Stress relaxation as processability indicator.

Processability is a rather loosely defined term in the rubber industry. For the synthetic rubber producer it generally involves the mixing characteristics during pigment incorporation into the rubber. Mixing cycle times are often fixed. Therefore, variations in the consistency of the synthetic rubber can result in poor or dirty mixes or mixes that require abnormally long times to complete. Laboratory testing for the variations leading to inconsistent mixing has not been entirely satisfactory. The standard ASTM Mooney test is important but does not give a complete picture. The delta Mooney test (15 minute reading less the 1.5 minute reading) is often used and specified for SBR products. This is an empirical test which seems to detect gross changes within narrow Mooney ranges. Over a wide Mooney range some rather contradictory situations can occur. Low Mooney polymers, which are specifically produced for improved processing, may typically have "poorer" delta Mooney values than their higher Mooney counterparts. Determining the mixing characteristics in a Brabender laboratory mixer correlates with factory mixing but is difficult to use as a control test in a production plant. In their 1993 review of processability quality control instrumentation White and Soos (ref. 1) outlined the history of rheological testing in the rubber industry. Specifically White and Tokita (ref. 2) related stress relaxation data to polymer properties related to processability. Molecular weight and variations in composition of EPM and EPDM were related to stress relaxation data by Tokita and Scott (ref. 3). However, the time-consuming and tedious procedures for most rheological tests have limited their usefulness in quality control situations. The advent of relatively quick stress relaxation testing has opened up the possibility of using it for processability quality control testing of synthetic rubbers.

Experimental

Two instruments were used in this study: the BFGoodrich Dynamic Stress Relaxometer Model 1 (DSR) and the Monsanto MV 2000E Mooney machine. The initial work was done with the DSR but work later focused on the MV2000E when the stress relaxation feature became available.

The DSR has been described in detail by Moghe (ref. 4) and Harrell (ref. 5). Briefly it consists of a conical stator and rotor which are thermally controlled. A 30-35 gram sample is inserted and the rotor is lowered so that a gap of 6.25 mm remains. The sample is allowed to preheat and to relax from this stress for a pre-set time. The rotor is then twisted 0.035 radians very rapidly and released. As the rubber sample is relaxing, the stress is recorded at intervals of 0.2 log time from 0 to 400 seconds. From these 21 data points, the coefficients for the following equation were calculated. (1) G(t) = [At.sup.-alpha] + B

Harrell has shown that the stress relaxation data generated by the DSR are equivalent to other more traditional rheological tests. While the temperature can be varied, most of the testing was done at 70 [degrees] C because some samples relaxed too fast at 100 [degrees] C. The preheat timer was set for six minutes.

The MV 2000E is a standard Mooney machine with the extra feature that a stress relaxation test can be run after the standard test. As with the DSR the coefficients for the log time vs. log stress regression are calculated. Also the time for the stress to decay or relax to a specified percentage of the original stress (final Mooney) is determined as the decay time (t-X%). While the temperature and times can be varied, these were set to conform with the standard ASTM Mooney test: 10 mill pass sample preparation, 100 [degrees] C, one minute preheat, four minute running time. Two additional conditions were set for the stress relaxation portion of the test: Time to 80% decay and two minutes for the total relaxation time.

Molecular weight data were generated using a Waters gel permeation chromatograph (GPC) Model 150CV with a refractometer and a viscosity detector. The rubber sample solution, 0.1 weight/volume %, was pumped at 1.1 ml/minute through a bank of four 10-micron columns with pore sizes of [10.sup.3], [10.sup.4], [10.sup.5] and [10.sup.6]A. The detectors and columns were maintained at 70 [degrees] C and the pumps at 40 [degrees] C. The data were collected on a Vax System 845 from Waters.

The lab mixing studies were done with a Brabender PL 2000 instrument equipped with a 60 cc Banbury mixing head and charging chute. The initial temperature was set at 60 [degrees] C and the rotor speed at 60 rpm. The 40 grams of polymer were added over a 30 second time period and masticated for 30 seconds before the addition of 20 grams of IRB #6 carbon black. The mix was removed nine minutes after the charging started and a torque vs. time graph was generated. From this graph the black wetting time (BWT) and black incorporation time (BIT) were determined. The BWT is the time from the black addition to the first minimum; the BIT is the time to the first maximum following the first minimum. The mix was removed from the Brabender and pressed flat. Then the Mooney viscosity (ML4) was run and reported as the Brabender processability index (BPI).

All the rubber samples represent production material. However, special runs and a few samples of off-specification material made under upset conditions are included.

Results

A little bit of the chronology of our experimentation might be helpful at this point. In the beginning only the BFG DSR was available to us. To get a feel for what it might tell us, a product sweep was made in which a sample of each product was run. First of all it was found that Mooney and DSR "A" were highly correlated ([r.sup.2] = 0.96) as shown in figure 1. However, the correlation between Mooney and DSR alpha was poor ([r.sup.2] = 0.51) indicating that it describes something different than the Mooney or DSR "A." Plotting DSR "A" versus DSR alpha revealed that the known differences in these samples such as bound styrene, oil extension and polymerization temperature (hot vs. cold) explained much of this difference. So the DSR was detecting variations other than Mooney. This testing along with some occasional testing of "good" versus "bad" samples led to the conclusion that at the same viscosity (measured by Mooney or the DSR "A" factor) the higher alpha value seemed to correlate with better processing. For instance the normal bound styrene polymers have higher DSR alphas than the comparable E-BR and the lower Mooney rubbers have higher DSR alphas than their higher Mooney counterparts. A study was then initiated to test production samples of several SBR types on a random basis to determine the extent of the variation within each product type. Indeed the samples generally fell within a fairly narrow band on the DSR "A" versus DSR alpha plot as shown for SBR 1712 in figure 2. Several samples known to be poor processing fell below this band. Ultimately this led to the identification of several causes of variation in the processability of SBR 1712. The upper line in figure 3 shows the DSR data for an improved product compared with the earlier product as represented by the lower line. The lower line is the same as figure 2 without the data points shown. While this proved to be a good internal tool, it had limited utility outside the company.

In 1991 the stress relaxation feature was installed on our Mooney machine. This promised a quick processability test by extending the Mooney viscosity test 1 to 2 minutes. While the testing conditions were much different, it was found that the Mooney stress relaxation data were similar to the DSR data. Therefore attempts to correlate Mooney stress relaxation testing with other processability tests were pursued.

During the production of SBR 1502 a situation occurred which was thought to be a potential cause of poor processing product. Samples of this material were compared with typical SBR 1502 and SBR 1507. These are similar with the primary variable being Mooney viscosity. Experience has shown that many of the processability tests are affected by Mooney viscosity. A broad range is desirable for good statistical analysis. The data were studied by plotting the various Mooney and stress tests against Mooney viscosity or other viscosity measurements. The samples of typical SBR 1502 and SBR 1507 generally fell into a narrow band. The atypical 1502 samples generally outside this band. The Mooney versus delta Mooney data are shown in figure 4. Delta Mooney data are interpreted that the lower the value, that is the more negative it is, the better the processing. The first thing to be noted here is the overall relationship which would indicate that the high Mooney rubber is better processing than the low Mooney rubber. This does not jibe with experience. This relationship has been seen in other sets of data. The only reason that delta Mooney specifications are somewhat valid is that the Mooney range is relatively narrow for each type of rubber. A better interpretation would be the lower the value the better the processability at the same Mooney viscosity. The next thing to note is that the Mooney versus delta Mooney relationship for the atypical samples is slightly different than the relationship for the typical samples. The atypical samples seem to be slightly poorer than the typical samples although most would be within the specification for SBR 1502.

The Mooney versus Mooney stress relaxation time to 80% decay (MSR t-80) shows that the atypical samples are much different than the typical samples. GPC molecular weight distribution or dispersity data and Brabender black incorporation time also show a significant difference between the two sets of rubber samples. Figures 5 and 6 show the correlation between MSR t-80 and GPC molecular weight distribution and Brabender BIT respectively.

The Mooney stress relaxation data for unextended SBR correlate well with several processability indicators. However, the oil extended SBRs are much more prone to variations in processability than the unextended types. SBR 1712 and SBR 1778 are the two high volume oil extended products. Both use a high Mooney 23.5% bound styrene polymer extended with 37.5 phr (27.3%) oil. SBR 1712 contains an ASTM Type 101 (highly aromatic) extender oil while SBR 1778 contains an ASTM Type 103 (naphthenic) extender oil. Several versions of SBR 1712 and SBR 1778 are available with different Mooney specifications within a 30 to 58 ML range. Table 1 shows the correlation between Brabender Black Incorporation Time and the several other processability tests for 23 samples of 1712. Note that the correlation with delta Mooney (DML) is only good when the Mooney viscosity is included. The correlation between MSR t-80 and BIT is good. For these oil extended rubbers the GPC molecular weight data do not correlate well. To prevent clogging of the columns, the GPC sample solutions were filtered. Unfortunately this filtration removed some of the high molecular weight fractions. When this happens the data represent only the portion passing through the filter. The tell-tale sign that this has occurred is an abnormally low Intrinsic Viscosity. This phenomenon is difficult to quantify. However, the general trend is toward a lower MSR t-80 with a narrower molecular weight distribution.

SBR 1778 shows the same types of correlation as SBR 1712. Figure 7 shows the correlation between Mooney Stress relaxation data and the combination of delta Mooney plus Mooney. The correlation between MSR t-80 and Black Incorporation Time is shown in figure 8. This relationship is not as good as for 1712 possibly because the second maximum is so broad that an exact time is difficult to quantify. A better correlation is seen between MSR t-80 and Brabender Black Wetting Time.

Conclusion

Mooney Stress Relaxation testing, particularly the Time to 80 Decay, is a good candidate to replace delta Mooney and Brabender Processability Index data as specification tests for processability of emulsion styrene-butadiene rubbers.

[Tabular data omitted]

References

[1.] J.L. White and I. Soos, Rubber Chem. Technol. 66, 435 (1993). [2.] J.L. White and N. Tokita, J. Applied Sci. 11, 321 (1967). [3.] N. Tokita and R. Scott, Proceed. of 5th Rheological Congress, 3, 29 (1970). [4.] S.R. Moghe, Rubber Chem. Technol. 39, 247 (1976). [5.] E.R. Harrell, J.P. Porter, and N. Nakajima, Rubber Chem. Technol. 64, 247 (1993).