EFFECTS OF REAL-TIME THERMAL AGING ON

GRAPHITE/POLYIMIDE COMPOSITES*

J.F. Haskins and J.R. Kerr

General Dynamics Convair Division

As part of a program to evaluate high-temperature

advanced composites for use on supersonic cruise

transport aircraft, two graphite/polyimide com-

posites have been aged at elevated temperatures for

times up to 5.7 years. Work on the first, HT-S/710

graphite/polyimide, was started in 1974. Evalua-

tion of the second polyimide, Celion 6000/

LARC-160, began in 1980. Baseline properties are

presented, including unnotched and notched ten-sile data as a function of temperature, compres-

sion, flexure, shear, and constant-amplitude

fatigue data at R = 0.1 and R = - 1.

Tensile specimens were aged in ovens where

pressure and aging temperatures were controlled

for various times up to and including 50,000 hours.

Changes in tensile strength were determined and

plotted as a function of aging time. The HT-S/710

composite aged at 450F and 550F is compared tothe Celion 6000/LARC-160 composite aged at

350F and 450F. After tensile testing, many of the

thermal aging specimens were examined using ascanning electron microscope. Results of these

studies are presented, and changes in propertiesand degradation mechanisms during high-temper-

ature aging are discussed and illustrated using

metallographic techniques.

INTRODUCTION

Advanced composites will play a key role in the

technology emerging for the design and fabrication

of future supersonic vehicles.

Research and development during recent years

has led to advances in fabrication techniques and

characterization of short-time properties and has

* Sponsored by the National Aeronautics and SpaceAdministration, Langley Research Center, Hampton,Virginia, under Contract NAS 1-12308, monitored byMr. Bland A. Stein.

315

provided limited supersonic flight experience forthese materials.

However, information on the effects of long-

time exposure to service environments represen-

tative of supersonic cruise aircraft on composite

materials has not generally been available. An ex-tensive program to generate such information has

been in progress at General Dynamics ConvairDivision under NASA Contract NAS1-12308 since

1973 (Ref. 1).

Figure 1 illustrates the overall NASA study

from which the material for this paper was taken.

Changes in mechanical properties that occur over

very long periods of time are being determined for

ambient and thermal aging conditions and for ran-

dom cyclic loading with cyclic temperature varia-tions. These latter tests, the flight simulation ex-

posures, are intended to provide data on the effect

of 10,000, 25,000 and 50,000 hours of simulated

supersonic flight service on residual properties of

the composites. The purpose of the thermal aging

study (conducted at constant temperature without

load) was to assist in understanding the results ofthe more complex flight simulation program.

While flight simulation tests are still in progress,

the original thermal aging specimens have com-

pleted the required 50,000 hours of exposure, andthe residual strength data is now available. The sec-

ond polyimide, introduced later in the program,has completed thermal aging tests out to 10,000hours.

This paper presents the results of these aging

tests for two graphite/polyimide systems. Earlier

work on thermal aging of HT-S/710 (Ref. 2) iscompared to more recent work on Celion 6000/

LARC-160. All exposures were conducted at am-

bient pressure. Previous work had Shown a direct

correlation of thermal aging and oxygen pressure

on residual strength of resin-matrix composites

(Ref. 1). In this paper, ambient pressure was

https://ntrs.nasa.gov/search.jsp?R=19860001816 2020-03-10T12:28:40+00:00Z

chosen to compare the two systems, since most of

the data has been generated at ambient pressure.

In actual use, a supersonic cruise vehicle

would be at a very high altitude during much of thetime at which the structure has reached its max-

imum temperature. Thus, if reduced oxygen pres-sure were taken into account, the composite mate-

rials could likely be used at high temperatures forlonger periods of time.

Table 1. Material systems.

MaterialSystem Cellon 60001LARC- I eo HT.8/710

Vendor Flbedte Hece.,ukm

Orientation [0 =:1:45, ]= [0 • :t:45" Is[o"+45");2 (o"±46.]sa

Fiber Content 67% 70%

Specific Gravity 1.56 1.48

EXPERIMENTAL

The graphite/polyimide composite systems em-ployed in this program were HT-S/710 and Celion

6000/LARC-160. Six-'and twelve-ply [0°+45]crossplied laminates of each system were fabricated

by Convair Division from vendor-supplied prepreg

material using conventional autoclave processing

methods. Quality assurance testing was conducted

on both the as-received prepreg and the fabricated

laminates. These acceptance tests included fiber,

resin, and volatile contents, resin flow, and process

gel for the prepreg material and ultrasonic C-scan,

fiber content, specific gravity, and metallographicexaminations for the panels. Table l lists informa-

tion for the two composites.

Test specimens were cut from the panels usinga diamond impregnated saw. Details of the

specimen configurations for the various types of

tests are presented in Table 2. Polyimide-quartzdoublers were bonded to all but the flexure and

short beam shear specimens using HT-424, amodified epoxy-phenolic film adhesive with an

aluminum filler. For the thermal aging specimens,the doublers were attached after exposure.

Conventional test methods were used for the

tensile, compression, flexure, and shear tests. The

notched tensile specimens contained a center hole

0.25 inch in diameter, giving a theoretical stress

concentration (Kt) of 2.43. Compressive strength

was determined using a Celanese-type test f'Lxture

with an 18-ply specimen prepared by bonding"together three six-ply panels. A similar 18-plyspecimen was also used for the short beam shear

tests.

Constant-amplitude fatigue tests were con-

ducted at a constant frequency of 30 Hz. Forelevated temperature tests, clamshell and ring fur-

naces were used. Temperature was monitored by a

thermocouple attached to the specimen at the

center of the gage section. A total" of nine

specimens was used to obtain an S-N curve for anygiven set of conditions. Test conditions were:

• Cycles: 103 to 107

• R values: - 1 and 0.1

• . Temperature: 75 and 350 or 450F

• Specimen configuration: unnotched andnotched

Thermal aging exposures were conducted in

specially constructed aging furnaces similar to the

sketch in Figure 2. The heater plates consist of in-sulated wire sandwiched between two thin

aluminum plates. All furnace temperatures were

equilibrium-controUed; i.e., a constant amount of

power was supplied to the heaters. The various

aging temperatures were generally maintained to

+ 5F with infrequent excursions to a maximum of+ 10F.

Table 2. Details of test specimens

Length Width OoublmlSpecimen Type (in.) (in.) Plies Required

Unnotched Tensile 9 0.5 6 YesNotched Tensile 9 I 6 YesCorn pressive 5.5 0.25 18 YesUnnotched Fatigue 9 0.5 6 and 12 YesNotched Fatigue 9 1.0 6 and 12 YesFlexure 3 0.5 12 NoShort Beam Shear 0.6 0.25 18 NoThermal Aging 9 0.5 6 Yes

316

The aging temperatures were: Celion 6000/

LARC-160, 350F and 450F; and HT-S/710, 450and 550F.

The procedure followed for the exposures was

to cut the specimen blanks to final size, heat at

250F for 24 hours to remove absorbed moisture,

and load into the aging furnaces. The specimens

were supported at their ends by narrow strips ofstainless steel so that almost the entire surface was

exposed to the air atmosphere within the furnaces.

At the required time intervals, the ovens were shut

down, Opened, and the specimens removed and

stored in a desiccator until doubler bonding and

tensile testing. Residual strength testing was per-formed at 350F for the Celion 6000/LARC-160

and at 450F and 550F for the HT-S/710. The data

points shown in the residual strength-versus-time

plots are averages obtained from three tensile

specimens per test condition.

After tensile testing, many of the thermal ag-

ing specimens were sectioned and mounted for

study using a scanning electron microscope. These

studies were intended to detect changes that occur-

red in the composites during exposure to assist in

identifying degradation mechanisms.

A more detailed account of the fabrication,

quality assurance, baseline testing, and thermal ag-ing procedures can be found in Reference 1.

RESULTS AND DISCUSSION

positioned to 350F; the resulting Weibull distribu-

tion fit is shown in Figures 5 and 6. Each polyimideis compared for both the unnotched and notchedconditions. The Weibull and normal distribution

parameters for each of the four curves are listed inTable 3.

The data was pooled in this manner to provide

the maximum use of the tensile data in setting load

levels for the flight simulation tests shown in Figure1. The loads in Table 3 can be converted to stress

by dividing by the laminate thickness (approx-imately 0.03 inch).

Other baseline properties are given in Table 4.

The data given compares flexural strength, shear

strength, and compressive strength and modulus of

HT-S/710 and Celion 6000/LARC-160. As was the

case for tensile strength, the Celion 6000/

LARC-160 composite is considerably stronger.

Results of the fatigue testing portion of the

program are presented in Figures 7 through 10.

These S-N plots compare the effects of stress ratio,R, notch configuration, and temperature on the

fatigue strength of the two graphite/polyirnide

systems.

The data shows the fatigue life for a stressratio of R = - 1 to be much lower than for R = 0.1.

The effect of test temperature or the presence of a

center notch (hole with Kt = 2.43) was slight com-pared to the stress ratio effect.

Table 4. Baseline mechanical property data.To establish sufficient baseline data, a number of

tensile tests were performed. In Figures 3 and 4, the

tensile properties for the two polyimide composites Flexural strength (ksi)75F 94 171 137 196

are shown for both the unnotched and notched 350F 79 127 124 145configurations. Tests were performed at a number sheers_,0th (_)

75F 5.4 7,0 8.8 9.0of temperatures between - 67 and 650F. Each data 350F -- -- 7.5 8.8

Compressive strength (ksl)point represents the average of at least three tests. 75F 55 -- a5 --A regression line was fitted to the data, as shown in Compressivemodulus(msi)

75F 6.7 -- 7.5 --Figures 3 and 4. Each set of data points was super-

HT-$/710 Cellon $O00/LARC-160

(O-deg :=iS) (O-deg) (O-deg=4S) (O-deg)

16033232-7

Table 3. Comparison of Weibull and normal distribution parameters for tensile loads of two polyimide composites.

Welbull

Material & Walbull /_ Standard Average Coof. of No. of

Specimen Configuration (z (Ib/In.) Deviation (Ib/In.) Variation Data Points

Celion 6000/LARC- 160 28 3,245 147 3,180 0.046 35

Unnotched

Celion 6000/LARC-160 14 2,377 192 2,290 0.083 35

Notched

HT-S/710 Unnotched 13 2,396 237 2,301 0.103 26

HT-S/710 Notched 12 1,675 145 1,611 0.090 29

317

The lines on each of the four graphs are

regression lines drawn by computer. Differences in

slope or slight changes in position are not con-

sidered significant. As for the mechanical proper-ties, in general, the Celion 6000/LARC-160 S-N

curves are about 25o7o higher than those forHT-S/710.

Residual tensile strength of the HT-S/710

system is plotted as a function of aging time inFigure 11. Curves for aging at 450 and 550F in 14.7

psi air are included. The material aged at 450F

showed very little change in tensile strength for

times up to 25,000 hours. However, when the agingtemperature was raised to 550F, significant

strength decreases were observed after 5,000 hoursof exposure. After 50,000 hours, almost no resin

was left in the specimens, only graphite fibers. The

specimens were badly warped and delaminated,

and were no longer suitable for tensile testing. Alltensile tests were performed at 350F after the in-

dicated aging times. Each point consists of the

average of three tensile coupons.

Residual tensile strength of the Celion

6000/LARC-160 system is plotted as a function of

aging time in Figure 12 for ambient air pressure. As

in the case of the HT-S/710, each point represents

the average of three tensile tests. Agingtemperatures were reduced for the second

polyimide to 350 and 450F. There was not enough

difference in the data for the two temperatures to

draw separate curves. As shown in Figure 12, thereis some loss in strength for aging times as short as

500 hours. After 10,000 hours of exposure at 450F,

a comparison of the two graphite/polyimidesystems shows that each has about the same

residual tensile strength. Based on the shape of the

aging curves, however, the HT-S/710 systemwould appear to be superior for time periodsgreater than 10,000 hours.

Scanning electron microscope photo-micrographs in Figures 13 through 15 show results

of metallographic examination of the graphite/polyimide aging specimens. For the HT-S/710

system (Figure 13), specimens aged at 450F ex-hibited no oxidation or matrix degradation effects

for the first 25,000 hours. After 50,000 hours,relief polishing around the graphite fibers, increas-

ed porosity, and more fiber-matrix separation were

observed. Raising the aging temperature to 550F

greatly increased the degree of matrix degradation

of the HT-710 system. Figure 13 shows that, in

318

one-atmosphere air, 10,000 hours at 550F has a

slightly greater e:tect on the microstructure than

50,000 hours at 450F.

For the Celion 6000/LARC-160 system,metallographic effects were much more evident for

the thermal aging specimens. At lower magnifica-

tions, the degree of microcracking and relief

polishing was found to be related to both the

temperature and length of time of aging. This can

be seen in Figure 14. As pointed out earlier,

however, the results of the tensile tests did not in-

dicate a significant temperature effect for at least

10,000 hours of exposure. At higher magnification,

evidence of oxidation, similar in nature to that first

observed in the A-S/3501 graphite/epoxy system

(Ref. 3) was found. Figure 15 shows the results of

the higher-magnification examinations. When the

polyimide resin matrix oxidized, it was more proneto crumble and resulted in an increased amount of

relief polishing around the individual fibers. After

10,000 hours of aging at both temperatures, con-

siderable oxidation has occurred in the outer plies

but almost none in the center plies. This would beexpected for an oxidation mechanism where attack

begins at the outer surfaces and proceeds inward.Careful examination of Figure 15 also reveals an

effect of temperature on oxidation where the

degree of relief polishing around the graphite fibersis slightly greater at 450F than at 350F.

CONCLUSIONS

The mechanical properties of the Celion 6000/

LARC-160 system were, in general, at least 25°7o

stronger than the HT-S/710 system. The fiber-

dominated properties of the two polyimides were

nearly independent of temperature for the regionexamined. The matrix-dominated properties show-

ed a slight decrease in strength as the temperaturewas increased.

The fatii_ue results clearly show that thestrength at 10 / cycles for a stress ratio of R = - 1 is

50o70 of that for R = 0.1. Differences between

room and elevated temperature S-N curves were

not statistically significant. Also, notching had lit-

tle effect on the fatigue results.

During thermal aging of HT-S/710 at 450F in

an air atmosphere, there was little change in tensile

strength for times out to 25,000 hours. At 550F,

the strength was reduced significantly by 5,000hours. Aging of Celion 6000/LARC-160 at both

350Fand 450F resulted in a 2007o loss in strength

after 5,000 hours and a 30°7o loss after 10,000

hours. Tests on these specimens are being con-

tinued to 25,000 hours.

Metallographic studies have shown an increase

in microcracking with aging time for the Celion

6000/LARC-160 system. These examinations havealso revealed evidence of conventional oxidation

damage to the matrix with the effect initiating at

the surface and progressing inward with time. Asimilar effect had been observed earlier for

A-S/3501 graphite/epoxy. For the HT-S/710

system, metallographic examinations showed very

little during the early stages of exposure. For the

longer times, increased porosity and fiber-matrix

separation accompanied by numerous fine cracksat the fiber-matrix interface were revealed. How-

ever, visual effects starting at the edges and movinginward as seen in the Celion 6000/LARC-160 and

A-S/3501 systems were not observed.

REFERENCES

1. Kerr, J.R. and Haskins, J.F., "'Time-

Temperature-Stress Capabilities of Composite

Materials for Advanced Supersonic

T_.chnology Application," General DynamicsConvair Division, Report No. NASA

CR-159267 (GDC-MAP-80-001), Prepared for

NASA-Langley Research Center, April 1980.

2. Kerr, J.R. and Haskins, J.F., "Effects of

50,000 Hours of Thermal Aging on

Graphite/Epoxy and Graphite/Polyimide

Composites," AIAA/ASME/ASCE/AHS23rd Structures, Structural Dynamics, and

Materials Conference, New Orleans, Loui-

siana, May 1982.

3. Haskins, J.F., Kerr, J.R., and Stein, B.A.,

"Flight Simulation Testing of Advanced Com-

posites for Supersonic Cruise Aircraft Applica-

tions," AIAA/ASME 18th Structures, Struc-

tural Dynamics, and Materials Conference,

San Diego, California, March 1977.

319

¢o

1o

o_c

u)

Aging of typical composite at max useTemperature 505K (450F) typical for B/AI

Ambient aging roomThermal

Cyclic thermalCyclic load/thermal aging

I0 10 20 30 40 50

Thousands of hours

Test en_ronments

218K (-67K) to 700K (800F)Single-variable testsMulUple-variable testsMaterialsProven composite systemsTested in severe supersonic environment

"B/E, G/E, B/PI, G/PI, BIN

Types of testsLaboratory tests

Tensile-fatigue-creep-agingFlight simulation

Temp profile-random loadingAnalysisLamination theoryWeerout modelDamage assessmentData base

ExistingExtendingConfidence ,6o33z32-2

Figure 1. Characterization of composite material for up to 50,000 hours of supersonic cruise aircraft environment.

PLYWOOD INSULATION

2 PSI

SPACERS

HEATER POWER

l_gure 2. Thermal aging furnace configuration.

PLATE

NTAINER

IRE SPECIMENS

ATM SPECIMENS

PLATE

0" THERMOCOUPLE

663217-72

320

120

0lO0 u

80

4'

ULTIMRTE TENSILE........ " .... 0

STRENGTH (KSI) ............

40---i o CELION6000/LRRC160

L;14_'-g7_fD ................

41''0

20

-loo 0 loo 200 300 400 r,oo eoo 7ooTEHPERRTURE F 1(1033232.3

Figure 3. Baseline tensile results comparing unnotched Celion 6000/LARC-160 and HT-S/710 (0+ 45) s polyimidecomposites.

120

loo

0 I

ULTIMBTE TENSILE _--_'-"""

STRENGTH (KSI) /// o

w ..2_..._

o CELION6000/LRRC160

"_"R_':gT_'fi_............

J

0

Y

20

-Ioo o loo aoo 3oo 400 soo coo 7ooTEMPERRTURE F' 16033232-4

Figure 4. Baseline tensile results comparing notched Cefion 6000/LARC-160 and HT-S/710 (0+ 45) s polyimide com-posites.

321

PERCENT F_JLED

I I I I I I

[] CELION 6000/LRRC 160

-_'-_=_?_i'O-....."...............

I00

9oq

eo ql

D7o

b

so J

_o _

3O

20 .

/o

I I I I I I I I I _4"_-_= I I I I I I I I I I I | I I I

I It I

i

i[]IE

,1

i_ iiii iiii iiii I|,11 ,,,,

ZOO01200 iiO0 1600 1800 2000 2200 2400 2SO0 2SO0 3000 3200 3400 3SO0 3800 4000LORD IN POUNDS/INCH 16033232.5

Figure 5. Comparison of unnotched tensile data for HT-S/710 and Celion 6000/LARC-160 (0+ 45) s (test pointssuperpositioned to 350F for pooling).

PERCENT FAILED

I00

_o _ ION .

_---_'r-__ro---_...............:80 _ oOi I Y70

_ _60 _ -_

I I

so_ _ ]. (4o

- f3o _

20 1 _ ';'-' !

f

-i_o ......_'.,,.,,_..........................................I000 1200 liO0 i600 1800 lO00 2300 2400 2600 2800 3000 3200 3100 3600 3800 iO00

LORD IN POUNDS/INCH 16033232.6

Figure 6. Comparison of notched tensile data for HT-S/710 and Celion 6000/LARC-160 (0 + 45) s (test points allsuperpositioned to 350F for pooling).

322

9O

MAXIMUM 5TRESS

PER CYCLE (KSI)

60

7O

6O

SO

v

R=O. 1 75F

]_D

"" • _ "_ -.=___. ix2o -- <> R=-I 150F |

J <>

1o ........ ] ,, ,l,,--_,, ........................................

I0' Id I_ 10' 10 s I0' lff 10'

NUMBER OF" CYCLE516033232-10

Figure 7. Axial fatigue data for HT-S/710 (0 + 45) s G/PI, unnotched, R = O. 1, R = - 1.

9O

8O

7O

6O

MAXIMUM STRESS

PER CYCLE {KSI} so

Io

3o

2o

(

0r

[] R=0.1 75P

.........................

o R=0.1 450F

_ R=-I 450F

-ILl

0

• -.-...__ °_

D

v 0-"

II0 I i l I l ill i l i IIIIll l I I Illll i I I I Iiii

10' 102 10: I0' 10sNUMBER OF" CYCLES

i i i illil I i i i till

0I Id

i i ,_,I,,

o'16033232-11

Figure 8. Axial fatigue data for HT-S/710 (0+45)s G/PI, notched, R =0.1, R = - 1.

323

9O

8O

7O

6OMAXIMUM STRESS

PRR CYCLE {KSI]5O

_0

3O

2O

I0

I0_

o R--0.1 350F" --_

I(3= 103 10' 10= 1_ 10_ 10jNUMBER OF CYCLES

16033232-8

Fi&ure 9. Axial fatigue data for Celion 6000/LARC-160 (0 + 45) s G/P/, unnotched, R = O.1, R = - 1.

go

8o

7o

6o

MRXIMUM STRESS so

PER CYCLE (KSI)

40

I.

.............. [3 0

.........................o R=0.1 350F30 --

A R---.-1 7SF

2o --- o R=-I 350F

10 J J , i,l,,i I

101 10a

_.. 0

.... _

o

(

<>

I I I IIXJ ! I i i i|1 I I i I i illJ 1 ! I Illll J I t elltl l ¢ i i llll

io' lo' ld' lo' id IdNUMBER OF CYCLES 10033232.0

Figure 10. Axial fati&ue data for Celion 6000/LARC-160 (0 ± 45) s G/P/, notched, R = O. 1, R = - 1.

324

1oo

ULTIMATE TENSILESTRENGTH (KS])

go

6o

4o

i2O 0 AFTER AGING AT 450F

"_°"_'P_-_E""h'_'i_"'_"_ _;b'P"...............

10= 10'TIME IN HOURS

0 I I l i i i ii I i I i i i ii

10= 10:

)

_1 i _q_l i I

16033232-12

Figure 11. Tensile strength at aging temperature of liT-S�710 (0 + 45) s G /PI after thermal aging in 14. 7psi air at in-dicated temperatures.

IiO

ULTIMATE TENSILE

STRENGTH IKS] J

6o

7(I

6O

0 RFTER AGING RT 350F

lO * * i l * l ,l I l I i I III I i i i i i ii i I i i i iii i I I i 1 i i

10= 10= 10= 10= 10' 10=TIME IN HOURS 16033232-13

FigUre 12. Tensile strength of Celion 6000/LARC.160 (0 + 45) s G/PI at 350F after thermal aging in 14. 7psi air at theindicated temperatures.

325

As received 7 1 10,000 hours, 450F, 14.7 psi t

10,000 hours, 550F, 14.7 psi

16033232-1 5

50,000 hours, 450F, 14.7 psi

Arrows indicate specimen edges

Figure 13. HT-S/710 graphite/polyimide after thermal aging at indicated conditions (100 x).

As received

5,000 hours, 350F, 14.7 psi 5,000 hours, 450F, 14.7 psi

10,000 hours, XOF, 14.7 psi 10,000 hours, 450F, 14.7 psi 16033232- 1 7

Figure 14. Celion 6000/LA RC-160 graphite/polyimide after thermal aging at indicated condirions (13 X ).

326

10,000 hours, 350F, 14.7 psi Inner ply

7 0,000 hours, 350F, 14.7 psi Outside ply

10,000 hours, 450F, 14.7 psi 10,000 hours, 450F, 14.7 psi

Inner ply Outside ply 16033232-1 6

Figure IS. Celion 6000/LARC-160 graphite/polyimide after fhermal aging at indicated conditions (1000 x).

327