Predicting the life of automotive power steering hose materials.

The expectations for the life of elastomeric components on vehicles has changed dramatically over the last few years. Previously, many rubber components were viewed as items that would be replaced once or more during the life of the vehicle. Components such as hoses and belts are good examples. However, the automotive industry has recently stopped viewing elastomeric components as items that would be replaced at regular maintenance intervals. Instead, elastomeric components are expected to last the life of the vehicle, which is currently defined as 10 years/150,000 miles for both passenger cars and light trucks.

This new design philosophy places a new emphasis on testing and especially on accelerated testing, as it is now much more time consuming and costly to prove out designs and materials on a real time basis. In developing new test methods to verify component life in a much shorter time span, the critical properties of the component for that application must be considered. For instance, tensile strength and elongation to break are useful easily generated properties, but their utility in predicting whether a gasket will perform adequately after 10 years is questionable. Thus, a means must be found for choosing the most critical properties of a material and then devising a method by which property changes after short aging times can be extrapolated to the longer times that are representative of a vehicle's life.

For many elastomeric components on vehicles the most critical property is the fracture resistance of the elastomer, as crack growth can often degrade the performance of elastomeric components. The growth of cracks in elastomers is governed by the critical tearing energy. The tearing energy depends on the structure of the elastomeric network and on the rate of tearing (ref. 1). In general as the rate of tearing decreases, the tearing energy decreases (ref. 2). Because an elastomer's network structure is changed during aging (either additional crosslinking, chain scission, or both) the elastomer's critical tearing energy will also change.

Predicting the changes in the tearing energy as ar elastomer ages over long periods of time would be useful in predicting failures of many automotive elastomeric components. For instance, one possible failure mode of power steering hoses is by the formation and propagation of a crack from the inner tube to the outside. Knowing how the tearing energy changes as the hose ages would help in predicting the life of this component. We have previously developed a model of elastomer aging which uses data taken from laboratory aged specimens and from thermal analysis to predict the tearing energy of aged elastomers (ref. 3). This article demonstrates how the results of that study can be applied to predict the life of automotive elastomeric components.

Experimental

Materials

Specimens for critical tearing energy tests were prepared from production power steering hose, supplied by either Preferred Technology Group or Dayco-Swan Co. Hose covers were either chlorinated polyethylene (CPE) compounds or chlorosulfonated polyethylene (CSM) compounds. Hose tubes were always CSM compounds. Specimens were made by disassembling the tube and cover from each other. The reinforcing fibers were then removed by passing the disassembled pieces through a skiving machine, that reduced the specimen thickness to 0.5 mm.

Critical tearing energy

Critical tearing energies were determined using a standard method which is detailed elsewhere (ref. 3). In brief, uncracked pure shear specimens were loaded in tension at a strain rate of 0.001 [s.sup.-1] and their stress-strain curve recorded. A crack of length, a, was then introduced into one side with a lubricated scalpel. The specimen was then extended again at the same rate and the crack tip observed through a microscope. When the crack began to propagate, the strain in the specimen away from the crack tip was noted and the load removed. The crack was then lengthened and the same process repeated. In this manner, many measurements were made on one specimen. The critical tearing energy, T, was calculated from the equation

(1) T= Wh where W is the strain energy density in the specimen when crack propagation begins and h is the specimen height (ref. 4). The strain energy density is calculated from the area under the stress-strain curve of the uncracked specimen.

Tearing energy values were determined for new hose covers and tubes, specimens aged in a forced air oven at various temperatures for various times, specimens aged in two different power steering fluids (designated fluid A and fluid B), and specimens made from hoses removed from impulse testing. impulse testing is an industrial test used by hose manufacturers to estimate the life of power steering hoses. It involves pressurizing a power steering hose with power steering fluid at elevated temperatures, and cycling the pressure from zero to approximately 1,600 psi. The tearing energy of specimens made from hoses removed from high mileage vehicles was also determined.

Results

Figures 1 and 2 show the inner tubes of power steering hoses after they have been subjected to 440,000 and 910,000 impulse cycles respectively. Little damage can be seen in the first hose with 440,000 cycles. The small fibers are an artifact of the disassembly process. The hose with 910,000 cycles shows large cracks running parallel to the hose axis. Upon disassembling, some of the cracks were found to penetrate all the way to the reinforcement that separates tube and cover. This cracking pattern is typical for hoses and is due to the stress distribution in the hose which causes the cracks to grow in both length and depth.

Aging causes the critical tearing energy of both CPE and CSM compounds to decrease, regardless of the specifics of the aging conditions. Table 1 shows the results for CSM specimens aged at 150[degrees]C for 62.5 hours. All values are normalized to the original tearing energy of the CSM tube compound (790 J/m2). As can be seen, aging in fresh power steering fluid A causes the greatest loss of tearing energy. Air is only slightly less damaging than fresh fluid A. Fresh fluid B is much less damaging than either air or fluid A. Previously aged power steering fluids, in this case aged in air for 500 hrs at 150[degrees]C before addition of the elastomer specimens, causes less damage than the corresponding fresh fluids, particularly fluid A. Aging the elastomer in a solution of zinc dialkyldithiophosphate (ZDP) in mineral oil causes severe embrittlement leaving the specimens unfit for mechanical testing. ZDP is an antifriction and antiwear additive in some power steering fluids.

[TABULAR DATA 1 OMITTED]

Table 2 shows values of the tearing energy of both tubes and covers removed from impulse testing after various numbers of cycles. Tearing energy decreases for both cove] s and tubes as the number of cycles increases. Table 3 shows the tearing energy of hoses removed from 12 high mileage vehicles. Vehicles are coded by letter, with the same letter being assigned to vehicles of the same model. Models for which there was more than one vehicle tested have a number after the letter. In general, tearing energy decreases with mileage for the cover, while the tearing energy of the tube is less sensitive to mileage.

[TABULAR DATA 2 & 3 OMITTED]

It should be noted that the thermal environment experienced by individual power steering hoses can be quite different from vehicle to vehicle due to operating conditions and package considerations.

Discussion

The critical tearing energy of the elastomers from which power steering hose is made decreases with aging. Tearing energy decreases for specimens aged in all ways: aged in air, aged in fresh power steering fluid, aged in degraded power steering fluid, aged due to impulse testing, and aged due to use on a vehicle. In all cases tearing energy decreases due to an increase in the crosslink density of the material. The increase in crosslink density is confirmed by an increase in the modulus of the aged material as compared to the unaged elastomer. The increase in crosslink density of those specimens exposed to air is mainly attributed to a reaction with molecular oxygen, as the polymers stiffen much more slowly in an inert atmosphere (ref. 5). The increase in crosslink density of the specimens aged in power steering fluid is attributed to a reaction with an additive in the fluid.

We have previously shown that the rate of aging of an elastomer can be predicted by combining results of critical tearing energy experiments on oven aged specimens with the measured oxidative induction time determined using DSC (ref. 3). The decrease in the tearing energy was modeled as a first order chemical reaction. The temperature dependence of the reaction rate was assumed to follow Arrhenius behavior, meaning the rate of reaction increases exponentially with temperature. From that study it was determined that the rate of aging of the CSM elastomer compound can be expressed:

(2) ln(r) = - 10545(1/T) + 11.93 where r is the decrease in tearing energy per second of aging, and T is the absolute temperature. The apparent activation energy for the process is 88 kJ/mol. A similar expression can be derived for the CPE compound,

(3) ln(r) = - 12982(1/T) + 16.6 where the symbols have the same meaning. The apparent activation energy for oxidative degradation of this CPE compound is 108 kJ/mol. Activation energies for both CPE and CSM agree well with those previously determined by other methods (117 kJ/mol for CPE and 88kJ/mol for CSM) (refs. 5-7). The usefulness of the model can be judged by comparing the results predicted by the above model to the results from hoses aged in fluid and taken from impulse testing and high mileage vehicles.

Impulse tested hoses

The thermal environment of impulse testing is reasonably well controlled. Specimens used in this study were cycled at 35 cycles per minute from zero to 1,600 psi. Both air and fluid temperature were maintained at 145[degrees]C - 5[degrees]C. Fluid A was used for all impulse testing as well as in all high mileage vehicles. Figure 3 shows the theoretical line predicted from equation 3 for the CPE cover at 145[degrees]C. An excellent correlabon is observed between the theory and experimental results. Figure 4 shows the tearing energy of the CSM tubes taken from the same hoses that the covers were taken from for figure 3 along with some additional hoses, for which the covers were not available. The theoretical line is determined from equation 2 using 145[degrees]C. Initially the tubes appear to follow the same behavior as is predicted for oxidative degradation, but the rate of aging appears to slow as the number of cycles increases. This appears to be due to the nature of the power steering fluid changing as the test progresses.

The changing nature of the fluid is likely due to the depletion of certain additives in the fluid as the impulse test is conducted. Power steering fluid is a mineral oil based fluid that contains approximately 20 additional additives to control viscosity, oxidative stability, friction and foaming. Different additives affect the elastomer in different ways. ZDP is an antifriction and antiwear compound that is also sold as a curative for elastomers other than CPE and CSM. Zinc is purported to be detrimental to the aging performance of both CPE and CSM. Fluid A contains ZDP while fluid B does not. Aging fluid A depletes it of ZDP, most likely through reaction with oxygen. ZDP also gets consumed by reacting with any elastomer present.

Because aging in a solution of ZDP degrades the elastomer thoroughly and because fluid A is much more aggressive than fluid B it appears that the ZDP in fluid A is the reason fluid A induces more changes in the elastomer than fluid B. Both fluids, once they are aged, are more benign towards the elastomer, although for fluid B the change is minimal. Thus the continual depletion of ZDP from fluid A causes the fluid to be of continually decreasing aggressiveness during impulse testing or on vehicles, which qualitatively explains the decreasing aging rate seen in figure 4. This behavior is not without precedent. It has been previously shown that certain additives in engine oil cause the degradation of fluorocarbon elastomer seals and that after the additive is depleted the oil is much less aggressive towards the elastomer (ref. 8).

High mileage vehicles

Power steering hoses removed from high mileage vehicles also show a decrease in their tearing energy as compared to unaged materials. As mentioned previously, the thermal environment on vehicles varies between models and also between individual vehicles, depending on how and where they are driven. Lower average speeds, such as found in city driving, tend to increase underhood temperatures. Increased ambient temperatures also increase underhood temperatures. Thus, generalizing about all vehicles will necessarily reduce the accuracy of the predictions.

Remembering the above caveat, it is possible to use estimates of the average operating temperature and speed for some vehicles to test the oxidative aging model. If for vehicle models A, B, C, E and F the average speed is 30 mph, and the average hose temperature is 113[degrees]C (235[degrees]F) then the predicted aging rate of the CPE covers falls near the predicted line, as seen in figure 5. Again, different thermal environments on individual vehicles makes this type of generalization tentative, as individual data points, such as the point from vehicle B, can deviate from expected trends due to unusual thermal histories. Figure 6 shows the results for the CSM covers on vehicles D and G. These models operate at slightly lower temperatures than the previous models. The data fall on a single line if the average temperature is assumed to be 93[degrees]C (200[degrees]F) and the average speed 27 mph. Thus, the results predicted by the oxidative aging model appear to work well for modeling the decrease in the tearing energy of power steering hose covers from high mileage vehicles.

At this time, results from the tubes removed from high mileage vehicles are only understood in a qualitative manner. From table 2 we know that the decrease in tearing energy for tubes removed from high mileage vehicles appears to be significantly less than that done by impulse testing. For instance, the tube from vehicle C had the same reduction in tearing energy at 150,000 miles as we would expect only 32,000 cycles of impulse testing. The main difference in environment between impulse testing and high mileage vehicles is the volume of fluid in the system. In impulse testing a large reservoir is used to supply the pressurized fluid. However, in a vehicle the total volume of fluid is less than one gallon. Thus, a greater amount of ZDP is present in the system during impulse testing, which will degrade the elastomer faster and to a greater extent than on the vehicle due to the smaller volume of fluid in a vehicle power steering system. Also, in vehicles the fluid stays hot throughout the system, while during impulse testing the majority of the fluid stays cool, eliminating the possibility of consuming the fluid through reaction with oxygen.

Implications

The biggest roadblock to using the results of this study to predict the life for hoses is the lack of knowledge of the minimum tearing energy required to prevent a hose from undergoing significant amounts of crack growth. After extended service, small cracks will be present in the tube of the hose. These cracks will grow over time by fatigue to larger cracks such as in figure 2. However, from a practical standpoint, once the cracks begin to propagate, the useful life of the hose is near its end. Thus, the real issue is when does the tearing energy reach a low enough value that small, perhaps preexisting, flaws begin to propagate.

Because the stress-strain behavior of elastomers is highly nonlinear and because the hose is a composite structure, calculating the stresses in the hose is not easy. Therefore calculating the pressure at which a crack of a given size will propagate is not possible. Thus the only practical way to estimate the minimum tearing energy necessary for a hose to be functional is to examine hoses removed from high mileage vehicles. This requires a fairly large number of hoses to determine which hoses have lowest tearing energy but still remain macroscopically crack free. From this a lower bound on the necessary tearing energy can be determined. From our initial studies of high mileage vehicles it appears this value is somewhere near 300 350 J/m2 for the compounds used in these hoses.

Conclusions

The critical tearing energy of the elastomers from which power steering hose is made decreases during aging at elevated temperature. A previously developed model predicts well the decrease in tearing energy for hose covers that were removed from hoses subjected to impulse testing and also from high mileage vehicles.

The decrease in tearing energy is due to attack by molecular oxygen which leads to additional crosslinking of the elastomeric network. Behavior of the hose tubes is more complicated. Interactions with the power steering fluid dominate the aging behavior. Initially, an additive in fresh power steering fluid causes more crosslinking of the elastomer than does oxygen. However, after the additive is depleted the fluid acts as a protectant and causes less degradation of the rubber than either fresh fluid or air. Taking this into consideration, tearing energy decreases observed for impulse tested and high mileage vehicle tubes correlate well with the results expected from the previously developed model.

[Figure 1, 2, 3, 4, 5, & 6 ILLUSTRATION OMITTED]

References

(1) A.K. Bhowmick, J. Macromolec. Sci., Rev., C(28), 339 (1988). (2) R.G. Stacer, E.D. vonMeerwall and F.N. Kelly, Rubb. Chem. and Tech., 58, 912 (1985). (3) M.E. Nichols and R.A. Pett, Rubb. Chem. and Tech., 67, 619 (1994). (4) R.S. Rivlin and A.G. Thomas, J. Polym. Sci., 10, 291, (1953). (5) F. Haaf and P.R. Johnson, Rubb. Chem. and Tech., 44, 1410 (1971). (6) I. Abu-Isa, Polym, Engin. and Sci., 15, 299 (1975). (7) K.T. Gillen and R.L. Clough, Polym. Deg. and Stab., 24, 137 (1989). (8) I.A. Abu-Isa and H.E. Trexler, Rubb. Chem. and Tech., 58, 326 (1985).