.9 vol%
0.7 ~ 0 1 %
1.o vol%
2.0 vol%
3.8 ~ 0 1 %
0
30 ~ 0 1 %
30 ~ 0 1 %
+2
+5
+7
+10
0
-2
+2
-3
+3
+3
+2
+4
-16
+6
-11
26 ~ 0 1 %
10 vol%
41 Vol%
10 vol%
40 ~ 0 1 %
10 vol%
38 ~ 0 1 %
+2
+2
+3
-3
+4
+1
+3
0.05 pbv
0.172 pbv
0.328 pbv
0.05 pbv
0.16 pbv
0.31 pbv
+6.5
+13
+16
+3.4
+8
+10.8
Technique
Condition
Torsion
pendulum
Dilatometry
Torsion
pendu Iu m
Vibron:
110 Hz
4.64 wt% water
dry
2.83 wt% water
Dilatometry
Table 4b. Effect of Fillers to the Glass Transition Temperature (T,) of Thermoplastics
~~
Literature
no.
Polystyrene
143
Poly(ethy1enepropylene)
Polyethylene glycol
Poly(butadienestyrene)
143
135
Filler
Matrix
143
Filler
content
Silica
(200 mVg)
41 VOI%
Carbon black
(80 m'ig)
(8 m'ig)
123
Poly(viny1 chloridevinyl acetate)
(87% vinyl chloride)
TiO, (0.22 Fm)
Abbreviations: wt%
Polyvinyl acetate
=
per cent by weight; voI%
TiO, (0.2 p n )
=
per cent by volume; pbv
Technique
0
Torsion
pendulum
(heating rate
1"Cim in)
+ 20
31 ~ 0 1 %
50 phr
100 phr
50 phr
100 phr
105 phr
105 phr
3.38 wt%
8.86 wt%
+ 32
+18
+1
+2.5
+5.8
+6.7
+0.1
+4.9
-4
-15
Dilatometry
Dilatometry
- 29
12 vol%
=
~~
AT,, "C
41 ~ 0 1 %
31 VOI%
13.77
17.90
124
~
0
Condition
Surface
modification
by (CH,),SiCI,
Untreated
Untreated
Untreated
Not vulcanized
Not vulcanized
Vulcanized
Vulcanized
Not vulcanized
Vulcanized
DueAo specific
interaction
between TiO,
and acetate
T, disappeared
A new peak
between T, and
p-transition
Dilatornetry
parts per volume; and phr = parts per hundred (by weight).
with the N M R study of filled polymethylmethacrylate
where the apparent activation energy of the a-methyl
group relaxation was decreased by fillers (131). The
existence of the rarefied interphase is also supported
by the density measurements (131, 146, 150-152). F i g tire 31 show75 the density profile of un epoxy resin filled
with ground quartz with respect to the distance from the
filler surface. This measurement was done by an etching
technique and the density difference was assumed to be
proportional to the etching depth. Up to 0.1-0.2 p m , the
density increased gradually and approached a constant.
In other words, the density of the interphase is lower
than the bulk and shows a density gradation. The effect
of heat treatment on the density of fiber-glass reinforced
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
141
H . lshida und J . L. Koenig
&
I-
3
2
.
I
I
0
0 4 8 12 16
F i g . 31. Density profile c!f'epox~/resin icith increasing distance
R,from the filler p r t i c l e surfme.
epoxy was studied by means of the luminescence spectra
produced by anthracene used a s a "molecular probe"
(151). Thc result showed that the interphase can be
densified and even the density difference can lie eliminated with proper heat treatment. Considerations of
density may be very important when water susceptihility is concerned. A loosely packed layer would lie expected to take up more moisture than the bulk. Studies
by Lipatov et al. (141, 153)utilizing gas chromatography
with polymer coated solid particles (glass and teflon) as a
column showed that vapor retention was increased by
these looscly packed surface layers. On the other hand,
vapor sorption study by Kwei and Kumins showed reduction of the sorption of organic vapors at low relative
pressures. They postulated that the sorptive properties
of the interphase are analogous to those of glassy polymers (124).
The ahove discussions can lw summarized as follows.
Due to the fillerhatrix interaction, the possible number
ofconformations of polymer chain is restricted, resulting
in loosely packed layers around the filler particles. The
nature of this special layer, interphase, is rather insensitive to thtx nature of the filler surface when there is no
specific interaction at the filler/ matrix interface. However, introduction of chemical interaction at the interface affects the morphology of the interphase. The
thickness of the interphase, therefore, varies depending
upon the filler/matrix system. Rut, in many cases, it is
reported in the range between 200 A and 20000 A (23,
24).
the hydrolysis reaction of the polyester resin and water
penetration by tensile stress. Compressive stress, on the
other hand, has an opposite effect on water penetration
(160). The higher the compressive stress, the smaller
the amount of water penetration. Thus, restrained
layers around glass fibers, if there are any, can prevent
water penetration as a vapor, rather than as a liquid
(161). Ashbee et ul. (162) reported that thermosetting
resins like polyesters first swelled and then shrunk due
to leaching of so1ul)le materials and continued polymerization through residual reactive groups. A water immersion test of polyester resin plate at 80°C for 5000 h
followed the same process described above for the first
few thousand hours. Then, the weight of the specimen
started increasing again. While leaching and polymerization were taking place, hydrolyzable linkages of polyester resin were attacked by water, and the resulting
terminal groups became the driving force of further
water penetration (163).Thc formation of O H sites as a
result of hydrolytic attack was confirmed spectroscopically (113).The C--0 bonds in the crosslinked structure
were attacked by water molecules and formed COOH
and C O H groups-which are highly hydrophilic. Raman
spectra of the polyester resin before and aftcr hydrolysis
are shown in Fig. 32 where a new line appears at 1106
c x - ' due to hydrolytic attack.
CONCLUSIONS
As a resnlt of urgent needs for a water resistant material with high specific strength, FRP technology has
made great advances. Following the technology,
phenomenological observations had heen made arid
theories of water action proposed. Owing to the extremely small relative amount of the materials that constitutes the interfaces, studies o n compositc systems on
the molecular level are far behind phenomenological
studies. Recent advances in spectroscopy. however,
provide a high sensitivity and selectivity. It ma)- no\\.
thus be possible to study these difficult regions.
1030 1000
nn
863 ?52
MATRIX
Mechanical properties of thermoplastics in liquid
media under stressed condition were extensively
studied by Okuda et al. (154-158) using creep, stress
relaxation, and fatigue testing. Almost no literature is
availalile on the mechanical properties of thermosetting
plastic in liquids or in media other than air. Water, such
as humid air with relative humidity about 70 percent,
has a marked effect on short term creep strength of
unsaturated polyester resins under stressed conditions
at 40 and 8WC, but very little effect at 20°C (159).
Nevertheless, there are no noticeable changes in the
Raman spcctra of the polyester resin for a composite
system exposed to the air at 38°C with 100 percent
relative humidity for one month (113).This discrepancy
can he explained by inechanochemical enhancement of
142
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
The Reinforcement Mechanism of Fiber-Glass Reinforced Plastics Under W e t Conditions: A Reciew
Reinforcement through chemical bonding at both interfaces seems to he well established. In the future, the
quantitative determination of the relationship of chemical bonds to the strength of an FRP should develop. To
this end, studies based on macroscopic and microscopic
viewpoints must be correlated.
REFERENCES
1. D . A. Scola, Proc. 30th Ann. Tech. Conf., Reinforced
Plastics/Composites Inst., SPI, Sec. 22-C (1975).
2. R. Wong, Adhesion, 4, 171 (1972).
3. L. Holland, “The Properties of Glass Surface,” p. 180,
Wile)., New York (1964).
4. G. D. Nichols, D. M. Hercules, R. C. Peek, and D. J.
Vaughan, Appl. Spect., 28, 219 (1974).
5. R. G . Adams, S. J. Milletari, and A. Roth, Mod. Plastics, 49,
74 (Mar. 1972).
6. P. C. Yaks and J. W. Trebilcock, Proc. 16th Ann. Tech. clr
Manag. Conf., Reinforced Plastics Div., SPI, Sec. 8-B
(1961).
7. 0 .K. Johanson, F. 0 .Stark, and R. Baney, AFML-TRI-65303, Part l(1965). G. E. Vogel, 0. K. Johanson, F. 0. Stark,
and H. M. Fleishmann, Proc. 22nd Ann. Tech. Conf., Reinforced Plast. Div., SPI, Sec. 13-B (1967). 0. K. Johanson,
F. 0. Stark, G. E. Vogel, and R. M. Fleishmann,J. Comp.
M a f . , 1, 278 (1967).
8. W. J . Eakins, Proc. 17th Ann. Tech. rW Manag. Conf., Reinforced Plastics Div., Sec. 10-C (1962).
9. V. A. Berstein and V. V. h’ikitin, Dokl. A k a d . S S S R , 190,
823 (1970).
10. V . A. Berstein, Y u . D. Varfolomeev, and V. V. Nikitin, Fiz.
Tcerd. Telu, 13, 693 (1971).
11. V. A. Berstein and V. V. Nikitin, Proc. 10th Intern. Cong.
Glass, Part 11, K>oto, Japan (July 1974).
12. F. R. I,. Schoenig, J . Appl. Phys., 31, 1779 (1966).
13. W. F. Thomas, Phy.r. Chem. Glasses, 1, (1960).
13. R. T. Trichane and G . B. Carrier, J . Am. Ceramic Soc., 44,
606 (1961).
15. S. Kurosaki, S. Saito, a i d G . Sato,J. C h e m . P h y . ~ .23,
, 1846
(1955).
15. S. Kurosaki, S. Saito, a r r d G . S a t o , J . Cheni. Phys., 23, 1846
(1955).
16. 51. M .Egonov, V. F. Koselev, and K. G. Krasil’nikov, Zh.
Fiz. Khim., 35, 2234 (1961).
17. M. J. Rand, Electro. Cheni. Soc., 109, 402 (1962).
18. R. J. Huang, T. Demirel, and T. D. Mc<>ee,J.Ani. Cerum.
Soc., 55, 399 (1972).
19. J . H. Frazer, Ph!/.r. R e c . , 33, 97 (1929).
20. \’. R. Dcitz, NRL R e p . 6812, Naval Research Lab.,
\Va\hington, D.C. (1968).
21. K. Strolenl)erg, E. Schluescner, and L. W. J . Damen,
“G1.‘is.f,xeiverstaerktt.
. .
Kunstoffe,” p. 137, ed. h y P. H. Seldon, Springer-Verlag, Berlin (1967).
22. 11. Fratzscher. Mod. Plu.rt., 49, 82 (MU.1972).
23. T . K . Kwci, J . P~l!/tti.Sci., 3, 3229 (1965).
2 1 , E . V, Trostyanskaya. A . hl. Poimanov, and Y r i . K .
1 1 , Xlckhmnika Polinleroc, 1, 26 (1965).
Ku
, . m \.k”
25. H . \Yeisl)art. Proc. 20th A n n . Tech. Conf.. Keinforced Plastics Di\.. SPI. Sec. ,‘3-D(1965).
, 54 (1964).
26. t-I. \t-t.isIxirt. l i r t r i . ~ t . r t o f f ( ~12,
27. P.J . hlacGarr!, Proc. Sixth Sag;iniore Ordnance hlaterials
Research C O I I ~(Atig.
,
18-21, 1959).
28. I,. L. Yaeger, Proc. 6th 4nn. T e c h Session, Reinforced
Plast. D i v. , SPI, Scc. 12 (1951).
29. J . Bjorksten, Proc. 7th .Ann. Tech. Session, Reinforced
Plast. Div., SPI, Sec. 15 (1952).
,
P l m t . , 29, 124 (1952)).
30. 1. Bjorksten and L. L. E ’ i ~ g e rMod.
31. P. \V. Erickson and I. Silvci-, Proc. 8th A n n . Tech. Q
\lanag. Coiif’.,Reinforced Plast. Div., SPI, Sec. 14 (1
:3” P. \I., Erickwn. I. Silvci-. a i i t l 11. A . Pcrr!, Jr.. Am. Chcni.
Soc.. Div, ofPaint. P l a h t . , and Priiit. Ink Chem.. 125tll . A n n .
Lleetiiig.
,
33 J. M. Stevens and P. M. Goodwin, Proc. 10th Ann. Tech. &
Manag. Conf., Reinforced Plast. Div., SPI, Sec. 7-1 (1955).
34 K. Gutfreund, H. S . Weber, and C . Brown, Proc. 15th Ann.
Tech. clr Manag. Conf., Heinforced Plast. Div., SPI, Sec.
10-C (1960).
35 B. M. Vanderbilt and J. P. Simko, Jr., Proc. 15th Ann. Tech.
& Manag. Conf., Reinforced Plast. Div., SPI, Sec. 10-D
(1960).
36 J. A. Laird and F. W. Nelson, Proc. 19th Ann. Tech. Conf.,
Reinforced Plast. Div., See. 11-C (1964).
37 T. E. White, Proc. 20th Ann. Tech. Conf., Reinforced Plast.
Div., SPI, Sec. 3-B (1965).
38. M. E. Schrader and I. Lerner, Proc. 20th Ann. Tech. Conf.,
Reinforced Plast. Div., SPI, Sec. 3-C (1965).
39. K. E. Kolb and J . G . Koelling, Proc. 21st Ann. Tech. Conf.,
Reinforced Plast. Div., SPI, Sec. 13-D (1966).
40, M. E. Schrader, I. Lerner, F. J. D’oria, and L. Deutsch,
Proc. 22nd Ann. Tech. Conf., Reinforced Plast. Div., SPI,
Sec. 13-A (1967).
41. E. P. Plueddemann, Proc. 25th Ann. Tech. Conf., Reinforced Plast. Div., SPI, Sec. 13-D (1970).
42. E. P. Plueddemann, Natl. SAMPE Tech. Conf., p. 453
(1969).
43. B. M.Vanderbilt, M o d . Plust., 37, 125 (Sept. 1959).
44. B. M. Vanderbilt and J. P. Simko, Jr., Mod. Plast., 38, 135
(Dec. 1960).
45. H. A. Clark and E. P. Plueddemann, Mod. Plast., 40, 133
(June 1963).
46. R. L. Kaas and J. L. Kardos, PoZym. Eng. Sci., 11, 11 (1971).
47. V. I. Alksne, V. Zh. Kronberg, and Ya. A. Edius, Mekhanika
Polimerou, 3, 601 (1967).
48 V. I. Alksne, K. P. Grinevich, a n d I. E . Yakobson,
Mekhanika Polimerou. 3. 787 (1967).
49. J. L. Koenig and P. T. K. Shih,]. CoZZoid Interface Sci., 36,
247 (1971).
50. A. N. Gent and E. C . Hsu, Macromolecules, 7,933 (1974).
51. A. V. Kisclev, V. I. Lygin, and I. N. Solomonova, Kolloidn.
Zh., 26, 324 (1964).
52. J. L. Koenig, P. T. K. Shih, and P. Lagally, Mat. Sci.Eng.,
20, 127 (1975).
53. V. 1. Lygin and A. L7.Kiselev,Kolloidn., Zh., 23,299( 1961).
54. I,. I. Golubenkova, A. M . Shabadash, S. N . Nikonova, and
M .S. Akritin, Vy.sokomo1. Soed., 4, 1354 (1962).
5s. G . Wirzing and H. W. Kohlschuetter,Z.Anal. Chemie, 198,
270 (1963).
56. M. G. Voronkov, E . A. Lasskaya, and A. A. Pashchenko,Zh.
Prikl. Kliim., 38, 1483 (1965).
57. J . Gahde and A. Wende, Plaste Kuutschuk, 13,23 (1966).
58. S . N. Nikonove, L. I. Goluhenkova, A. N. Shabadash, and
M .S. Akutin, Plmticheskie Massy, 2, 27 (1966).
59. W. A. Zisman, NKL Rep. 6083 (June 9, 1964).
60. M. E. Schrader, Proc. 25th Ann. Tech. Conf., Reinforced
Plast. Div., SPI, Sec. 13-E (1970).
61. P. W. Erickson, Proc. 25th A n n . Tech. ConE, Reinforced
Plast. Div., SPI, Sec. 13-A (1970).
62. E. J. Eakins, PLASTEC Rep. No. 18, AD609526 (Sept.
1964).
63. W.D. Bascom, Proc. 25th Ann. Tech. Conf., Reinforced
Plast. Div., See. 13-C (1970).
64. i V . J . Eakins, SPE Trans., 1, 199 (1961).
65. I,. Pauling, “The Nature of the Chemical Bond,” Cornell
Univ. Press, Ithaca, N.Y. (1948).
66. K . C . iisthoi‘i’, A. X1. Bueche, axid W. T. Grubb, J . Am.
Chem. Soc., 76, 4659 (1954).
67. A . C . Matellock, Natl. Meeting Am. Chem. Soc., Div.
Pol! ni. Chein., Boston, Massachusetts (1959).
68. L. A. Spitze and D. 0. Richards, J . Appl. Phys., 18, 904
(1947).
69. W. M . Arnleim, T. K. Kwei, and C. A. Kumins, Proc. 20th
Ann. Tech. Conf., Reinforced Plast. Div., SPI, Sec. 15-D
( 1965).
70. \V. D . Bascom, Proc. 20th Ann. Tech. Conf., Reinforced
Plast. Div., SPI, Sec. 15-€3(1965).
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
,
I
143
H . Zshidu and I . L. Koenig
71. W. D. McNeil, B. Bennett, and R. I. Leininger, Proc. 19th
Ann. Tech. Conf:, Reinforced Plast. Div., SPI, Sec. 11-B
(1964).
72. J. Delmonte and E. Sarua, Proc. 18th Ann. Tech. & Manag.
Conf., Reinforced Plast. Div., SPI, Sec. 20-A (1963).
73. E. P. Pleuddemann, Mod. Plast., 47, 92 (1970).
74. D. J. Vaughan and E . L. McPherson, Proc. 27th Ann. Tech.
Conf., Reinforced Plast. Ijiv., SPI, Sec. 21-C (1972).
75. S. Sterinan and H. B. Bradley, SPE Trans., 1, (1961).
76. S. Sterman and H. B. Bradley, Proc. 16th Ann. Tech. &
Manag. Conf., Reinforced Plast. Div., SPI, Sec. 8-D(1961).
77. L. H. Lee, Proc. 23rd Ann. Tech. Conf., Reinforced Plast.
Div., SPI, Sec. 9-D (1968).
78. H. Ishida and J. L. Koenig, Proc. 31st Ann. Tech. Conf’.,
Reinforced Plastic /Composite Inst., SPI, Sec. 6-C (1976).
79. L. V. Sergeyev, A. Baigozhim, and S. G. Fattakhov, Vysokomol. Soed., 4,977 (1962).
80. D. L. Chamberlain, M. V. Christensen, and M. Bertolucci,
Pro:. 24th Ann. Tech. Conf., Reinforced Plast. Div., SPI,
Sec. 19-C (1969).
81. A. Ahagon, Ph.D. Dissertation, Univ. of Akron, Akron,
Ohio (1974).
82. V. I. Alksne, V. Zh. Kronberg, andYa. A. Eidus,Makhanika
Polimerou, 4, 182 (1968).
83. H. Ishida, Ph.D. Dissertation, Case Western Reserve
Univ., Cleveland, Ohio (1977).
84. G. Schreiber, Spectrochimicu Acta, 22, 107 (1966).
85. D. J. Vaughan, R. C. Peek, Jr., Proc. 29th Ann. Tech. Conf.,
Reinforced Plast. Div., SPI, Sec. 13-D (1974).
86. E. P. Plueddemann, H. A. Clark, L. E. Nelson, and K. R.
Hoffmann, Mod. Plast., 39, No. 8, 139 (1962).
87. H. A. Clark and E. P. Plueddemann, Mod. Plus., 40, No. 6,
133 (1963).
88. E. P. Plueddemann, Proc. 24th Ann. Tech. Conf., Reinforced Plast. Div., SPI, Sec. 19-A (1969).
89. P. T. K. Shih and J . L. Koenig, M a t . Sci. Eng., 20, 137
(1975).
90. L. R. Daley and F. Rodriguez, Polym. E r g Sci., 9, 428
(1969).
91. €I. Ishida a n d J . L. Koenig, Proc. 32nd Ann. Tech. Conf.,
Reinforced Plastics /Composite Inst., SPI, Sec. 4-B (1977).
92. R. V. Viventi, H. T. Plant, and R. If. Maher,Mod. Plast., 45,
129 (Jan. 1968).
93. P. V. Wright and J. A. Semlyen, Polymer, 11,462 (1970).
94. E. P. Pleuddemann and G. L. Stark, Proc. 28th Ann. Tech.
Conf., Reinforced Plast. Div., SPI, Sec. 21-E (1973).
95. 1f. E. Schrader and A. Block, ]. Polym. Sci., C , 34, 281
(1971).
96. F. J. McGarry, Proc. 13th Ann. Tech. & Manag. Conf.,
Reinforced Plast. Div., SPI, Sec. 11 (1958).
97. E. P. Plueddcmann,J. Puint Techno/., 40, 1 (1968).
98. W. J . Eakins, 20th A n n . Tech. Conf’,, Reinforced Plast.
Div., SPI, Sec. 13-C (1965).
99. C. A. Kumins and J. Roteman,J. P o l y m . Sci., 1, 527(1963).
100. R. C. Hooper, 11th .4nn. Tech. & Manag. Conf’.,Reinforced
Plast. Div., SPI, Sec. 8-B (1956).
101. J. 0. Outwater, Jr., 11th Ann. Tech. a n d hlanag. Conf.,
Reinforced Plast. Div., SPI, Sec. 9-B (1956).
102. J. 0. Outwater and M. Yrndrzeski, 20th Ann. Tech. Conf’,,
Reinforced Plast. Div., SPI, Sec. 19-B (1965).
10G . J. 0. Outwater, F. I,. Sole, P. W. Erickson, and J. \’. Duffj,
Proc. 23rd A n n . Tech. Conf., Reinforced Plast. Div., SPI,
Sec. 9-A (1968).
104. E. P. Pl~icddemann,11. A. Clark, L. E. Nelson, and K. R.
Hoffinan, Proc. 17th Ann. Tech. & Maiiug. Conf., Reinforced Plast. Div., Sec. 14-A (1962).
105. H. A. Clark and E. P. Plueddemann, Proc. 18th Ann. Tech.
& Manag. Coiif., Reinforced Plast. Div., SPI, Sec. 20-C
(1963).
106. E. P. Plueddemann, Proc. 20th A n n . Tech. Conf., Reinforced Plast. Div., SPI, Sec. 19-.4 (1965).
107. E. P. Plueddemann, H. A. Clark; L. E. Nelson. ant1 K . R.
Hoffinan, M o d . Pla.Pt., 39, 135 (Aug. 1962).
.
108. E. P. Plueddemann, Mod. Plast., 43, 131 ( h r ~ g1966).
144
109. S. Sterman and J. G . Marsden, Mod. Plast., 43, 133 (July
1966).
110. S. Sterman, Polyrn. Eng. Sci., 6, 97 (1966).
111. M. S. Akutin, M. L. Kerber, I . 0. Stal’nova, N . L .
Grodskaya, and E. E. Alekseev, Mekhanila Polimeror;, 8,
1048 (1972).
112. B. M.Vanderbilt, Mod. Plast., 37, No. 1, 125 (1959).
113. P. T. K. Shih, Ph.D. Dissertation, Case Western Reserve
Univ., Cleveland, Ohio, Ch. VII (1973).
114. W. D . Hascom, Macromo/ecuks, 5, 792 (1972).
115. R. G. Greenler, ]. Chem. Phys., 44, 310 (1966).
116. F. J. Boerio, L. H . Schoenlien, and J. E. Greivenkamp,
Proc. 32nd Ann. T e c h . Conf., Reinforced Plastics/
Composite Inst., SPI, Sec. 4-A (1977).
117. E . P. Plueddemann, ed., “Composite Materials,” Vol. 6, in
“Interfaces in Polymer Matrix Composites,” Academic
Press, New York (1974).
118. E . P. Plueddemann, private communication (July 1976).
119. H. Ishida and J. L. Koenig, to be published.
120. J . A. Manson and L. H. Sperling, “Polymer Blends and
Composites,” Plenum Press, New York (1976).
121. Yu. S. Lipatov and L. M. Sergeeva, “Adsorption of Polymers,” John Wiley & Sons, New York (1974).
122. Yu. S. Lipatov, T. E. Lipatova, Y. P. Vasilenko, and L. M .
Sergeeva, Vysokomol. Soed., 5,290 (1963).
123. C. A. Kiimins and J. Roteman,]. Polyrn. Sci., A1,527(1963).
124. T. K. Kwei and C. A. Kumins,]. Appl. Polym. Sci., 8, 1483
(1964).
125. T. K. Kwei, J. Polym. Sci., A3, 3229 (1965).
126. I. Galperin and T . K. Kwei, ]. A p p l . Polym. Sci., 10, 673
(1966).
127. Yu. S. Iipatov and T. E. Celler, Vysokomol. Soed., 9, 241
(1967).
128. F. R. Dammont and T. K. Kwei, A m . Chem. Soc. Polym.
Prep., 8, 920 (1967).
129. I. Galperin, 1.Appl. Polym. Sci., 11, 1475 (1967).
130. Yu. S. Lipatov and F. C. Fabul!ak, Vysokomol. Soed., 10,
1605 (1968).
131. Yu. S . Lipatov and F. G. Fabulyak, Vysokoinol. Soed., 11,
708 (1969).
132. S. Krause, ]. Polym. Sci., A-27, 249 (1969).
133. D. H. Droste and A. T. DiI3enedetto.J. Appl. Polyrn. Sci.,
13, 2419 (1969).
134. .4. Yini and L. E. St. Pierre,]. Polym. Sci., B7,237 (1969).
135. C:. Kraus and J . T. Gruver,]. Polym. Sci., A-2, 8,571 (1970).
136. T. B. Lewis arid L. E. Nielson,]. Appl. Polynt. Sci., 14,1449
(1970).
137. V. P. Solomko and S. P. Pas’ko, Vysokoniol.Soed., 12,681
(1970).
138. S.Kaufman, W. P. Slichter, and D. D. Davis.]. Polym. Sci.,
A-2, 9, 829 (1971).
139. Y u . S. Lipatov and V. P. Privalko, V!/sokon1o/. Socjd., 14,
1643 (1972).
140. J . L. Kardm. W. L. hlcDonncl1. and J. R;lisoni, J . ,Cicrcromol. Sci. Phys., B6, 397 (1972).
141. A. Ye. Nesterov arid Yu. S. Lipatov, V!/sokomol. Soed., 15,
2601 (1973).
142. R. J . Roe, D. D. Davis, and T. K. Kwei, .4m.C h e m . Soc.
Polyni. Prep. 14, 587 (1973).
143. A. Yim, R. S. Chahal, and L. E. St. Pierrc..j. Colloid Z r t l e r f m e Sci., 43, 583 (197.3).
144. J. A. blanson a n d E. H. Chiu,Am. Chent. S o c PolIym. Prep.,
14, 469 (1973).
145. T. Hirai and D. E. Kline, J . Camp Mut., 7, 160 (1973).
146. Ye. B. Trost!anska)a, A. 11. Poim;ciiov, ;ind Ye. F. no so\^,
V!/.sokomol.Soed., 15, 612 (1973).
147. R. 1. ILIorg>\n,J . M a t . Sci., 9, 1219 (1974).
148. \’. P. Solo~i~ko,
N. F. Vo\ k(ltrtll), S. P. Piis’ko, ~ i r r d\’, I.
Surovtsev, V!/sokomol. Soed, 16, 519 (1974).
149. Yu. S. Lipatov, L’. F. Bal)ich, and V. F. Ro\ovzk!, J . .4ppl,
P o l q m . S c i . , 18, 1213 (1974).
POLYMER ENGINEERING AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
The Reinforcement Mechanism of Fiber-Glass Reinforced Plastics Under W e t Conditions: A Review
151. Ye. G . Moisya, G. M. Semenovich, and Yu. S. Lipatov,
Vysokomol. Soed., 15, 1337 (1973).
152. V. M . Startsev, A . Ye. Chalykh, S. A. Nenakhov, and A. T.
Sanzharovskii, Vysokomol. Soed., 17, 836 (1975).
153. A. Ye. Nesterov and Yu. S. Lipatov, Vysokomol. Soed., 17,
671 (1975).
154. S. Okuda and S. Nishina, Proc. 9th Cong. Testing Mater.,
Japan, 109 (1966).
155. S. Okuda and T. Iguchi, Proc. 9th Cong. Testing Mater.,
Japan, 113 (1966).
156. S. Okuda and S. Nishina, Proc. 10th Cong. Testing Mater.,
Japan, IS7 (1967).
157. S. Okudaand T. Iguchi, Proc. 5th Int. Cong. Rheology,Vol.
3, p. 357 (1970).
158. S. Okuda, S. Nishina, and T. Iguchi, Proc. Int. Cong. Mech.
Behavior Mater., Kyoto, Japan (1973).
159. H. Ishida, Master Thesis, Doshisha Univ., Kyoto, Japan
(1973).
160. E. M. Filyanov, 0. G . Tarakanov, and I. V. Shamov,
Mekhanika Polimerou, 10, 163 (1974).
161. D. I. James, R. H. Norman, and M. H. Stone,Plast. Polym.,
31 (Feb. 1968).
162. K. H. Ashbee and R. C. Wyatt, Proc. Roy. Soc., A312, 553
(1969).
163. S. Okuda, H. Ishida, and €1.Hukuyama, unpublished data.
164. A. K. Rastogi, J. P. Rynd, and W. N. Stassen, Proc. 31st Ann.
Tech. Conf., Reinforced Plastics/Composite Inst., SPI,
See. 6-B (1976).
POLYMER ENGlNEERlNG AND SCIENCE, MID-FEBRUARY, 1978, Vol. 18, No. 2
145