Innovative peroxide and coagent cure systems for use with HNBR elastomers.

Hydrogenated nitrile butadiene rubbers (HNBR) have been cured with peroxides since their commercial introduction in 1984 (refs. 1, 2, 3 and 5). Peroxide cures can be used with all the different saturation levels of HNBR elastomers. Typical peroxides used in HNBR elastomers are dicumyl, 2,5-dimethyl-2,5-di(t-butylperoxy) hexane, and [alpha], [alpha]'-bis(t-butylperoxy) diisopropylbenzene. Recently manufacturers (refs. 6 and 7) of peroxides have, introduced scorch retarded versions of these materials which will be investigated in direct comparison to the standard, non-retarded peroxides. Coagents have been used in peroxide cures to improve processing, physical properties and increase cure speed. These coagents also showed a tendency to increase the scorch sensitivity of the compounds. Recently scorch inhibited versions of these materials have been introduced to the market.

This investigation will involve looking at the benefits and drawbacks of using scorch inhibited peroxides and coagents. Their effect on processing properties such as Mooney viscosity, scorch and cure rate will be evaluated. Tradeoffs in physical properties such as heat resistance, compression set, elongation, tear, water and oil aging will also be evaluated. In the coagent study, the effect on the 10% modulus and the Young's modulus will be investigated and the potential benefits discussed.

Experimental

One elastomer was utilized in this study, Zetpol 2010, a 36% ACN, iodine value of 12, which corresponds to a 95% hydrogenated acrylonitrile butadiene copolymer. Six different peroxides were used in this study. They are:

* Di-Cup 40C, a dicumyl peroxide, which will be designated DC40;

* Peroximon DC 40 KEP SR, a scorch-retarded dicumyl peroxide, which will be designated DC40SR;

* Luperco 101-XL, a 2,5-dimethyl-2,5-di(t-butylperoxy) hexane, which will be designated L101XL;

* Luperco 101-XL SR, a scorch-retarded 2,5-dimethyl-2,5-di(t-butylperoxy) hexane, which will be designated L101XLSR;

* Vul-Cup 40 KE, a [alpha], [alpha]'-bis(t-butylperoxy) diisopropylbenzene, which will be designated VC40;

* Retilox F 40 KEP SR, a scorch-retarded [alpha], [alpha]'-bis(t-butylperoxy) diisopropylbenzene, which will be designated VC40SR.

These peroxides were chosen because they are commonly used in a variety of HNBR applications. Peroxides have different activation temperatures and each standard is shown in direct comparison to its retarded counterpart. They are compared at equivalent peroxide levels taking into compensation the level of retarder present. Table 1 shows these recipes and basic properties.

[TABULAR DATA OMITTED]

Six different coagents were used in this study. They are:

* HVA-2, a N-N'-m-phenylenedimaleimide, which will be designated HV2;

* Triallyl isocyanurate, which will be designated: TAIC;

* SR 350, a trimethylolpropane trimethacrylate, which will be designated TMPT;

* Saret SR 500, a nitrosamine retarded trifunctional crosslinking agent, designated SR500;

* Saret SR 517, a non-nitrosamine retarded trifunctional crosslinking agent, designated SR517;

* Saret SR 519, a non-nitrosamine retarded trimethylolpropane triacrylate crosslinking agent, designated SR519.

The coagents used in this study all come from published literature (refs. 6 and 7) and the scorch retarded coagents are recent additions which have found their way into industrial uses. All of the coagents are compared at 4 phr and 8 phr levels. Table 2 shows these recipes and their basic properties.

[TABULAR DATA OMITTED]

All of the formulations were mixed in an internal mixer. The compounds were then transferred to a 16"x8" water-cooled mill for sheeting and blending. After the compounds were allowed to cool, they were again placed into the mixer for the addition of the curatives and additional mixing. Mooney viscosity and scorch values were measured in accordance with ASTM D 1646. Rheometer measurements were taken using the Monsanto R-100 (ODR) and the Monsanto MDR-2000E as outlined in ASTM D 2084. All compounds were molded at 170[degrees]C. All of the cure times were in excess of their t'90 times as measured on the ODR. ASTM standards were used to generate the data found in this evaluation.

Results and discussions

Peroxide study

This study was designed to determine the potential benefits which could be attained by using scorch resistant peroxides and the potential sacrifices which may be required to attain these benefits. Three different types of peroxide were evaluated, all at equivalent milli-molar ratios. The VC40 at 8.00 phr is the standard to which the other peroxide levels were adjusted. The retarded version of this material VC40SR, was tested at 8.42 phr. The DC40 was evaluated at 12.78 phr and 6.39 phr to evaluate different levels of the same peroxide. This information was compared with a retarded version, DC40SR, which was tested at 13.46 phr and 6.73 phr. The slightly higher levels were used to compensate for the retarder while providing for the same peroxide level. L101XL was tested at a 6.10 phr level with the scorch retarded version, L101XLSR, at 6.39 phr.

Rheological properties

One of the prime reasons to use a scorch resistant peroxide is to improve the Mooney scorch and [t.sub.s.2] times of a compound. Additional scorch safety can aid in processing and knitting of a material during molding. After evaluating the Mooney scorch at 125[degrees] C, we found that all of the scorch resistant peroxides have [t.sub.5] times in excess of 30 minutes. This made the greatest difference when using high levels of DC40, where the [t.sub.5] scorch time improves from 15.8 minutes to greater than 30 minutes by changing to the scorch resistant version.

The ODR properties demonstrate the tradeoffs occuring when using a scorch resistant peroxide. In figure 1, we see that the [t.sub.s] 2 times of the DC40 and VC40 have shown increases of 18 to 30 seconds when DC40SR and VC40SR are used. However, figure 2 illustrates that this also corresponds to an increase in cure time defined as the t'90 time minus the [t.sub.s] 2 time. The DC40SR time of 5.6 minutes is 0.4 minutes longer than the 5.2 minutes of DC40 and the VC40SR time of 9.7 minutes is 0.7 minutes longer than the 9.0 minutes of VC40. This is an 8% increase in the time needed to reach t'90. The [M.sub.H]-[M.sub.L], as shown in figure 3, demonstrates that the DC40SR and VC40SR give lower crosslink densities at equal levels versus their non-scorch resistant counterparts.

The exception appears to be the L101XLSR which shows no loss in cure speed and actually has a higher maximum torque than the standard L101XL. The disadvantage is this material only yields a six second improvement in scorch measured by [t.sub.5] times from the Mooney scorch testing.

Physical properties

The Shore A hardness and tensile strength show no significant difference between the standard and scorch resistant peroxides. The modulus and elongation showed some minor changes, with the modulus being lower and the elongation higher. As expected, when the elongation increases so do the tear properties. The information in figure 4 shows the die C tear increasing from 29.6 to 43.4 kN/m and from 43.1 to 53.6 kN/m for the high and low levels of DC40, respectively, when using the scorch resistant version, DC40SR. The tear of the VC40 compound increased from 36.6 to 42.4 kN/m when VC40SR was used, while the L101XL showed an increase of less than 3 kN/m when the L101XLSR was used.

Aged properties

Compression set resistance is where the greatest sacrifice is made when using the scorch resistant peroxides. After aging for 70 hours at 150[degrees] C DC40 and VC40 have compression set values of 19.2% and 18.6%, respectively. The DC40SR and VC40SR have values of 27.6% and 25.0%, a sacrifice of 8.4% and 6.4% respectively. The L101XL has compression set values of approximately 35% and shows no loss while using the L101XLSR version.

The other aged properties reported, air, water and oil, exhibit very little difference whether the scorch resistant peroxides were used or not. The only exception to this is the air-aged elongation change and these values are higher for the scorch resistant version, which could be attributable to the higher initial elongation as the final elongation values are all very similar.

Coagent study

This study was undertaken to determine the potential benefits that could be attained by using a scorch resistant coagent and the potential sacrifices which may be required to attain these benefits. Coagents are frequently used to improve flow through lower viscosity, to improve the compression set through higher crosslink density and to improve the cure speed. The coagents selected were each tested at both a 4 phr and 8 phr level and compared to each other as well as against a non-coagent containing standard.

Rheological properties

As with the peroxides discussed previously, we would expect to see an improved [t.sub.5] value from the Mooney scorch when evaluating an inhibited coagent versus its standard counterpart. The control compound that contains no coagent has a [t.sub.5] value greater than 30 minutes while the inhibited coagents at the 4 phr level are all less than 30 minutes, with the SR519 less than 15 minutes. The HV2 and the TMPT both show [t.sub.5] values between 10 and 15 minutes and is one of the sacrifices which must be made to improve cure speed and crosslink density.

Coagents are frequently used to lower the Mooney viscosity of a compound. The information in figure 5 demonstrates that while a majority of the coagents will lower the viscosity, HV2 shows no effect. At an 8 phr level the TMPT has lowered the viscosity by over 20 points and this will improve flow and reduce shear forces on the material as it moves through the mold.

Another reason that coagents are used in HNBR formulations is increased cure speeds and this can be evaluated by measuring the change in torque over time using an ODR. The information in figure 6 is a plot of cure speed for each of the different coagent/level combinations. The most significant observations were the increase in cure speed of the HV2 coagent. The cure speed increased from the control value of 0.14 to 0.21 and 0.25 dN[center dot] m/sec with 4 phr and 8 phr HV2, respectively. TMPT at the 4 phr level shows no effect on cure speed, however a 35% increase from 0.14 to 0.19 dN[center dot] m/sec is observed at the 8 phr level. SR500 and SR517 gave moderate increases ranging between 0.15 and 0.17 dN[center dot] m/sec depending on addition level. Of the acrylate coagent types, SR519 shows the most significant increase with a rise to 0.17 dN[center dot] m/sec at the 4 phr level. The only compound which had a drop in cure rate with the addition of coagent is the TAIC. The cure rate dropped to 0.13 at 4 phr and to 0.12 dN[center dot] m/sec at 8 phr and this demonstrates the vast differences in cure kinetics which can be achieved through the use and selection of specific coagents.

Physical properties

The physical properties are commonly changed by the increase in crosslink density which occurs when coagents are used. Typical changes seen are an increase in modulus and a decline in elongation (ref 4). The data generated in this study show that only the HV2 material caused any significant change in Shore A hardness of the material. The control compound started at a hardness of 68 Shore A, and went up to 74 and 80, respectively, when adding 4 phr and 8 phr of the HV2 coagent. The other coagents showed only modest increases of 2 to 4 points. The 50% modulus exhibited a similar trend with the HV2 material increasing more than twice the value of the control (2.1 MPa to 4.8 MPa) for the compound with the 8 phr level.

There was virtually no change in tensile strength, with all but one sample falling between 25 and 30 MPa. The elongation did show a significant change when coagents were used. The HV2 and the TAIC dramatically decreased the elongation. The control compound had an initial elongation of 327% compared to 164% elongation with the HV2 compound at the 8 phr level. The TAIC based compounds had elongations ranging from 222% to 239%, which is a significant drop considering there was no increase in modulus. The acrylic coagents showed very little change at the 4 phr level, yielding elongations around 300%. The 8 phr level of the acrylic coagents dropped the elongations to the 250% level, but unlike the TAIC, these did correspond to increases in the modulus.

Young's modulus

Young's Modulus is the measurement of resistance to deformation at very low extensions. It was decided to measure this property and how it changed with the addition of coagents. It was determined that the acrylic coagents generally cause a decline in the Young's modulus value. The HV2 shows a significant increase in Young's modulus, with a 4 phr level providing an increase from 5.1 MPa to 7.4 MPa. The TAIC also caused an increase in the Young's modulus, as did the 8 phr level of TMPT (table 2).

Aged properties

A number of aged properties were evaluated, however our discussion here will be limited to compression set and hot air aging. With addition and variation of coagents the compression set resistance had much less effect than previously anticipated. The addition of coagents at the 4 phr level, as shown in figure 7, was more likely to provide a worsening of the long term compression set than an improvement, although, in the short term, improvements could be achieved. At 4 phr the TAIC had the most dramatic effect with a compression set value of 59.8% versus the control value of 41.3%. At an 8 phr level, the HV2 and the TMPT showed significant improvements over the control compound. After 1,000 hours the HV2 material had a compression set value of 32.1% versus the control of 41.3%. The TMPT had a value of 37% and the TAIC at 8 phr had a value of 69%. Figure 8 effectively demonstrates that the choice of coagent can have a dramatic effect on the compression set of a material.

Air aged properties were evaluated in a similar fashion to the compression set. The 4 phr and 8 phr coagent levels were evaluated as independent groups. At coagent levels of 4 phr, it was determined that some short term heat aging, 70 hours at 150[degrees] C, showed improvements in elongation retention: -14% for the control compound versus -5% for the HV2 compounds and -7% for the SR519 compound. After 1,000 hours all coagent containing materials had slightly higher elongation loss than the control compound, with the exception of TAIC which had elongation loss between 79% and 89% compared to the control at 59%. When the 8 phr compounds were evaluated, TMPT, SR500 and SR517 had improved elongation retention versus the control compound after 70 hours at 150[degrees] C. After 1,000 hours at 150[degrees] C, only the HV2 and SR517 materials could equal the -59% performance of the control compound. This leads to the conclusion that not all coagents improve long term heat aging, but may be used to meet specific requirements when only short term testing is required.

Conclusion

Peroxides

Three different peroxides and their inhibited counterparts were evaluated and it was determined that these retarded peroxides do offer improved scorch safety for better processing. This benefit comes at the cost of higher compression set values, although these higher values are below what would be expected by lowering the amount of peroxide. Scorch-retarded peroxides also show an improvement in the balance between tear and elongation without much other sacrifice. These materials could be quite useful in applications where some sacrifice in crosslink density could be tolerated for improvements in processing.

Coagents

The coagents evaluated all showed some distinct advantages and drawbacks. The HV2 material gave the greatest improvements in compression set, modulus and cure speed, with its major drawbacks being a short scorch time and low elongation. The TAIC material showed no advantages over the other coagents. However, it has shown advantages in high temperature water aging in other studies. The TAIC material was also the only material to show a significant drop in compression set resistance when used versus the control compound. The TMPT material provided distinct improvements in viscosity reduction, modulus and cure speed versus the control material. TMPT was also the scorchiest of all the compounds tested. The scorch inhibited coagents did show definite improvements in scorch safety versus the non-inhibited coagents, except for the SR519. The SR519 also showed the fastest cure rate and the highest modulus of all of the inhibited coagents. The three scorch inhibited coagents show no significant change in heat aging, compression set or oil aging versus the standard methacrylate coagent and offer a definite improvement in scorch safety. If cure speed is the primary concern, then HV2 is the best choice, however the scorch inhibited coagents offer a new alternative for scorch safety with only minor sacrifices in properties.

References

[1.] Hashimoto, K. et al, 124th Rubber Division Meeting, ACS, Houston, Texas, USA, October 25-28 (1983). [2.] Hashimoto, K et al, The Annual Meeting of the Swedish Institute of Rubber Technology, Gothenburg, Sweden, May 17-18, (1984). [3.] Hashimoto, K. et al, 127 Rubber Division Meeting, ACS, Los Angeles, California, USA, April 24 (1985). [4.] Hofmann, W. "Vulcanization and vulcanizing agents," Maclaren and Sons Ltd., London, England [5.] Nakagawa, T et al, The Annual Meeting of the Swedish Institute of Rubber Technology, Tylosand, Sweden, May 5-6 (1988). [6.] Sartomer Corporation, Technical Literature SB-04, SB-05. [7.] Du Pont Chemicals, Technical Literature No. 59A.