Continuous compressive stress relaxation of elastomers used in engine sealing applications.

The use of compressive stress relaxation (CSR) measurements for screening elastomers in sealing and gasketing type parts in the automobile industry has been slowly gaining acceptance. Stress relaxation is a phenomenon which comes about upon application of a constant strain (tensile or compression modes) to a rubber sample. The force that is necessary to maintain that applied strain is not constant, but decreases with time (ref. 1). In compression, there is a decrease in sealing contact force over time at a constant strain (ref. 2). The processes that are taking placing in the rubber matrix are both physical and chemical in nature. There is an immediate physical relaxation of molecular chains and fillers due to the deformation. The whole process of chain flow and resulting movement of their entanglements is reversible upon removing the strain from the system. Afterwards, there are chemical processes that may take over, either in the absence (thermal degradation) or presence of oxygen (oxidative degradation), both leading primarily to chain scission type reactions and loss of polymer molecular weight. The chemical component of relaxation is completely irreversible. Both the physical and chemical aspects of stress relaxation will cause a reduction in counterforce during measurement. Under special circumstances, degradation mechanisms may lead to both crosslink density increases and chain scission reactions. In this case, only the newly formed network chains that are load-bearing may impact counterforce measurements (ref. 3). Theoretically, the greater the stored elastic strain energy of the seal, the greater the resistance to relaxation effects during service (ref. 4). An excellent review outlining compounding principles for reducing stress relaxation effects seen in sealing materials can be found in reference 2.

The ultimate goal in using CSR testing is to relate the stress relaxation data to lifetime predictions. In the case of a seal or gasket, elastomeric seal life ends when the seal starts to leak, as the sealing force at this point equals or falls below the system pressure (ref. 2). Arrhenius plots (in time versus 1/T) can be used to make lifetime predictions, but one needs to run tests at at least three different temperatures. Gillen et al. (ref. 5), points out the importance of diffusion-limited oxidation anomalies in CSR testing leading to non-Arrhenius behavior, and thus proposes the need to determine, using modeling techniques, the correct size and geometry of the test piece to eliminate this effect.

Stress strain immersion testing has always been the major tool used in deciphering the resistance of a particular rubber compound to the test media aged at elevated temperatures. As a general rule of thumb, a loss of 50% in both the elongation and tensile strength of a material is used as a benchmark to decipher its usefulness in that particular test media. Many rubber parts, however, do not see strains beyond 25% in their lifetime; with the majority being under 10% (refs. 2 and 6). In the case of thicker products, such as gaskets and seals, it is suggested that stress relaxation in compression is much more suitable (ref. 1).

ASTM D-6147, ISO 3384, GMNA 3922 TP and Ford FLTM BP 116-02 provide the general guidelines for performing CSR measurements on rubber compounds. Numerous methods have been used to measure stress relaxation in compression over the last few decades. These methods differ in jig design and in the measurement of the sealing force. Hardware or fixture design influences the stress relaxation response and overall sensitivity of the measurement. The advantages and disadvantages of continuous (Elastocon, modified Jamak, ARDL jig) versus discontinuous (Lucas, Jamak, Jones-Odom, Shawbury-Wallace, Wykeham-Farrance, Dyneon test jig) test methods have been elucidated (refs. 7-10). Sample geometry (large or small compression set buttons, o-ring, die cut washer, micro-pellets, molded part) is in itself a huge issue, as it will have an impact on the extent of diffusion-limited oxidation anomalies, as well as test-to-test variability.

The aeration of motor oil results in an emulsified oil system caused by the input of air through the operation of the engine valve system. This process changes not only the character of the oil (pH and oxidation products), but also the compatibility of the oil with the rubber seal (ref. 11). Earlier, Dinzburg (refs. 12 and 13) introduced the importance of including aeration in immersion testing of robber in motor oils in order to achieve a reasonable correlation with field conditions. A low level of aeration (2mL/min.) three hours on and 21 hours off and weekly oil changes were among his recommendations. FTIR spectroscopy was used to follow oxidation effects in the oil. Oil aggressiveness depended primarily on the additive package and not the base oil. This testing was carried out in immersion, and stress strain tensile change was used as the indicator of rubber mechanical property deterioration. Walker (ref. 7) has applied a wide range of aeration techniques (Ford method, air stone, jet and bubble-bar) on rubber samples in modified Jamak jigs, with each technique providing differing levels of aeration.

Dinges, et al. (ref. 14), tested the effectiveness of aeration using the Elastocon continuous CSR system on HNBR and silicone rubbers. It was found that a continuous 2mL/min. air rate was insufficient to generate noticeable changes in the rubber, especially in HNBR, so a maximum flow rate of 7 mL/min, was used on a continuous basis. After 300 hours of testing, non-aerated or aerated HNBR had the same stress relaxation behavior, whereas aeration in silicone had an important effect in reducing sealing force in the same time frame. Oil additives were found to affect the sealing force behavior of HNBR.

This article is a further investigation into the understanding of material behavior using continuous CSR under both aeration and non-aeration environments. Besides HNBR, the properties of other automotive sealing elastomers, polylacrylate (ACM), ethylene acrylic elastomer (AEM), fluoroelastomer (FKM) and silicone robber (VMQ), used in maximum service temperature conditions of 150[degrees]C will be presented and contrasted. These results will update the most recent work in studying specialty elastomers used in sealing applications (refs. 15-17).

Experimental section

HNBR, ACM, AEM, FKM and VMQ were analyzed in this study. Based on ASTM D2000/SAE J 200 classifications, the maximum heat resistance of the above elastomers is given as: HNBR--150[degrees]C; ACM and AEM--175[degrees]C; FKM--250[degrees]C; and VMQ--200[degrees]C. The 5W-20 motor oil (factory fill 2015 for Ford) is described, according to the provided MSDS sheet, as a lubricating oil consisting of a mixture of saturated and unsaturated hydrocarbons from paraffinic distillate. Additives are also contained in the oil mixture. Standard ASTM laboratory procedures were followed for all compound testing.

Compressive stress relaxation is a method of testing in which a rubber sample is compressed to a constant deformation in a fixture that consists of two parallel plates made of a corrosion-resistant material. The plates have holes drilled through their center if a ring test piece is used, which allows fluid inside the ring and equalizes the pressure. A ring or cylindrical test piece is compressed to, typically, 25% strain. The final compression is fixed and maintained throughout the test. The sample is aged at predetermined environmental conditions, and the decay in the force exerted by the specimen on the fixture is measured over time. A compression device (or jig) and a counter-force measuring mechanism are the requirements for CSR testing.

The preferred counter-force-measuring device is one that continuously monitors the decay in force throughout the duration of the test. One such apparatus has been designed by the Swedish company Elastocon. The Elastocon design is such that the load cell is incorporated in the jig, which is placed in an Elastocon or modified oven. The jig is connected to a computer, which records the force and temperature data. This allows continuous measurements of the resisting force at the aging temperature. ISO 3384 describes both the discontinuous and continuous methods of determination of stress relaxation in compression. An inter-laboratory test program was completed within ISO TC 45 for this procedure. The results demonstrated a much better repeatability and reproducibility for the continuous method over intermittent methods.

Washer samples were punched out from a tensile macro sheet with an outside diameter of 19 mm, an inside diameter of 15 mm and a thickness of approximately 2 mm. A volume of 200 ml of test oil was used for each jig. All samples were tested with and without aerating the test oil. The aeration assembly that is included with the Elastocon equipment was used in the testing in which the samples were exposed to aerated oil. Figure 1 shows that the Elastocon apparatus includes a single plastic tube that introduces air into the sample container. The flow rate of the air was 7 ml/min. The samples were compressed by 25% of their room temperature thickness and aged in an oven at a temperature of 150[degrees]C. Counter-force measurements were taken during the entire test period at the aging temperature.

[FIGURE 1 OMITTED]

The physical relaxation process at a particular temperature T can be adequately described by the following equation (refs. 18 and 19):

(1) E(t) = [E.sub.e] {1 + [(t/[[tau].sub.T]).sup.-m]}

where [E.sub.e] is the equilibrium modulus, m is an empirical, temperature-independent constant and [[tau].sub.T] is a temperature-dependent time constant. This long-term process is due to diffusion of chain branches or dangling ends within a polymer network comprising both physical entanglements and chemically crosslinked chains. An increase in crosslink density will decrease the time necessary to achieve the value of the equilibrium modulus. Equation 1 is valid for temperatures below which no chemical reactions or degradation processes are taking place in the system.

Chemical stress relaxation can be represented by a Maxwellian decay pattern which, besides being a function of time, can be expressed by the well-known Arhennius equation:

(2) K(T) = A x exp (-[E.sub.a]/RT)

where [E.sub.a] is the activation energy of reaction (energy/mole), R is the general gas constant, T is the temperature and A is a measure of collision frequency and the probability of a reaction occurring (refs. 20-22). The Arhennius approach assumes that the failure process consists of chemical reactions where the rate of reaction will increase as the temperature increases. Types of reactions occurring include chain scission due to oxidation, crosslinking caused by oxidative chain reactions and crosslinking caused by vulcanizating agents not used up in initial cure. Chain cleavage could also be taking place either randomly along the chain backbone or directly at the crosslink sites. Typically, a ten degree temperature increases gives a two- to three-fold increase in reaction rate for organic chemical reactions.

Ideal stress relaxation behavior for gasketing applications is represented by some initial relaxation followed by a relatively constant residual stress as a function of time (ref. 4). In other words, after the initial physical relaxation of the polymer chains, any chemical relaxation process is minimized, providing a steady and constant seating counterforce as a function of aging. Any material presenting a continuous relaxation behavior would signify a steady loss of pressure rating as a gasket material.

Results and discussion

Table 1 displays the stress-strain characteristics of all five compounds measured at room temperature. All samples possess a hardness of approximately 65 [+ or -] 5 pts., which is the range of a typical compound used in an engine sealing application such as a seal or a gasket. HNBR clearly possesses the highest tensile strength of all the samples, with silicone being the lowest. All five compounds have elongation values varying from near 200% up to 300%. FKM and silicone have higher stress values at low elongations, while AEM and HNBR have higher stress values for elongations beyond 100%. The tear strength of HNBR is also superior to the other compounds.

Figure 2 shows the compressive stress relaxation behavior up to 2,800 hours aging in 5W20 oil at 150[degrees]C under non-aeration conditions. The most obvious drop in sealing retention as a function of aging is seen for the silicone sample, which retains only a few percent reduced force after 1,360 hours of the testing. The other four compounds show similar CCSR behavior with more mild distinctions between them. The ACM sample tails downward to approximately 58% reduced force after 1,360 hours. The FKM curve shows a very gradual decline in reduced force up to 1,360 hours. Testing of the VMQ, ACM and FKM samples was stopped at 1,360 hours. AEM causes an initial rise in the sealing force at the beginning, then a general decline to a plateau behavior which starts to break around 1,100 hours, at which point the curve starts to gradually descend to approximately 55% reduced force at 2,800 hours. Similar CSR test results have been seen in SF-105G test oil at 150[degrees]C for both ACM and AEM type elastomers (ref. 23). The HNBR sample shows a more consistent sealing force behavior, with an initial decline to a plateau value of 80% after 200 hours, which is retained up to 2,800 hours of testing.

[FIGURE 2 OMITTED]

Figure 3 illustrates the sealing force retention behavior up to 2,500 hours aging in 5W20 oil at 150[degrees]C under aeration conditions. Parallel to the behavior seen under non-aeration conditions, the silicone sample shows a steady decline and has about 10% sealing retention at 1,360 hours. The sealing force retention of silicone is actually slightly better under aeration conditions. Compared to the other samples, the ACM compound gradually declines to 62% force retention at the end of testing. Both the HNBR and AEM compounds were tested to 2,500 hours. AEM again experiences an initial rise and gradually declines to overlap with the ACM curve at longer hours. Its sealing force retention is less at 2,500 hours under aeration versus non-aeration. Both the FKM and HNBR samples show very similar compressive stress relaxation behavior up to 1,360 hours of testing, while HNBR presents a constant sealing force retention (78%) up to 2,500 hours.

[FIGURE 3 OMITTED]

Figure 4 shows the nominal values of force, measured in Newtons, for the same testing conditions given in figure 3. As it can be seen, the initial counterforce exerted by the compounds is highest in the case of HNBR, followed by FKM and VMQ, then by ACM and AEM. Excellent retention of mechanical properties at high temperatures is critical for gasket functioning. It is clearly shown that the HNBR compound exerts the highest actual counterforce throughout the whole duration of testing.

[FIGURE 4 OMITTED]

The CCSR behavior of HNBR under both aeration and non-aeration conditions has been overlapped in figure 5. For all intents and purposes, both sealing retention curves exhibit the same behavior, showing that HNBR performs comparably in aeration and non-aeration environments. This conclusion was also seen in earlier work on HNBR (ref. 14).

[FIGURE 5 OMITTED]

Besides examining the CCSR behavior of the five elastomers, it was decided to test the same samples also in the tensile mode after aging in hot air, ASTM #1, IRM903 and in 5W20 motor oil.

The hot air stress strain aging results for 504 hours at 150[degrees]C are presented in figure 6. ACM, VMQ, AEM and HNBR all exhibit hardening

due to the oxidative aging. The oxidative reaction mechanisms operating in HNBR have been summarized (ref. 24). According to the ASTM classification of the heat resistance of each elastomer, the hot air aging results fall within the expectations of their maximum service temperature.

[FIGURE 6 OMITTED]

In figure 7, the hot air resistance was measured under compression up to 1,008 hrs. at 150[degrees]C for the five compounds. The best resistance to compression set is demonstrated by FKM, followed closely by HNBR and AEM. VMQ takes on high values of compression set. Compared to the results under tensile (figure 6), HNBR performs much better in hot air resistance under compression type conditions. These results do not fall in line with the normal ASTM classification of the maximum heat resistance under service conditions, underestimating the good properties shown by elastomers such as HNBR under compression conditions. The order of ranking of the best sealing elastomers from this testing does bear a certain resemblance to the CCSR (aeration or non aeration) data which predicted the worst behavior from silicone and the best behaviors from both HNBR and FKM. Anderson (ref. 10) has put forth that there exists a strong correlation between compression set and CCSR testing in the case of hot air aging and the use of small sample buttons.

[FIGURE 7 OMITTED]

Figure 8 illustrates the stress strain behavior after aging 70 hours at 150[degrees]C in ASTM test oil #1. It can be observed that all elastomers show little change due to aging in this oil. Both ACM and VMQ soften during aging, while AEM and HNBR harden. ACM and AEM show little or no volume change, while VMQ has a slight positive volume change. HNBR possesses a slight negative volume change due to leaching out of the plasticizer.

[FIGURE 8 OMITTED]

The stress strain behavior after aging 70 hours at 150[degrees]C in IRM903 oil, shown in figure 9, differentiates the elastomers to a greater degree compared to ASTM #1 oil. All elastomers soften to a certain extent, with HNBR showing the greatest resistance to softening. VMQ and ACM possess quite high volume swells (between 30 and 40%) while HNBR has minimal volume change (about 5%) and excellent retention of physicals in this highly ,aromatic test oil.

[FIGURE 9 OMITTED]

In figure 10, the stress strain behavior of the four elastomers is shown for 504 hours aging at 150[degrees]C in 5W20 motor oil. It appears that VMQ exhibits the most change in all properties compared to the other three elastomers, as it absorbs high amounts of oil and softens at the same time. The softening of the silicone and loss of modulus seen in stress strain aging in 5W20 conditions explain the continuous and gradual drop in sealing force as a function of time (figures 2-4). According to these results, both ACM and AEM possess very good general resistance to this low viscosity motor oil. However, their results in compression (CCSR) show a gradual loss of sealing force as a function of aging. HNBR also displays good general resistance to 5W20 oil, as well as excellent volume swell resistance. These results were also seen during the CCSR testing under both aeration and non-aeration conditions.

[FIGURE 10 OMITTED]

Excellent low temperature properties are also crucial in the good functioning of seal and gasketing applications. Good sealing retention at low temperatures prevents problems like premature seal leakage, especially during cold start-up conditions. Figure 11 illustrates the behavior of tan delta over a wide temperature range (-50 to 150[degrees]C) for the five compounds. Silicone rubber possesses the lowest tan delta in this temperature range, presenting stable and quite constant elastic properties for a seal compound. The peaks of the tan delta data can be correlated to glass transition temperatures, the lowest of which (after silicones) are shown for ACM, then HNBR, AEM and finally FKM. Retention of elastic characteristics (storage modulus) is critical in part use. It can be clearly observed that both VMQ and HNBR possess the most stable elastic properties after the glass transition zone towards the higher temperatures, while FKM and AEM show the most variation in their dynamic properties. Temperature retraction measurements on the five compounds are illustrated in figure 12. It can be seen that silicone possesses the best low temperature behavior, followed by ACM, HNBR, AEM and then FKM. These results concur with the tan delta maxima observed in figure 11. Besides the high temperature behavior, it is clear that low temperature sealing is equally important. Both these properties can be measured by using CCSR under temperature cycling conditions (ref. 14).

[FIGURE 11 OMITTED]

Conclusions

This study has clearly compared the properties of the main automotive elastomers used in sealing and gasketing applications. Besides considering only tensile aging data, this study shows the importance of testing the gasketing material in the compression mode by using continuous compressive stress relaxation techniques. HNBR, in particular, presents both a high initial sealing force and an excellent sealing force retention as a function of fluid aging in motor oil. This study also demonstrates that testing solely stress strain behavior after hot air aging or liquid immersion testing may dismiss other important elastomer properties that are vital to the application.

Table 1--stress strain properties (die C)
measured at 23[degrees]C of tested compounds

Properties                   ACM      VMQ      AEM       FKM     HNBR

Hardness duro. A2 (pts.)     60       67       61        65       67
Ultimate tensile (MPa)        9.4      6.2     14.1      10.7     20.2
Ultimate elongation (%)     267      306      219       266      244
Stress @ 25 (MPa)             1.1      1.6      1.0       1.4      1.1
Stress @ 50 (MPa)             1.7      2.3      1.8       2.0      2.0
Stress @ 100 (MPa)            3.8      3.5      5.0       3.6      5.7
Stress @ 200 (MPa)            8.2      4.9     13.2       7.8     17.2
Stress @ 300 (MPa)                     6.1
Tear strength (kN/m)         21.0     21.3     23.3       Not     31.2
                                                       tested

References

(1.) G. Spetz, "Applying stress relaxation tests," Rubber Technology International (1997).

(2.) K. Smith, Rubber & Plastic News, Jan. 31, 15 (1994).

(3.) W. McKnight, "Chemical stress relaxation," in "Introduction to polymer viscoelasticity," Ed.: J.J. Aklonis, Wiley, Interscience, New York (1983).

(4.) M. Brown, "Seals & sealing handbook, 4th edition," Elsevier Science Publishers, Oxford, U.K. (1995).

(5.) K.T. Gillen, M. Celina and M.R. Keenan, Rubber Chem. Technol., 73, 265 (2000).

(6.) A.N. Gent, "Rubber elasticity" in "Science and technology of rubber," Ed.: ER. Eireich, Academic Press, New York, p. 2, 6 and 16 (1978).

(7.) F.J. Walker' "CSR testing of elastomers," Advanced Elastomers Conference, Dearborn, MI, June 2002.

(8.) P. Tuckner, SAE paper 2001-01-0742, SAE 2001 World Congress, Detroit, MI, March 2001; SAE paper 2000-01-0752, SAE 2000 World Congress, Detroit, MI, March 2000.

(9.) A. Pannikottu, C. Lu and M. Centea, 3rd International Symposium on Finite Element Analysis of Rubber and Rubber-like Materials, Akron, OH (1999).

(10.) Anderson, R. Bruner and P. Manley, SAE paper 2003-01-0947, SAE 2003 World Congress, Detroit, MI (2003).

(11.) F.J. Walker, "Automotive sealing applications of elastomers," (in preparation).

(12.) B. Dinzburg, ACS Rubber Division Meeting, Nashville, TN, November 1992.

(13.) B. Dinzburg, SAE International Congress and Exhibition, Detroit, MI, March 1-5, 1993.

(14.) U. Dinges, R.J. Pazur and D.W. Wall Automotive Elastomers Conference, Dearborn, M1, June 2002.

2(15.) M.T. Gallagher, W. von Hellens and H. Bender, ACS Rubber Division, paper no. H, Chicago, IL, April 1999.

(16.) M.E. Wood and M.J. Recchio, ACS Rubber Division, paper no. 7, Louisville, KY, October, 1996.

(17.) A.S. Farid, ACS Rubber Division, paper 57, Chicago, IL, April 1999.

(18.) J.G. Curro and E.A. Salazar, J. Appl. Polym. Sci., 19, 2571 (1975).

(19.) J.G. Curro and P Pincus, Macromolecules 16, 559 (1983).

(20.) "Lifespan of rubber materials and thermoplastic elastomers in air' water and oil, 1st edition," Editor: C. Andersson, Swedish Institute for Fibre and Polymer Research (1999).

(21.) K.T. Gillen, M.R. Keenan and J. Wise, Die Angewandte Makromolekulare Chemie 261/262, 83 (1998).

(22.) A. Pannikottu "Service life prediction of elastomeric components used in severe service environments," paper 23, ACS Rubber Division Meeting, Providence, RI, April 2001.

(23.) PE. Manley and C.T. Smith, ACS Rubber Division, paper 127, Cincinnati, OH, October 2000.

(24.) H. Bender and E. Campomizzi, Kautsch. Gummi Kunstst. 5,14 (2001).