Post vulcanization stabilization for NR.

Post vulcanization stabilization for NR

The physical properties, performance characteristics and durability of vulcanized rubber compounds are directly related to the number (crosslink density) and type of crosslinks formed during the vulcanization reaction[1,2 and 3]. The relationship between physical property development and crosslink density is depicted in figure 1.

The mechanism for accelerated sulfur vulcanization of natural rubber[4,5 and 6] involves complex interactions between the sulfur, accelerator, activators and polymer.

Ideally, all of the available sulfur is consumed in the formation of "effective crosslinks" which join together two polymer chains and enhance the overall strength of the polymer matrix. Unfortunately, due to the complexity of these interactions a variety of different crosslink types are formed during the vulcanization reaction[7,8,9] as indicated in figure 2.

Crosslink types [S.sub.1], [S.sub.2], and [S.sub.x] are referred to as "effective crosslinks" because they have transformed the two-dimensional polymer matrix into a three-dimensional crosslinked network which is capable of bearing significantly higher stresses under either static or dynamic conditions as compared to the unvulcanized polymer. The main chain modifications such as cyclic sulfides, pendant sulfides and conjugated diene/triene units do not add to the overall strength of the vulcanized compound and may in fact contribute to increased rates of degradation.

The bond strength of S-S bonds is inherently lower than that for C-C bonds and is further reduced as the number of adjacent sulfur atoms increases in the chain as shown in table 1.

When sulfur vulcanized natural rubber compounds are exposed to a thermal aging environment significant changes in physical properties and performance characteristics are observed[10,11 and 12]. These changes are directly related to modifications of the original crosslink structure as shown in figure 3.

All three of these crosslink modification reactions are occuring simultaneously, however the decomposition reactions tend to predominate and thus lead to a reduction in crosslink density and physical properties as observed during overcure. When using higher curing temperatures the effect becomes more pronounced[13].

Over the years, the rubber industry has developed two compounding approaches to address the changes in crosslink structure during thermal aging. These are, the use of high accelerator/low sulfur ratios or sulfur donors, both of which provide a sulfur content in the crosslinks found during vulcanization and thus improved thermal stability.

As with many formulation changes in rubber compounding, there is a compromise that must be made when attempting to improve one performance characteristic. Improving the thermal stability of vulcanized natural rubber compounds by reducing the sulfur content of the crosslink through the use of a more efficient vulcanization system will reduce dynamic performance properties such as fatigue resistance, as indicated in figure 5.

The challenge becomes to define a way of improving thermal stability while maintaining performance characteristics. Work in our laboratories over the last few years has been directed toward addressing this issue and has led to the development of hexamethylene-1,6-bisthiosulfate disodium salt, dihydrate (D-HTS) a rubber chemical that promotes the formation of flexible hybrid crosslinks[14].

By reducing the average sulfur chain length at the points of attachment of the polymer backbone, thermal stability is improved. Maintaining a long chain within the crosslink structure provides enhanced flexibility under dynamic conditions. The use of D-HTS eliminates the compromise between thermal aging and dynamic properties as illustrated in table 2.

This article presents the results from a statistically designed experiment which was undertaken to develop a systematic approach for optimizing the use of D-HTS to provide improved thermal aging in resistance while maintaining an acceptable level of performance properties in natural rubber compounds. Comparison is also made to other commonly used heat resistant systems.

Experimental details

Formulation

A natural rubber masterbatch consisting of Standard Malaysian Rubber (SMR 5CV) (100), carbon black (50), aromatic oil (5), zinc oxide (5) and stearic acid (5) was prepared by a custom mixer. Antidegradant (6PPD), sulfur, accelerators and the post vulcanization stabilizer were added to portions of this masterbatch on a 15 cm x 30 cm mill at 60 [degrees] C - 70 [degrees] C. The formula parts of sulfur, accelerators and post vulcanization stabilizer were variable.

Performance testing

Processing characteristics were measured on the Monsanto Mooney Viscometer (ASTM D-1646-89). Curing properties were determined on the Monsanto Rheometer model R-100 (ASTM D-2084-88). Test specimens were cured to [T.sub.90] @ 140 [degrees] C according to ASTM D-1382 and stress strain properties measured according to ASTM D-412-87. Goodrich heat build-up data was obtained using the Goodrich Flexometer (ASTM D-623-88). Fatigue-to-failure tests were carried out according to ASTM D-4482-85 on the Monsanto fatigue-to-failure tester at 100% strain. Heat aging was conducted according to the test tube method ASTM D-865-88. Compression set was measured according to ASTM D-395-89, Method B. Abrasion was measured using the Pico abrasion tester ASTM D-2228-88. Tear resistance was determined according to ASTM D-624-86. Rebound decay (measure of hysteresis) data was obtained using the Lupke rebound test which measures the rebound with successive impacts of a pendulum. The rate of decay as defined as the slope of the line when plotting log rebound versus the number of impacts. The lower the hysteresis, the less the rebound decay.

Design of experiment

The design of experiments is a planning and evaluation process which allows us to establish the relationship between independent variables (inputs) and dependent variables (responses) by varying the inputs in a controlled manner such that the responses can be predicted through the use of statistical analysis techniques.

In this work, it was desired to systematically evaluate the effects of changes in the concentration of the independent variables: sulfur, accelerator (MBS) and post vulcanization stabilizer (D-HTS)

The evaluation was carried out using response surface methodology[15]. The following variables were considered within the concentration limits indicated.

  Independent variables      Range
  [X.sub.1]  sulfur     0.375 - 4.625 phr
  [X.sub.2]  MBS        0.400 - 2.100 phr
  [X.sub.3]  D-HTS      0.300 - 4.000 phr

The experimental design was a modified three variable central composite design evaluated in three blocks. The combinations of independent variables tested are shown in table 3.

In each case the following dependent variables were measured.

[Y.sub.1] Mooney scorch ([t.sub.5] at 121 [degrees] C)

[Y.sub.2] rheometer ([t.sub.90] at 140 [degrees] C)

[Y.sub.3] reversion at 180 [degrees] C (% torque decrease at [t [prime].sub.max] + 30

min)

[Y.sub.4] tensile modulus at 300%

[Y.sub.5] Goodrich blowout time (at 286 psi, 22.5%

deflection, 100 [degrees] C)

[Y.sub.6] fatigue-to-failure (kilocycles to failure at 100% ex

tension)

[Y.sub.7] rebound decay

[Y.sub.8] tear resistance at room temp. (Die C)

[Y.sub.9] compression set

The data from these were analyzed by multiple regression analysis. This procedure produces predictive equations of the following form. [Mathematical Expression Omitted] When [Y.sub.i] = predicted values of the independent variable i

[X.sub.1] = phr sulfur

[X.sub.2] = phr MBS

[X.sub.3] = phr D-HTS

The resulting regression equations were then used to produce contour plots, to optimize the overall results and to predict the results of confirmatory experiments.

Table 4 indicates excellent relationships (measured by the overall correlation coefficient) were obtained between the independent variables and Mooney scorch time, rheometer cure time, 300% modulus and oxidative aging. Somewhat poorer relationships were obtained for reversion, Goodrich blowout time, fatigue, rebound decay, compression set and tear.

The general relationships between the independent variables and the dependent variables have been summarized in table 5 and the resulting contour plots are available from the author.

The following summarizes the trends observed in physical properties and performance characteristics as changes are made to the sulfur, accelerator and D-HTS concentrations.

* Mooney scorch time - Mooney scorch time, [t.sub.5] minutes, measured at 121 [degrees] C is more sensitive to adjustments in curatives than in D-HTS levels. Increased sulfur and MBS concentrations reduced scorch time with sulfur level changes having the greatest effect. Increased D-HTS concentration reduces scorch time slightly.

* Rheometer cure time - Increased MBS concentration significantly reduces rheometer [t.sub.90] cure time. The reverse is true with D-HTS. The effect of sulfur to cure time is minor.

* Reversion resistance - Reversion resistance (% torque decreases at t [prime] max + 30 [prime] @ 180 [degrees] C) is improved with increased D-HTS and MBS concentrations. Increased sulfur levels reduce reversion resistance.

* 300% tensile modulus - Higher concentrations of sulfur and MBS increase the 300% modulus as expected, while increased levels of D-HTS tend to lower modulus.

* Goodrich blowout time - Increasing the concentration of accelerator and D-HTS increases the blowout time as measured by the Goodrich flexometer. Sulfur level has a minor effect on Goodrich blowout time under these conditions.

* Fatigue-to-failure - Improved fatigue properties can be obtained by increased sulfur concentration and reduced MBS level which result in more flex resistance polysulfuric crosslinks. D-HTS improves fatigue at intermediate levels.

* Rebound decay - Increased concentrations of sulfur, MBS and D-HTS give lower rebound decay values which can be translated into lower hysterisis. This effect is more pronounced at sulfur levels greater than 2.50 phr.

* Compression set - As expected, compression set improves at low sulfur levels which generate more stable mono and di sulfidic crosslinks. The concentration of MBS and D-HTS has minor effect on compression set.

* Tear resistance - High die C tear resistance was obtained at low sulfur and D-HTS concentrations and at high MBS levels.

* Thermal oxidative aging - Higher oxidative aging resistance is obtained at low levels of sulfur and MBS. At low sulfur levels (1.25 phr), the addition of D-HTS improves thermal oxidative aging. As sulfur level increases (above [congruent] 3.0 phr), the effect of D-HTS is reduced.

Optimization

Contour plots provide a good picture of the possible changes in a single dependent variable. These plots can be used to pick the combination of independent variables producing the desired result of that variable. Overlaying contour plots of two dependent variables can be used as a means of determining the best compromise between these two variables. However, when more than two dependent variables must be considered, the determination of the best compromise can be quite difficult. In order to address this problem, Harrington[16] developed the concept of a desirability coefficient. This concept has been converted into a useful computer program by Derringer and Geary [17] and has been used to optimize the results of this study.

Harrington proposed that it is possible to assign some measure of desirability to each of the dependent variables considered in a response surface experiment. In fact it is possible to develop a function for each dependent variable such that the desirability coefficient ([d.sub.i]) for that variable will have the following characteristics.

[d.sub.i] = 0 to 1

[d.sub.i] = 1 when the value for [Y.sub.i] is completely satisfactory

[d.sub.i] = 0 when the value for [Y.sub.i] is unsatisfactory

[d.sub.i] >0<1 when the value for [Y.sub.i] is usable, but not completely satisfactory.

The various desirability coefficients for the individual dependent variables are then combined into an overall desirability coefficient in the following manner. [Mathematical Expression Omitted] where [w.sub.i] = a weight factor for each [d.sub.i]

The overall desirability coefficient has the property that it is zero if any of the individual coefficients is zero, and is equal to one only if all of the individual coefficients are equal to one. If the overall desirability coefficient is maximized, one has the best combination of the dependent variables as defined by the user's definition of desirability. Also, if a suitable computer program is available (such as the Derringer-Geary program), it is a comparatively simple matter to perform optimizations on more than one combination of individual desireabilities, or to change the relative importance of the individual dependent variables by changing the weight factors.

In the case of the current study, the desirabilities for the various dependent variables were chosen according to the limits shown in table 6.

Maximization of the overall desirability coefficient under these constraints produced the results shown in table 7.

This combination of independent variables provides the highest value for the overall desirability coefficient, but does not indicate the amount of variation which can be tolerated without significant loss of properties. The surface plots shown in figure 6 indicates that changes in the overall desirability coefficient are not greatly affected by changes in HTS concentration but that it is necessary to maintain the sulfur concentration close to the 1.9 phr level. The MBS concentration is less critical, but increases in its concentration result in somewhat lower desirabilities.

Should another compounder desire different desirability functions for the various dependent variables, the optimization program permits these changes to be made quickly. In this manner, it is possible to optimize the combination of independent variables to suit specific physical properties of performance characteristics for different applications.

Confirmatory experiments

An experiment such as this is only as good as the ability of the resulting regression equations to predict further results. Therefore, the contour plots were examined to choose several combinations of sulfur, MBS and D-HTS which should produce interesting combinations of properties. The three combinations shown in table 8 were chosen as confirmatory points. Both the actual results obtained in these conformation tests and the values predicted by the regression equation are shown. In most cases, the agreement between actual and predicted data is quite good. The Mooney scorch time ([t.sub.5]) is higher for all the actual values. This may be explained as a block effect since the compounds for |predicted' and |actual' were mixed at different times. However, the relative difference betwen the three stocks were similar, and therefore a high confidence level is achieved for using this experimental design to choose compound modifications which provide significant changes in physical properties and performance characteristics.

Comparative evaluation

As discussed earlier, the reversion and thermal oxidative aging resistance of natural rubber can be improved through the use of efficient and semi-efficient vulcanization systems. Due to the increased nmber of monosulfidic and disulfidic crosslinks and the reduced number of polysulfidic crosslinks in these vulcanization systems, fatigue properties are reduced. The addition of post vulcanization stabilizer D-HTS can improve reversion resistance and enhance dynamic properties. Seven vulcanization systems were selected to compare processing, curing and performance properties. Table 9 shows the systems under evaluation. Included in this evaluation is an "optimized system" (OPT) from our Design of Experiments.

The results of this comparative evaluation are:

* Reversion resistance - When measured as % torque reduction at t [prime] ma x + 30 minutes at 180 [degrees] C, the efficient vulcanization system (E.V.) shows the smallest drop (indicating the highest resistance to reversion) and the control the largest drop, as expected. The other vulcanization systems are at the in-between levels (figure 7).

Addition of D-HTS to the control (D-HTS/II) and the Hi/Lo semi-E.V. systems (D-HTS/I) gave significant torque reduction. This is the result of the formation of the more stable hybrid crosslinks with the D-HTS.

The optimized system has the least torque reduction among the three D-HTS systems.

* Fatigue properties - In general, the high sulfur compounds give the highest unaged fatigue resitance. Addition of D-HTS significantly increases the fatigue of both the conventional and semi-E.V. vulcanization system. The OPT has the best fatigue property among the E.V. and semi-E.V. systems. The OPT system is even slightly better than the control (figure 8).

* Thermal oxidative aging - Basically, aging resistance can be divided into two groups: high sulfur (2.5 phr) and low sulfur (1.2-1.8 phr). As expected, the low sulfur compounds give the higher resistance to thermal oxidative aging than the high sulfur compounds. Addition of D-HTS does not affect oxidative aging (figure 9).

* Compression set - As expected, the E.V. system provides the lowest compression set. The Hi/Lo and S-D systems are similar. D-HTS (at 2 phr) has no effect on compression set. However, the OPT system with D-HTS at 3.1 phr shows a significant improvement in compression set.

* Processing and curing properties - The S-D system gives the longest [t.sub.5] and the E.V. system the shortest. Addition of D-HTS to a conventional or semi-E.V. system reduces scorch time and increases [t.sub.90] cure time. Minor modulus variation among the system is observed. The OPT system shows good scorch safety with the longest [t.sub.90] cure time.

Summary and conclusions

* The physical properties and performance characteristics for natural rubber compounds will be adversely affected when exposed to a thermal aging environment.

* These changes are directly related to modifications of the vulcanizate structure (e.g., decrease in crosslink density and change in crosslink type).

* The use of efficient and/or semi-efficient vulcanization systems will provide improved thermal stability but at the expense of dynamic performance properties (e.g., fatigue resistance).

* A new rubber chemical, hexamethylene-1,6-bisthio-sulfate (D-HTS) provides improved thermal stability and maintains dynamic performance properties by promoting the formation of flexible hydrid crosslinks during the vulcanization reaction.

* Post vulcanization stabilization through the use of D-HTS is effective with a variety of curing systems.

* Optimization of D-HTS based cure systems can be achieved through statistically designed experiments, giving the compounder flexibility to select the vulcanization system which best fits high performance requirements. [Figures 1 to 9 Omitted] [Table 1 to 9 Omitted]

References

[1]Coran, A.Y., "The art of sulfur vulcanization," Chemtech, 1983, 13, 106. [2]Studebaker, M.L., "Effect of curing systems on selected physical properties of natural rubber vulcanizates," Rubber Chemistry and Technology, 1966, 39, 1359. [3]"The chemistry and physics of rubber-like substances," Bateman, L. Ed; Moe Laren: London, 1963. [4]Coran, A.Y. in "Science and technology of rubber," F.R. Eirich, ed., Academic Press, New York (1978), Chap. 7. [5]Campbell, R.H. and Wise, R.W., Rubber Chemistry and Technology, 1964, 37, 650. [6]Scheele, W. and Hu, P.L., Kautsch Gummi, 1962, 15, 440. [7]Moore, C.G., Proc. Nat. Rubber Producers Research Association, 1964, 167, London: Moe Laren & Sons Ltd. [8]Porter, M. Rubber Chemistry and Technology, 1967, 40, 866. [9]Sovelle, B. and Watson, A.A., Rubber Chemistry and Technology, 1967, 40, 100. [10]Morreson, N.J., Porter, M., "Temperature effects on the stability of intermediates and crosslinks in sulfur vulcanization," Rubber Chemistry and Technology, 1984, 57, 63. [11]Davis, K.M., "Practical consequences of vulcanizate structure changes at high cure temperatures," Plastics and Rubber Processing, 1977, 87. [12]Cunnein, J.I. and Russell, R.M., "Occurrence and prevention of changes in the chemical structure of natural rubber tire tread vulcanizates during service," Journal of the Rubber Research Institute of Malaysia, 1969, 22, 300. [13]Walker, L.A., and Helt, W.F., Rubber Chemistry and Technology, 1986, 59, 285. [14]Anthoine G., Lynch, R.A., Mauer, D.E. and Moniotte, P.G., "A new concept to stabilize cured NR properties during thermal aging," ACS Rubber Division Meeting, 1985. [15]Box, George E.P., Hunter, William G., Hunter J. Stuart, "Statistics for experimenters," John Wiley & Sons. [16]Harrington, E.C., Jr. Ind. Qual. Control 21, 494 (1965). [17]Derringer, George C., Geary, Carroll L, "Optimization of processes and products," presentation at Canadian