Comparison of shear thinning behavior using capillary and rotorless shear rheometry.

During processing, the flow of a raw rubber or rubber compound is shear rate dependent. A higher shear rate will reduce the viscosity of rubber. This is known as non-Newtonian behavior, pseudoplasticity or simply shear thinning. Different raw rubbers or rubber compounds have different degrees of this shear thinning quality (ref. 1). Being able to accurately measure this shear thinning quality is very important because it relates to how well a rubber compound will process, the throughput of the process, its stability and the quality imparted to the rubber product. Shear thinning characteristics are particularly important for extrusion and injection molding operations.

The degree of shear thinning behavior is determined by several characteristics of a rubber compound. Prior studies reported the effects of different filler-oil systems on the shear thinning characteristics of rubber compounds (refs. 2 and 3). Also it is reported that raw rubber characteristics greatly affect the shear thinning characteristics of the respective rubber compound. Some raw rubber characteristics which influence shear thinning dependence are:

* Average molecular weight (ref. 4);

* molecular weight distribution (ref. 5);

* long chain branching (ref. 18);

* gel (ref. 6); and

* chain entanglements (ref. 7).

The purpose of this study was to quantify some of these polymer effects on the respective rubber compound shear thinning characteristics.

Experimental

Table 1 shows the 20 different raw elastomers which are selected for this study. As can be seen, the selected polymers include four different EPDMs, two fluoroelastomers, three polychloroprenes, two chlorinated polyethylenes, two NBRs, and one each for HNBR, ECO, star branched bromobutyl rubber and polyacrylate elastomer. These elastomers were tested as raw rubber and were tested in two different formulations. These formulations were prepared through BR internal mixing followed by milling. Since the specific gravity for these elastomers was different, all comparisons were made on an equal volume basis. One formulation consisted of mixing each elastomer with N660 carbon black at 25 parts by volume. The other formulation designated that each elastomer be mixed with N650 carbon black at 55 parts by volume. In addition, two types of thermoplastic elastomers (SBS and TPV) were also studied outside of the compounding study. (TPEs are not usually compounding ingredients, but are normally used "as is".)

Test instruments

These rubber compounds were tested for shear thinning behavior with a capillary rheometer (hereafter referred to as the MPT) and a rubber process analyzer (hereafter referred to as the RPA). The MPT technology was introduced in the 1970s. This test method is typical of a capillary rheometer which can vary the applied shear rate by changing the piston speed of the instrument. The RPA technology has been widely used worldwide since 1992. This instrument is a special rotorless oscillating shear rheometer, which is capable of test conditions not normally available with traditional oscillating rheometers.

For this study, the MPT capillary rheometer was programmed to move its piston in the barrel at predetermined speeds in order to apply sequentially increasing shear rates of 30, 100, 300 and 1,000 [sec.sup.-1] to the rubber formulation as it extrudes through a 1.5 mm (0.059 inch) diameter orifice die with an L/D of 20:1. The test temperatures selected were 100, 125, 150, 200, 215 and 225[degrees] C. (The higher temperatures were used to compare TPEs.) The MPT output used in this study was apparent viscosity ([[eta].sub.app.]) in kPa * sec. or Pa * sec.

The RPA was used as a processability tester. The RPA molds the uncured rubber formulation in a sealed pressurized cavity and has a special direct drive motor, which oscillates the lower die sinusoidally over a range of preprogrammed strains, frequencies and temperatures. RPA shear rate sweeps (through variations in frequency) were performed at 7% and 14% strain to measure the drop in complex ([[eta].sup.*]) and real ([eta].sub.']) dynamic viscosity with increasing shear rate for raw polymers and test compounds. Also, very high strain sweeps were performed with the RPA. Other outputs from the RPA include G' for storage (elastic) modulus, G" for loss (viscous) modulus, S' for elastic torque and S" for viscous torque. Also, another output, tan [delta], is derived from either dividing G" by G' or S" by S'. This parameter is dimensionless, without units.

Quantifying non-Newtonian flow

The viscosity of a Newtonian fluid, such as water, will remain constant with an increase in applied shear rate. However, raw rubber and rubber compounds are non-Newtonian. The viscosity of an uncured rubber compound will decrease with an increase in applied shear rate. Typically this decrease will be displayed as a power law drop. Different base elastomers impart different magnitudes of shear thinning behavior to a given compound. Therefore, they impart different slopes. So when the log of viscosity is plotted against the log of applied shear rate, a different linear slope results for different types of elastomers. Therefore, all of these non-Newtonian slopes from increasing shear rates for either the MPT or the RPA are derived from the power law formula given in equations 1 and 2.

y = [kt.sup.[-alpha]] (1) log y = - [alpha] log t + log k (2) Where y = [[eta].sub.app] for MPT testing or [[eta].sup.*] for RPA testing; k = intercept; [alpha] = slope; t = time.

Correlation coefficients (r) for the statistical power law regressions for both the MPT and RPA data were calculated. Also, an alternate method to compare non-Newtonian behavior is to simply calculate the percent drop in viscosity between a pre-designated low shear rate and a pre-designated high shear rate.

The relationship between viscosity measurements from capillary rheometers vs. dynamic mechanical testers with sinusoidal deformation should also be addressed. Cox and Merz in the early 1950s published the following empirical relationship they found between capillary rheometer apparent viscosity measured under conditions of steady state shear rate and dynamic complex viscosity which is measured by sinusoidal deformations (and thus a constantly changing shear rate) by using a dynamic mechanical rheological tester (such as the RPA) (ref. 8). This relationship is given in equation 3.

[[eta].sub.app.]([gamma]) = [[eta].sup.*] ([omega])|[omega]=[gamma] (3) Where: [[eta].sub.app] is the apparent (uncorrected) capillary viscosity at a steady shear rate of [gamma] in ([sec.sup.-1]); and [[eta].sup.*] is dynamic complex viscosity measured at an oscillatory frequency of [omega] (in radians per second).

Also, the same relationship is reported to work when comparing the corrected viscosity from the MPT vs. the real dynamic viscosity [[eta].sup.'] from a DMRT. This empirical relationship sometimes works quite well; however, it does not have a theoretical basis, therefore it may not always work (ref. 9).

Results and discussion

Overall comparison of viscosity values

Figure 1 gives a direct comparison of [[eta].sub.app] viscosity measurement from the MPT vs. the [[eta].sup.*] complex dynamic viscosity measurements from the RPA for compounds loaded with 25 parts by volume of N660 carbon black. As can be seen, the agreement in Pascal * Second units is excellent and the slope of the regression line is very close to unity. Thus, the agreement with the Cox Merz role is very good at this filler loading level. Figure 2 makes the same comparison with the compounds containing higher carbon black loadings of N650 at 55 parts by volume. At these higher filler loadings, the correlation is still very good; however, the regression slope is less than 1.0, showing some deviation from the Cox-Merz rule at this higher filler loading. This supports past studies, which have also found very good correlation between the MPT apparent viscosity vs. the RPA complex dynamic viscosity at low and somewhat high filler loadings, but some calculable deviations from the Cox-Merz rule at higher filler loadings (refs. 10 and 11). It should be noted, that because of concerns for higher variation of the capillary rheometer viscosity measurements vs. that experienced with the RPA, all comparisons to the MPT data were done with averages from duplicate testing. If the MPT duplicate test results did not agree, then two more tests were performed and used to calculate the average. In some cases where selected polymers displayed noticeable slippage or "slip stick" from the MPT capillary rheometer testing, these results were deleted from the statistical regression analysis.

[FIGURE 1 AND 2 OMITTED]

Repeatability comparison

In one series of tests, a significant number of repeats was performed for both the MPT and RPA testing in order to measure the test repeatability of each instrument. This repeatability was measured by calculating the coefficient of variation (CV) as shown in equation 4. CV = 100 * (std. dev./mean) (4)

This standard deviation is a pooled standard deviation Sp derived from replicate testing of the same samples, as shown in equation 5.

Sp = [([S.sup.2.sub.1] + [S.sup.2.sub.2] + [S.sup.2.sub.3] + ... [S.sup.2.sub.n])/n][sup.1/2] (4) Where: Sp is the pooled standard deviation; and [S.sub.1], [S.sub.2], [S.sub.3],.... [S.sub.n] are calculated standard deviations from replicate testing of mixed batches 1, 2, 3, ... n.

Figure 3 shows clearly that the repeatability of the RPA viscosity measurements in this study was twice as good as the repeatability for the MPT capillary rheometer measurements.

[FIGURE 3 OMITTED]

General comparison of shear thinning measurements

Figures 4 and 5 illustrate the shear thinning characteristics of the selected polymers loaded at 25 parts by volume with N660 carbon black as measured by the MPT and the RPA. Figure 6 shows the very good correlation between the shear thinning measurements from the RPA vs. the MPT for the compound series containing 25 parts by volume of N660 carbon black. Figure 7 compares the magnitude of shear thinning that each polymer imparts to the compound. It is interesting to note that in many cases the percent drop in viscosity is about the same whether measured by the MPT or the RPA. The reason for this similarity is that in each case the comparisons were made between the same shear rate conditions, i.e., 30 to 100 [sec.sup.-1] for the MPT and 30 to 100 rad/sec. for the RPA. In general, a word of caution is noted when trying to measure shear thinning characteristics with the MPT vs. the RPA. The RPA can measure at very low shear rates, while the MPT can not. To get the best correlation, one normally should not include very low shear rate measurements from the RPA which are below the MPT's normal shear rate range. Including these very low frequency measurements is unnecessary and may introduce noise. Also, the MPT shear thinning calculations for CSM and CPE had to be left out of this comparison because a close examination of the MPT raw data confirmed that these polymers displayed "slippage."

[FIGURE 4, 5, 6, 7 AND 8 OMITTED]

Figure 8 shows a good correlation between the shear thinning measurements from the RPA under high strain conditions vs. the conventional MPT measurements for the compound series containing 55 parts by volume of N650. This correlation could have been better; however, from examination of the MPT raw data, there was some "slip stick" and slippage that occurred with these loadings, even at 125[degrees] C. Figure 9 compares again the shear thinning imparted by the different polymers to this series of compounds. Because of the significantly higher loading of carbon black in this comparison, the high strain test conditions on the RPA gave a definite advantage. Previous studies have found advantages in applying higher strains to the closed sealed cavity of the RPA in order to break up the carbon black aggregate network (refs. 12-14) and/or gel and chain entanglements (refs. 15 and 16), which is exactly what happens to the rubber during factory processing.

[FIGURE 8 AND 9 OMITTED]

Effects of filler loading on shear thinning

It was observed that increasing loadings of filler impart a greater shear thinning quality to the rubber compound. The power law regression slopes ([alpha] in equation 2) were calculated for [[eta].sup.*] from RPA frequency sweeps of the raw rubbers, the rubbers mixed with 25 parts by volume of N660 carbon black and the rubbers mixed with 55 parts by volume of N650 carbon black. This power law regression slope can be used as a "shear thinning index." The greater the value means the more "shear thinning" the rubber. Figure 10 illustrates the effects of increasing carbon black loading on shear thinning of the rubber compound. As can be observed for each polymer, its calculated [alpha] increases from when it is tested in the raw state, to higher values with increased loading of carbon black. This filler effect on shear thinning has also been observed in earlier studies (ref. 17).

[FIGURE 10 OMITTED]

Effects of temperature on measured viscosity

Increasing the processing temperature for a rubber compound will reduce its viscosity. A rule of thumb is that a 10[degrees] C rise in temperature will reduce the viscosity by approximately 10%. In this series of experiments we examined the magnitude of this drop in viscosity with increasing test temperature. Figure 11 shows the results from this testing. This figure shows excellent agreement in measuring the percent drop in viscosity with temperature rise as measured by the MPT vs. the RPA. Also this figure shows that different types of elastomers possess significantly different degrees of temperature dependency regarding viscosity drop with a given temperature rise. For example, EPDM compounds show less change than fluoroelastomer compounds.

[FIGURE 11 OMITTED]

Figure 12 compares the RPA temperature sweep results for selected rubbers loaded with 25 parts by volume N660 carbon black. All compounds display the same characteristic drop in [[eta].sup.*] viscosity with temperature rise in a linear plot; however, it will be noted that some of these polymer compounds drop more in viscosity than other polymers. In some cases, there is even cross over.

[FIGURE 12 OMITTED]

Polymer comparisons

EPDM comparison - two pairs of EPDM elastomers were compared for rheological differences. One pair differed mainly in long chain branching (LCB). Previous studies have reported that LCB has the greatest effect on the uncured tan [delta], especially at very low frequencies (ref. 18). Figure 13 illustrates this quite well. At the very low frequency of 0.1 rad/sec., this difference is greatest. The greater the LCB content means the greater the reduction of the low frequency tan [delta].

[FIGURE 13 OMITTED]

The effects of varying ethylene content were also studied. An EPDM containing greater than 65% ethylene was compared to another EPDM which only contained 50% ethylene. We found the higher ethylene content imparted a lower complex dynamic viscosity at very low shear rates at 150[degrees] C (a temperature where any crystallites would have melted).

FKM comparison - a pair of fluoroelastomers was also examined in this study. These FKM polymers differed mainly in their molecular weight distribution (mwd). Figure 14 shows the difference seen in shear thinning behavior between these two polymers when tested in the raw state. Also, the same behavior was observed when these two polymers are each compounded with 25 parts by volume of N660 carbon black. Also, MPT capillary rheometer shear thinning data show, the same shear thinning relation observed with the RPA.

[FIGURE 14 OMITTED]

CR comparison - three different polychloroprenes were also included in this study. Neoprene W and WRT are mercaptan modified types, while Neoprene GRT is a sulfur modified type polychloroprene. It was observed that the G type imparted a significantly different shear thinning profile than the other two W types studied. This same pattern was observed for the RPA testing of the raw CRs, and the filled polychloroprenes at 25 and 55 parts by volume carbon black. Also, the same patterns were seen for MPT capillary rheometer testing of the filled CR polymers at 25 and 55 parts by volume carbon black.

CPE comparison - two chlorinated polyethylenes with identical % chlorination, but different average molecular weights and Mooney viscosity values, were tested as raw rubbers and mixed compounds. Dynamic complex viscosity [[eta].sup.*] moved upward from increased in average mw. This same pattern was observed for all RPA and MPT testing. Also, this pattern was observed for the raw polymers as well as the two sets of compounds with 25 and 55 parts by volume, respectively, of carbon black.

CSM comparison - two chloro-sulfonyl-polyethylene polymers with the same Mooney viscosity but different chlorine contents and crystallinity were compared. The CSM with the higher chlorination imparted a higher dynamic viscosity than the lower chlorination did to the filled compound at 55 parts by volume carbon black. Again, this same pattern was noted for all RPA and MPT testing of both the raw polymers and mixed stocks.

NBR and HNBR comparison - two different NBR polymers and one HNBR (hydrogenated NBR) were compared in this study. Hydrogenating the butadiene backbone of NBR polymers should render the HNBR more thermoplastic. It was observed that HNBR is more shear thinning than the NBR polymers tested. This pattern was observed with both RPA and MPT testing.

Polyacrylate and polyepichlorohydrin comparison - these polymers were compared by MPT and RPA testing. Both test methods showed the polyacrylate is significantly more shear thinning than the polyepichlorohydrin. This pattern was observed for both levels of carbon black loading, as well as for the RPA testing of the raw elastomers.

SBS comparison - these styrene-butadiene-styrene block polymers are commonly used as thermoplastic elastomers (TPEs). The two samples we examined in this study (SBS type 1 and SBS type 2) are actually SBS polymers that were preformulated with proprietary oils and other ingredients to make these polymers more injection moldable. Type 1 is reported to be a film grade which can be used in injection molding, while type 2 is formulated to have a higher viscosity and is also commonly used in injection molding. Figure 15 gives a comparison of shear thinning data from both the RPA and the MPT. From a power law regression through both the RPA and MPT viscosity data points, an [R.sup.2] of 0.99 was achieved for both types of SBS polymers tested.

[FIGURE 15 OMITTED]

TPV comparisons - three different thermoplastic vulcanizates based on EPDM, polypropylene and other proprietary ingredients (possibly including oils) were tested with both the RPA and MPT at 200[degrees], 215[degrees] and 225[degrees] C. Figure 16 compares the different shear thinning behaviors observed through RPA testing. As can be seen, different TPVs have significantly different shear thinning characteristics. Moreover, compared to conventional rubber compounds, TPVs appear to be significantly more shear thinning. At very high shear rates, the TPVs drop to much lower viscosity levels than the conventional rubber compounds we studied. MPT testing measured viscosity values at less than 100 Pascal-seconds (1,000 poises) at 1,000 [sec.sup.-1] shear rate for TPVs. This is much lower than what is measured for conventional rubber compounds. Figure 17 gives a comparison of the calculated shear thinning index ([alpha] from equation 2) for some of these TPVs vs. other conventional rubber compounds tested. However it was also observed that the viscosity values for TPVs do not decrease nearly as much with a rise in temperature. The percent drop with a 10[degrees] C change is only about 4% compared to significantly higher percent drops noted earlier in this study for conventional rubber.

[FIGURE 16 AND 17 OMITTED]

Conclusions

The Cox-Merz Rule for relating viscosity from the RPA and the MPT was very successful across a broad range of specialty elastomers, provided the filler loading levels are not excessively high and no slippage occurs during the MPT testing.

With higher filler loadings, the RPA and MPT viscosity values still correlated quite well, even though there was a shift in data.

Shear thinning measurements from the RPA agreed well with the MPT shear thinning measurements when data from each instrument were from within the same shear rate range.

For high filler loaded compounds, RPA high strain testing correlated well with MPT shear thinning measurements.

Higher carbon black loadings increased shear thinning behavior.

The RPA and MPT gave good agreement in percent drop in viscosity values due to a given rise in temperature.

The RPA easily measured rheological changes due to the following polymer variations: Molecular weight; molecular weight distribution; long chain branching; % halogenation; % hydrogenation; % crystallinity.

From RPA and MPT testing, TPV viscosity values were found to be more dependent on shear rate, but less affected by temperature variation than the other specialty elastomers tested.

RPA measurements were found to have repeatability two times better than the MPT measurements.

Figure 17 - MPT comparison of "shear thinning"
index [alpha] (equation 2) for TPVs vs. other
conventional rubbers filled with 25 parts by
volume of N660 carbon black

Shear rate index,   Calculated shear thinning
  alpha               index (alpha)

CPE, ML80                     0.81
FKM, narrower
  mwd                         0.66
NBR, ACN =
  36%, ML35                   0.59
EPDM, less
  LCB
TPV No. 3
TPV No. 2                     0.81
TPV No.1                      0.83

Note: Table made from bar graph.

References

(1.) J.S. Dick and Martin Gale, Processability Tests, Polymer Testing Handbook, R. Brown, Editor, Marcel Dekker, 1999, p. 206.

(2.) J.S. Dick and H.A. Pawlowski, "Application of the rubber process analyzer in characterizing the effects of silica on uncured and cured compound properties," ITEC '96 Select by Rubber and Plastics News, Sept. 1997.

(3.) W.G. DePierri, Jr., "The effect of oil and black on the injection molding of EPDM compounds," Rubber Chem. & Tech., Vol. 42 (1969), p. 1,321.

(4.) L.L. Blyer, Jr., "Some aspects of the relationship between molecular structure and flow behavior for polymer melts," Rubber Chem. & Tech., Vol. 42, (1969), p. 823.

(5.) K.P. Beardsley and W.A. Wortman, "Dynamic testing as a quality control tool for EPDM polymers and compounds," paper No. 21 presented at the Rubber Div. ACS meeting in Nashville, TN, Oct., 1998.

(6.) J.S. Dick, C. Harmon and A. Vare, "Quality assurance of natural rubber using the rubber process analyzer," J. of Polymer Testing, 18 (1999)pp. 327-362.

(7.) V.L. Folt, "The effects of mechanical degradation on rheological properties of elastomers," Rubber Chem. & Tech., Vol. 42 (1969), p. 1,294.

(8.) W.P. Cox and E.H. Merz, J. of Polym. Sci., 28, 619 (1958).

(9.) Y.P. Khanna, "Dynamic melt rheology. 1: Re-examining dynamic viscosity in relationship to the steady shear flow viscosity," Polymer Engineering and Science, March 1991, Vol. 31, No. 6, p. 440.

(10.) J.S. Dick and H. Pawlowski, "Application of the rubber process analyzer in characterizing the effects of silica on uncured and cured compound properties," ITEC '96 Select by Rubber and Plastics News, Sept. 1997.

(11.) J.S. Dick, H. Pawlowski and J. Moore, "Viscous heating and reinforcement effects of different fillers using the rubber process analyzer, "paper No. 7, Rubber Div., ACS, Meeting, Chicago, April, 1999.

(12.) J.S. Dick and H. Pawlowski, "Applications for the rubber process analyzer," Rubber and Plastics News, Apr. 26 and May 10, 1993.

(13). J.S. Dick and H. Pawlowski, "Applications of a new dynamic mechanical rheological tester in measuring carbon black and oil effects on rubber compound properties," J. of Elastomers and Plastics, Vol. 27, 1995.

(14.) J.S. Dick and H. Pawlowski, "Rubber characterization by applied strain variations using the rubber process analyzer," Rubber World, Jan. 1995.

(15.) J.S. Dick, C. Harmon and A. Vare, "Quality assurance of natural rubber using the rubber process analyzer," Polymer Testing, Sept. (1999).

(16.) C. Stevens and J. Dick, "Factory testing and control of raw natural rubber and mixed batches using the rubber process analyzer," paper No. 4, Rubber Div. Meeting, Orlando, FL, Sept. 21, 1999.

(17.) J.S. Dick and H.A. Pawlowski, "Application of the rubber process analyzer in characterizing the effects of silica on uncured and cured compound properties," ITEC '96 Select by Rubber and Plastics News, Sept. 1997.

(18.) H.J.H. Beelen, L.R. Maag and J.W.M. Noordermeer, "Understanding the influence of polymer and compounding variations on EPDM extrusions," Rubber World, July, 1998.