Compound selection for dynamic shaft seals.

The sealing mechanism of the dynamic shaft seal is complicated. There can be various reasons for seal leakage. Using a long term, worldwide experience, the following most important rubber properties have to be taken into consideration to select or develop a rubber compound:

* Elasticity;

* resistance to lubricated wear;

* magnitude of swelling in oil;

* resistance to the formation of a hard deposit on the lip contact area (carbonization);

* initial stiffness;

* the ability of seals to pump; and

* cold resistance.

It is necessary to take into account that the dynamic shaft seal can leak if any of the above mentioned rubber properties does not meet certain specifications or limits. Limit of all properties depends on the seal application, seal design and operating conditions. Currently it is possible to use different methods and procedures to evaluate rubber properties. Each tested method or the procedure can provide different results. To select the test methods and procedures for each rubber property, it is necessary to know the conditions and type of deformation that mostly represents the real seal function.

The basic aim of our investigation was to select or develop test methods that can replace seal tests with rubber slab tests that will reduce the time and cost of development or selection of rubber compounds.

Elasticity

Elasticity is the ability to sustain a large deformation and promptly return to the original configuration when the deforming force is removed. For dynamic shaft seals, it is necessary to maintain the rubber elasticity during the full period of seal usage. Elasticity of the rubber component provides a seal with followability or the ability of the seal lip to follow the shaft during shaft vibration and shaft surface imperfection. If a seal does not have a proper followability, the gap created between the shaft and seal lip can be the major cause of seal leakage.

There are various types of methods to measure rubber elasticity in extension, compression, twisting, etc. For dynamic shaft seals, the most typical deformation is bending. The test method "Recovery From Bending" (RFB) was developed which represents a seal followability and has a good correlation with seal life. Currently, the Recovery FKM Bending test is approved as standard ASTM D6515 test.

A strip 100 mm long, 2 mm thick and 10 mm wide was bent and clamped with a paper clip. The clamped sample was immersed in oil with aeration. In a previous investigation, it was found that for engine seals only the oil immersion test with aeration can duplicate real field conditions (ref. 2). For engine oils, the oil was changed every 168 hrs. (one week). For gear oils, the immersion was done without oil change, and for transmission fluid the immersion test can be performed without aeration and oil change.

After aging, the sample was released from the clamp and cooled. The distance between the end points is measured and RFB calculated. Using the kinetics of Recovery From Bending change and equation 1 the heat resistance index can be calculated (equation 2).

(1) R = exp([-bt.sup.1/2])

(2) HRI-r = -1/b

where: R - Recovery From Bending; t - aging time, hours; b - coefficient, the slope of the straight line; and HRI-r - Heat Resistance Index, calculated from the kinetics of recovery from bending.

Figure 1 demonstrates that the HRI has good correlation with seal life. For rubber compounds that have very high volume swell in oil, for instance silicone rubber compound, it is necessary to multiply the HRI-r by inverted normalized modulus (equation 3).

[FIGURE 1 OMITTED]

The new parameter is called "sealing force" (SF):

(3) SF = HRI-r * [M.sub.0]/[M.sub.t]

Where: SF - sealing force; [M.sub.0] - modulus 10% before testing; and [M.sub.t] - modulus 10% after aging in oil.

Figure 1 demonstrates good correlation of the RFB measurement with the dynamic seal bench test. Using the equation in figure 3 and factoring that a passenger car ran 15,000 miles per year with average speed of 45 miles per hour, it is possible to calculate the seal life for different compounds: NBR - 0.32 years; HNBR-1 - 1.34 years; HNBR-2 - 1.71 years; AEM - 1.34 years; ACM - 2.13 years; FKM-9 -8.86 years; and FKM-5 - 10.92 years.

This procedure can reduce the time of development and selection of compound for seal applications.

Resistance to lubricated wear

For the last 50 years, essential progress was attained in the area of investigation of surface friction and wear. Most of the research work was dedicated to the investigation of the interaction of solid hard substances, for instance, between metals in dry and lubricated conditions. Many investigations were done for friction and wear of rubber with asphalt or concrete, and different theories about wear mechanisms were created. The area of friction and wear of elastomers and metal in lubricated conditions was not properly investigated and published. Some very limited information can be found in the D. Moore book (ref. 3). It was shown that absorption of oil by a rubber surface creates a thin layer that has different properties than oil and rubber. The properties of this layer are functions of the type of rubber compound, chemistry of the lubricant and swelling conditions.

At the present time, only one method is available to measure the seal wear in lubricated conditions; the seal bench test. To compare rubber compounds it is necessary to mold and test seals during various periods of time and obtain correlation between test times and wear widths on the seal lip contact area. It was shown that kinetics of lubricated wear can be described by the following equation:

W = [at.sup.b]

where: W - wear width; t - running time. hrs.; and a,b - coefficient.

By knowing the correlation for each compound, it is possible to calculate the wear width at certain periods of time.

An attempt to replace the lubricated seal wear test with a dry seal wear test such as the PICO abrasion resistance test (D 2228) or Tabor abrasion resistance test (D 1630) did not show any correlation with the seal lubricated test. Quite the opposite, some rubber compounds that have poor resistance to dry wear such as FKM compounds demonstrated excellent resistance to lubricated wear.

To develop a laboratory lubricated rubber test, the automated computer-controlled Tribometer LRI-1a was selected. The modified LRI-1a instrument allowed testing of wear resistance in oil at various speeds in rotational and oscillatory conditions at different pressures. The instrument can measure the kinetics of wear, friction coefficient and also oil and "under lip" temperatures.

Figure 2 shows the principle scheme of the LRI-1a instrument. The experimental data demonstrate the ability to correlate laboratory data with the seal bench test (figure 3). This experiment shows a potential way to use this laboratory lubricated wear test as a replacement for the dynamic seal bench test.

[FIGURES 2-3 OMITTED]

Volume swell

The magnitude of swelling rubber in oil is one of the very important factors that can have an effect on seal life. The interaction between rubber and oil has been very widely studied by different scientists over long periods of time. It is possible to make a statement that the rubber swelling phenomenon has a more theoretical basis than other rubber properties. In most investigations, the rubber swelled in pure oils and some correlation was established between solubility parameters and volume swell.

Swelling phenomena for fully formulated oil is a lot more complicated. First of all, fully formulated oil for the automotive industry contains a lot of additives that change the volume swell. It is not possible to predict the volume swell for fully formulated oil using the theoretical equations that have not taken into account the influence of all oil additives used in fully formulated oils. It is also necessary to take into account that rubber swelling in oil is a process of diffusion (ref. 4), and consists of two procedures; one is the polymer swelling and the second is an extraction of low molecular components from the rubber compound. The sum of these two opposite processes form the readings of experimental volume swell. When rubber swells in oil at elevated temperatures, the third process takes place, that is rubber aging. The process of rubber aging begins on the rubber surface. From one side, aging increases the crosslink density on the rubber surface and reduces oil penetration in the rubber bulk, but at the same time, temperature increases the kinetic of oil penetration. The complicated process of swelling rubber in oil can be reproduced in lab conditions using a standard ASTM test D471. For engine seals, the best result was demonstrated by an immersion test with aeration. Usually, swelling in oil reduces the mechanical properties. However, an investigation of dynamic properties on dry and swelled rubber demonstrates that in a swollen state, the rubber is less sensitive to temperature and crosslink density changes (ref. 5).

What is the influence of swelling on seal life? Very often the magnitude of swelling rubber in oil is related to the so-called "oil seal compatibility." In reality, it was proven that there is no direct correlation between volume of swell and seal life. It is not possible to reach a conclusion that if rubber swells less in oil, it will provide longer seal life. For each application and seal design, there is a limit of volume swell.

The following example demonstrates this statement (ref. 6). Diminishing of the acrylonitrile content (ACN) leads to the increase of heat resistance of rubber compound which has a positive effect on seal life, but in the same way enlarges the swell in oil which has negative effects on seal performance.

Figure 4 demonstrates the seal life of compounds with different ACN content. As expected, when the ACN content went down, the heat resistance increased. This provided higher elasticity and, as a result, higher followability and naturally longer seal life. When the ACN content diminished less than 22%, the seal life dramatically decreased. Analysis of seals with 18% ACN content after leakage indicated the reduction of radial load and the creation of waviness on the seal contact area. The experiment demonstrates the necessity not to exceed a certain limit of volume swell. It does not mean that rubber compound with 18% of ACN content can not be used for a tested application. To use this compound, it is probably necessary to increase the radial load or change the seal design.

[FIGURE 4 OMITTED]

Carbonization

Dynamic shaft seals may leak during their usage, while the basic rubber member remains flexible and still has a good shaft followability. One of the reasons for seal leakage in this case could be the carbonization created on the lip contact point. The carbonization phenomenon on the seal lip usually appears on the seal compounds that have high heat resistance, especially those based on fluoroelastomers. However, in some cases, seals that have carbonization on the lip did not leak and ran a very long time. It was not possible to identify if the carbonization is really the cause of seal leakage.

In our investigation, it was found that the carbonization phenomenon is not only the process of sticking carbonized oil to the rubber surface. In some cases, the oil penetrates into the rubber bulk and carbonizes inside the rubber lip area. The carbonization depends on oil thermostability, resistance of rubber to the interaction with aggressive oils and oil additives at elevated temperature.

A laboratory method was developed to evaluate the resistance of rubber with oil and evaluate oil aggressiveness (ref. 7).

The 1 x 2 samples punch off from 6 mm slab and put it in oil for aging. After aging, the 6 mm strips were cut off from the aged sample and the micro hardness was tested on two surfaces. First, on the side surface, and the second on the cross-section of the sample. The differences between these two measurements indicate the difference in hardness on the surface and in the rubber bulk and represent carbonization. The difference in IRHD was named "carbonization index."

Five fluoroelastomer compounds were selected (table 1) and aged in commercial aggressive gear oil 75w140 with 6% of friction modifier at 166 [degrees] C.

Figure 5 shows the kinetic of the difference in IRHD between outside and inside surfaces versus aging time. By analyzing the test results, it is possible to conclude that the FKM-5 compound did not demonstrate the hardness change after aging in oil up to 504 hours. This compound is most resistant to the process of carbonization. The remaining compounds are very dependent on aging time. The differences in hardness grew when the aging time increased. For all compounds in both oils, there is a satisfactory power correlation, which in log-log coordinates gives a straight line. The slope of the straight line can represent the characteristic of oil carbonization. Carbonization can also be characterized with hardness change after a certain period of aging in oil.

[FIGURE 5 OMITTED]

Table 2 demonstrates the comparison of carbonization indexes for different FKM compounds. This rating of rubber compounds correlates with the test results obtained on seal bench tests and under field conditions. Thus, this method can be used to evaluate resistance to carbonization for different rubber compounds, and also the level of oil aggressiveness.

Stiffness

One of the important seal characteristics is radial force or radial load, the total force acting on the seal lip that tends to maintain lip contact on the shaft. Radial load depends on the interference between the seal lip and the shaft, the pressure of the garter spring, if present, and the rubber stiffness.

To select a compound with the required stiffness, the measurements of hardness or extension modulus are usually used. Principle of hardness measurement is process of compression or more precisely penetration of needle into rubber bulk. The tension modulus, first of all, represents the rubber property in tension mode, and second this modulus is usually measured at high deformation of 50, 100 or 300%. Measurement of tension modulus at small deformation in the same way as hardness measurements shows very high variations that do not allow the use of these methods for compound selection.

In reality, dynamic shaft seals, like most rubber articles, work at small deformation. It is well known that rubber properties at small and high deformations are different. Figure 6 demonstrates the stress-strain curves for two compounds which are usually called high and low modulus compounds. It is possible to see that a so-called high modulus compound has lower modulus at small deformation.

[FIGURE 6 OMITTED]

Because seals utilize small deformation during usage and the typical mode of deformation is bending, a meth-od was developed to measure the rubber stiffness in bending (ref. 8). A typical deflection curve in bending is demonstrated in figure 7. Two characteristics can be calculated from this curve; maximum bend stress, and stiffness modulus in bending.

[M.sub.bd] - maximum bend stress, kg/[cm.sup.2](psi)

[M.sub.bd] = 3PL/2b[d.sup.2]

[S.sub.mb] - stiffness modulus in bending, kg/[cm.sup.2](psi)

[S.sub.mb] = [L.sup.3]M/4b[d.sup.3]

[FIGURE 7 OMITTED]

Where: P - maximum force, kg (lb.); L - support span, cm (in.); b - width of the beam, cm (in.); m - slope of the tangent to the initial straight line portion, kg/cm (lb./in.); and d - depth of the beam, cm (in.).

Figure 8 demonstrates the correlation of experimental radial load and four rubber properties. The stiffness shows the best correlation with radial load and can better predict the seal radial load.

[FIGURE 8 OMITTED]

The stiffness measurement in bending provides more information, particularly for seal compounds, than hardness or tension modulus, and can be recommended for seal design, rubber compound development and rubber selection.

Seal pumping ability

The ability of rotational lip seals to pump oil from the air side to the oil side is another important seal characteristic. The pumping mechanism has been investigated for more than 40 years.

Currently, there is no theory that can fully explain all the experimental results. The most advanced review of the investigations in this area was published recently by R. Salant (ref. 9).

Although the seal pumping theories are not fully disclosed, from a practical point of view it was established that the ability of a seal to pump depends on two main factors: Seal design and rubber properties.

The aim of this article is not to discuss theories and factors that can influence the pump rate. It is the necessity to establish the level of pump rate that will guarantee the required seal life. Unfortunately, it is not possible to test the ability of a seal to pump without testing the molded seals. Therefore, seals with two compounds were tested on the research pump tester. An NBR compound with an initial pump rate of 14 x [10.sup.-2] ml per minute, and an HNBR compound with an initial pump rate of 4 x [10.sup.-2] ml per minute were tested. During the continuous pump test (measurements provided every 24 hours), the NBR seals with high pump rate leaked 2-1/2 times earlier than the HNBR seals (figure 9).

[FIGURE 9 OMITTED]

This experiment demonstrates that the initial pump rate is not a predominant factor responsible for seal life. The HNBR rubber has better heat resistance and maintains elasticity and suitable followability for a longer period of time. When the NBR rubber lost followability, a gap was created between the seal lip and shaft, and the seal leaked.

Thus, seal pumping ability is a very important factor. If a seal does not pump, the seal will instantly leak. It is not necessary to have a very high initial pump rate. There is a limit for the initial pump rate. This limit depends on seal design, interference, speed and spring load. In the future, all theories about the pumping mechanism will have to take into account the gap between the seal lip and shaft. The seal can pump only when the gap between the seal lip and shaft does not exceed a certain level.

The experiment described above was done with plain lip seals. If the seals were made with hydrodynamic fixtures, like helices, the requirements for the seals, particularly for rubber compounds, dramatically changed. Seals with two NBR compounds were tested; one is the same compound that was shown in figure 9, the second is a compound for reciprocational seals which has zero pump rate on regular plain lip seals. Figure 10 demonstrates that seals with hydrodynamic helices with rotational compound increase the initial pump rate from 14 to 110 x [10.sup.-2] ml per minute, and seal life increases from 700 to 1,200 hours. The second fact, that is very impressive, is that seals with reciprocational compound which did not pump at all in the plain lip seal design have a high initial pump rate and the seal life is even higher than for the rotational compound. A very important conclusion can be drawn from this experiment. For rotational dynamic lip seals with hydrodynamic helices, it is not necessary to develop a compound that is able to provide pumping ability. In this case, it is possible to create a compound with higher elasticity, better resistance to wear and higher physical properties.

[FIGURE 10 OMITTED]

Resistance to cold temperature

Currently, resistance of seals to cold temperature is measured in the so-called "cold box" test according to ASTM J110.

Testing one seal using the cold box requires one week. To reduce testing time and avoid the necessity to mold and test seals, a project was established to correlate the standard cold rubber slab test with the test results obtained with the cold box test.

First of all, the parameters for the cold box test have to be established which introduce the seal conditions in the field. It was presumed that during seal rotation at high speed, the friction increased between the seal and shaft. When the friction increases, the under lip temperature also increases, defrosting the rubber and suitably increasing rubber elasticity and seal followability. A thermocouple has been mounted in the lip area, about 0.5 mm from the seal contact point, and the seal was installed in the cold box tester.

The test was started with an FKM compound with a glass transition temperature ([t.sub.g]) of -12.8 [degrees] C (9.8 [degrees] F) in mineral SAE 5w30 oil. Before the test began, the seal was soaked at -40 [degrees] F for at least eight hours. Rotation started at 700 rpm that represents the idle speed for warm-up period. After five minutes, the speed was increased to 1,500 rpm that represents the beginning of actual driving. After 35 minutes, the speed was increased to 2,000 rpm, and after 90 minutes to 3,000 rpm.

The test results shown in figure 11 demonstrate that after 1.25 minutes the under lip temperature rises dramatically, and after two minutes the temperature increases from -40 [degrees] F to -10 [degrees] F. In this case, seal followability is high enough to prevent seal leakage. In the same period of time, the oil sump temperature is still low. The oil temperature reaches the under lip temperature only after 15 minutes, and after 2.5 hours the temperature was 275 [degrees] F with a tendency for future escalation.

[FIGURE 11 OMITTED]

Oil leakage also depends on oil viscosity. If the oil viscosity at low temperatures is very high, even with the gap created between the shaft and seal lip, the seal will not leak. To investigate the influence of oil viscosity, additional experiments were performed with lower oil viscosity at -40 [degrees] C. Mineral SAE 5w30 with an oil viscosity of 3 x [10.sup.3] mPa*s, synthetic oil Mobil 1 with 2.1 x [10.sup.4] mPa*s, and automatic transmission fluid (ATF) were used. Although the ATF is not used in engines, it was taken as a model because it has a very low viscosity of 0.9 x [10.sup.3] mPa*s which is 23 times lower than the viscosity of Mobil 1. The experiments with all three oils did not show oil leakage and did not demonstrate any substantial differences in the under lip temperature, even with ATF.

The tests described above were done with seals containing hydrodynamic helices. The concern was brought up that helixes increase pumping ability and prevent oil leakage. To verify this statement, a test was done with a seal without helixes. The FKM compound was selected with the highest glass transition temperature ([t.sub.g] = 0 [degrees] C or 32 [degrees] F). The test results were the same - no leakage occurred on seals without helixes with very high glass transition temperatures. The conclusion can be made from this experiment that under test conditions, which correspond to field conditions, at -40 [degrees] C and high speed, seals did not leak, even when the glass transition temperature is 0 [degrees] C and the oil viscosity is as low as 0.9 x [10.sup.3] mPa*s at -40 [degrees] C.

At the high rotational shaft speed, seals can crack and leak only when a car is started at very low temperatures without a heating up period. In this case, a very important seal property is the resistance of the rubber to the impact conditions at low temperature, such as the brittleness point. The RMA investigation (ref. 10) established a good correlation between seal leakage and ASTM D 2137 brittleness point test. Using this experimental data, an equation was derived between seal leakage temperature ([t.sub.sl]), brittleness point ([t.sub.bp]), and dynamic run-out ([d.sub.ro]), the difference between centers of shaft and seal rotations (eccentricity).

[t.sub.sl] = -25 + [t.sub.br] + 12.6 x [d.sub.ro]

Table 3 demonstrates excellent correlation between experimental and calculated test results.

If a seal is tested at low speed (60 rpm), typical for axle seals, in gear oil with high oil viscosity at low temperature flexibility, the seal starts leaking even when the seal test is started at -15 [degrees] C (5 [degrees] F).

Figure 12 shows the cold box test for three inch FKM seals with Tg -12.8 [degrees] C (8.9 [degrees] F) in gear oil Emery Emgard. After 12 minutes, the lip temperature is 9 [degrees] F, and in spite of the oil temperature being very low a meniscus of oil appears between seal lip and shaft. This temperature correlates with the glass transition temperature.

[FIGURE 12 OMITTED]

This experiment demonstrates that for seals with low rotational speed, the rubber cold temperature performance is very important. In this case, to characterize the seal cold resistance the glass transition temperature, the ASTM D 1053 "Stiffening at Low Temperatures" or the ASTM D 1329 "Retraction at Low Temperature" tests have to be used to select rubber compound.

Conclusions

To develop or select a rubber compound for dynamic shaft seals, the following rubber properties and laboratory test methods can be used.

* Rubber elasticity - elasticity is one of the most important rubber properties responsible for seal followability. To measure the elasticity, it is recommended to use the recovery from bending (RFB) test ASTM D 6515 and measure the kinetics of RFB change in oil with aeration (where appropriate) and calculate the heat resistance index (HRI-r).

* Resistance to lubricated wear - to measure the resistance to lubricated wear (RLW) the modified laboratory LRI-1a tester can be used which shows a satisfactory correlation with the seal bench test. Note: HRI-r and RLW have direct correlation with seal life.

* Volume swell - rubber volume swell does not have a direct correlation with seal life. For each seal design and application, there is a limit to volume swell. The ASTM D 471 test method can be employed in oil used for this application.

* Resistance to carbonization - resistance to carbonization is a very important rubber property, especially for heat resistant compounds such as fluoroelastomers when used in oils with aggressive oil additives. To measure carbonization, the ASTM D 1415 international hardness test method can be used with the procedure described in the reference (ref. 7). If rubber has insufficient resistance to carbonization, the seal can leak no matter how good the remaining properties are.

* Stiffness - stiffness is a very important rubber property to select compound with a required seal radial load. Stiffness is described more precisely as rubber rigidity than hardness or modulus. To measure the stiffness, it is necessary to use the device described in reference 8.

* Pumping-ability - pumping-ability is a very important seal property. Unfortunately, it is not possible to test seal pumping-ability using any slab test method. If a seal does not pump, it will leak instantly. There is no direct correlation with the magnitude of initial seal pump rate and seal life. Most important is for kinetics of pumping-ability. Seal pumping-ability can compensate the shaft imperfections and shaft lead.

* Cold resistance - for seals with a high speed application, the ASTM D 2137 brittleness point test demonstrates good correlation and can be used to predict seal cold temperature performance. For low speed seals, the ASTM D 1053 stiffness at low temperature test can be recommended.

Table 1 - characteristics of tested compounds

Compound   Polymer                Fluorine     Curative
   FKM     characteristic         content, %

    1      Copolymer                 66        Bisphenol
             V[F.sub.2],HFP
    2      Terpolymer                68        Bisphenol
             V[F.sub.2],HFP,TFE
    3      Terpolymer                70        Peroxide
             V[F.sub.2],HFP,TFE
    4      Terpolymer                56        Bisphenol
             V[F.sub.2],TFE,P
    5      Copolymer TFE,P           52        Peroxide
Table 2 - comparison of carbonization indexes for
FKM compounds

Compound     Difference in     Carbonization
  FKM           hardness            index
             after 504 hrs.      from slope

   2               22               0.63
   1               21               0.63
   4               10               0.55
   3                7               0.49
   5                1                 0
Table 3

Material     [d.sub.ro],     [t.sub.bp],      [t.sub.sl],
                 mm          [degrees] C      [degrees] C
                                              experimental

A               0.5             -30.7           -49.5
                0.75            -30.7           -46.5
                1               -30.7           -42.5

B               0.5             -15.9           -34
                0.75            -15.9           -30.6
                1               -15.9           -27.5

C               0.5              -4.3           -23.2
                0.75             -4.3           -19.9
                1                -4.3           -17

Material     [t.sub.sl],     Difference,
             [degrees] C     [degrees] C
             calculated

A               -49.4            0.1
                -46.2            0.3
                -43             -0.1
B               -34.3           -0.3
                -31.2           -0.6
                -28             -0.5
C               -23             -0.2
                -19.9            0
                -16.7            0.3

Acknowledgements

"An improved method for measuring filler dispersion of uncured rubber" is based on a paper given at the October, 2001 meeting of the Rubber Division.

"Wing function technology - a new rotor technology for the Farrel Banbury mixer" is based on a paper given at the October, 2001 meeting of the Rubber Division.

"Compound selection for dynamic shaft seals" is based on a paper given at the October, 2001 meeting of the Rubber Division.

References

(1.) B. Dinzburg, "Measurement of rubber elasticity and correlation with seal life," SAE Congress, Detroit, Michigan, February 1997, Paper 970547; ACS Rubber Division, Cleveland, October 1995, Paper #71; Rubber & Plastics News, October 7, 1996.

(2.) B. Dinzburg, "Influence of lubricant additives on rubber properties in conditions similar to the field," Lubrication Engineering 51,10, 796-804, 1995.

(3.) D. Moore, "The friction and lubrication of elastomers," Pergamon Press, 1973.

(4.) J.M. Bouvier, M. Geus, "Diffusion of heavy oil in a swelling elastomer," Rubber Chem. & Technology, 59, 233-240, 1986.

(5.) C. Neogi, A.K. Bhattacharya, A.K. Bhowmick, "Dynamic analysis of carbon black filled vulcanizate under swollen conditions," Rubber Chem. Technology, 63, 5, 651, 1990.

(6.) B. Dinzburg, R. Bond and R.W. Keller, "Heat resistance evaluation for rubber compounds," Rubber World 197, 5, 28-32, 1988.

(7.) B. Dinzburg, "Carbonization phenomena on dynamic shaft seals and test method," SAE World Congress, Detroit, MI, March 2000, paper 2000-01-0682.

(8.) B. Dinzburg, R.Bond, "Stiffness measurement for evaluation heat resistance of rubber," Rubber World, 201, 4, 20-24, 1990; Patent #517801Z

(9.) R.F. Salant, "Theory of lubrication of elastomeric rotary shaft seals," Proc. Inst. Mech. Engrs. 213, part 1, 189-201, 1999.

(10.) "Oil seal performance at low temperature and ASTM test method," RMA Technical Bulletin Oil Seal Subdivision OS-10.

Boris Dinzburg, Chicago Rawhide, SKF Seal Division