CREEP AND FATIGUE OF LONG GLASS REINFORCED

POLYPROPYLENE COMPOUNDS
Charles C. Mentzer
Peter H. Foss
General Motors

Abstract

This paper details the creep and fatigue characteristics of several long glass
reinforced polypropylene compounds that have been used in the development of a
thermoplastic liftgate. Both the creep and fatigue data show stress levels associated
with a composite part should be limited to 30 % of the failure strength at the maximum
use temperature. Parts where there is considerable flow and orientation of these
composites should limit stress levels in the cross-flow direction to less than 20 % of the
failure strength in that direction.

Introduction

Polypropylene reinforced with various types of long glass fibers have been used in
several automotive applications especially in Europe [1]. A paper [2] presented at the
2004 SPE Automotive Composites Conference documented the development of a
liftgate fabricated using long glass fiber reinforced Polypropylene (PP). In normal use
the liftgate subjects these PP compounds to both creep and fatigue. The gas struts for
the liftgate and glass introduce a constant load which could lead to localized creep at the
attachment points. Vibrations during operation and opening and closing of the liftgate
could lead to fatigue failure. Creep and fatigue of long glass reinforced PP were
investigated to determine what stress would lead to creep rupture and fatigue failure.
Fatigue of injection molded glass reinforced thermoplastics has been investigated by
many researchers [3, 4]. Their results indicate fatigue data could be normalized by the
initial modulus of the test specimen [4]. The current fatigue experiments were
undertaken to determine if the data could be normalized by the failure stress obtained
from a standard tensile test. These long glass reinforced PP compounds have much
lower elongation at break than the injection molded resins. (1 to1.5% versus 2 to 3%)
which means the initial modulus and overall modulus based on the failure stress are
similar.

Molding
Four long glass reinforced polypropylene (PP) materials were used in the creep and
fatigue experiments. Two GMT Azdel products R401 and C467 were used as
standard materials. R401 consists of PP containing 40% by weight random continuous
glass fibers. C467 consists of PP containing 35% by weight 12.5 mm random chopped
fibers and 10 % by weight talc. These two materials were molded into 22 by 29 cm
plaques. To minimize flow orientation during molding, the plaques were molded using
an 80 % initial charge (20 by 28 cm). Two configurations of tensile bars routered from
these plaques, one with a 12.5 mm reduced section and another with a 38 mm reduced
section.
The creep experiments also were conducted using a PP compound containing 40 % by
weight random chopped 25 mm glass fibers. Tensile bars (12.5 mm) were routered from
45 by 60 cm plaques molded at Composite Technologies, Corp. (CTC) in Dayton, OH.
The plaque configuration and charge placement are shown in Figure 1.
45 mm

Flow

60 mm

Flow

Figure 1. CTC liftgate material plaques showing charge placement and flow direction.

This charge placement resulted in orientation in the 60 cm direction. Tensile bars

were cut parallel and perpendicular to flow. The liftgate material used in the fatigue
experiments was molded at Composite Product Inc. (CPI) in Winona, WI using 40 % by
weight 12.5 mm glass fibers. The tensile bars were cut from the window portion of the
outer upper panel of the 360 liftgate (Figure 2). Figure 3 shows how the 38 mm tensile
bars were routered from the window section.
 Window Opening Trim Line

Flat Region

Pressure Transducers

Figure 2. Upper outer liftgate panel showing region used for tensile bars.

Figure 3. Tensile bar layout for 38 mm CPI liftgate material.

Results

Tensile data for the four materials is shown in Table 1. Only the GMT R401 tensile
properties were measured at 80°C. Five samples were used for each tensile test.
Creep and fatigue stress levels were set based on the tensile strength for each material.
 Table I
Tensile Properties of Glass Filled Polypropylene used in Creep and Fatigue
Material Tensile Strength, Mpa Standard Deviation
GMT R401, 23°C 124 7.8
GMT R401, 80°C 100 9.1
GMT C467, 23°C 83.1 2.1
CTC Liftgate, Flow, 23°C 71.4 15.1
CTC Liftgate, Cross-Flow, 23°C 38.3 14
CPI Liftgate, Flow,23°C 87C 19

Creep

All creep experiments were conducted at Datapoint Labs in Ithaca, NY. Standard
ASTM 638 (12.5 mm reduced section) tensile bars of the CTC liftgate material were
tested first. Creep results for the CTC liftgate material in the flow direction at 23°C are
shown in Figure 4.

0.6

0.5 10 Mpa
10 Mpa
0.4
15 Mpa
Strain, %

15 Mpa
0.3
20 Mpa
0.2 20 Mpa
30 Mpa
0.1 30 Mpa

0
0.01 0.1 1 10 100 1000 10000
Time, hrs

Figure 4. Tensile creep curves for CTC liftgate material in flow direction, 23°C.

Two samples were tested at each load level. The lowest load level of 10 MPa was
14% of the failure stress and the highest load of 30 MPa was 41% of the failure stress.
The 10, 15 and 20 MPa samples increase linearly with log time, where as, the 30 MPa
samples show a nonlinear response to time, indicating these samples may stress
rupture with increased time. One way to get a better measure of the potential for stress
rupture is to plot the data as an isochronous stress-strain curve, Figure 5 [5].
 35

30

Stress, Mpa
25 1000 hr
100 hr
20
10 hr
15 1 hr
1 min
10

0
0 0.1 0.2 0.3 0.4 0.5 0.6
Strain, %

Figure 5. Isochronous stress-strain curves for CTC liftgate material in flow direction,
23°C.

If the time plots are linear with stain, stress rupture is not imminent, however, if the
time curve is nonlinear, stress rupture is probable with greater load and time. The
Isochronous curves for the CTC liftgate material are not smooth; however, the 1000 hr
curve is slightly more nonlinear than the other time curves. Longer time and higher
stress levels could lead to stress rupture.

The tensile creep curves for the CTC liftgate material in the cross-flow direction at
23°C are shown in Figure 6 and the isochronous curves in Figure 7.

9
8
7 1000 hr
Stress, Mpa

6 100 hr
5 10 hr
4
1 hr
3
2 1 min
1
0
0 0.2 0.4 0.6 0.8 1
Strain, %

Figure 6. Tensile creep curves for CTC liftgate material in cross-flow direction, 23°C.
 9
8
Stress, Mpa 7 1000 hr
6 100 hr
5 10 hr
4
1 hr
3
2 1 min
1
0
0 0.2 0.4 0.6 0.8 1
Strain, %
Figure 7. Isochronous stress-strain curves for CTC liftgate material

Based on the isochronous curves, it appears the strain measurements at 6 MPa are
incorrect. Neglecting the data at 6 Mpa, Figure 8, only at the lowest times (7 hrs or less)
do the curves appear to be linear, indicating all the cross-flow samples will stress rupture
with higher stress and longer creep times.

9
8
7 1000 hr
6
Stress, Mpa

100 hr
5 7 hr
4 1 hr
3 1 min
2
1
0
0 0.2 0.4 0.6 0.8 1
Strain, %

Figure 8. Isochronous stress-strain curves for CTC liftgate material in cross-flow

direction, 23°C, 6 MPa data removed.
 The 80°C tensile creep flow and cross-flow data are shown in Figures 9 and 10 and
the isochronous plots in Figures 11 and 12. The two cross-flow samples at 5 MPa
(Figure10) show considerable scatter, resulting in inflection in the isochronous plot
(Figure 12). At all the load levels used, the 80°C isochronous plots indicate stress
rupture has occurred or will occur with increasing time.

1.6
6 Mpa
1.4 6 Mpa
1.2 9 Mpa
1 9 Mpa
Strain, %

12 Mpa
0.8
12 Mpa
0.6 15 Mpa
0.4 15 Mpa
0.2 18 Mpa
18 Mpa
0
0.01 0.1 1 10 100 1000 10000
Time, hrs

Figure 9. Tensile creep curves for CTC liftgate material in flow direction, 80°C.

0.8
0.7
0.6 2.5 Mpa
0.5 2.5 Mpa
Strain, %

3.75 Mpa
0.4
3.75 Mpa
0.3 5 Mpa
0.2 5 Mpa
0.1
0
0.01 0.1 1 10 100 1000 10000
Time, hrs

Figure 10. Tensile creep curves for CTC liftgate material in cross-flow direction, 80°C
 20
18
16
14 1000 hr
Stress, Mpa

12 100 hr
10 10 hr
8 1 hr

6 1 min

4
2
0
0 0.2 0.4 0.6 0.8 1
Strain, %

Figure 11. Isochronous stress-strain curves for CTC liftgate material in flow direction,
80°C

4 1000 Hr
Stress, Mpa

100 Hr
3 10 Hr
1 Hr
2 1 Min

0
0 0.2 0.4 0.6 0.8
Strain, %

Figure 12. Isochronous stress-strain curves for CTC liftgate material in cross-flow
direction, 80°C.

A second set of tensile bars to Datapoint Labs. These tensile were made using GMT
R401, a continuous random glass mat containing 60% by weight PP. Both the liftgate
material and the GMT R401 material contain the same % glass, however, the GMT
R401 glass is much more uniform in concentration throughout the tensile bars. Four
samples were used for each load level at 23°C and 80°C. The tensile creep curves for
the GMT 401 material are shown in Figures13 and 14 and the isochronous plots in
Figures 15 and 16, the data indicates that over the load range of 15 to 40 MPa stress
rupture of the GMT material is not imminent. Only at 40 MPa (32% of the failure stress)
and 23°C is there any nonlinearity (Figure 15).
 1.6
1.4

Average Strain, %
1.2
1 15 Mpa
20 Mpa
0.8
30 Mpa
0.6 40 Mpa
0.4
0.2
0
0.01 1 100 10000
Time, hrs

Figure 13. Tensile creep curves for GMT R401, 23°C

1
0.9
0.8
Average Strain, %

0.7
10 Mpa
0.6
15 Mpa
0.5
25 Mpa
0.4
30 Mpa
0.3
0.2
0.1
0
0.01 1 100 10000
Time, hrs

Figure 14. Tensile creep curves for GMT R401, 80°C

 45
40
35
1000 hr
Stress, Mpa

30
100 hr
25
10 hr
20
15 1 hr
10 1 min
5
0
0 0.5 1 1.5 2
Strain, %

Figure 15. Isochronous stress-strain curves for GMT R401, 23°C.

35

30

25
1000 hr
Stress, Mpa

20 100 hr
10 hr
15
1 hr
10 1 min

0
0 0.2 0.4 0.6 0.8 1
Strain, %

Figure 16. Isochronous stress-strain curves for GMT R401, 80°C,

The creep response for GMT R401 and CTC liftgate material in the flow direction are
similar (Figures 4 and 13); however the strain at a given load is about two times higher
for the GMT material. Finally, GMT R401 is less sensitive to temperature than the CTC
liftgate material (Figures 11 and 16).
 Fatigue

As stated earlier, many automotive parts undergo fatigue. The materials being
analyzed here have been used to make liftgates. One test where fatigue is important is
liftgate slam durability where the liftgate is opened and closed for 3 life cycles. Cyclic
tensile-tensile fatigue tests were preformed on theses materials to determine their
sensitivity to cyclic loading. Since most fatigue work on thermoplastic composites has
been perform using very short glass fibers (0.01 to 0.1 mm) [3,4] compared to these
materials which contain glass fibers 12 mm and longer, two sizes of tensile bars were
used to determine if sample width effected the results. All fatigue tests were performed
using an Instron 8501 servo hydraulic test machine in a tensile-tensile mode. The load
was applied as a sine wave at 3 Hz with the minimum load set at 10% of the maximum
load. Cyclic fatigue at rates greater than 3 Hz can lead to heating of composite test
specimens [3].

Figure 17 shows the fatigue results in the form of an S-N plot for GMT R401 tensile
bars with a 38 mm reduced section. Typically, there is a high degree of scatter between
samples at a given load. Here three samples were tested at each load level. The linear
regression fit and its R² value of 0.9186 indicate a very good linear fit between stress
and cycles to failure. S-N curve for GMT R401 bars with a reduced section of 12.5 mm
is shown in Figure 18; S-N curve for GMT C467 bars with a reduced section of 38 mm is
shown in Figure 19; and S-N curve for the CPI liftgate material at 38 mm reduced
section is shown in Figure 20. The only poor linear fit was for the CPI liftgate material,
however, only a small number of samples were tested.

120.0

100.0 R2 = 0.9186

80.0
Stress, Mpa

60.0

40.0

20.0

0.0
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07
Log N

Figure 17. S-N curve for GMT R401, 38 mm bars.

 150
140
130
120 R2 = 0.9583
110
100
Stress - Mpa

90
80
70
60
50
40
30
20
10
0
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07
Log. N

Figure 18. S-N curve for GMT R401, 12.5 mm bars.

100.0
90.0
80.0
2
R = 0.9744
70.0
Stress, Mpa

60.0
50.0
40.0
30.0
20.0
10.0
0.0
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06
Log N

Figure 19. S-N curve for GMT C467, 38 mm bars.

 60.0

2
50.0 R = 0.7803

40.0

Stress, Mpa
30.0

20.0

10.0

0.0
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07
Log N

Figure 20. S-N curve for CPI liftgate material, 38 mm bars.

One way to compare these results is to generate an S-N curve for all the materials.
In Figure 21, the fatigue results were normalized by the failure stress for each material
(Table 1). All materials can be fit to one curve with a R² value of 0.9405. This means
there are no major differences between the four data sets. Glass length and the
presence of talc in the PP compounds affect the overall strength but do not affect the
fatigue characteristics of the PP compounds. Also, sample width does not appear to be
a factor.

100.0
Azdel R401 12.5 mm
90.0 2
R = 0.9405 Azdel R401 - 38 mm
80.0 Azdel C467 38 mm
70.0 CPI Liftgate 38 mm
% Max Stress

60.0 Composite Data

50.0
40.0
30.0
20.0
10.0
0.0
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07
Log N

Figure 21. S-N curve for all long glass polypropylene materials.
 Fatigue displacement or strain data can also be plotted against log cycles. In Figure
22, GMT R401 displacement data for the three loading levels was plotted versus log
cycles. Just before failure, displacement increases rapidly indicating creep is occurring
as the fatigue test progresses. This curve looks similar to the creep curves for GMT
R401, Figure 13, indicating the failure mode for creep and fatigue may be the same.

0.8 50% Max Stress

40% Max Stress
0.7
30% Max Stress
0.6
Displacement, mm

0.5

0.4

0.3

0.2

0.1

0
1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E+06 1.E+07
Log Cycle

Figure 22. Fatigue displacement curves for GMT R401, 12.5 mm bars.

Summary

Typically, load levels for composite structures are kept to 25 to 30 % of the failure
stress. The creep and fatigue data reported here for the long glass PP composite
material are consistent with these guidelines. Creep data for the GMT R401 and CTC
liftgate material in the flow direction show the potential for creep rupture at greater than
30% of the tensile failure strength. CTC liftgate material in the cross-flow direction is
somewhat weaker than the flow direction with nonlinear response as low as 20%. When
designing a part care should be taken about introducing a creep situation in the cross-
flow direction.

The composite fatigue data curve (Figure 21) showed the fatigue results can be
normalized by tensile strength and that glass length and the presence of talc in the PP
compounds do not affect the fatigue characteristics of the PP compounds. Another
general guide line for fatigue data is at least one million cycles are needed for good
material fatigue characteristics and this occurs at 30% of the tensile strength for the PP
compounds. The ability to use tensile strength to determine fatigue limits is important
since some material suppliers do not provide tensile modulus information.
 References

1. European Alliance for Thermoplastics Composites (EATC), www.eatconline.org.

2. F o ss, P .H ., M e n tze r, C .C ., F ra n klin , D .W ., “D e sig n a n d V a lid a tio n o f a

Thermopla stic C o m p o site L iftg a te “, P ro ce e d in g s o f th e S P E A u to m o tive
Composites Conference, September 2004.

3. W yzg o ski, M .G ., K ro h n , J.A ., a n d N o va k, G .E ., “F a tig u e o f F ib e r-Reinforced

Injection Molded Plastics. 1: Stress-L ife tim e D a ta ”, P o lym e r C o m p o site s, O ct.
2004, Vol. 25 No.5, 489-498.

4. W yzg o ski, M .G ., a n d N o va k, G .E ., “P re d ictin g fa tig u e S -N (stress-number of

cycle s to fa il) b e h a vio r o f re in fo rce d p la stics u sin g fra ctu re m e ch a n ics th e o ry”.
Journal of Material Science, January 2005, vol. 40 No. 2, 295-308.

5. Moo re , D .R ., a n d T u rn e r, S ., “M e ch a n ica l E va lu a tio n S tra te g ie s fo r P la stics”,

Woodland Publishing, 2001, pp 250-274.