Enhancing compound properties and aging resistance by using low viscosity HNBR.

Recent advances in HNBR technology have led to the development of low Mooney viscosity (39 MU, ML1+4@100[degrees]C), easy to process HNBR polymers. Compounds based on the low viscosity HNBR polymer have shown improved mixing characteristics (faster black incorporation time, lower mix temperature) and excellent processing characteristics (faster extrusion rates, shorter injection times, lower pressure). Furthermore, the narrower molecular weight distribution gives rise to better mechanical properties.

The 39 MU low viscosity HNBR grade contains 34% acrylonitrile and has been hydrogenated to less than 0.9% residual double bond content. This grade has been developed for applications where improved compound flow is particularly beneficial: wire and cable, oil well industry, as well as automotive applications (seals, gaskets and hoses). In addition to the excellent processing characteristics, the new low viscosity polymer enables the use of known compounding technologies that might otherwise be excluded because of processing limitations.

A screening experiment has been carried out to study the influence of five factors: added silicate filler, plasticizer level, antidegradant system, metal oxide and HNBR polymer viscosity, on the compound properties. The choice of polymer was found to be a statistically significant variable in many of the properties related to processing. The compound viscosities covered a large range, from a low of about 44 MU to a high of 125, while the compound scorch safety was doubled between worst and best case combinations. In general, the compound physical properties are not highly dependant on the polymer viscosity, but are more dependant on filler, plasticizer and antidegradant selection. The availability of the new low viscosity HNBR polymer has enabled alternate compounding techniques that have been previously excluded due to high compound Mooney viscosity and other significant processing considerations.

Hydrogenated acrylonitrile butadiene rubber (HNBR) is a high performance elastomer that is known for its excellent heat/oil resistance and for its superior mechanical properties. Commercially available HNBR polymers are produced by catalyzed hydrogenation of the unsaturation within the polybutadiene chain segment of the butadiene co-acrylonitrile precursor polymer. The hydrogenation process can be controlled to reduce the number of reactive double bonds in the backbone of the polymer chain to a predetermined level. The result is a high performance polymer with improved chemical and temperature resistance.

However, due to the changes in the polymer molecular structure during hydrogenation, the viscosity of the HNBR products are, in general, nearly twice that of the NBR feedstock from which they are produced. The NBR manufacturing process has restrictions that limit the minimum viscosity for polymers that can be practically produced. Consequently, most commercially available HBNR grades have relatively high viscosity; with the low end of the polymer Mooney viscosity in the range of about 60-65 (ML 1+4 at 100[degrees]C).

Efforts have been made to produce a low Mooney HNBR by applying high shear stress (at a shearing rate of 500-5,000 [s.sup.-1]) in the presence of an antioxidant and in the absence of an oxygen-donor. This approach can lower the Mooney viscosity by at least 15 points (ref. 1). However, such a process leads to a broader molecular weight distribution and, in general, the control of mechanical shear processes and chemical environments to give constant and reproducible properties can be rather difficult.

Thus, until very recently the production of a low molecular weight HNBR grade with much improved processability, while maintaining the superior mechanical strength and heat/ oil resistance, has been a challenge.

Guerin and Guo (ref. 2) have developed processes for the manufacture of low molecular weight HNBR products (HNBR-AT, Advanced Technology), which are characterized by narrow molecular weight distribution and significantly lower molecular weight. The nominal Mooney viscosity value for this polymer is 39 MU, a significantly lower viscosity than any previously available HNBR product. Initial studies with the material suggested that the low viscosity polymer permits compounded formulations with excellent improvement in rubber processing characteristics, while the slightly narrower molecular weight distribution allows for superior mechanical performance. Also, the low viscosity HNBR permits the development of formulations containing little or no internal plasticizers. Such compound formulations generally result in improved heat aging and compressive stress relaxation properties.

In addition to the excellent processing characteristics, the new low viscosity polymer enables the use of known compounding technologies that might otherwise be excluded because of processing limitations. The object of this investigation is to explore various known compounding strategies in combination with low viscosity HNBR in order to better understand the direction for development of HNBR formulations with improved physical properties and improved aging resistance.

Experimental

The compounds prepared for this study were mixed in a laboratory BR-82 internal mixer. Standard laboratory mixing procedures were used to incorporate the curatives in a separate mixing step on an open mill (10 x 20 inches). All physical tests were carried out according to ASTM methods. A listing of these tests is shown in table 1.

The compound formulations used in this assessment were based on a simplified peroxide test recipe in which three selected categorical (discrete) variables and two continuous variables were evaluated. The experimental design is a simple screening experiment, which is represented by a partial factorial design and includes two replicate points.

Table 2 illustrates the independent variables examined and the levels used in the study. A code identifier is indicated in parentheses. In order to simplify the discussions around the effect of the independent variables on the dependant property of interest, the simple code identifier is utilized to elucidate both the identification and the usage level for the variable under examination. The code identifiers are also used in the bar charts to identify the compounds.

The polymer type is indicated by either a (4) or (7) representing the viscosity (ML 1+4 @ 100[degrees]C = 39 or 70) for the low and high viscosity polymers, respectively. The type of antioxidant system used is indicated by (A) to represent a common standard A/O package (substituted diphenylamine/ zinc methylmercaptobenzimidazole) and (H) to indicate the use of Therban HT heat resistance technology. The two levels of magnesium oxide are indicated by (m) for the low level and (M) for the high level. Similarly, the plasticizer level is indicated by (lp) for the low plasticizer level and (P) for the high level. Finally, the filler system is identified as (B) or (W) to indicate a black or white filled system. Note that all formulations contain carbon black, but the compounds identified by (W) contain additional alumina magnesium silicate filler and silane coupling agent, which are not included in the compounds identified as (B).

The peroxide (40% active 2,2' bis [tert-butylperoxy] diisopropylbenzene on an inert carrier) and the co-agent (triallyl isocyanurate) that were used for vulcanization are commercial grade. The silane coupling agent addition rate was 3.5 phr (72% active on inert carrier) whenever the silicate was incorporated. Carbon black (N 660) and zinc oxide were common to all the compounds.

Results and discussion

Table 1 shows the ASTM test procedures that were used to evaluate the physical properties of the compounds produced from the experimental mixes. The resultant compound data have been analyzed using SAS JMP software, and the best-fit models were used to determine the most significant factors. The actual test results are shown graphically as bar charts in the figures. The statistically significant main effects are shown in the prediction profile plot and give a good visual indicator of the impact of the most important independent variables. The prediction profile plots are also shown in the figures, and the actual data points that most closely correspond to the low value and to the high value profile plot predictions are highlighted in the bar graphs. Two-factor interaction effects and higher are confounded in this design and are not considered.

Unaged physical properties

The effect of the compounding variables on the compound Mooney viscosity is shown in figure 1. The compounds that give the highest and the lowest result are highlighted. The bar chart shows the Mooney viscosity can range from a low of about 44 Mooney units to almost three times that value for the high Mooney viscosity value. The primary independent factors that contribute to this result are the polymer, the plasticizer and the filler system. The maximum and minimum measured values, as well as the predicted values, are shown. The low value is obtained when the plasticizer level is high at 15 phr (P), the polymer viscosity is low (4) and the filler system does not include the silicate filler (B). Note that in figure 1, three bars are highlighted. The two low viscosity bars are replicates that were included in the design, and these show good reproducibility. The profiler charts, respectively, show the low and high viscosity value predicted by the model. The independent factors with the greatest slope have the largest effect on the predicted value of the dependant variable.

[FIGURE 1 OMITTED]

Figures 2 and 3 illustrate the effect on compound scorch safety at 135[degrees]C and on the MDR cure time, respectively. The primary factors influencing these properties include polymer and metal oxide (magnesium oxide) loading, as well as plasticizer loading for the scorch safety, and filler selection for the cure time. In both instances, the polymer has a large influence, as can be visualized from the slope of the polymer line in the predictor profile plots. Improving the scorch safety of a compound is beneficial, especially for injection molding applications. These results suggest that, with a judicious choice of polymer and filler system, recipes with improved scorch safety and minimal increase in total cure time can be formulated.

[FIGURES 2-3 OMITTED]

The compound hardness (figure 4) and the modulus at 100% extension (figure 5) are influenced primarily by filler, metal oxide and plasticizer loading. These two properties are not highly affected by the use of low viscosity polymer or the use of the HNBR-HT heat resistance technology. The filler and metal oxide act in the same direction to cause increased modulus and hardness while, as might be expected, the plasticizer acts in the opposing direction. Exceptionally high stiffness compounds (M100 over 16 MPa) can be achieved by using the 7/H/W/M/lp combination.

[FIGURES 4-5 OMITTED]

Figure 6 shows that unaged tensile at break is influenced primarily by the plasticizer level and by the polymer selection. The filler selection has a smaller influence than either the polymer or the plasticizer level. The influence of HT on unaged tensile at break is slightly less than that of the filler selection, while the influence of metal oxide level is negligible.

[FIGURE 6 OMITTED]

Unaged elongation at break is most strongly influenced by the filler system, the metal oxide and the plasticizer loading levels. The use of high or low viscosity polymer or the choice of antidegradant package has little influence on the ultimate elongation. Figure 7 clearly shows that the strongest influence is the filler, and that the filler and metal oxide act in the same direction. Adding the silicate/silane in addition to the base black loading strongly reduces the elongation at break, while the metal oxide and plasticizer effects are relatively small and in opposing directions. The combined effect is to reduce elongation by a factor of almost half, from 283% to 138%.

[FIGURE 7 OMITTED]

In general, the compression set resistance for these compounds is good. The compression set is minimally influenced by polymer viscosity, but the short-term compression set is influenced by use of high levels of metal oxide and by the selection of the antidegradant package. This latter effect has been previously reported (ref. 3) and the benefit of the HNBR-HT antidegradant system is realized after long-term aging. Examining the slope of the lines in the profiler charts in figure 8 shows that only the filler selection line has a small negative slope, leading to improvement in compression set. This result is likely related to the increase in the number of crosslinks generated through the silane.

[FIGURE 8 OMITTED]

The factors influencing the Die B tear strength are highlighted in figure 9. There is more scatter in the test data, and this fact is reflected in the lower [R.sup.2] value, as well as the broader error bars in the prediction profile plots. Nevertheless, the data suggest directional changes for the independent factors examined in this study. The data suggest that the metal oxide, the antidegradant system and the polymer selection have the strongest influences on the Die B tear result. The prediction profile plots indicate that increasing the reinforcement loading decreases tear strength, while using the lower viscosity polymer and higher loading of plasticizer favor increased tear strength. Even though there is large scatter in the data and the impact on the tear strength for each independent variable is rather small, the general trends are in line with the expected effect of each variable: i.e.. increasing compound stiffness, whether through increased filler or lower plasticity of the compound, leads to decreased tear strength.

[FIGURE 9 OMITTED]

Cold temperature properties

The influence of the compounding variables on cold temperature properties is shown in figures 10 and 11. The retraction at lower temperature (TR) test has been used to characterize the relative performance of the compounds. The [TR.sub.10] value, which correlates to brittle point, is shown in figure 10; while figure 11 shows the [TR.sub.(70-10)] values that represent the recovery tendency for the compound as temperature rises. The [R.sup.2] value for the [TR.sub.(70-10)] is an acceptable 0.87 and somewhat better than the lower [R.sup.2] value of 0.56 for the [TR.sub.10]. The smaller error bars in the prediction profile plots of figure 11 reflect the improved correlation between the predicted values and the actual reported values.

[FIGURES 10-11 OMITTED]

The predictor profile plots suggest that the plasticizer level is the most significant factor for determining both the [TR.sub.10] and [TR.sub.(70-10)] values. This is in agreement with, and confirms the conventional wisdom regarding the use of plasticizer to improve cold temperature performance. For [TR.sub.10], the slope of the profile lines in the predictor plots indicate that filler, antidegradant, metal oxide and polymer play much lesser roles in determining retraction properties. The slopes for these lines are very shallow (slope is almost zero) and are within the spread of the error bars. Even though the second and third most significant factors are shown in the figure, their contribution to the overall effect is minor.

In the case of [TR.sub.(70-10)], the profile plots indicate that filler, antidegradant and metal oxide are not significant factors, and that polymer selection may also be a secondary contributing factor. The [TR.sub.(70-10)] value is a measure of the temperature range over which a frozen component recovers to the rubbery state. High mobility of the molecular segments is a condition of the rubbery state (ref. 4). Consequently, chain entanglements, chain length, crosslink density and crystallization rate all play a role in the elastic recovery of a compound from the frozen state. The dependence of each of these components on polymer selection, plasticizer selection, and the interaction of these two variables need to be explored further to clarify the role of each factor.

Although the variables studied here do not give a highly accurate model for predicting the cold temperature characteristics of a compound, it is still important to note that the actual test data show an 8 to 9[degrees]C spread in the [TR.sub.10] results. These data are useful to illustrate that the choice of compounding strategy enabled by the use of low viscosity polymer (at constant ACN level) may affect the final cold temperature properties of a compound.

Aged physical properties--hot air

Hot air aging performance was evaluated by determining stress strain properties after hot air aging for 1,008 hours at 160[degrees]C. According to ASTM D2000, HNBR is rated at 150[degrees]C service, but this work was carried out at the higher temperature to ensure that the end-points would be clearly distinguished.

Figures 12 and 13 show the tensile and elongation at break, respectively, for the compounds studied. In both instances, the analysis indicates that the choice of polymer is not a significant factor in determining the tensile or elongation at break after long-term aging. The slope of the parameter effect line in both of the prediction profile plots is essentially zero. This indicates that the use of low viscosity HNBR polymer, even after severe hot air aging, does not have a detrimental affect on these polymer properties.

[FIGURES 12-13 OMITTED]

The antidegradant, plasticizer and filler selections, however, are significant factors for aged tensile at break. A wide range of values, from approximately 13 MPa to 20 MPa, can be achieved by varying these parameters. Both the HNBR-HT antidegradant selection and high plasticizer level give rise to lower tensile strength, while increasing levels of both silicate filler and metal oxide promote higher tensile at break values. The use of HNBR-HT is known (ref. 3) to reduce the tensile at break after hot air aging, but nevertheless, a compound with a tensile at break of about 13 MPa is still very useful and functional.

The reverse trends are noted for the elongation at break after long-term aging at elevated temperature (figure 13). The use of HNBR-HT gives much-improved retention of elongation at break, and the level of plasticizer is not a significant factor. Increasing the levels of both silicate filler and metal oxide promotes lower elongation at break values. The values for elongation at break after aging range from a low of about 30% to a high value in excess of 200% elongation. Clearly, the latter material has retained much more of its viscoelastic properties compared to the compound that had a final elongation of only 30%.

Since the data indicate that polymer viscosity is not a significant factor in determining the final physical properties of the compound after long term hot air aging, it would be judicious to use the low viscosity polymer in an optimized compound. The 39 Mooney viscosity polymer, in combination with lower plasticizer level, higher filler loading and HNBR-HT may be a route to optimize the long term performance of the compound, while still retaining a low viscosity, easy processing compound for the production environment.

Aged physical properties--automatic transmission fluid

Automatic transmission fluid (ATF) was used to assess fluid immersion characteristics. The compounds were immersed in Dexron III ATF (PetroCanada RDL 2746) for 1,008 hours at 150[degrees]C. The effect on tensile at break is shown in figure 14, and the effect on elongation at break is shown in figure 15. The [R.sub.2] values are 0.781 and 0.758, respectively, which are somewhat low but can be useful to indicate trends.

[FIGURES 14-15 OMITTED]

Polymer viscosity does not appear to have a large effect on tensile at break, while metal oxide, plasticizer and antidegradant selection appear to have the greatest degree of influence. The lowest tensile at break value is observed when low metal oxide and high plasticizer loading is used in combination with a high viscosity polymer, the black filled system and using the standard antidegradant package. Conversely, the greatest tensile at break value is observed when using high metal oxide and low plasticizer levels in conjunction with the new low viscosity polymer and using the black filler with a heat resistant antidegradant package.

The factors that most influence the elongation at break values are antidegradant, filler and polymer; while metal oxide and plasticizer had almost no influence. The actual test results covered a wide range of values, from low values of about 40% to high values of almost 175%. The lowest value is given when the high viscosity polymer is used in combination with the standard antidegradant package and without silicate filler. The highest value is observed when the low viscosity polymer is used in combination with the HT antidegradant and without silicate filler.

The combination of low viscosity polymer with HT antidegradant and without silicate filler also gives volume swell values of -2.2% (a small shrinkage). Figure 16 shows the full range of volume swell results that were recorded for the various compounds mixed in this study. The predictor profile plot clearly shows that plasticizer loading is the primary independent variable that impacts on volume swell. The slope of the line is almost -1, indicating a strong dependence. High initial plasticizer loading leads to low swell or even some shrinkage. Although metal oxide and polymer are indicated as the second and third most significant factors, the profile plots clearly show that the slope of the effects line for these parameters is very nearly zero. The selection of an intermediate level of plasticizer and the inclusion of the silicate/silane, in conjunction with the low viscosity polymer, is a preferred combination for achieving a slightly positive volume swell in Dexron III automatic transmission fluid.

[FIGURE 16 OMITTED]

Conclusions

The availability of a new ultra low viscosity HNBR polymer offers the potential for alternate compounding strategies. Three discrete variables and two continuous variables have been screened in a simple experimental design to determine their effect on compound properties. Statistical analysis of the compounds mixed for this study suggests that use of the low Mooney polymer has a statistically significant effect for many of the measured properties of interest.

The screening study suggests that compounds with good scorch safety, fast cure times and low compound viscosity can be designed for injection molding applications. The data suggest that compounds that contain the combinations of low viscosity polymer and HNBR-HT technology can be judiciously formulated to improve heat aging performance, as well as fluid aging characteristics.

Compression set was not significantly affected by the choice of polymer. However, HNBR-HT, high levels of metal oxide and, to a lesser extent, high plasticizer level each contributed to slightly poorer compression set performance. A small adjustment in peroxide level may be considered to restore the compression set performance.

The [R.sup.2] correlation results for the Die B tear test, as well as for cold temperature tests, [TR.sub.10], were somewhat poor and may suggest either high variability in the original test data or that the parameters included in this model (screening design) were not sufficient to fully account for the variation in the data. Additional studies, with more detailed experimental designs, are needed in order to fully understand the factors influencing both tear and cold temperature characteristics.

The availability of the new low viscosity HNBR polymer has enabled alternate compounding techniques that have been previously excluded due to high compound Mooney viscosity and other significant processing considerations. This screening experiment has shown that tailoring of compound properties may be possible, and it is useful in suggesting possible directions for future compound studies.

References

(1.) Fujii Yoshinori, Ikeda Atsumi, JP2004-002893, 08. 01.2004, "Method for manufacturing nitrile group-containing highly saturated copolymer rubber having low Mooney viscosity."

(2.) S.X. Guo and F. Guerin, Rubber Division, ACS, paper no. 62, Columbus, OH, October 2004.

(3.) E.C. Campomizzi and H. Bender, "Improving the heat resistance of hydrogenated nitrile rubber compounds (Part 1)," Kautschuk Gummi Kunststoffe, v. 54, no. 1-2, p. 14 (2001).

(4.) Goran Spetz, "Review of test methods for determination of low-temperature properties of elastomers," SRC 1989, Polymer Testing, v. 9, 27-37 (1990).

Ezio C. Campomizzi, L.P. Ferrari and R.J. Pazur, Lanxess

Table 1--test methods

Test method                       ASTM procedures

Mooney viscosity and scorch           ASTM D 1646
MDR rheometer                         ASTM D 5289
Stress strain properties               ASTM D 412
Hardness-durometer A                  ASTM D 2240
Low temperature retraction            ASTM D 1329
Compression set               ASTM D 395 Method B
Tear strength - Die B                  ASTM D 624
Air oven aging                         ASTM D 573
Fluid immersions                       ASTM D 471

Table 2--variable types and levels examined in the screening experiment

Ingredient                      Phr         Low level    High level

HNBR type                     100.0       HNBR-AT (4)      HNBR (7)
Magnesium oxide            Variable           3.0 (m)      18.0 (M)
Stabilizer package         Variable   Standard AO (A)   HNBR-HT (H)
Carbon black                  50.00
Silicate/silane DLC        Variable           0.0 (B)      25.0 (W)
TOTM plasticizer           Variable          5.0 (Ip)      15.0 (P)
Zinc oxide                     3.00
Peroxide co-agent (TAIC)       1.50
Peroxide (40% active)          7.50