Printer Friendly

Correlation of results from curemeters of different designs.

Correlation of results from curemeters of different designs

Instruments for measuring the cure reaction of rubber have been used in the rubber industry for more than 50 years. The oscillating disk curemeter was first introduced commercially in 1963, and has undergone an evolutionary change from that time to the present. ASTM D-2084, standard test method for rubber property - vulcanization using oscillating disk curemeter, was adopted in 1971. It has become a standard test for checking the properties of rubber compounds in factory operations. The most common examples of the curemeter specified in ASTM D2084 have been the Monsanto TM-100, and its successor, the R-100 rheometer. More than 5,000 R-100 curemeters are in use worldwide.

In 1987 improvements in curemeter design significantly increased the precision and accuracy of the test, but changed the shape of the cure curve. For some users, a change in the shape of the cure curve is a concern when comparing test results to those from curemeters of older design. This article reviews the evolution of curemeter design, shows the changes in cure curve shapes resulting from those changes, and establishes guidelines for comparing test results from recently introduced curemeters to those from older models.

History of the oscillating curemeter

In 1963, the first commercial oscillating disk curemeters oscillated the disk at a frequency of 3 cycles per minute, typically at [+ or -] 3 degrees of arc. The dies for these early curemeters, called SCD dies, had a 50 mm square cavity 10 mm high, with a biconical rotor centered in the cavity (figure 1). A typical rubber sample of 1.15 specific gravity weighed 22g, and was loaded in two pieces, above and below the rotor. A 20-60 second preheat was required after closure before collecting data. The shear strain on the sample at 3 degrees arc was 21%.

Frequencies of oscillation of 10, 100 and 900 cycles per minute became available over the next five years. These created different curve shapes due to the heat energy added to the cavity in working the rubber, breakdown of polymer structure when curing under dynamic conditions, and the dependence of the rubber flow resistance on shear rate. Figure 2 compares the curemeter curves for the various speeds of oscillation using the SCD dies and an arc of [+ or -] 3 degrees.

When the oscillating disk curemeter was first proposed as an ASTM standard in 1968, a smaller, production-size table model of the curemeter was introduced along with a new die that was 50 mm in diameter (figure 1). This circular die, called the LPC die, had the same height and used the same disk as the SCD dies. The LPC die produced similar minimum and maximum torque values as with the SCD die. The sample could be loaded as one piece on top of the rotor for most stocks. The practice of adding a preheat, as commonly used with the SCD die, was eliminated. The elimination of the preheat time led to slightly faster cures for the LPC die. Another advantage of this die was the flat lower die surface which allowed easier removal of the cured sample for most stocks.

In 1971, a smaller version of the LPC die, called the MPC die, was adopted as the standard for ASTM D-2084. The MPC die (figure 1) used a rotor with a different conical angle. The ASTM standard MPC die produces higher strains than the larger dies at the same arc of oscillation. At 3 degrees of arc, the strain is 48% vs 21% for the larger dies and rotor. Higher strain leads to higher torques, and the smaller specimen (10-12g for a stock with a specific gravity of 1.15) gives a shorter cure time. The result is a cure curve with a significantly different shape, as shown in figure 3.

Higher torques have been shown to cause slippage at the surface of the rotor for many stocks, and that slippage is a potential source of variation in test results. Studies conducted with the LPC die and rotor indicated that slippage due to rotor contamination can occur above 50 dN-m of torque (ref. 1). As a result, the ASTM D-2084 specifies a 1 degree arc of oscillation as standard, with a 16% shear strain.

The curves obtained with MPC dies at 1 degree arc were used as standards until 1987, when further reductions in the temperature recovery time became possible. Another improvement introduced at this time was the reduction of the mechanical compliance of the rotor drive system. This improved reproducibility between instruments. The improved mechanical design raises maximum torques and lowers minimum torques by reducing friction. An example of these improvements in oscillating disk curemeter design is the Monsanto ODR-2000 rheometer.

As part of the continuing evolution in curemeter design, a number of manufacturers have introduced rotorless curemeters. The MDR-2000 rheometer is an example of this type of curemeter. In the MDR-2000, the lower die oscillates and the torque is transmitted through the sample to the upper die, where it is measured by a torque transducer. The MDR-2000 uses a smaller (4 to 5g) sample with a thinner cross section to obtain a more rapid temperature recovery and more uniform temperature throughout the specimen than is possible with the oscillating disk design. The MDR-2000 dies are directly heated and smaller in mass than for the oscillating disk curemeters. This means that faster die temperature recovery can be achieved. MDR-2000 torque levels are lower than torques measured by the oscillating disk curemeters, because the two MDR-2000 dies have less area under load than the two sides of the oscillating disk. Faster temperature recovery leads to faster cure times.

Faster temperature recovery has three advantages: First, the test is closer to the desired cure temperature for more accurate results; second, the potential for variation in results from the cooling of the rotor during loading and unloading is greatly reduced; and third, the cure time is shorter so more tests may be run in a time period. Table 1 illustrates the potential variation that may be caused by changes in sample loading procedures. Curemeters with faster temperature recovery were less sensitive to sample loading in these tests.

Comparison of three commonly used curemeters

As mentioned earlier, the most common curemeter in use today is the R-100 rheometer. Since 1987, many users have replaced R-100 rheometers with either the oscillating disc rheometer or the moving die rheometer. Test results from these three instruments are compared in this article.

Before comparing test results, a brief listing of the differences between the curemeters will describe the variables of interest in this study. The ODR-2000 uses the same die cavity as the R-100 rheometer, but the platen that heats the dies is smaller in the ODR-2000. This permits faster temperature recovery for the ODR-2000. An optional temperature control function in the ODR-2000 slows down the recovery time to simulate the temperature recovery of the R-100.

The MDR-2000 uses a different die configuration than either the R-100 or the ODR-2000. The torque transducer of the MDR-2000 measures torques applied to the die surfaces, with a smaller area than for the oscillating disk rotor surfaces. At the same arc of oscillation, the MDR-2000 will produce lower torque values than the ODR-2000 or R-100 rheometers. With a significantly reduced temperature recovery time, the MDR-2000 produces shorter cure times than either oscillating disk curemeter.

Some users select test conditions other than the 1 degree arc of oscillation specified in ASTM D-2084. For stocks with low torque values, a larger arc of oscillation produces higher torques for increased sensitivity. Torque values above 50 dN-m should be avoided in increasing the arc, however, because slippage may reduce sensitivity. Both the MDR-2000 and the ODR-2000 increase sensitivity by increasing the resolution of the torque signal to 0.01 dN-m versus a resolution of 0.1 dN-m for an R-100 with a computer data system. The possibility of using dynamic property measurements from the MDR-2000 has prompted many users to select 0.5 degrees arc of oscillation. Lower strains are often desirable for dynamic property measurements. While lower strains reduce the magnitude of the torque values, the increased resolution (0.01 dN-m) and additional information from the MDR-2000 can increase sensitivity for most stocks (refs. 2 and 3).

Table 2 lists the percent precision values for the R-100 based on an ASTM round robin test as published in ASTM D-2084-88. Data collected from an instrument user who had 14 R-100 rheometers in several locations showed a similar lack of precision, especially for reproducibility. After an extensive service program, the precision of the 14 rheometers improved to a level comparable to data collected from new R-100 rheometers before shipment. Data collected from new ODR-2000 and MDR-2000 rheometers indicate significantly improved precision versus new R-100 rheometers. The MDR-2000 precision values are higher than for the ODR-2000 because the values in table 2 are expressed as percent of the nominal results. Actual levels of variation for the ODR-2000 and MDR-2000 were the same in these evaluations. Design changes incorporated in the ODR-2000 and MDR-2000 have eliminated most of the service-related problems reported in the 14 R-100 rheometers. While the R-100 may gain better resolution by using computer data systems, the precision of the R-100 indicates no real improvement can be achieved beyond 0.1 dN-m.

The possible instrument options to be compared in a correlation study between the R-100, ODR-2000 and MDR-2000 include:

* R-100 at 1, 3 and 5 degrees arc;

* ODR-2000 at 1, 3 degrees arc, and with normal and slow temperature recovery times;

* MDR-2000 at 0.5 and 1.0 degrees of arc.

Experimental

To compare the three curemeters, stocks with sulfenamide cure systems were collected from three groups, as listed in table 3. Eight typical retread stocks were collected from the Rubber Manufacturer's Association (RMA). The cure temperatures for these stocks were 93 [degrees] C to 149 [degrees] C, following guidelines specified in RMA Shop Bulletin No. 14 (ref. 4). In addition, the same stocks were also tested at 177 [degrees] C, to duplicate a possible batch QA test condition. Seventeen stocks were mixed using various combinations of accelarators, sulfur and prevulcanization inhibitor to create a variety of cure curve shapes. Seven additional stocks were mixed with levels of carbon black varied from 0 to 80 phr. A total of 32 different compounds were tested in the three curemeters, with nine instrument setup configurations. Tests on each compound were run at 160 [degrees] C and 180 [degrees] C. A total of 532 different tests were run for this study.

The possible number of cure points that might come from a cure curve is infinite. To simplify the curemeter comparisons, the data points collected were: ML, minimum torque, dN-m; MH, maximum torque, dN-m; TS2, time to 2 dN-m rise in torque from the minimum, minutes (TS1 was used for the MDR-2000 at 0.5 degrees arc); TC50, time for 50% of the rise in torque from ML to MH, minutes; and TC90, time to 90% of the rise from ML to MH, minutes.

Results and discussion

The results from tests run in the three instruments were correlated by three types of regression analyses. A simple factor type equation (linear with zero intercept), a linear equation with non-zero intercept and a quadratic equation were all tried as models of correlation. The equations selected as the best correlation models for the stocks tested from the ODR-2000 and MDR-2000 to the R-100 are listed in table 4. The coefficients for the equations, the correlation coefficients and the standard error calculations are listed for each data point in this table.

The correlation coefficients were nearly all above 0.95, indicating a good fit to the data for the equations selected. Points that deviated from the model equations appeared to be isolated test results that may not be repeatable. The agreement in the equations among the three groups of stocks tested was very close, and the data were combined for the regression equations reported. The exceptions to the good correlations were the minimum and maximum torque values from the R-100 rheometer at 3 and 5 degrees of arc versus the MDR-2000 at 0.5 and 1 degree arcs.

The standard error values in table 4 are a measure of the deviation of individual test points from the predicted correlations. In most cases, the standard errors were of the same order of magnitude for repeatability as for tests from the R-100 rheometer, indicated in table 2.

The type of equation selected for table 4 depended on the differences between the instruments. In each case, the simplest equation that fit known differences between the instruments was chosen. The major differences in torque measurement between the ODR-2000 and the R-100 include reduced friction and less flexure in the disk drive for the ODR-2000. Reduced friction means that for a zero torque stock, the R-100 would have a higher torque value. For stocks with high torques, the ODR-2000 would have the higher values. The MDR-2000 has less friction than the ODR-2000, and produces less torque than either the R-100 or ODR-2000. Figure 4 shows a linear correlation equation for ML between the R-100 at 1 degree arc, the ODR-2000 at 1 degree arc (normal temperature recovery), and the MDR-2000 at 0.5 degree arc for this study. Lower minimum torques were measured with both of the new curemeters than for the R-100, but the equations in table 4 indicate a greater sensitivity to torque for the ODR-2000. The MDR-2000 has lower torques, but by eliminating more friction, may have increased sensitivity over the ODR-2000.

Figure 5 shows the same comparisons for maximum torque (MH) values. The ODR-2000 measured higher maximum torques, and the MDR-2000 measured lower torque values vs the R-100. A quadratic correlation equation is shown for MDR-2000 MH versus R-100 MH values. This is apparently related to more slippage occurring with the R-100 disk than for the MDR-2000 dies. Slippage should be less in the MDR-2000 due to lower torque values and the larger grooves than in the R-100 disk. Since the ODR-2000 and R-100 use the same disk design, their best correlation for maximum torque was a linear equation.

For cure time correlations between the ODR-2000 and the R-100, an excellent fit was obtained by using a simple factor, with a zero intercept. The thermal recovery properties due to the presence of the disk in the sample is common to both instruments. Figure 6 shows reduced TS2 scorch times for a normal temperature recovery ODR-2000 versus the R-100. Test results with a slow temperature recovery mode ODR-2000 were closer to the R-100 values.

The MDR-2000 has significantly faster heating of the stock at the beginning of the test than either the R-100 or ODR-2000. Therefore, a linear equation with an intercept on the R-100 axis had the best correlation for cure times between the MDR-2000 and the R-100. Figure 7 indicates reduced TS1 times for an MDR-2000 at 0.5 degrees arc versus TS2 results from a 1 degree arc R-100.

Figure 8 shows faster TC90 cure times for the ODR-2000 versus the R-100, and how much closer the ODR-2000 comes to R-100 values when the ODR-2000 is operated in the slow recovery mode. Figure 9 shows reduced TC90 cure times from the MDR-2000 versus the R-100 rheometer. The reduction in cure times for the MDR-2000 was greater for high temperature, fast cures than for low temperature, slow cure tests.

The non-linearity of torque correlations between the MDR-2000 and the R-100 increased when the R-100 was run at 3 or 5 degrees of arc. Above approximately 50 dN-m of torque, nonlinear variation occurred. Earlier studies (ref. 1) with the R-100 indicated slippage may be expected above this level due to contamination of the rotor surface. Tests at high arcs of oscillation should only be used for soft compounds that do not produce torques above 50 dN-m. The MDR-2000, with larger grooves than in the ODR-2000 rotor, and lower torque values, appears to slip less. Slippage may contribute to scatter in test data. The change in the coefficients of the equations in table 4 with curemeter configuration reflect differences in the conditions for each test. For example, a change of arc from 1 degree to 3 degrees increases the applied strain in the oscillating disk rheometers from 15% to 45%. Because the stress-strain response of rubber is non-linear, the 3X increase in strain approximately doubles the maximum torque, and gives a 50% increase in the minimum torque. The differences in thermal response of the rheometers further complicates correlations between rheometers for significant compound differences. In other words, the sensitivity to a major compound variation differs for each rheometer configuration. For example, compound B-13 had 1.25 phr of accelerator, and compound B-14 had 1.75 phr. Otherwise, both compounds were nearly identical. Table 5 compares the data points measured by each rheometer configuration for these three stocks. The magnitude of the differences in test results depended on the rheometer configuration selected. Sensitivity is measured in table 5 by dividing the spread in data by the resolution of each instrument. By selecting the rheometer configuration best suited for a compound, the sensitivity of the instrument to key compounding variations may be optimized.

To compare test results from one location to another, adopting standard rheometer configurations is useful. That is the reasoning behind selecting a recommended configuration for the R-100 (as listed in ASTM D-2084). Similar suggested standard configurations for the ODR-2000 include 1 degree arc, and the normal (rapid) temperature recovery time. For the MDR-2000, a suggested standard configuration uses an arc of 0.5 degrees.

The model equations in table 4 are tested in table 6. An SBR and a chloroprene compound fit the model, even though the chloroprene compound has no carbon black filler. A fluoroelastomer compound and an NBR compound did not fit the model as well, possibly due to higher torque levels.

Conclusions

For sulfenamide cure system stocks, correlation equations are useful to estimate test results from one type of curemeter to another. The equations in this report should serve only as a general guide. They are not a substitute for collecting actual test results on the compounds in question. A change in compound may lead to a fundamental change in correlations between curemeter results.

Curemeter users who wish to take advantage of the continuing improvements in instrument performance may use correlations of the type shown in this article to estimate changes in test results from one instrument to another.

Table 1 - operator effects on curemeter tests
 Cure times, minutes
Sample loading variable TS2 TC50 TC50


Slow temp. recovery ODR (R-100):
Optimum loading time (20 sec) 2.01 3.51 5.86
2 minutes loading time 2.33 3.90 6.33
% change with loading time +15.9 +11.1 +8.0


Rapid temp. recovery ODR (ODR-2000):
Optimum loading time (20 sec) 1.71 3.06 5.40
2 minutes loading time 1.86 3.23 5.60
% change with loading time +8.8 +5.6 +3.7


Rapid temp. recovery rotorless rheometer (MDR-2000):
Optimum loading time (20 sec) 1.21 2.22 3.53
2 minutes loading time 1.21 2.23 3.52
% change with loading time 0.0 +0.5 -0.5


Table 2 - precision of rheometer tests using ASTM D4483

Instrument reproducibility (r), % Repeatability (R), % R-100 ASTM D2084 precision
 ML 5.4 45.8
 MH 1.5 16.6
 TC50 6.7 15.0


R-100 14 units in use, before servicing
 ML 32.3
 MH 16.7
 TC50 24.2


R-100 14 units above, after servicing
 ML 6.7
 MH 8.3
 TC50 10.0


R-100 production run, 100 new units
 ML 5.4 10.9
 MH 1.3 4.2
 TC50 5.3 10.0


ODR-2000 production run, 50 new units
 ML 2.0 5.7
 MH 1.0 1.4
 TC50 1.7 3.6


MDR-2000 production run, 50 new units
 ML 3.2 6.6
 MH 1.5 2.8
 TC50 2.6 5.9


Table 5 - example compound variation
Stock no B-6 B-1 B-14
Phr TBBS 0.75 1.25 1.75 Sensitivity
160 [degrees] C curemeter tests: change/
 resolution


ML, minimum torque, dN-m
 R-100, 1 deg arc 6.2 6.2 6.0 -2
 ODR-2000, 1 deg arc 5.11 5.60 5.10 -1
 MDR-2000, 0.5 deg arc 1.64 1.54 1.48 -16
 MDR-2000, 1 deg arc 2.51 2.47 2.37 -14
 ODR-2000, 3 deg arc 8.81 9.03 8.88 +7
 R-100, 5 deg arc 12.2 12.4 12.0 -2


MH, maximum torque, dN-m
 R-100, 1 deg arc 28.6 32.0 34.5 +59
 ODR-2000, 1 deg arc 29.52 34.57 37.51 +799
 MDR-2000, 0.5 deg arc 11.64 13.38 14.86 +322
 MDR-2000, 1 deg arc 19.17 22.91 25.54 +637
 ODR-2000, 3 deg arc 55.46 64.48 70.73 +1527
 R-100, 5 deg arc 82.7 82.3 90.8 +810


Scorch time, minutes
 R-100, 1 deg arc 6.90 5.42 4.83 -124
 ODR-2000, 1 deg arc 6.21 4.84 4.29 -192
 MDR-2000, 0.5 deg arc 5.08 3.91 3.50 -158
 MDR-2000, 1 deg arc 5.89 4.34 3.75 -214
 ODR-2000, 3 deg arc 4.88 3.89 3.43 -145
 R-100, 5 deg arc 4.27 4.28 3.72 -33


TC50 cure time, minutes
 R-100, 1 deg arc 11.08 8.17 6.95 -248
 ODR-2000, 1 deg arc 10.76 7.66 6.61 -415
 MDR-2000, 0.5 deg arc 9.80 6.92 5.86 -394
 MDR-2000, 1 deg arc 9.52 6.77 5.66 -386
 ODR-2000, 3 deg arc 10.19 7.10 5.96 -423
 R-100, 5 deg arc 7.67 7.67 6.43 -74


TC90 cure time, minutes
 R-100, 1 deg arc 17.57 12.10 9.87 -462
 ODR-2000, 1 deg arc 16.96 11.23 9.35 -761
 MDR-2000, 0.5 deg arc 16.60 11.11 9.23 -737
 MDR-2000, 1 deg arc 15.58 10.56 8.62 -386
 ODR-2000, 3 deg arc 15.48 10.15 8.26 -722
 R-100, 5 deg arc 11.15 11.10 9.03 -127


Table 6 - regression model tests

Compound: FKM CR NBR SBR

ML, minimum torque, dN-m

R-100 6.1 4.1 9.2 6.3 Predicted from

MDR-2000, 1 deg arc: 5.49 4.15 9.98 6.27 Predicted from

ODR-2000, 1 deg arc: 6.54 4.30 9.37 6.31

MH, minimum torque, dN-m

R-100 51.0 21.0 47.9 32.5 Predicted from

MDR-2000, 1 deg arc: 54.83 21.81 58.36 32.56 Predicted from

ODR-2000, 1 deg arc: 54.09 22.30 48.37 32.76

TS2, scorch time, minutes

R-100 1.93 3.73 1.36 6.58 Predicted from

MDR-2000, 1 deg arc: 1.96 3.65 1.67 6.75 Predicted from

ODR-2000, 1 deg arc: 2.20 3.58 1.26 6.31

TC50, cure time, minutes

R-100 2.71 7.15 2.23 10.25 Predicted from

MDR-2000, 1 deg arc: 2.18 7.22 2.38 10.60 Predicted from

ODR-2000, 1 deg arc: 2.80 7.01 2.10 9.85

TC90, cure time, minutes

R-100 2.93 18.69 3.66 16.34 Predicted from

MDR-2000, 1 deg arc: 2.73 19.68 3.30 17.40 Predicted from

ODR-2000, 1 deg arc: 2.80 18.90 3.44 15.54 [Tabular Data 3 and 4 Omitted] [Figures 1 to 9 Omitted]

References

[1]R.W. Wise and W.R. Deason, "Oscillating disk curemeters part I die-rotor design and optimization of operating conditions," Monsanto Co., December 8, 1969. [2]P.J. DiMauro, J. DeRudder and J. P. Etienne, "New rheometer and Mooney technology," Rubber World, January, 1990. [3]D. Lederer, J. Sezna, H. Pawlowski and C. Sholley, "Determination of minimum detectable levels using curemeters," Rubber Division, ACS, Poster Session, Spring Meeting, 5/90. [4]U.S. Rubber Manufacturer's Association Publication, "Shop Bulletin No. 14."
COPYRIGHT 1992 Lippincott & Peto, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1992, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Author:Sezna, John A.
Publication:Rubber World
Date:Jan 1, 1992
Words:4023
Previous Article:Quality management in the latex laboratory.
Next Article:Valley Rubber Mixing.
Topics:


Related Articles
Polyurethane applications for the vibrating needle curemeter.
Evaluation of TBSI.
Viscoelastic characterization of rubber with a new dynamic mechanical tester.
Rubber testing for injection molding.
Curing time software in theory and practice.
Simulation of the injection molding process with a special capillary rheometer.
Innovations in curemeter and Mooney viscometer technology.
Storage stability of FKM compound based on a bisphenol AF/onium cure system and its potential as a standard reference compound.
Optimization of the production of EPDM sponge rubber seals for automotive use.
Ten ways to improve test productivity and reduce testing costs.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters