Fatigue Characteristics of High Glass Content Sheet

Molding Compound (SMC) Materials

P. K. MALLICK"
Department of Mechanical Engineering
The University of Michigan-Dearborn
Dearborn, Michigan 48128

Static and fatigue properties of high glass content sheet

molding compound (SMC) materials were studied at various
temperatures. It was shown that the matrix plays an important
role in both fatigue characteristics and failure mechanisms of
such randomly oriented short fiber composites. Specifically,
vinyl ester matrix shows better fatigue properties and post-
fatigue performance than the polyester system considered
here.

INTRODUCTION tent; however, i n this case, the notninal fiber length is

w ith the introduction of high glass content, high

strength sheet molding compounds (HSMC) in
recent years, fiber reinforced plastics are being con-
1/2 in. compared to 1 in. for the polyester and vinyl ester
systems. None of the systems has any filler in it. All
these inaterials are commercially available and are being
sidered in many structural applications in automotive considered for various structural applications by the
and other non-aerospace industries. This class of mate- automotive industry.
rial is basically a modified SMC, containing 50-65 per- The compounds were inolded into flat plaques in
cent b y weight of randoinly oriented chopped glass fiber laboratory size presses at or near a mold temperature of
strands and little or no filler whereas conventional SbIC 300°F. Test specimens were first saw cut from the
contains only 25-30 percent glass fiber and 30-40 percent inolded plaques and then polished on a sand belt to
filler, usually calcium carbonate. HSMC is coinpounded smoothen the edges. All specimens were approximately
and prepared in a standard SMC machine and coinpres- half-inch wide and at least 5 in. long.
sion inolded at pressures ranging up to 3,000 psi. Be-
cause of their relatively high strength and niodulus EXPERIMENTAL PROCEDURE
(which arise from higher glass content) and ease of mold-
ing, complex structural shapes can now be designed arid Fatigue tests were peiformed on an AITS servo-
manufactured froin these inaterials. Consequently, hydraulic machine with a three-point constrained flex-
there is a great need for characterizing the mechanical ure fixture. A schematic of the fixture is shown in Fig. I
and physical properties of HS MCs and developing de- in which the specimen is clamped at the ends and loaded
sign guidelines. However, most of the studies dealing at the center of the span. In this mode of loading, equal
with such inaterials have involved formulation and coin- iiiaxiinum stress occurs under the center loading point
pounding (1,2), static mechanical properties (3)and re- and near the end constraints. Fatigue specimens were
cently, evaluation of curing characteristics (4). Very lit- subjected to completely reversed, sinusoidal alternating
tle attention has been given to fatigue properties even stresses (tensioii-compression, R = - 1)at 300cpm. The
though some of the applications under consideration stroke of the actuator, set to provide a predetermined
may lie subjected to cyclic loading. initial peak load, was maintained at a constant level
In this paper, static and fatigue properties of high during the test. At intervals, cycling frequency was re-
glass content SMC inaterials with three different matrix duced to 3 cprn and load-displacemetit curves were
systems are reported. In addition, inechanisnis offailure recorded. Dynamic niodulus values were ol)tained from
are also presented. the initial slopes of these curves. Failure was defined as a
15-20 percent drop in the dynamic modulus from its
IMATERIALS value measured between 1-10 cycles. At the end of the
The principal inaterials itivestigated in this study are fatigue tests, each specimen was tested in static flexure
polyester (isophthalic) and vinyl ester HMSCs with a to nieasure its residual strength. Fatigue tests were run
noniinal glass fiber content of 60-65 percent by weight. i n an envii-onmental chamlier at RT, 130, 220, and
In addition, some tests were also pelformed on an epoxy 320' F .
niolding compound which has a sitnilar glass filwr coil- All static tests, including those after fatigue louding,
were peifortned on a siinply supported three-point flex
'Fuimrily with Pl.iiticr Drvelc,pment m d .Applic,itions Office. Fold Mntcu Co.
fisture on an Instron universal testing machine. Ele-
Drtioit. Michig.m. vated temperature static tests were run i n the Instron

18 POLYMER COMPOSITES, JANUARY, 1981, Vol. 2, No. 1

 Futigue Churucteristics ( f H i g h Gluss Content Sheet Molding Compound JSMC)Materiuls

TO LOAD CELL

-} 1.5” 1.5”

1 a
TO ACTUATOR Y
a
z
20 - A EPOXY -

I I I I I I

b
Fig. 1 . Schrmutic (fcxperimentul set-up i u ) futiguc.. test [ b )
stutic j e x t u r e test uftrr futigue.

environmental chamber. A two-inch s p a n was used in all

”I-
2 1 I
these tests. Thus the damaged ends of the fatigue speci- .POLYESTER
.VINYL ESTER
mens near the clamps were excludrd from the post- z 20 A EPOXY
fatigue static test area.

RESULTS 0
Static Properties 50 150 250 350
TEST TEMPERATURE, ‘F
Room temperature static flexural .>ropertiesof high Fig.3. Retmtion of stutic modulus inJi,xure us u function oftest
glass content inaterials (HSMC) are presented in Table temperutures.
1 . For coinparison. typical values for c:onventional SMC
are also included. Higher values for the present mate- is mostly a result of softening of the matrix resin, a clear
rials are direct results of their higher glass fiber content. superiority of one system over another cannot be estab-
The load-deflection curves are initi;tllv linear, but, at lished from these figures. However, for design purposes
stresses lwtween 35-50 percent of the ultimate values where short-term high temperature is encountered,
they deviate from linearity indicating the occurrence of these diagrams will be helpful.
sotile irreversible daiiiage in the 111 iterial. Geiierally,
this datiiage takes the forin of “microc,racks” or “crazes”, Fatigue Properties
possilily a result of fil)er-matrix del)onding ( 5 , 7 ) .Final
failure takes place in a step-wise fasl-lion starting with a Definition of Failure. Since materials under consid-
failure on the tension suifilce followec, I)\ tlelaminations. eration fail in a progressive tnanner through gradual
Figures 2 and 3 show the percent retention of flexural softening a s the test progresses, a clear definition of
strength and iiiotlulus a s a function of test temperature. failure is not possible. Therefore, a 15-20 percent drop
As expected, both strength and modulus decrease with in modulus from its value near the 10th cvcle was arbi-
increasing temperature. Even though this reduction trarily selected as a criterion of failure. Figure 4 is one
such plot of percent drop in modulus vs iiuml)er of
Table 1. Room Temperature Static Flexui*alProperties of High cycles. It can be seen here that the vinyl ester cotnpos-
Glass Content SMC Mat’brials ites show a lesser drop in niodulus than the polyester
~ ~~~~~~~~

Nominal composites for an equal number of fatigue cycles.

glass Flentural Flextural The loss in modulus is, i n general, a result of the
Matrix content, strength modulus appearance and growth of damage zones at the areas of
material (Yo wt) (F si) (10” psi) tnaxiinum stresses. The damage is in the forin of mi-
~ ~ ~ ~ ~~~

Polyester* 30 35 000 1.8 crocracks which first appear o n the outer suiface, then
Polyester 65 55 660 2.30 extends into the interior of the specimens. The cracks
Vinyl ester 59 43,200 1.92 are. in most cases, oriented in ii direction iiorinal to the
EPOXY 63 64 970 2.10
principal stress, but do not always cover the whole width
‘Conventlonal SMC. of the specimen. As shown inFig. 5 a , the intensity of the
POLYMER COMPOSITES, JANUARY, 1987, Vol. 2, No. 1 19
 P. K. Mallick

Rooin temperature data for polyester composites are

1
a.) u a= 2 .56 uuc
b.) U, = 2 3 4 UUc also plotted inFig. 6 to indicate the effect of temperature
e.) u, = 2 .20 UUc
on fatigue life. Table 2 lists the fatigue properties for
40
three different matrices at or near 200,000cycles at a test
temperature of 220°F. Although based on limited data,
it indicates that vinyl esters have superior fatigue per-
forinance than polyesters and nearly equal to the epoxy
system considered. It should be noted, though, that the
epoxy system had significantly shorter fibers than the
other two systems.
Post-Fatigue Performance. After testing for a certain
number of cycles, middle sections of some of the
fatigued specimens were subjected to static flexure to
0 2 4 6 deterinine the residual strength of the material. These
LOG CYCLES
values are presented i n Table 3 . I n general, residual
Fig.4. Drop in modulus us number of cycles during fatigue at static strength decreases with increasing numl,er of cy-
130°F. cles i n contrast to some high perforniance unidirectional
composites where ininiinal loss i n static strength is often
observed (6). The niicrocracks formed on the suiface of
thc HShlC specimens duringfatigue loading act as sharp
notches. Some of these niicrocracks extend and/or join
to form a critical crack, thereby causing a premature
failure at a lower static load. Harris, et nl. (7)reported a
similar result on dough molding compound (DMC) ma-
terials where a large and rapid drop in strength was
observed after 100-1000cycles at cyclic stresses of about
75 percent of the ultimate strength of the material. At
Fig. 5. (0)Fatigue dumuge i n polyester specimens ut L- 15,000 lower stresses, however, the tests were carried to only
psi with (A) 2,000 (€3) 40.000 and (C) 367,000 cycles. ( b )Fatigue
damugt. in vinyl ester sprcimcns-(D)89.5.000 cycles at 15.000 -
psi, ( E ) 1.5,OOO cyclcs. and ( F ) 376,000 cyclcs (it t 20,000 psi.
1,000 cycles; even then, a loss in static strength was
observed, but only after an initial increase at about 10
cycles.
damage zone, indicated b y the number of cracks, in- Another observation to be niade from Table 3 is that
creases with number of cycles. Figure .% shows that for vinyl ester material retains more of its static strength
equal number of cycles, at the sanie stress level, there than polyester after fatigue. This is in accordance with
are iiiore niicrocracks on the pol>-esterspecimen than on the fact that there were more fatigue microcracking on
the vinyl ester specimen. polyester specimens than the vinyl ester ones.
Fatigue Life. Figure 6 shows a S-N plot of fatigue
stress level vs numher of cycles to failure for polyester DISCUSSION
and vinyl ester composites at 13O'F. \Vithin the range of From the data presented, it is evident that specimens
cycles investigated, these materials do not exhihit any with vinyl ester resin show better resistance to fatigue
fatigue limit and the numher of cycles to f'ailure gradu- than those with polyester. The reason for the dif€vreiice
ally increases with decreasing stress level. However. at can be found in the intensity of niicrocracks that form
loticycles, the endurance limit ofvinyl cster composites during cycling on the suifaces of these t\vo tlifferent
is approximately 60 percent higher than the polyester resins. The niicrocracks are prol)al)ly results of !loth
composites. niatrix cracking and failure of fil)er-matrix inteifucial
1.
OPOLYESTER, R.T.
1)onds ( 5 ) .
OPOLYESTER. 13OoF Since random filler composites are not its much fiber
AVlNYL ESTER, 130'F - dependent a s unidirectional composites, the properties

'.'
of the matrix resin and/or fiher-matrix inteiface are es-
\\ \
pectrtd to play a major role i n determining the inechani-
.6
cal properties of such mwteriils. Vinyl ester resins have
higher tensile strength and straiii-to-f~Iil~I1.e than poly-
4 ester. I n atltlition, the secont1;iry hytlroxyl group i n the
vinyl ester chain provides a site for hydrogen 1)ontling
with a similar gl-oup on the sutface of glass fihers. result-
ing i n good wet-out and adhesion to fillers (8). The
isophtalic polyester chain does not contain any such
hydroxyl group and therefore, adhesion aiitl wet-out can
LOG CYCLES he promoted only through the binder 0 1 1 tht. glass sur-
Fig. 6. Futiguc stra.i.s vs numbcr ofcyc1e.s in rec;cmc~d~cmrul face. Thus vinyl ester resins \vould prol)al)ly have a
futiguc. higher resistance t o debontling a i d matrix cracking than

20 POLYMER COMPOSITES, JANUARY, 1981, Vol. 2, No. 1

 Fatigue Churucteristics OJ. High Gluss Content Sheet Molding Compound (SMC) Muterids

Table 2. Fatigue Properties of High Glass Content SMC at 220°F

Residual
properties
Percent Percent
Matrix Stress UTS No of drop Strength Modulus
material level ai 220°F cycles in modulus (Psi) (10” psi)

Polyester 9,000 18.8 175,000 12.7

Vinyl ester 10,000 25 201,260 5 41,620 2.02
12,500 31 200,000 10.9 35,380 1.99
15,000 37 210,000 10.6 41,880 2.205
EPOXY 10,000 21.7 202,150 0 56,960 1.97
12,500 27 201,720 11.2 50,390 1.94
15,000 32.5 200.000 20.1 36,850 1.47

Table 3. Residual Strength at Room Temperature after Fatigue

Test Percent
temp cr) Matrix aa UTS N uRes -
uRes
CF) material (Psi) at T (c yc Ie s) (Psi) ault

130 Polyester 9,030 20 1.72 x l o h 39,010 0.70

10,ODO 22.5 1.63 x l o 6 57,640 1.03
12,500 28 1.30 x lo6 40,710 0.73
15,000 34 2,000 53,290 0.96
15,COO 34 18,000 44,850 0.80
15,COO 34 48,000 32,390 0.58
15,COO 34 367,000 36,900 0.67
20,C’OO 45 40,000 35,990 0.65
25,(100 56 4,110 36,460
Vinyl ester 15,(100 38 895,000 38,660 0.89
20,(100 51 15,000 42,940 1.99
20JtOO 51 55,020 44,090 1.02
20,000 51 376,000 33,660 0.78
25,000 64 40,000 39,280 0.91
30,000 76 10,000 40,160 0.93
320 Polyester 9,000 25 200,400 47,800 0.88
9,000 25 438,000 42,670 0.77
Vinyl ester 9,1100 36 306.890 42,440 0.98
EPOXY 9,300 28 41 0,000 54,950 0.85

polyesters. Pires et a1 found evidence of better adhesion ture and number of cycles.
with vinyl esters in their work on fiber-matrix iiiterac- At low fatigue stress level (30-50 percent of the
tion effects 011 fracture toughness on structural coinpos- UTS), the fatigue damage is progressive, rather
ites (9). than catastrophic.

CONCLUSION ACKNOWLEDGMENT
B asc d o 11 the 1i in it ed expe r i i n e 11t a1 ob e I-vat io 11s The author wishes to thank Dr. S. Newman of the
presented, following conclusions can be made regarding Ford Motor Company for his valuable suggestioiis dur-
the fatigue characteristics of high a s s content SXIC ing the preparation of this paper.
materials:
Up to 10”cycles, no fatigue limit could be observed REFERENCES
with any of the two systems coilsidered.
1. J. Maaghul and E. J. Potkanowicz, 31stAnnual Tech. Confer.,
In rairdoni fiber composites, I I atrix plays an im- SPI, 7-C (1976).
portant role in determining the fatigue characteris- 2. R. S.Sapp, Reinforced Thermoset Press Molding Committee
tics of the material. More spec- fically, viiiyl ester Mtg., SPI (1976).
inatrix has shown fatigue propel ties which are bet- 3. R. E. Thomas and J. H. Enos,33rdAnnual Tech. Confer., SPI
(1978).
ter than polyester and may be equivalent to the 4. P. K. Mallick and N. Raghupathi, Symposium for Thermosets,
epoxy systein considered here. SPE RETEC (1978).
Fatigue damage i n these materids is manifested by 5. M. J. Owen, Compos. Muter., 5 , 313 (1974).
a cluster of Inicrocracks, whow severity depends 6. J. B. Sturgeon, 28th .4nnual Tech. Confer., SPI, 12-B (1973).
on the choice of matrix. As a result of such mi- 7. B. Harris, ‘4.0. Ankara, D. Cawthorne and S. M. T. Bye,
Composites, 185 (1977).
crocracks, the static strength of the inaterial is 8. Derakane Vinyl Ester Resins for Corrosion Resistance, Dow
reduced significantly. The inagnitude of damage Chemical Co. (1975).
tlcyxnds 011 the fatigut. stress level, test tempera- 9. D. G. Pires, et d., AMMRC CTR 75-15 (July 1975).

POLYMER COMPOSITES, JANUARY, 1981, \‘ol. 2, No. I 21