Rubber characterization by applied strain variations using the rubber process analyzer.

The RPA 2000 rubber process analyzer is a new dynamic mechanical rheological tester (DMRT) designed to measure the dynamic properties of raw polymers, uncured compounds and final cured compounds. The RPA strains a sample in shear by oscillating the lower die sinusoidally. Oscillation frequency can be set from 0.1 to 2,000 cycles per minute (cpm). The magnitude of the lower die movement can be set by the angular oscillation of the lower die or by the required strain on a sample. The lower die can oscillate from [+ or -]0.05 degrees of arc to [+ or -]90.00 degrees of arc. This angular oscillation corresponds to a strain of [+ or -]0.7% to [+ or -]1,256%. Figure 1 illustrates the conversion of angular oscillation to strain and percent strain for the RPA biconical die system. These formulas also apply to a cone and plate die system except that the cone angle is not multiplied by two (ref 1). The available combinations of frequency and strain in the RPA are limited to a maximum shear rate of 30 [sec.sup.-1]. Table 1 lists some examples of allowable combinations.

The strains available for testing in the RPA can be split into two major groups: low strains which commonly correspond to the linear viscoelastic region and high strains that correspond to the non-linear viscoelastic region.

In the non-linear viscoelastic region, it was suspected that there might be an irreversible movement of material within the sample because the strain exceeded some critical value. Testing with strips of black and white stocks in the non-linear viscoelastic region has shown that the movement of material is laminar and in the direction of the straining movement. There is no significant movement of material from one die to the other.

The RPA analyzer is the only DMRT which can test with repeatability and reproducibility at very high strains. Other DMRTs do not have sealed and pressurized sample chambers. As figure 2 shows, the material at the edge of an unsealed sample cavity can distort under high strains in a non-repeatable way. This will result in poor performance. An RPA 2000 sample is confined at the edge and prevented from distorting in a non-repeatable way (figure 3). An RPA sample is deformed during high strains only as much as the seals and seal plates will allow, but the sample distortion at the edge of the lower die is not measured at the torque transducer located at the upper die.

Linear viscoelasticity 

Linear viscoelasticity occurs when the ratio of stress to strain is a function only of time and temperature. This ratio is not a function of the magnitude of the strain itself (ref. 2). Generally this behavior is exhibited for a polymer below some strain level called the critical strain. In the RPA, the ratio of stress or torque (S) to applied strain (deformation) is proportional to the shear modulus (G). The elastic shear modulus G' and viscous modulus G" are calculated as follows:

G' = K S' / strain

G" = K S" / strain Where

S' is elastic torque

S" is viscous torque

K is a constant related to the geometry of the dies and sample cavity Also

Tan 8 = S"/S' or G"/G'

Figure 4 shows the RPA viscoelastic responses for a variety of raw synthetic elastomers versus various applied strains. Note that G' remains constant with increasing strain up to the critical strain (also G" and tan [delta] remain constant as well). After this critical strain is exceeded, the shear moduli starts to decrease. This critical strain denotes where the nonlinear viscoelastic region begins. Most of the raw synthetic elastomers tested showed linear viscoelastic behavior at strains under [+ or -]14% strain (1 degree arc). The linear viscoelastic region in a strain sweep is very useful in detecting many subtle differences among polymers.

Figure 5a and 5b show G' and G" responses respectively versus strain for eight different natural rubbers from an RPA low strain sweep. The G' response gave better separation of these raw Nrs than the G" response. In the linear viscoelastic region, the differences in the elastic response for NRs are greater than the differences in viscous response because of differences in molecular weight and chain entanglements.

Uncured mixed rubber compounds have a shorter linear viscoelastic region compared with raw elastomers. This is primarily due to the incorporation of carbon black and the establishment of a carbon black aggregate-aggregate network which is altered with rising strain (ref. 3). Previous studies in the literature have indicated that the critical strain for a carbon black loaded rubber compound is typically about 0.2% (ref. 4).

The non-linear viscoelastic quality due to carbon black is illustrated in figure 6, where a series of SBR compounds given in table 2 were analyzed in an RPA strain sweep. The G' and G" moduli rise rapidly as the carbon black loadings increase. The more carbon black that is present means the more non-linear the viscoelastic response. However, at the lowest carbon black loading level of 30 phr, the viscoelastic response approaches linearity within a portion of the applied strain sweep.

Non-linear viscoelasticity

Non-linear viscoelasticity is observed when the modulus (the ratio of stress to strain) changes in a strain sweep under constant conditions of temperature and frequency. Non-linear viscoelasticity occurs at an applied strain above acritical strain discussed earlier. Most factory processing applications are in the non-linear viscoelastic region for both raw rubber as well as mixed stocks. Therefore, rheological behavior in the non-linear region should be very important in predicting down-stream processability in the factory.

Raw polymers

Previous studies (refs. 5-7) have shown that raw rubber processability can be characterized very well in the linear viscoelastic region with the RPA. However, figure 7 shows the non-linear G' response from a high strain sweep applied to a variety of raw elastomers. These strain sweeps were all well above the critical strain. Note that the discriminating power of G' appeared to diminish as strain increased. Figure 8 shows the elastic torque response (S') for the same high strain test shown before in figure 7. By using S' instead of G', greater differences are seen in these polymers at high strains. Typically, the S' passes through a maximum with increasing strain. Experience has shown that changes in S' at the high strain region relate to processing differences in the factory. Therefore, the torque response (S' and S") is possibly better for studying rheological changes occurring at high strains in the non-linear viscoelastic region.

Raw rubber characterization - An important source of variation in rubber compounding and mixing is the raw rubber. Variations in the quality of rubber can greatly affect processing and the performance of the final product. Of all the raw material variations that must be controlled, the consistency of the raw rubber is one of the most important considerations.

For more than 40 years, Mooney viscosity (ASTM D1646) has been the standard for measuring the quality of raw polymers in the rubber industry. However, good Mooney viscosity lot to lot uniformity of a rubber does not by itself assure consistent processing. Variations that may not be detected by Mooney viscosity alone include molecular weight distribution, branching, monomer ratios or a variety of other micro and macro characteristics which can affect processing and end-use performance of the compound (ref. 8). A European conference synthetic rubber concluded that raw rubber shipments cannot be sufficiently characterized in relation to processability by a final Mooney viscosity alone (ref. 9). Previous work (refs. 6, 7 and 10) showed RPA frequency and strain sweeps can effectively characterize a variety of elastomers. Examples of strain sweeps on raw polymers have already been shown in figures 4 and 8.

Besides differences from changes in elastomer type, significant differences were also seen with the "same" elastomer from different sources. For example, figure 9 illustrates the differences seen in the S' response generated by an RPA high strain sweep for two different sources of SBR 1006 that have essentially the same Mooney value (ML 1+4 of 48.1 vs. 48.2). Likewise, figure 10 also confirms rheological differences with the RPA tan [delta] response from a frequency sweep (ref. 7) in the same experiment. (The uncured tan [delta] at low frequencies is an accepted method for predicting a raw polymer's processability characteristics).

Another example of rheological differences observed with a high strain sweep is shown in figure 11. Both of the natural rubbers shown are technically classified TC10 grade. However, the NR sample A processes normally while NR sample B processes poorly. Rheological differences can clearly be seen from the S' response. However, when a frequency sweep was performed on the same two NR samples, very little difference could be seen. This is an example where a high strain sweep on a raw rubber gave rheological information which could not be obtained in a normal frequency sweep.

Mixed rubber stocks

The rheological behavior of mixed rubber stocks in a factory operation is different from that of raw polymers. This is because most mixed stocks contain carbon black as well as other filler loadings. Carbon black loading greatly affects the rheological properties of a rubber compound. One of the reasons for this is that carbon black forms an aggregate-aggregate network in the rubber medium. It is many times necessary to destroy this network through the high strain testing on the RPA in order to predict the viscoelastic behavior of the rubber compound down stream in the production plant. This is an important advantage of high strain testing in predicting processability characteristics of a rubber compound.

Tire compounds - Figure 12 shows the S', S" and tan [+ or -] from an RPA high strain sweep of 14 different tire compounds selected from the literature (ref. 11). These 14 tire compounds differ greatly in their polymer base, carbon black type and loading, filler-oil ratio, etc. These variations represent the diversity of compounds one might normally find in a tire plant. Figure 12 shows that the S' & S" responses give-better separation for these 14 stocks at the higher strain levels on the RPA. Also these diverse compounds reach the S' maximum values at different strains.

Mixing study - A series of controlled mixing studies was conducted in which the state-of-mix was measured with the RPA. Laboratory simulation of factory mixing techniques was applied to an internal mixer to achieve different states-of-mix (ref. 11a). A power integrator recorded the total amount of work in kilowatt hours at dump for each batch.

From the recipe shown in table 2, a series of SBR tread stocks were mixed with different energies at dump to represent different states-of-mix and % carbon black dispersion. The S' response in the high strain sweep shown in figure 13 separates the SBR tread stocks very well in order of increasing energy input at dump.

Applications of high strain testing of rubber compounds

The amount and type of carbon black and oil used in a rubber compound greatly affects its processing characteristics and cured physical properties. Carbon black surface area, structure and loading levels all affect a compound's processing behavior and final cured properties. Above a threshold loading, carbon black forms an aggregate-aggregate network when mixed into a rubber compound (ref. 12). Applied strain breaks down this network. After a sufficient relaxation time, the network reforms. The nature of the breakdown and reformation of this network affects a compound's processability in the uncured state and its mechanical properties in the cured state (refs. 7, 13 and 14). The following discussion shows the benefits of applying high strain test conditions.

Effects of carbon black types

A rubber compound study (ref. 8) with different carbon black types illustrates the usefulness of the RPA's capability to combine low and high strains in the same test. Twelve different carbon black types were mixed in separate batches using the standard ASTM SBR test recipe given in table 4. These twelve carbon blacks differed in both average particle size and structure. The carbon black average particle size can be determined from the inverse of the carbon black surface area. Nitrogen adsorption as measured by ASTM Test Method D 3037 is an effective method for measuring the surface area of a carbon black. Figure 14 shows a scatter plot of the RPA G' response for the twelve SBR stocks at very low strain (0.7%) to the nitrogen adsorption surface area for each of the carbon blacks used. The result was a good correlation between G' and particle size.

Carbon black structure can be measured by DBP absorption (ASTM D2414). Figure 15 shows the correlation of RPA G" response at very high strain (1,256%) for the twelve SBR test stocks to the DBP absorption of the carbon blacks used. Again the results show good correlation.

Figure 16 shows that the carbon black structure (DBP absorption measurement) does not correlate to the carbon black particle size (nitrogen adsorption) for the twelve carbon blacks used in this experiment. From this comparison we conclude that the RPA can give an indication of both particle size and structure by testing compounds at both low and high strain conditions. This conclusion is based on the assumption that RPA testing at very low strain did not destroy the carbon black network. Since the density of the carbon black network is inversely proportional to average particle size, the G' response is also inversely related to carbon black particle size. Conversely, under high strain test conditions, the carbon black network is destroyed (ref 15). This leaves only the carbon black structural characteristics which have a great effect on down stream processing performance such as die swell and dimensional stability. This illustrates why it is important to apply high strain testing to a rubber compound in order to effectively predict the compound's processability down stream.

Capillary rheometer comparisons

Many processes in the factory occur at high shear rates which cause the destruction of the carbon black networks discussed earlier.

The Monsanto MPF capillary rheometer has been used for many years to impart a high shear rate to a rubber compound and predict its processability. An experiment was run to determine the best RPA conditions to correlate with MPT shear stress values at different shear rates. The 14 tire compounds discussed earlier were selected for this experiment (ref. 6). This selection of tire compounds represents the diversity normally found in a tire plant. RPA comparisons were made with the MPT at shear rates of 30, 100, 300 and 1000 [sec.sup.-1]. In every case, the correlation of the RPA with the MPT improved greatly as the applied strain on the RPA was increased. Figures 17, 18 and 19 show the good correlations of the RPA uncured elastic modulus G' with the MPT shear stress at 30, 100 and 300 [sec.sup.-1] respectively. In every case, a good correlation was achieved under RPA high strain conditions that broke up the carbon black network. Figure 20 shows how the correlation coefficients improved between the RPA and MPT as the RPA's applied strain was increased.

References

[1.] Whorlow, R. W., Rheological techniques, Wiley (1980) pp. 255, 274. [2.] G.R. Hamed, "Viscoelasticity," Designing with elastomers - rubber as an engineering material, Educational Symposium, Energy Rubber Group, Sept. 24, 1990. [3.] A. Coran, J.B. Donnet, Rubber Chem. and Tech., vol. 65, No. 5, (1992), p. 1037. [4.] M. Gerspacher, "Carbon black response to dynamic strains," Elastomerics, Nov., 1990. [5.] H. Pawlowski, J. Dick, "Viscoelastic characterization of rubber with a new dynamic mechanical tester," Rubber World, June, 1992. [6.] J.S. Dick, H.A. Pawlowski, "Applications for the rubber process analyzer," Rubber & Plastics News, April 26 and May 10, 1993. [7.] J.S. Dick, H.A. Pawlowski, "Applications of the rubber process analyzer in predicting processsability and cured dynamic properties of rubber compounds," presented at the Rubber Division meeting at Denver, Colorado, Spring, 1993, pp. 17-22. 7a. ibid. [8.] H.A. Pawlowski, J.S. Dick, "A new dynamic mechanical tester designed for testing rubber," presented at Rubber Division, ACS meeting at Louisville, KY, Spring, 1992, p. 8-9. 8a. ibid. [9.] Reference to Polymer Conference of April 16-17, 1991 at Luven, Belgium, cited in European Rubber Journal, June, 1991, p 28. [10.] H. Pawlowski, J. Dick, "Viscoelastic characterization of rubber with a new dynamic mechanical tester," Rubber World, June, 1992. [11.] J. S. Dick, H.A. Pawlowski, "Applications for the rubber process analyzer," presented at the ACS Rubber Division Meeting at Nashville, Tennessee, Fall, 1992, p. 20. 11a. ibid., pp. 11-15. [12.] A. I. Medalia, "Effects of carbon black on dynamic properties of rubber vulcanizates," Rubber Chem. and Tech., vol. 51, No. 3, (1978), pp. 457-465. [13.] A. Coran, J.B. Donnet, Rubber Chem. and Tech, vol. 65, no. 5, (1992), p. 1037. [14.] H.G. Burhin, "Improved techniques for characterization of polymer and compounds, before, during and after cure, Presented at Brighton, UK, 1992, p. 6. [15.] A.R. Payne, "The dynamic properties of carbon blackloaded natural rubber vulcanizates," Rubber Chem. and Tech., vol. 36, 532 (1963).