POLIMERY 2005, 50, nr 3 213

ZBIGNIEW DOBKOWSKI
Industrial Chemistry Research Institute
ul. Rydygiera 8, 01-793 Warszawa
e-mail: [email protected]

Lifetime prediction for polymer materials using OIT measurements

by the DSC method

RAPID COMMUNICATION

Summary — The standard procedure of OIT measurements by DSC method

was applied for prediction of lifetime of polymer materials. Investigations of
unmodified and modified, i.e. grafted and filled PE-LLD have been described,
as an example. OIT values were measured in the temperature range (in gene-
ral) from 180 to 240 oC, then the activation energy and the lifetime values were
calculated. The following sequence of increasing thermal stability and increas-
ing lifetime values has been found PE-LLD grafted < PE-LLD unmodified <
PE-LLD filled compositions. The described approach can be recommended for
the lifetime prediction of other polymer materials, in particular those suscep-
tible to the thermooxidative degradation.
Key words: oxidation induction time OIT, activation energy from OIT, life-
time values of polymer materials, modified PE.

The oxidation induction time (OIT) is commonly or in the logarithmic form

used for assessment of the thermooxidative stability of ln k(T) = –Ead/RT + CA (2)
polymeric materials. Usually, the differential scanning where: k(T) — degradation process rate constant, Ead — acti-
calorimetry (DSC) method is applied for OIT determina- vation energy, T — absolute temperature in K, R — gas con-
tions, as it is currently recommended by standards stant, A and CA — constants.
[1—2] and described elsewhere [3—6]. Moreover, OIT The plot of ln k(T) vs. 1/T is a straight line with the
measurements can also be used for calculation of the slope equal to -Ead/R.
activation energy of thermooxidative degradation pro- Time-to-failure tf(P) can be calculated from the follow-
cess [4, 7, 8], and then the lifetime of polymer materials ing equation [11, 12, 14, 15]:
can be predicted. tf(P) = {f(P)/A exp(Ead/RT)} (3)
Lifetime (or service life, or long term performance) of where: f(P) — function dependent on the considered polymer
a polymer material, as well as of a polymer end-use property but independent on temperature.
product, at ambient temperature or at elevated tempera- Then the lifetime at specific temperature, or permissi-
ture is a crucial property, which determines the scope of ble temperature of service during a specified time range,
application of polymer materials. Reasonable lifetime can be estimated from the plot of ln tf vs. 1/T [12].
predictions can be obtained from accelerated ageing OIT value corresponds to the initial time-point of the
tests when results of such tests are extrapolated to ser- oxidation curve, i.e. to the time-to-failure and thermal
vice conditions, using an appropriate time-temperature stability of a given material [tf(OIT) = OIT] [4, 6, 8, 15]. It
model [9]. Various degradation tests, usually based on means that OIT is a given property P in eq. (3). There-
determination of time-dependent mechanical properties, fore, time-to-failure is given by
e.g. tensile strength [10] or ultimate elongation [11], as tf(OIT) = Cox exp(Ead/RT) (4)
well as based on weight loss [10—13], can be applied. where: Cox — constant corresponding to the preexponential
Then, the Arrhenius model is used to evaluate accele- term in eq. (3).
rated test results involving the effect of temperature. The lifetime measured as the time-to-failure can be
Therefore, kinetics of the thermooxidative degradation estimated from the plot of ln tf(OIT) vs. 1/T [4, 8, 15, 16].
process can be described by the Arrhenius equation In this paper, the comparative thermal stability as-
k(T) = A exp(–Ead/RT) (1) sessment and the lifetime prediction for unmodified and
214 POLIMERY 2005, 50, nr 3

chemically modified PE and its compositions is de- 5

scribed.
4
EXPERIMENTAL
3

ln(OIT )
Materials
2
The following polymer materials (some of them pre- C A
pared in the laboratory of polymer compositions in our 1 B
Institute [16]) were used in this work:
— commercially available linear low density poly- 0
ethylene (PE-LLD), Stamylex PE-LLD, DSM; 0,0019 0,0020 0,0021 0,0022
— chemically modified (grafted) PE-LLD; 1/T , deg-1
— physically modified (filled) PE-LLD, i.e. compo-
sites of grafted PE-LLD with a flame retardant filler (alu- Fig. 1. Plot of ln (OIT) vs. 1/T for: A — PE-LLD unmodified,
minum hydroxide). B — PE-LLD grafted, C — PE-LLD filled (in all cases three
series of measurements of OIT were done-denoted by *, o
Method and +)

A differential scanning calorimeter DSC 7 Perkin-El-

mer with GSA 7 Perkin-Elmer automatic gas flow switch T a b l e 2. Activation energy of thermooxidative degradation
(Ead) determinated by correlation and calculated lifetime (tf(OIT))
was used for OIT measurements. The standard proce- values of PE-LLD samples
dure [5] was used both for the instrument calibration
PE-LLD
and for OIT measurements, however appropriate mea- Calculation results
surement temperatures were selected to make the sub- unmodified grafted filled
sequent calculations of activation energy possible. Activation energy (Ead),
kJ/mol 149 150 180
RESULTS AND DISCUSSION Correlation coefficient (r2) 0.817 0.924 0.992
tf(OIT) at 210 oC, min 9 4 75
tf(OIT) at 100 oC, min 1.0 0.5 80
OIT values were measured for unmodified and
modified PE-LLD samples at various temperatures:
190—220 oC, 180—210 oC and 210—240 oC in the case of
unmodified, grafted and filled PE-LLD, respectively (Ta- The activation energy values (Table 2) were calcu-
ble 1). The temperature range was selected according to lated from OIT data using logarithmic form of eq. (4) as
the reasonable time of measurement. As can be seen a respective linear regression equation. Large dissipa-
from Fig. 1 the thermal stability of tested materials can tion of results was observed for unmodified and grafted
be compared according to the temperature range, and materials (the correlation coefficients in Table 2, see also
the following sequence of increasing thermal stability Fig. 1) because of matrix (unmodified) material differ-
was found: ences (different lots of PE-LLD) and probable variations
PE-LLD grafted < PE-LLD unmodified (matrix) < in subsequent grafting reactions. In the case of filled
< PE-LLD filled composites. PE-LLD, the material is expected to be more uniform
then in the previous cases since it has been stabilized and
T a b l e 1. OIT experimental values for unmodified, grafted and homogenized during preparation of the composites.
filled PE-LLD As regards calculated Ead values, an intrinsic, rather
OIT values at given temperature, min significant, error has to be taken into account as the
Num-
PE-LLD ber slope of Ead/R is determined within a narrow range of
180 190 200 210 220 230 240
of lot o
C o
C o
C o
C o
C o
C o
C
temperature (in this work 30 oC) and the number of
experimental points (n, in this work n = 4) is usually
1 — 50 28 13 7.6 — —
unmodi- limited.
2 — 83 24 12 4.7 — —
fied
3 — 30 11 5.1 2.8 — —
The values of lifetime (or of time to failure tf(OIT))
were calculated [from equation (4)] (Table 2) at tempera-
1 55 14 11 2.9 — — —
ture 210 oC to compare with experimental data under
grafted 2 43 14 8.2 4.9 — — —
thermooxidative conditions and at arbitrarily taken tem-
3 60 31 7.0 5.2 — — —
perature 100 oC.
1 — — — 79 32 12 5.7 The same sequence of increasing thermal stability as
filled 2 — — — 70 32 14 5.9
for the testing temperature range has been observed for
3 — — — 70 30 13 4.4
the lifetime values:
POLIMERY 2005, 50, nr 3 215

PE-LLD grafted < PE-LLD unmodified < 4. Bair H. E.: “Thermal Analysis of Additives in Poly-
< PE-LLD filled (composites). mers” in: “Thermal characterization of polymeric
materials” (Ed. Turi E. A.), 2nd ed., Academic Press,
CONCLUSIONS San Diego 1997, pp. 2263—2410.
5. Dobkowski Z.: Polimery 2004, 49, 343.
The procedure of OIT measurements by DSC 6. Schmid M., Ritter A., Affolter S.: Polimery 2004, 49,
method, applied for thermal stability characteristics of 333.
polyolefins, can also be used for predictions of their life- 7. Budrugeac P.: Polym. Degrad. Stab. 2002, 68, 289.
times. In principle, it can be applied for various polymer 8. Dobkowski Z.: “Activation energy of thermooxida-
materials, in particular those susceptible to the ther- tive degradation from OIT for lifetime prediction of
mooxidative degradation. polymer materials” in: Proceedings of POLMATIN
The following lifetimes have been predicted for un- Workshop “Standardisation, Measurement and Test-
modified and modified PE-LLD materials at 100 oC (ar- ing of polymer materials for better quality of life”,
bitrarily taken as an example), in years (in parentheses): Warsaw, 9—10 October 2003.
PE-LLD grafted (0.5) < PE-LLD unmodified (1.0) < 9. Jakubowicz I.: Polimery 2004, 49, 321.
PE-LLD, filled (80). The calculated lifetime values seem 10. Montanari G. C.: IEEE Trans. Electr. Insulat. 1990, 25,
1046.
to be reasonable, but the experimental data which could
11. Langlois V., Audouin L., Verdu J., Courtois P.: Polym.
confirm the results for tested materials are not available.
Degrad. Stab. 1993, 40, 399.
12. Dobkowski Z., Rudnik E.: J. Thermal. Anal. 1997, 48,
REFERENCES 1393.
13. Gupta Y. N., Chakraborty A., Pandey G. D., Setua D.
1. PN-EN 728:1999, “Plastic piping systems and K.: J. Appl. Polym. Sci. 2004, 92, 1737.
ducting systems. Polyolefin pipes and fittings. De- 14. Montanari G. C., Lebok F. J.: IEEE Trans. Electr. Insu-
termination of oxidation induction time”. lat. 1990, 25, 1029.
2. ISO 11357-6:2002, “Plastics — Differential Scanning 15. Dobkowski Z., Rudnik E.: J. Therm. Anal. Cal. 2002,
Calorimetry (DSC) — Part 6: Determination of oxi- 69, 693.
dation induction time”. 16. Dobkowski Z., Pe³ka J.: ICRI Ann. Report 2001, ICRI
3. Mason L. R., Reynolds A. B.: J. Appl. Polym. Sci. 1997, Warsaw 2002, 109.
66, 1691. Received 1 IX 2004.