A simplified approach to QC and testing.

Sensitivity to compound variance and speed of testing have always been crucial to the rubber manufacturer. As early as 1967, Jones, of the Gates Rubber Company, indicated that testing should keep up with the time required to mix a batch (ref. 1). He then decided that it would require a 3-1/2 min-ute test, with minimal set up time, and that the test needed to have an instrument for each internal mixing line. Traditional means of testing would not have achieved this. Gates Rubber was one of the first companies to introduce the oscillating disk rheometer (ODR) into its QC procedures.

Prior to the ODR, two time-consuming steps were used for testing, neither of which can fully characterize batch-to- batch consistency. A Mooney viscometer was used for the first step to characterize how a rubber compound would process or take shape up to the point of scorch (ref. 2). At processing temperatures between 75[degrees] and 125[degrees]C, the Mooney viscometer used a rotor embedded in a rubber specimen surrounded by heated platens. With a one-minute preheat, a Mooney viscosity test generally ran either five minutes or nine minutes, depending on the polymer. Though the viscosity and scorch time can be easily plotted using this method, the system does not allow for accurate results after scorch. The rubber, after scorch, will slip through the dies, since the rotor rotates. Accurate measurements after scorch are, therefore, not possible.

The second step, measuring how a compound would cure, was a much more laborious process, relying on tensile modulus of a cured compound. The time required to cure a compound was determined experimentally by curing multiple tensile slabs of a compound for various times. Stress strain properties were measured at each of the different cure times, and the modulus values were plotted against time. This generated a curve which indicated cure time. These tests were useful in the development of a compound, but because of the time required to perform the test, it was not practical for quality control.

The Gates study in 1967 used some of the very early oscillating disk rheometers to provide a quality control test for each batch. The results were successful, and provided 20% to 50% more sensitivity than traditional tests (ref. 3). In 1972, a study by Dunn of Polysar, and Bennet of Goodyear, further elaborated on the advantages of rheometer testing over stress strain properties of vulcanizates. They found that the oscillating disk rheometer "provided more useful information at a higher level of reliability and in a shorter time" than traditional testing (ref. 4). Another validation of rhe-ometer testing was conducted in 1977, when Cottrell of Dow Coming did a similar evaluation on silicone rubber, comparing it with plastometer, durometer and tensile modulus. The study showed that the ODR tests were "at least as effective as traditional tests in detecting compounding errors and material variance. Furthermore, these greatly reduced testing time and increased the precision of testing (ref. 5)."

Though the industry had clearly progressed, allowing for cost savings and higher quality, there was a continuing need for, as Sezna and Dick explained in 1991, instrument results that were "reproducible" "sensitive to the process of concern" and "capable of measuring the variation of the process (ref. 6)." The introduction of the moving die rheometer (MDR) accomplished these results, while decreasing the test time. The MDR also added the capability of examining properties beyond the traditional parameters that were recorded with the ODR. Measuring loss torque and tangent delta, along with the standard elastic torque, provided a better understanding of viscoelastic properties. The MDR made testing faster and more accurate by improving on existing technologies.

The progression towards a more discriminating test and one that relates to the process continued with the introduction of the Monsanto RPA 2000 in the early 1990s (ref. 7). The RPA provided programmable test sequences to allow combinations of different strains, frequencies and temperatures to be used in a single test to better define the material. RPA type instruments may be used to test according to ASTM methods D5289, D6204 and D6601 (refs. 8-10). However, as more information was gathered in a single test, the test time often became longer. Batch control had become more precise and more sophisticated. An industry accepted fast test was not defined as clearly as the RPA's rheological predecessors, the ODR and the MDR.

This article will focus on using a combination of varying frequency, strain and temperature to provide more detailed processability information, as well as standard curing characteristics of the rubber compound, while completing testing in a time that is still acceptable for quality control.

Experimental

Overview

The experiment was divided into two parts. The first was to establish the effect of temperature, strain and frequency on the cure curve as determined by a specially designed moving die rheometer. The second part of the experiment was to apply these principles to establish a new production control test that is both fast and able to discriminate flow and scorch differences.

Test equipment

All tests were made using the rheoTECH M[D.sub.PT] manufactured by Tech Pro. The M[D.sub.PT] utilizes a sealed die system as described in ASTM D5289. In addition to being capable of testing at standard conditions of 1.67 Hz (100 cpm) and 7% strain (0.5[degrees] arc), the M[D.sub.PT] may be programmed to run at various frequencies, strains and temperatures in a single test. The M[D.sub.PT] meets the requirements of ASTM D5289, D6204 and D6601, and is user programmable for specialized testing.

The oscillatory strain in the M[D.sub.PT] is produced by using a variable eccentric drive. The variable drive resembles standard ODR and MDR fixed eccentric drives. The variable eccentric, however, is not fixed as in previous instruments, but can be programmed in the test configuration and automatically changed during a test. The variable eccentric is driven by a brushless, DC servo motor which uses a high frequency PID function to precisely control the speed of rotation. Varying frequencies are obtained by specifying the frequency in the test configuration, which in turn automatically sets the motor speed. The temperature may also be programmed in the test configuration.

The software has been configured with a special parameter to continuously look at the torque curve and to determine the point of scorch in real time. When scorch occurs, either based on an increase in slope of the torque curve or on a predetermined increase in torque, the test conditions change to new settings based on frequency, strain and temperature (FST). For example, a cure test may begin at a low temperature and high strain up to the point of scorch, ts1, and then change to a lower strain and higher temperature to finish the cure cycle.

Compound

The test formula, compound TPI, for the first part of the study was mixed in a laboratory internal mixer, Size B. The nominal capacity of the mixer is 1.5 liters. Rotor speed was set at 77 rpm for both the masterbatch and final passes. The formula is listed in table 1, and the mixing procedures in table 2.

For the second part of the study, 60 batches from two production runs of an SBR/natural rubber compound, PD 1, were obtained and used for evaluation.

Test conditions

To illustrate the use of the FST parameter, test conditions were set to vary the frequency, strain and temperature of a test before and after scorch (based on an increase in torque). These test conditions are shown in table 3.

The test conditions for part 2 of the study are shown in table 4.

Results and discussion

The first part of the study was to demonstrate the use of the FST parameter. Since the final objective was to establish new test conditions to better define flow and scorch characteristics, a series of tests was made varying strain and frequency. The results of testing at three different strains are shown in figure 1, and the numerical data in table 5. It is visually obvious that higher strains result in higher torque measurements in the processability region of the curves.

[FIGURE 1 OMITTED]

The observation regarding scorch time, however, may not be as clear. To define scorch, it can be seen that the use of a constant rise in torque may be misleading as to the actual time of scorch. According to the scorch time at any given increase above minimum, it would appear that the compound scorches faster at higher strains. However, the percent of cure for any scorch time determined by a rise in torque is represented by:

(1) where: % cure = 100[S.sub.x]/(MH-ML) [S.sub.x] = increase in torque above ML; MH = maximum torque; and ML = minimum torque.

As seen in table 6, a 1 dNm increase in torque at 7% strain represents a significantly higher percent cure than a 1 dNm increase at 50%. An increase of 0.25 dNm at 7% strain results in approximately the same scorch time as a 1.0 dNm increase at 25% and a 1.5 dNm increase at 50%. This is because they are representative of approximately the same percent cure. For evaluation of the FST parameter, it is necessary to differentiate the torque rise at different strains to define scorch.

The results of varying frequency only are shown in figure 2. Here it is seen that the minimum torque increases with increased frequency, but the scorch time is not significantly affected. The same torque increase to indicate scorch is approximately the same for all frequencies.

[FIGURE 2 OMITTED]

The objective of looking at conditions of varying strain and frequency was to determine a new set of conditions to use in the FST parameter. No attempt is made here to optimize these conditions, but instead to select one set of conditions where information in the processing area of the cure curve is amplified. Therefore, the middle strain, 23%, and middle frequency, 10 Hz, were selected to use with the FST parameter.

The FST parameter was set up to test the compound at 10 Hz and 25% strain up to the point of scorch as determined by a 1 dNm increase in torque. This test at the constant temperature of 165[degrees]C is shown in figure 3, along with the control. The FST scorch time, ts1, was 1.04 minute and compared well to the control's scorch time, ts0.25, of 0.96 minute. The minimum torque, however, was amplified approximately four times.

[FIGURE 3 OMITTED]

The final step in selection of the FST parameter was to determine the temperatures necessary to lengthen the scorch measurement while still maintaining a test of approximately the same length of time as the control. The effect of temperature is given in table 7. The object was to at least double the scorch time as compared to the control. To achieve this, it was necessary to drop the temperature at the beginning of the FST test by 10[degrees]C. It was also desired to try to test in nearly the same length of time as the control. To determine the second temperature, the time to 100% cure was examined. From table 7, it can be seen that a 10[degrees]C increase in temperature resulted in only a slight increase in cure time. The time required to cool the dies for the next test adds about one minute to the test turnaround time.

To test the repeatability of the processability portion of these conditions, 25 repeat tests were made according to ASTM D5289 at 175[degrees]C, 7% strain and 1.67 Hz (STD) and compared to 25 repeats of the FST parameter using 155[degrees]C, 25% strain and 10 Hz. Since the magnitude of these measurements differs under these conditions, the repeatability was expressed in terms of the coefficient of variation. These results are shown in table 8. The data were normalized so that minimum torque and scorch values from both the FST and isothermal (standard) tests were expressed as a percentage of the average and plotted in figure 4. The bars on these graphs equal +/- three normalized standard deviations from the mean. The smaller the deviation, the better the repeatability and more precise the measurement (ref. 11). It can be seen that for minimum torque and scorch time, the FST is nearly twice as repeatable. This is significant when looking for differences in these properties. Basically, the more repeatable a test is, the better that it is at discriminating differences. To say this in another way, FST scorch values that lie outside +/- 3 CV from a given mean are significantly different than the mean value.

[FIGURE 4 OMITTED]

With the experimental background obtained in part one, the same principles were used to set up test conditions for two productions runs of the material PD1 (an SBR/NR compound). Here, the normal quality control test was an oscillating disk rheometer run six minutes at 175[degrees]C.

Moving from the oscillating disk to moving die test reduced the test time from six minutes to four minutes. The goal was to select FST conditions to complete a test in four minutes or less, so that the test turnaround time could be minimized. As in the first part of the experiment, the frequency and strain were set at 10 Hz and 25%, respectively. In order to obtain a test that was less than four minutes, the temperature for the processing section of the curve was set to 155[degrees]C. At 1 dNm rise above minimum, the temperature was raised to 185[degrees]C, frequency reduced to 1.67 Hz and strain was reduced to 7%. The comparison of the standard isothermal moving die (MD) and FST test is shown in figure 5.

[FIGURE 5 OMITTED]

The 60 batches were run consecutively on the M[D.sub.PT using the FST parameter and separately run on an M[D.sub.PT] at constant conditions of 175[degrees]C, 7% strain, 1.67 Hz. The cure curves are shown in figure 6. The run charts for the data from these curves are shown in figure 7. The limits for the run charts are based on the coefficient of variation determined from the results of part 1 of this study.

[FIGURES 6-7 OMITTED]

Examination of the minimum torque values for the FST run chart shows a general downward trend that is also seen in the standard isothermal chart. When compared to the limits, however, it can be seen that the FST flags 14 batches as being significantly different than the mean. The standard test flags only four batches. It should be noted that just the appearance of a significant difference does not mean that the batches will be problematic in further processing. It may, however, provide a better tool for controlling the process through mixing, and perhaps give useful information about the downstream processability of the compound.

Examination of the scorch time values for FST versus standard tests shows even more significant differences. A data shill is seen in the FST test from the first run to the second. This data shift is not seen in the standard isothermal run chart. The second observation is the significance of the reading. Based on the coefficient of variation (CV) data for the standard isothermal tests, it can be seen that values that lie within a 0.038 minute range are considered to be equal. Accordingly, 54 of the 60 isothermal scorch measurements appear to be equal. The CV of the FST tests indicates that data that lie within a 0.048 minute range are equal. The centerline placement of these limits definitely shows the significant differences between the first and second run, with smaller differences seen within each run.

Run charts for cure time and maximum torque are also shown in figure 7. Here it is seen that there is no significant difference in the capabilities of the FST test compared to the standard isothermal test for cure time. The maximum torque values are similar, but the standard tests here show slightly better discrimination. This may be due to the nearly 25% higher torque at maximum when the standard test is used. This higher torque is, at least in part, a result of lower temperature used for the isothermal test. It may also be a function that the higher strains for the FST test up to the point of scorch may have an effect on maximum torque.

The above analysis closely follows the process that has been used in interpreting rheometer curves for the past 30 years. This makes the transition from standard curemeter testing to testing using the FST parameter seamless. New instruments such as the M[D.sub.PT] also are capable of providing viscoelastic properties of the compound before, during and after cure. Further study is needed to investigate the applications of these additional parameters.

Summary

A new test was specifically designed to give a better measurement of flow of the uncured material and a better measurement of scorch time. The basic consideration was to test the material starting at a temperature lower than would be typically used for a standard moving die rheometer quality control test, and to raise the temperature at a time determined by scorch. This would effectively lengthen the scorch time. In addition, while at this lower temperature, the test conditions of strain and frequency were raised to provide an amplified view of the viscosity. After scorch, the temperature was raised to finish curing, and the frequency and strain returned to standard ASTM D5289 conditions.

Repeat tests were made on a control material, comparing the new FST test to the standard rheometer test. The FST test was approximately twice as discriminating for measurement of viscosity and scorch as compared to the standard test. After the occurrence of scorch, the FST parameter automatically reduced the strain and frequency and raised the temperature to complete the test. The repeatability of the cure time and maximum torque was approximately the same for both the FST and standard test. The FST test was one minute longer than the standard test, due the necessity of reducing temperature between tests.

For the purposes of decision making, more precise and discriminating information is often available only at an increase in cost or increase in time. The rheoTECH M[D.sub.PT] with FST control is a new instrument that can provide this information in a timely and cost effective manner.

Table 1--formula for compound TPI

Ingredient           Phr

Masterbatch
SS MR 5             100.0
N330 black           45.0
6 PPD                 1.5
TMQ                   1.5
Stearic acid          2.0
Zinc oxide            3.0
Masterbatch total   153.0

Final
Masterbatch         153.0
Sulfur                1.5
MBS                   1.5
TMTM                  0.2

Final total         156.2

Table--2 mixing procedure

Mixing event                        Add at:
Masterbatkch
Rubber                               0 min.
Filler and misc. ingredient          1 min.
Sweep at:                          5.5 min.
Drop batch at:                       9 min.

Final
MB                                   0 min.
Cue system                           1 min.
Drop batch at:                125[degrees]C

Table--3 part 1 testing conditions

                   Before scorch

Test        Temp.   Freq.   Strain    FST
       [degrees]C      Hz        %    tsx

TP01          165    1.67        7     na
TP02          165    1.67       25     na
TP03          165    1.67       50     na
TP04          165       3        7     na
TP05          165       5        7     na
TP06          165      10        7     na
TP07          165      20        7     na
TP08          160      10       25   1.00
TP09          155      10       25   1.00
TP10          155      10       25   1.00
TP11          155      10       25   1.00

               After scorch

Test        Temp.   Freq.   Strain
       [degrees]C      Hz        %

TP01          165    1.67        7
TP02          165    1.67       25
TP03          165    1.67       50
TP04          165       3        7
TP05          165       5        7
TP06          165      10        7
TP07          165      20        7
TP08          165    1.67        7
TP09          165    1.67        7
TP10          170    1.67        7
TP11          175    1.67        7

Table 4--part 2 testing conditions

                 Before scorch

Test     Temp.      Freq.   Strain   FST
       [degrees]C    Hz       %      tsx

TP12      175       1.67      7        na
TP13      155        10       25     1.00

              After scorch

Test     Temp.      Freq.   Strain
       [degrees]C    Hz       %

TP12      165       1.67      7
TP13      185       1.67      7

Table 5--effect of strain on the cure of characteristics

              MH,    ML,    t'c10,   t'c90,   ts0.25,
              dNm    dNm     min.     min.     min.

7% strain    13.01   1.46    0.95     2.04     0.96
25% strain   38.76   3.55    0.99     2.01     0.87
50% strain   61.71   5.35    0.99     1.88     0.83

             ts0.5,   ts1,   ts1.5,   ts2,
              min.    min.    min.    min.

7% strain     1.06    1.19    1.24    1.30
25% strain    0.95    1.04    1.09    1.14
50% strain    0.90    0.98    1.02    1.06

Table 6--comparison of scorch rise to percent cure

             % cure @   % cure @   % cure @   % cure @   % cure @
             0.25 dNm   0.5 dNm    1.0 dNm    1.5 dNm     2 dNm

7% strain      2.16       4.33       8.66      12.99      17.32
25% strain     0.71       1.42       2.84       4.26       5.68
50% strain     0.44       0.89       1.77       2.66       3.55

Table 7--variation of temperature for FST test

                          MH      ML    Scorch   tc100
                          dNm    dNm     min.    min.

Control, 165[degrees]C   13.01   1.46    0.96    3.27
160/165 *                11.7    6.15    1.39    3.41
155/165 *                11.6    6.34    1.83     4.0
155/170 *                11.7    6.34    1.82     3.5
155/175 *                11.2    6.35    1.82     3.4

* FST beginning/ending temperatures

Table 8--repeatability of standard test vs. FST test

                                ML       Scorch time

                           Std    FST    Std    FST
                           dNM    dNM    dNM    dNM

Mean                       1.38   6.22   0.69   1.79
Standard deviation         0.02   0.05   0.01   0.01
Coefficient of variation   1.70   0.79   1.22   0.59

                                 MH         Cure time

                            Std     FST    Std    FST
                            dNM     dNM    dNM    dNM

Mean                       12.56   12.57   1.30   2.69
Standard deviation          0.06    0.08   0.01   0.02
Coefficient of variation    0.45    0.65   1.60   0.73

Note: Scorch - ts0.25 used for STD tests, ts1.0 used for FST tests

Note: Cure Time - t90 used for STD and FST tests

References

(1.) R.W. Jones, "Batch control testing rheometers," Rubber Age, Sept. 1968, p. 53.

(2.) G.E. Decker, R. W. Wise and D. Guerry, "An oscillating disk rheometer for measuring dynamic properties during vulcanization," Rubber Chemistry and Technology, v. 36-12, p. 451 (1963).

(3.) R.W. Jones, "Batch control testing rheometers," Rubber Age, Sept. 1968, p. 53.

(4.) J.R. Dunn and H.J. Bennett, "Oscillating disk curemeters," Rubber Age, October, 1972.

(5.) K. Cottrell, "Rheological measurements as a quality control tool for silicone rubber formulations," paper #48, presented at the 112th Meeting of the Rubber Div., ACS (1977).

(6.) J.A. Sezna and J.S. Dick, "The use of rheometers for process control," paper #44 presented at the October, 1991 meeting of the Rubber Division of the ACS.

(7.) J.S. Dick and H.A Pawlowski, "Applications of the rubber process analyzer in predicting processability and cured dynamic properties of rubber compounds," paper #2 presented at the May, 1993 meeting of the Rubber Division.

(8.) ASTM D5289-95(2001), "Standard test method for rubber property--vulcanization using rotorless cure meters," ASTM International, West Conshohocken, PA, Vol. 9.01 (2002).

(9.) ASTM D6204-01, "Standard test method for rubber--measurement of unvulcanized rheological properties using rotorless shear rheometers," ASTM International, West Conshohocken, PA. Vol. 9.01 (2002).

(10.) ASTM D6601-00, "Standard test method for rubber properties--measurement of cure and after-cure dynamic properties using a rotorless shear rheometer," ASTM International West Conshohocken, PA., Vol 9.01 (2002).

(11.) W. Mendenhall, D. Wackerly, R. Scheaffer, Mathematical Statistics with Applications, PWS-KENT Publishing Company, MA, 1990, p. 331.