Processability by Mooney relaxation for isobutylene elastomers.

Complete processability characterization of a polymer may require a battery of rheological and structural tests. For practical reasons simple viscosity indicators are commonly used to describe and specify polymers. Mooney viscosity and melt index are the two primary processability parameters used in the rubber and plastics industry. These parameters are generally related to the molecular weight of the polymer, most likely to the weight average molecular weight. They can usually provide a good processability indication for a polymer that belongs to a family of polymers which the user is familiar with. However, when some structural features of a polymer are changed beyond what is expected within the family, some aspects of the viscoelastic nature can be affected, resulting in a change of the polymer processability even if the Mooney viscosity or melt index are unchanged. There is therefore a desire to replace the viscosity parameter by a viscoelastic one or alternatively add another simple indicator which is an elastic measure. Since viscoelasticity is strongly dependent on polymer structure, the added parameter can be either a rheological or a structural one.

This article addresses isobutylene based elastomers which are primarily linear polymers. For linear polymers it is understood that viscoelasticity is strongly dependent on the molecular weight distribution (MWD) which was the reason for some reported attempts to derive MWDs from measured rheological behavior (refs. 1 and 2). Simple MWD parameters such as the various molecular weight averages and their ratios can be used to describe some of the characteristics of the distribution. A popular parameter for distribution width is the ratio of weight to number molecular weight averages, Mw/Mn. MWDs for most linear polymers can be reasonably well described by just two parameters as long as the distribution is a simple mono-modal one. The most popular parameter pairs for MWD consist of one of the molecular weight averages, Mw or Mn, and the width parameter Mw/Mn. The weight average molecular weight, Mw, is a good measure of polymer viscosity while Mw/Mn is a reasonable measure of elasticity. It is felt that the current viscosity parameters, Mooney viscosity and melt index, are highly established in the industry and can not be replaced. Therefore, improved processability parameters can consist of a combination of a viscosity parameter and a second parameter which is a measure of elasticity. The second parameter can be a structural one, for example Mw/Mn obtained by an analytical method such as gel permeation chromatography (GPC) or an elastic parameter obtained by rheological testing.

Simple rheological tests have the advantage of being relatively fast, fairly reliable and simple. For processability indication they also have an added advantage of being a result of a test that is closer to a processing situation. Some of the simpler rheological tests which measure an elastic response include stress relaxation, strain recovery and flow induced swell (die swell). The last principle is not well suited for uncompounded elastomers since they tend to melt fracture in flow. Relaxation and recovery are therefore the tests of choice for elastomers. A number of relatively simple instruments were built to exploit these two methodologies. Strain recovery tests date back to the Williams Plastometer and its more recent improved version named the Defo Elastometer (refs. 3 and 4). Instruments that measure stress relaxation after a step strain include the DSR (ref. 5) and RPA (ref. 6). Mooney relaxation (MLR) is another stress relaxation test categorized as a cessation of flow test since Mooney torque decays following an abrupt braking of the Mooney viscometer rotor. MLR was recognized as a possible test for measuring an elastic response shortly after the Mooney viscosity test was first introduced, however, it suffered from poor sensitivity and repeatability and was not pursued commercially. Recent improvement in equipment, including abrupt motor braking, reduced friction, computer correction for differences in friction before and after motor braking and online data analysis resulted in the development of a Mooney viscometer with MLR capabilities (refs. 7 and 8). Such an instrument should be able to provide information on both viscosity and elasticity in one test. This led Friedersdorf et al (ref. 9) to develop test parameters for quality control based on MLR which were later incorporated into the ASTM D1646 test for Mooney viscosity. A different MLR parameter is proposed in this article to resemble the MWD width parameter Mw/Mn. This parameter named Mooney relaxation index (MRI) was specifically developed for isobutylene based polymers. The combination of Mooney viscosity and MRI, obtained in one test, can offer an improved processability measure since they capture some aspect of the viscoelastic nature of the polymer.

Stress relaxation

Stress relaxation of polymer melts after a small constant step in strain is an established technique for the determination of linear viscoelastic properties (ref. 10). The stress relaxation behavior of a polymer contains information on its MWD which as mentioned above lead to attempts to derive MWDs from stress relaxation tests on linear polymers (refs. 1 and 2). A polymer consisting of single length molecules tested well above its glass transition temperature is expected to exhibit a stress decay following a single exponential function

G(t) = G exp(-t/[Pi])

where G(t) is the relaxation modulus as a function of time and [Pi] is a relaxation time constant. G and [Pi] are constants for the specific molecular weight. When the MWD is wider, the relaxation modulus can be presented by a sum of exponentials with pairs of constants [G.sub.i] and [Pi.sub.i] known as the "relaxation spectrum." The total area under a stress relaxation curve is the zero shear viscosity (ref. 10) of the tested polymer, [Eta.sub.0]

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

The long time tail of the relaxation function can be fitted with a single exponential having the longest relaxation time constant [Pi.sub.m]. The longest relaxation time or the zero shear viscosity are known to be a power function of the weight average molecular weight to an approximate power of 3.4 for linear polymers. This relation was proven to hold for many linear polymers, including polyisobutylene (refs. 11-13). Chung et al (ref. 14) used this relation for molecular weight determination of polyisobutylene with the RPA instrument mentioned above.

Stress relaxation for polymers of various molecular weights and various MWDs tested well above their glass transition temperature can be generally described by figure 1. Since relaxation time needs to be followed over many decades of time it is customary to plot stress relaxation using logarithmic scales. The initial decay in figure 1 represents a transition from a glassy to a rubbery state resulting from the instantaneous step strain which can not be practically measured due to the very short time of this transition. Stress relaxation for very high molecular weight polymers is slowed down significantly following a 3.4 power as discussed above. A wider MWD polymer contains elements of low and high molecular weight resulting in a distribution of decay times. As shown in figure 1, a polymer with a wide MWD will show a faster relaxation rate at short times due to the presence of smaller molecules and a slow relaxation rate at longer times due to the presence of larger molecules. Comparing relaxation rates at long and short times can therefore provide some information on MWD width or alternatively on the elastic nature of the polymer. Wider MWD polymers exhibit higher elasticity but they also exhibit a larger degree of shear thinning both of which have processability consequences.

Mooney relaxation (MLR)

Mooney relaxation is the decay of Mooney torque after braking the motor of a Mooney viscometer. The general description and the parameters that can be calculated from a Mooney relaxation test are shown in figure 2. Since MLR is a cessation of flow test it does not obey the step strain stress relaxation analysis described above but the general arguments demonstrated in figure 1 are generally valid. If MLR stress decay is plotted against time in a log-log plot it resembles the flow portion, which is the final portion, in figure 1. Being a cessation of flow test, where smaller polymer molecules can relax during the steady shear flow period, the initial decay for MLR is much faster than it would be in a step strain test. But the initial decay rate is still a measure of low molecular weight species while a slow decay at the longer time is a measure of high molecular weight species as shown schematically in figure 3. The area under the actual (linear) Mooney relaxation curve is also a measure of a higher average molecular weight but it is not identical to zero shear viscosity as in a step strain relaxation described by equation 2. Some ratio of the initial decay to MLR area should therefore provide MWD width information similar to Mn/Mw.

It has been empirically observed that MLR torque plotted against time in a log-log scale is close to a linear line for most polymers as shown in figure 4 during most or some of the relaxation time. This corresponds to the flow region in figure 1 where logarithmic decay is accelerated. For some polymers, such as EPDM, linearity starts at about one second, but for others linearity may start at longer times as is the case for isobutylene based polymers exemplified for IIR in figure 5. A regression line of the MLR data for such a power law model can be written as:

log(m) = a log(t) + log(k)

where M is Mooney torque, a is the regression slope, t is time in seconds and log(k) is the regression intercept. k is therefore the regression intercept in Mooney units at one sec. A, MLR area under the linear Mooney vs. time curve from an initial time to [t.sub.0] a final time [t.sub.f] calculated from equation 3 is:

A = [k/(a+1)][[t.sub.f.sup.(a+1)] - [t.sub.0.sup.(a+1)]]

MLR area has test repeatability advantages over single point MLR parameters (ref. 9) due to data smoothing obtained by the regression fit to equation 3. MLR area for times between one sec. and a final time [t.sub.f] was adopted as the MLR area parameter in ASTM D 1646. However, for those polymers that exhibit a power law relaxation behavior at longer times a regression fit can be started at a higher to and the MLR area for the desired [t.sub.0] to [t.sub.f] time range.

Experimental

Polymers

Isobutylene elastomers used in this study consisted of the following commercially available polymers: polyisobutylene (IM), butyl rubber (IIR), halobutyl (CIIR and BIIR) and isobutylene-co-p-bromomethylstyrene elastomers (BIMS). These polymers were selected to have different molecular weights or Mooney viscosities and some variation in MWD widths as measured by gel permeation chromatography (GPC). The polymers were obtained from a variety of sources.

Mooney and Mooney relaxation tests

Mooney and MLR tests were performed on a Monsanto MV2000E. This instrument was capable of properly braking the rotor, correcting for a proper electronic zero signal during and after rotation, obtaining a short time Mooney torque at 0.6 s after braking the motor and performing a power law regression fit to relaxation data between any assigned starting time [t.sub.0] and final time [t.sub.f]. The tests were performed at either 100 [degrees] C or 125 [degrees] C. Instrument zero was established before starting each test. Final MLR conditions established for isobutylene polymers were: Mooney test at (1+8) minutes before relaxation, 60 s of relaxation and 4 s "hold off" time forcing a regression fit to relaxation data between 4 and 60 s. For these conditions MLR area was:

A = [k/(a+1)] [[60.sup.(a+1)] - [4.sup.(a+1)]]

where `a' is the regression slope (a negative number!) and `k' is the regression intercept in terms of MLR torque in Mooney units at 1 s.

Results and discussion

A variety of isobutylene elastomers was tested by MRI as shown in table 1. Preliminary tests showed that isobutylene polymers relax faster than other elastomers due to the absence of branching. Branching has a stronger effect on slowing down relaxation at longer times, and therefore on polymer elasticity, than does a wide MWD in linear polymers. In an MLR relaxation test there is therefore a tendency to approach a zero torque at much shorter times for isobutylene elastomers than there is for a polymer such as EPDM. Since the zero noise sensitivity of the MLR instrument is approached for some isobutylene elastomers at 60 s after relaxation start it was decided to terminate the tests at 60 s. It was also observed that relaxation data for some isobutylene polymers do not obey a power law model from 1 to 60 s but there is a reasonable fit to a power law from 4 to 60 s of relaxation as shown in figure 5. The regression fit was therefore limited to the time range of 4 to 60 s giving a more accurate model for the longer relaxation time. MRI area was also determined accordingly for the 4 to 60 s range by equation 5. MLR area was preferred over any other single point MLR parameter due to an expected improvement in analysis reliability as found by Friedersdorf et al (ref. 9).

In order to obtain an elastic parameter from MLR it was decided to normalize the data in an attempt to sort polymers in a similar ranking to that of Mw/Mn. A logical index was a ratio between the MLR area of equation 5 to an initial decay parameter since this would mimic a ratio between a higher and a lower molecular weight average. It would have been ideal to choose an initial decay area. However, since the instrument microprocessor was not able to handle two regression fits, it was decided to work with the Mooney torque at 0.6 s after relaxation start, [M.sub.0.6], which is the shortest relaxation torque reliably displayed by the instrument. When ratios of area to [M.sub.0.6] were examined for the various polymers it was obvious that other corrections had to be introduced to smooth out the sought after parameter and force it to follow Mw/Mn. One correction involved the normalization of MLR area by Mooney viscosity. Since MLR area is expected to be a higher power function of Mw as discussed above the normalization had to be done with some power of the Mooney viscosity larger than 1.0. This power and a second constant had to be found such that the MLR measurements at the two temperatures used, 100 [degrees] C and 125 [degrees] C, will produce interchangeable results and such that the proposed index will produce numbers which are similar in magnitude to Mw/Mn. After empirically attempting a variety of functions the following index, named the Mooney relaxation index (MRI), was proposed

MRI = (A/[M.sub.0.6])(40/ML)(1.3)

MRI results are listed in table 1 demonstrating that they generally follow Mw/Mn values for all isobutylene elastomers. MRI data were not yet generated for other polymer families but MRI is not expected to follow Mw/Mn for branched polymers as was explained above.

Some repeatability data of MRI for isobutylene polymers are shown in table 2. It is felt that the standard deviation may improve significantly if an area parameter at low relaxation time can replace the single point measurement of [M.sub.0.6] since the standard deviation for [M.sub.0.6] is relatively much poorer than it is for the MLR area, A.

Table 1 -- MRI values for various isobutylene elastomers

Polymer       Test Temp.      Mooney Vis.   MRI   Mw/Mn
              ([degrees] C)     ML 1+8            (approx.)
IIR               125             52        4.1      3.5
IIR               100             48        4.4      3.5
BIIR              125             33        5.7      4.5
BIIR              125             47        5.4      4.5
CIIR              125             39        4.8      4.5
CIIR              125             50        5.3      4.5
BIMS              125             39        2.7      2.7
BIMS              125             47        3.0      2.7
[IM.sup.*]        180             57        2.2      2.2

(*) approx. viscosity average ME = 900,000

Table 2 -- approximate MLR standard deviations for IIR

Parameter                  Value   Standard    Relative
                                   deviation   STD (%)

ML 1+8 (125 [degrees] C)     52       0.35       0.67
M @ 0.6s                     43       1.5        3.5
MLR area                    25O         5          2
MRI                         4.1      0.15        3.7

Conclusions

A Mooney Relaxation Index, MRI, that resembles the MWD width function Mw/Mn was proposed for isobutylene elastomers. MRI values change between about 2 and 10 with higher values indicating wider MWD and therefore higher elasticity. The precision and repeatability of MRI can be improved when a dual regression capability for relaxation data will be available for the instrument at which case two MLR areas can be defined to indicate both low and high molecular species in the polymer.

References

[1.] S.H. Wasserman and W.W. Graessley, J. Rheol 36, 543 (1992).

[2.] W.J. McGrory and W.H. Tuminello, J. Rheol. 34, 867 (1990).

[3.] W.G. Schramm, Kautch. Gummi Kunstst. 40 (8), 756 (1987).

[4.] R. Koopman and G. Marwede, Kautch. Gummi Kunstst. 45 (2), 133 (1992).

[5.] D.F. Hasman and D. W. Askea, paper No. 65 presented at a meeting of the Rubber Division, American Chemical Society, Detroit, Michigan, October 8-11, 1991.

[6.] J.S. Dick and H.A. Pawlowski, Rubber World, 215, No. 4, 21, (Jan. 1997).

[7.] P.J. DiMauro, J. DeRudder and J.P. Etienne, Rubber World, 201, (No. 4), 25 (1990).

[8.] H. Burhin, W. Spreutels and J. Sezna, paper No. 74 presented at a meeting of the Rubber Division, American Chemical Society, Detroit, Michigan, October 17-20, 1989.

[9.] C.B. Friedersdorf and I. Duvdevani, Rubber World 211 (4), 30 (1995).

[10.] J.L. White, in "Science and Technology of Rubber, " 2nd ed., J.E. Mark, B. Erman and F.R. Eirich, Eds., Academic Press, San Diego, CA, 1994, ch. 6

[11.] R.D. Andrews and A. V. Tobolsky, J. Polym. Sci. 7, 221 (1952).

[12.] G. Berry and T G. Fox, Adv. Polym. Sci. 5, 261 (1968).

[13.] L.J. Fetters, W. W. Graesley and A.D. Kiss, Macromolecules 24, 3136 (1991).

[14.] W.P. Chung and I. Duvdevani, "The application of stress relaxation to quality control of polyisobutylene," Rubber World, 215, No. 1, (October, 1996).

Acknowledgements

"Processability by Mooney relaxation for isobutylene elastomers," is based on a paper given at the October, 1996 meeting of the Rubber Division. "New test method evaluates release agents," is based on a paper given at the May, 1997 meeting of the Rubber Division. "Effective processability measurements of acrylonitrile butadiene rubber," is based on a paper given at the October, 1997 meeting of the Rubber Division.