Effect of power steering fluid type on the performance of radial lip seals.

Rubber compounds are increasingly being used to make critical engineering components for under the hood applications in the automotive industry. This has increased the demand on rubber compounds to withstand and perform in very hostile service conditions that include exposures to chemically aggressive fluids (refs. 1-3), high ambient service temperatures and the mechanical working of the surrounding metals. The elastomers that are chosen to be used under these conditions have to be simultaneously resistant to the high ambient service temperature and the low molecular weight working fluids that are in contact with the rubber components. The direct effect of low molecular weight fluids in prolonged contact with a rubber compound is either swelling or hardening of the rubber; or migration into and possible dissolution by extraction of rubber (refs. 4-6). In any case, excessive swelling or hardening degrades performance and can lead to component failure. A swollen o-ring, for example, becomes softer and tends to extrude or creep under its designed pressure. If it becomes hardened, it shrinks and creates pathways for air or fluid leakage. The nature and behavior of chemical additives and their mechanism of migration from rubber compounds were discussed by Ignatz-Hoover, et al (ref. 7). An excellent review of the durability of elastomers for severe fluid duties was published in Rubber Chemistry and Technology (ref. 8), highlighting several service conditions in which rubber compounds can be used and how such different service conditions and fluids can affect elastomer properties and performance.

The present work focuses on rubber compounds as used in the sealing components of most power steering (PS) systems in mobility vehicles. We particularly discuss the results of the degradation of sealing performance if the wrong fluid is used. A wrong fluid could be defined as any fluid that is not recommended for a PS system. We start by briefly explaining the power steering (PS) system for mobility vehicles, its sealing component, the fluids of interest (figures 1a and b), and the diffusion of fluids in rubber compounds. The work concludes with the discussion of the results of our laboratory tests and a few cases of seals from the field where non-ATF fluid was used.

[FIGURES 1a-1b OMITTED]

Power steering system

The power steering system for mobility vehicles consists of several parts that work together. These include the PS fluid reservoir, the hydraulic pump, the rack and pinion gear box, and a system of hoses that includes suction, pressure and return hoses. This work concentrates on radial rack and pinion seals typically made from hydrogenated nitrile (HNBR) because of its general resistance to fluids and chemicals, its thermal stability and its flexibility over a wide spectrum of temperature.

Radial lip seals are integral to automotive power steering (PS) systems. They are located on the rack and pinion components of the steering gear. The pinion shaft has the input and output seals, while the rack shaft has the rack seals. The seals are, on one side, in constant contact with pressurized PS fluid at its service temperature, while the other side is exposed to the atmosphere. The purpose of the seals is to prevent pressurized PS fluid from leaking, thereby maintaining pressure that assists vehicular steering. The seal also prevents ingress of contaminants external to the steering system. A typical example is illustrated in figures 2a and b. The actual working of such a seal is illustrated in figure 2c, showing the lip in contact with a shaft. A radial lip seal is required to effectively perform these fundamental functions under sometimes-hostile ambient conditions. Its environments include combinations of irregular dynamic motion, high temperature (135-140[degrees]C), or extreme cold temperatures (-40[degrees]C), in the presence of pressurized PS fluid.

[FIGURES 2a-2c OMITTED]

Automatic transmission fluids (ATF)

Automatic transmission fluids (ATF) are engineering fluids that are designed specifically for transmission gears. The history of the development of manual or automatic transmission gears and the associated fluids is briefly described in reference 9. Most ATF fluids are derived from petroleum base stock with additional modifying formulations for specific gear and functional applications. This requires careful and balanced additives and ingredients to ensure its performance. ATF fluids are formulated to be resistant to service-induced oxidation and sludge formation (ref. 10), to provide sufficient friction for locomotion and to maintain long-term thermal stability (ref. 11).

ATF fluids have good anti-wear and corrosion resistance. They are thermally stable and can conduct heat away from metal parts, in addition to serving as lubricant to hot metal parts, without degrading. Performance additives in engineering fluids, however, can be detrimental to rubber seals. When properly formulated, ATF can provide its services without significant adverse effect on rubber or thermoplastic compounds and seals that are integral to gears. Due to these and other desirable properties, ATF is the hydraulic fluid of choice for use in hydraulic or power assist steering systems for mobility vehicles. A typical ATF is Dexron, which is widely used in many platforms (figure 1a). Other important engineering requirements of ATF have been described elsewhere (refs. 12-14).

Automotive engine oil

Commercial SAE 15W40 multi-grade automotive engine oil was chosen for this work because of its comparative lower cost to typical ATF, availability and potential temptation to use it in the place of ATF.

Prolonged contact between fluids and rubber seals: Diffusion of fluids into rubbers seals

When a lower molecular weight fluid is in contact with seals, there is the tendency of the fluid to diffuse into and penetrate the polymer. This is particularly true when molded rubber seals are in contact with aggressive fluids at elevated service conditions. The speed and degree of penetration depends on the interactions between the seal and the liquid, the length of time, the existing temperature and whether the contact is in static or dynamic conditions. The general physical description of fluid absorption into materials has been given (ref. 17) and described mathematically, as shown in equation (1). The steady state solution is obtained when the right hand side of equation (1) is zero. The complexity of the mathematical description depends on the geometry (refs. 1-4) and many geometry types have been described in reference 17. The prediction for simple diffusion of liquid into a flat rubber sheet, for example, is shown in figure 4.

[FIGURE 4 OMITTED]

The non-steady state mathematical description of the diffusion of liquids into the seal is given in reference 17 for radial diffusion in a cylindrical geometry.

1/r d/dr [rD dC/dr] = [partial derivative]c/[partial derivative]t for a < r < b (1)

Where: r = radius of the cylindrical seal; D = diffusion coefficient (a constant defined by the fluid penetrating power); C = fluid concentration; t = time; and a and b are the minimum and maximum possible depth (radius) of fluid penetration into the seal.

For components with regular solid boundaries, the solution to equation (1) is readily available in the literature, at r = a (minimum penetration), C = [C.sub.1], and at r = b (maximum penetration), C = [C.sub.2]. However, for a typical rotary seal, [C.sub.1] = [C.sub.2], the concentration of PS fluid in contact with a seal in the course of its service remains constant at its surface. A modification of the solutions that give the mass uptake of fluid in a hollow cylinder with similar boundary conditions, but with an initial fluid distribution of C(r) = [C.sub.o] in a seal continuum within the region r [less than or equal to] [xi] [less than or equal to] R, has also been given (ref. 17) in terms of zero order Bessel functions of the first ([Y.sub.o]) and second ([J.sub.o]) kinds, as shown in equation (2), with shape of the curves determined by the geometry.

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (2)

where: [m.sub.t] = mass of fluid that has diffused into the seal at any time t; [M.sub.[varies]] = mass of fluid that could possibly diffuse into the seal at time = infinity; R = outer maximum radius of the seal; r = the radius of the diffusing fluid front at any time t; D = the diffusion constant for the fluid; [alpha] = a constant; and [J.sub.o] = a Bessel function.

Experiment

Seal material

The seal material was made of hydrogenated nitrile, which is a copolymer of butadiene and acrylonitrile (ACN). The base elastomer is generally more resistant to grease and most ingredients and chemical additive packages in typical lubricating gear oils than the unsaturated nitrile. However, lubricant additives vary with the targeted product and intended application, but generally could include detergent, friction reducers, antioxidants, anti-wear agents, anti-foams, pressure reducers, corrosion inhibitors and dispersants.

Power steering fluids

Two fluids, 15W40 oil and Dexron ATF, were used in this work. These two commercial fluids were obtained from commercial retail stores. One additional ATF was also tested with compound slabs.

The performance and physical properties of seals that were evaluated included the inner diameter (ID), hardness and sealing force as measured on the lip as lip load (figures 2a and b). The physical properties were measured before and after soaking the seals in the respective fluids at the temperature and time indicated.

We first obtained the mass of fluid uptake in cured slabs made from an HNBR compound at 100[degrees]C with time. This was followed with testing commercial seals. Three replicate pinion seals were soaked in each fluid at each temperature of 93[degrees], 104[degrees] and 121[degrees]C for the respective time duration of 250, 500 and 1,000 hours. Additional pinion seals were heated in the air-circulated oven, but without contact with fluid. The choice of temperature is dictated by the expected service conditions of steering gears in a typical passenger car or long distance haulage truck.

The physical dimensions were measured using a calibrated non-contact laser vision system attached to a UMM-500 Zeiss measuring instrument. The sealing force was measured using a strain gauge electronic device, and the hardness was measured using an Exacta 2000 micro-hardness tester (Newage). The viscosity of each fluid was measured before and after heating using spindle #2 of a Brookfield Viscometer (with guard) at room temperature.

Discussion

Figure 1a shows the two fluids on a side-by-side visual comparison before they were heated. Figure 1b shows the appearance of the respective fluids after 1,000 hours at 121[degrees]C. We defined a thermal viscosity degradation index VDI as ([Vis.sub.i] - [Vis.sub.f])/[Vis.sub.i]. [Vis.sub.i] and [Vis.sub.f] are the single point viscosity measured "before" and "after" heating the fluid, respectively. The VDI of the ATF is half that of the 15W40 oil, indicating the ability of the ATF to maintain its performance properties at service temperature for a prolonged time.

The cross-section functional diagram of the seal is shown in figure 2c, illustrating the sealing lip, the garter spring, the metal shaft, the inner (fluid side) and outer (air side) environment. When a fluid causes performance degradation in a seal, it could be in the form of excessive swelling or excessive shrinkage of the exposed seal. Figure 3a illustrates when a seal becomes swollen. A swollen lip seal accommodates a significant amount of fluid that makes it become spongy, softer, and decreases its load bearing capacity. Excess diffusion of fluid into seals creates additional residual stress on the rubber molecules. This leads to premature lip wear, because of the additional apparent increase in seal interference. On the other extreme are the fluids that extract performance-enhancing ingredients out of the seal. At elevated service temperature, some non-ATF fluids act as a solvent to rubber compounds, with high penetrating power. Such fluids extract unattached ingredients away from the rubber, leaving the seal to shrink excessively (figure 3b). Excessive shrinkage leads to increased hardness and loss of elasticity of the lip. It could also lead to increases in ID and leakage. These could be further made worse by low temperature.

[FIGURES 3a-3b OMITTED]

We have tested slabs of two seal compounds in three fluids at 100[degrees]C. Figure 4 shows the prediction of fluid migrating into a fiat slab of rubber compound. It predicts that at infinite thickness, fluid penetration into the slab decreases asymptotically to almost zero value. Figure 5 is the actual observation using two ATF fluids and one non-ATF fluid. It suggests differences between the two ATF and a non-ATF fluid. The ATF fluids migrate into the slab and stay until an equilibrium value is reached, while the non-ATF fluid appears to extract all unattached ingredients from the slab. This effect can lead to stiffness of the lip and loss of ability of the lip to follow the contour of the shaft.

[FIGURE 5 OMITTED]

Figure 6 shows the prediction of fluid migrating into a pinion seal, with its partially cylindrical geometry, as described by the solution to the diffusion equation (equation 2). It predicts that as r [right arrow] R, the amount of fluid increases to its maximum possible level. We used two seal formulations, compounds 1 and 2, and measured the mass of fluid intake. The plot of the result with time is shown in figure 7 for compound 2. In figure 7, the ATF fluids show an increase in fluid uptake with time, but the non-ATF fluid shows a decrease of weight with time. The effect of differences in compounding of seal materials is illustrated in figure 8. While the observations of figure 7 hold for both compounds, we further observe a narrow window of fluid uptake for compound 1 when the plots are combined in figure 8. We observed increases in weight, with time of immersion, for both compounds in ATF, but a decrease in the weight of both compounds when immersed in non-ATF fluid. From figure 8, it is obvious, and further confirms what is already known in practice, that seal compounding can have an effect on the seal performance with different fluids. Whereas compound 1 had very narrow variation in fluid sorption in the three fluids (non-ATF 15W40, ATF-1 and Dexron), compound 2 showed small swelling in the two ATF fluids but shrinkage in the non-ATF fluid. It further shows that the choice of fluid can determine the performance of seals in PS gears, once the choice in compounding has been selected. On the other hand, perhaps if the fluid is determined a priori. the seal makers could engineer their sealing compounds to accommodate the fluid.

[FIGURES 6-8 OMITTED]

The physical characteristics that are normally used to characterize a seal returned from warranty include: simple visual observation of the lip area of the seal under a microscope, measurement of hardness if possible, determination of the dimensional change in ID and OD of the seals, and measurement of the lip sealing force. We have performed some of these basic measurements on commercial pinion seals soaked in two commercial fluids (one ATF and one non-ATF). Figure 9 shows the change in micro hardness of the seals after they were heated in each fluid at 93[degrees], 104[degrees], and 121[degrees]C for 250, 500. and 1,000 hours. Consistently, the seals that were soaked in 15W40 fluid became harder than the seals that were soaked in Dexron ATF fluid. We have measured the change in the inner diameters (change in ID = [final ID - initial ID]) as functions of time and temperature. When the ID of a working seal increases significantly more than the built-in interference, the seal loses its sealability. A small decrease in ID indicates moderate swelling in the region of about 5% or less. This amount of swell is needed to compensate for the small, but inevitable, initial wear that the pinion shaft would cause in seal lips. Small decreases in ID, indicating moderate swell, were observed among the seals that were soaked and heated in ATF (figure 10). The data in figure 10 further confirm the observations for slab compounds in figure 8. Figure 10 indicates that if the seals were used in conjunction with 15W40 fluid in a steering gear, they would likely lose their sealing capability sooner than the seals used in ATF-filled steering gears. The hardening of rubber compound can be linked to migration by extraction of unbonded components, which could include plasticizers, anti-degradation additives and waxes (refs. 7 and 8). Depletion of such additives leads to early degradation of properties and leakage. This is a potential result of using non-ATF fluid in gears.

[FIGURES 9-10 OMITTED]

The ultimate test for sealability is the sealing force or lip load. It is a measure of the grip of the seal lip on the shaft. Figure 11 shows the change in the sealing force for pinion seals that have been soaked in Dexron and 15W40 fluids at the temperature and time indicated. For each fluid, the temperature effect is obvious. With increases in temperature and time of exposure, the hardness of the lip increases. A combination of long exposure time and high temperature has a detrimental effect on seal performance, particularly when non-ATF fluid is used. However, it could be seen that when the ATF is used, the combination of long time and high temperature exposure may not totally degrade the seal performance to the level that can cause failure. This is the result of the difference between the ATF and non-ATF abilities to conduct heat away from the seal and metal surroundings. Figure 11 shows that soaking a seal in Dexron fluid for 1,000 hours at 121[degrees]C is not as detrimental as soaking in 15W40 fluid at 93[degrees]C for the same period of time. The significantly higher lip load values for the seals soaked and heated in 15W40 fluid could be attributed to the hardening of the seals and by shrinkage. Hardening increases the modulus and degrades the flexibility and the ability of the lip to seal properly.

[FIGURE 11 OMITTED]

While we did not focus much attention on the heated fluid itself, we measured the respective viscosity before and after heating them for 1,000 hours at 121[degrees]C. Table 1 shows the reduction in viscosity, contrary to the expectation of a small increase in viscosity due to removal of volatiles during heating. It is, however, important to note that the heating was in a static condition, where the fluids were not worked or put under shear that could further degrade their molecular chains into smaller lengths. This absence of shearing could explain the small reduction of viscosities.

We have further measured how the physical properties of seals were affected when heated in air, and compared the results to when heated in the two fluids. As expected, figures 12-14 show the adverse effect of increasing temperature on the physical properties of seals. At 121[degrees]C, the seals that were heated in 15W40 fluid became harder than those heated in air or ATF fluid. Again, the effect of heating is minimal on the seals heated in ATF. It is important to note, too, that all the tests were conducted in static conditions. The combined effect of time, temperature and dynamic articulation is a subject for future consideration. We, however, have produced road data from returned seals that were suspected of losing seal capabilities because non-ATF fluid was used in the gears. Figure 15 shows the relevant physical data (ID, hardness and sealing force) that were measured on the returned seals in a case-by-case basis and compared with that of a new seal. In all cases, the use of non-ATF fluid had significant detrimental effect on all indicators of good sealing properties of seals.

[FIGURES 12-14 OMITTED]

Conclusions

The safe and satisfactory performance of most power assist vehicular steering systems for mobility vehicles depends on using the recommended power steering fluid in the gear. This is because the rubber seals that are integral to the inner working of gears can be adversely affected if the power steering fluid is not compatible with the seals. In the working conditions of a typical steering gear, rubber seals can swell or shrink excessively, or can be in other ways degraded by the contacting fluid. Hydrogenated nitrile (HNBR) rubber compounds and seals were studied in this work. HNBR was chosen because of its being a choice elastomer for making PS gear seals.

The ATF fluids were less destructive with the two seal compounds at the temperatures investigated. The non-ATF fluid tended to extract unattached and un-bonded compounding ingredients from the rubber matrix. This leads to excessive shrinkage, increasing hardness, reduced lip flexibility, reduced interference and seal leakage.

References

(1.) L.R.G. Treloar, "Effect of network breakdown and reformation on the swelling of rubber in compression," Rubber Chem. & Tech., vol. 42, no. 2, p. 589 (1969).

(2.) E. Southern and A.G. Thomas, "Diffusion of liquids in crosslinked rubber," Rubber Chem. & Tech., vol. 42, no. 2, p. 495 (1969).

(3.) K. Ono, A. Kaeriyama and K. Murakami, "Effects of diffusion in oxidative degradation of vulcanized rubbers: I. Rate of chain scission in the steady state," Rubber Chem. & Tech., vol. 50, no. 1, p. 43 (1977).

(4.) J.M. Bouvier and M. Gelus, "Diffusion of heavy oil in a swelling elastomer," Rubber Chem. & Tech., vol. 59, no. 2, p. 233 (1986).

(5.) H. Oikawa and K. Murakami, "Some comments on the swelling mechanism of rubber vulcanizates," Rubber Chem. & Tech., vol. 60, no. 4, p. 579 (1987).

(6.) J.T. South, S.W. Case and K.L Reifsnider, "Effects of thermal aging on the mechanical properties of natural rubber," Rubber Chem. & Tech., vol. 76, no. 4, p. 785 (2003).

(7.) F. Ignatz-Hoover, B.H. To, R.N. Datta, A.J. De Hoog, N.M. Huntink and A.G. Talma, "Chemical additive migration in rubber," Rubber Chem. & Tech., vol. 76, no. 3, p. 747 (2003).

(8.) R.P. Campion, "Durability review of elastomers for severe fluid duties," Rubber Chem. & Tech., vol. 76, no. 3, p. 719(2003).

(9.) Automotive Lubricants Reference Book, 2nd edition., ed. by R.F. Haycock and J.E. Hillier, p. 263, Society of Automotive Engineers, SAE International, Warrendale, PA (2004).

(10.) H. Ohtani, et al., "Oxidation stability of automatic transmission fluids," a study by the International Lubricants Standardization and Approval Committee (ILSAC) ATF Subcommittee, SAE Paper No. 2001-01-1991, Society of Automotive Engineers, Warrendale, PA (2001).

(11.) E.J. Friihauf, "Automatic transmission fluids--some aspects of friction," SAE Paper No. 740051, Society of Automotive Engineers, Warrendale, PA (1974).

(12.) R. Tourret and E.P. Wright, (eds.), "Performance and testing of gear oils and transmission fluids," Proceeding of the International Symposium of the Institute of Petroleum, (1980).

(13.) R.F. Watts, et al., "The impact of evolving automatic transmission fluids on base oil selection," SAE Paper No. 2001-01-1992, Society of Automotive Engineers, Warrendale, PA (2001).

(14.) H.E. Henderson and B. Swinney, "High quality base oil for next generation automatic transmission fluids," SAE Paper No. 982666, Society of Automotive Engineers, Warrendale, PA (1998).

(15.) R.W. Miller, Lubricants and their Applications, McGraw Hill Inc., NY, 1993.

(16.) SAE Fuels and Lubricants Standard Manuals, HS-23, Society of Automotive Engineers, Warrendale, PA (1997).

(17.) J. Crank, Mathematics of Diffusion, Oxford Univ. Press, London, 1975.

Table 1--single point viscosity (cp) of
heated 7 fluids measured using Brookfield
viscometer spindle #2 with guard
at room temperature

                                     15W40       Dexron

Original viscosity                   46.4 cp      11.89 cp
1,000 hrs. at 93[degrees]C           46.1         11.8
1,000 hrs. at 104[degrees]C          45.8         11.9
1,000 hrs. at 121[degrees]C          45.1         11.6
Average viscosity                    45.67        11.77
Viscosity reduction                   0.73         0.12
Vis. degradation. Index               0.02         0.01
(VDI) *

* VDI = (initial value - final value)/initial value

Figure 15--comparative physical
characteristics of new seal and
warranty seals exposed to non-ATF gear
fluids (hardness data on case-1 missing)

ID, lip load and IRHD hardness

             Spr. I.D.     Spr. lip load      IRHD hardness

New seal       25.05           23.33               71.0
Case-1         26.49           17.48           Lip load
Case-2         26.57           15.78               86.2
Case-3         26.45           15.02               88.3
Case-4         26.413          17.633              85.2

Note: Table made from bar graph.

by Gabriel Osanaiye, TRW Automotive

([e-mail])