Measuring the effects of filler variations on compound viscoelastic properties as measured by the capillary rheometer and RPA.
Die swell (also called extrudate swell or the Barus Effect) is a phenomenon that occurs when the extruded rubber compound leaves the extruder die with a greater dimensional diameter than the diameter of the extrader die head orifice itself. Die swell is a viscoelastic effect. This effect is demonstrated in figure 1.
This die swell phenomenon can be partially explained by entropy changes. Before the rubber compound enters the die, the coiled macromolecules are more random, perhaps more spheroidal with less directionality, moving at a slower speed. These rubber chains are also more entangled. Here, the rubber compound's entropy is relatively high. On being forced to enter the smaller diameter die orifice, the macromolecules increase their velocity and start to uncoil in the direction of flow. Also, these rubber chains start to disentangle with the higher velocity. Because of this forced confinement, the macromolecules are more elongated in the direction of flow, and the normal force (perpendicular to the direction of flow) increases significantly. With this forced confinement, the entropy level is lowered because of the forced directionality of these macromolecules. Then, upon exiting from the die orifice, this normal force is eliminated, the rubber macromolecules recoil and also become re-entangled, the entropy increases again resulting in extrudate swell in the normal direction, and corresponding dimensional contraction in the direction of flow (ref. 5).
[FIGURE 1 OMITTED]
Rubber compounds have elastic memory. Upon exiting confinement in a smaller space, they remember the dimensions that they had earlier before this confinement in the die orifice. If the length to diameter (L/D) of this die orifice is rather low, such as an L/D of 1:1, then a rubber compound's elastic memory is very good and the compound will possess a rather high die swell. On the other hand, if the L/D is rather high, such as 30:1, then the rubber compound will show relatively poor elastic memory and will display only a small amount of die swell. Since this article is focused on comparing compound differences in die swell, only an L/D of 1:1 was used. Rubber compounds have a property known as stress relaxation. So a longer residence time in the die orifice itself can explain this loss in elastic memory (refs. 6 and 7).
Time and temperature also affect die swell (ref. 8). If the die swell measurement is made shortly after the compound extrudes through the die, it is called running die swell. On the other hand, if you allow more time to pass after the extrudate exits the die before measuring, then it is called relaxed die swell. Figure 2 shows the relationship between running die swell and relaxed die swell.
In compounding, different types and amounts of fillers such as carbon black and silica have a definite effect on the rheology of the compound by occluding the rubber with the colloidal particles of filler, encapsulating the available rubber which affects the compounds' die swell and other rheological properties (refs. 9 and 10).
[FIGURE 2 OMITTED]
RPA strain sweeps, strain softening and linear vs. non-linear viscoelasticity
Most rubber compounds used in commerce contain significant quantities of colloidal particle sized carbon blacks, and sometimes precipitated silica as well. By using RPA very low strain sweeps that were developed in 2013 with the new extended dynamic range (EDR) (ref. 11), the effects of the critical strain (the separation point between linear and non-linear viscoelasticity) can be studied. One of the purposes of this study was to determine what effects changes in the applied strain (in relation to the critical strain) have on the accuracy of the Cox Merz model regarding capillary rheometer measurements (viscosity measurements under steady state flow conditions), as well as other processing properties. The results from our study of conformance to the Cox Merz rule were presented last year (ref. 12). It is well known that increasing the loading levels of colloidal particle sized fillers such as carbon black can significantly reduce the critical strain of the rubber compound (ref. 13).
A series of mixed rubber recipes with variations in the filler loading levels of N234, N330 and N660 carbon blacks in addition to variations in non-black fillers which included Hi-Sil 210 precipitated hydrated silica (with and without organosilane coupling agent TESPT), clay and ground calcium carbonate (whiting) are shown in variations of table 1.
Subject filler is N234 carbon black, N330 carbon black, N660 carbon black, Hi-Sil 210 precipitated hydrated silica, hard clay and finely ground limestone (whiting).
Silica was used with and without OFS-6945 organo silane (TESPT). When the organosilane was used with the silica, it was used in the following proportions: Hi Sil 210 (phr)--15,30, 45,60 and 75; and OFS-6945 silane/carbon black (50:50)--2.25, 4.5, 6.75,9 and 11.25 phr, respectively.
With these compounds, comparisons were made between the Alpha Technologies RPA 2000 with extended dynamic range (EDR) and the Alpha Technologies ARC 2020.
The RPA with EDR (extended dynamic range) was used to test these compounds at 100[degrees]C, at frequencies shown in table 2. The new RPA operates at 50 Hz (instead of a maximum of 33 Hz previously). The EDR feature allows the RPA to easily test at very low strains as well. Strain sweeps for measuring the Payne effect were routinely started at 0.07% strain.
Figure 3 shows both the Alpha Technologies RPA 2000 and ARC 2020 capillary rheometer that were used.
For the RPA, multiple frequency sweeps were performed at strains of 0.07, 1.0,7.0, 14, 50 and 300% strain, all at 100[degrees]C, as noted in table 2.
ARC 2020 capillary rheometer
With the introduction of the ARC 2020 capillary rheometer in the last decade, there was a complete redesign with the latest state-of-the-art mechanical design, electronics and software. The ARC 2020 was set up to use a barrel which is 0.5 inch (1.27 cm) in diameter which is designed for use with conventional rubber compounds. This barrel is 9 inches (22.86 cm) in length with a working length of 5 inches (12.7 cm), five times the working distance that the older MPT had. Also, with the much more flexible and fully programmable software, we created test configurations with nine working zones (compared to only four zones for the MPT). With the greater working length of the ARC 2020, one can create test configurations that have greater running distance at very high shear rates (ref. 14).
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
All ARC 2020 testing was performed at 100[degrees]C with a Y6001 die which has a diameter of 1.524 mm, an L/D of 1:1 and an entrance angle of 45[degrees].
Die swell (or extrudate swell) is another important measurement performed with the capillary rheometer. The die swell is directly measured by the shadow created from a laser beam hitting the extmdate. This percent die swell is calculated from Equation 1.
% die swell = ([D.sub.E] - [D.sub.O])/[D.sub.O] * 100 (1)
[D.sub.E] = diameter of extrudate
[D.sub.O] = diameter of the capillary
It is better to use dies with lower L/D values for better accuracy and test sensitivity to differences in die swell. Die swell is related to a compound's uncured elasticity or "nerviness." Figure 4 is a schematic diagram showing how laser light can be used to measure die swell of the extrudate.
Figure 5 shows the actual laser measurement attachment for the ARC 2020 capillary rheometer which allows the direct measurement of die swell. Figure 6 shows the specific die with an L/D of 1:1 that we used in these experiments.
Effect of filler loading on the Payne Effect
In order to quantify the linear viscoelastic region that remains from mixing different fillers at different concentrations, a series of strain sweeps was performed with the RPA 2000 starting at a very low strain of 0.07 % (a feature of the EDR). Figure 7 shows that as the concentration of N234, a fully reinforcing carbon black, is increased from only 15 phr to 75 phr, the critical strain (the dividing line between linear and non-linear viscoelasticity) rapidly decreases. With non-linear viscoelasticity, G' decreases with a rise in applied strain. The G' (elastic modulus) for the linear viscoelasticity region does not decrease with a rise in strain.
[FIGURE 5 OMITTED]
[FIGURE 6 OMITTED]
Hi-Sil 210 silica, shown in figure 8, has a particle size similar to N330; however, it shows a larger linear viscoelastic region than the fully reinforcing carbon blacks N234 and N330.
Using Hi-Sil 210 silica with organosilane coupling agent (TESPT) significantly reduces the hydrogen bonding between particles, causing a significant reduction in G' elastic modulus (figure 9). There is also a reduction in the critical strain with the addition of TESPT.
Comparison of running die swell vs. relaxed die swell
As discussed, running die swell from the capillary rheometer is measured for the extmdate while extruding, as shown in figures 4-6. In between extrusion zones, the piston is programmed to withdraw 1 mm. After 60 seconds, the relaxed die swell is measured as well before going into the next extrusion zone, typically at a logarithmically higher shear rate. It should be noted that the relaxed die swell measurement is made at what particular temperature the extrudate has cooled down to in exactly one minute of relaxation. As noted in figure 2, the relaxed die swell is typically higher than the running die swell.
[FIGURE 7 OMITTED]
[FIGURE 8 OMITTED]
[FIGURE 9 OMITTED]
Figure 10 shows the overall correlation between running die swell and relaxed die swell from the completion of these experiments with over 285 pairs of data.
Clearly, the correlation at [R.sup.2] of 0.72 is not perfect. So the relaxed die swell may very well provide additional information that the running die swell cannot provide.
Effects of differences in filler type, surface area and/or structure on running die swell
Figure 11 shows the resulting running die swell values for all fillers loaded at 75 phr in the model SBR test recipe.
[FIGURE 10 OMITTED]
Because N234 carbon black has the highest surface area per gram, it sets up the tightest filler network at 75 phr loading level, thus producing the lowest running die swell.
N330 is an "old technology" carbon black that is fully reinforcing with the second highest surface area and the second smallest particle size resulting in the second most dense filler network and the second lowest running die swell values of all fillers tested.
N660 carcass grade carbon black, with a larger particle size of about 60 nm, gives the third lowest running die swell.
Just straight precipitated hydrated silica (Hi-Sil 210) without any organosilane (TESPT) gives the fourth lowest die swell because of the formation of hydrogen bonding to form a filler network. However, the silanized precipitated hydrated silica (Hi-Sil 210 with TESPT) has no hydrogen bonding because of the silanization of the surface of the silica particles resulting in very high die swell, even higher than the die swell generated from the non-reinforcing fillers such as whiting (ground calcium carbonate) and hard clay (both with relatively large particle sizes).
[FIGURE 11 OMITTED]
[FIGURE 12 OMITTED]
Also, it should be noted that overall, die swell usually increases with a higher applied shear rate.
Figure 12 compares the running die swell values at only 45 phr filler concentration.
At 45 phr filler level, less dense filler networks form. As a result, about the same ordinal relationship among different fillers at the same concentration is maintained. However, the rheological differences are significantly less and the running die swell values are clustered in a much tighter band than we saw at 75 phr filler loading. However, still a higher applied shear rate causes some of the compounds with larger particle size fillers to rise in die swell.
Filler concentration effects on die swell
The filler concentration in a rubber compound has a large effect on the compound's rheology and specifically affects greatly its die swell characteristics. Generally, the higher the loading of filler, the lower the percentage of rubber hydrocarbon, and for colloidal particle sized fillers, the greater the density of the filler network. This network is formed from either Van der Waals forces, for carbon blacks, or hydrogen bonding, in the case of silica. These effects tie up the rubber and keep it from swelling. Even without these specific filler networking effects seen with carbon black and silica, there is still less swelling with large particle size diluent fillers, such as clays and whitings, simply because with a higher concentration of filler there is a lower concentration of rubber.
Figure 13 shows the large effect on running die swell from changing the concentration of a fully reinforcing carbon black such as N330.
RPA measurements of filler effects
Historically, RPA tan [delta] and G' have been used extensively to predict die swell for rubber compounds (refs. 15-17). While RPA parameters work quite well for predicting extrusion performance of factory recipes that contain relatively lower levels of fillers, they only work well with higher loading levels of fillers when careful consideration is made for the RPA strain amplitude that is applied. To understand this better, we will examine a series of RPA frequency sweeps performed at different applied strains, which has a large effect on the rheological measurements.
[FIGURE 13 OMITTED]
Figure 14 shows the results of an RPA frequency sweep at only 1% strain.
As can be seen, the N234 filler network is not disturbed greatly by applying only 1% strain during this RPA frequency sweep. Therefore, the 75 phrN234 carbon black shows the lowest uncured tan [delta] values even at the higher frequencies. The tan [delta] is equal to the viscous modulus divided by the elastic modulus (G"/G"). It is sometimes referenced to as the V/E ratio. Since the applied strain of only 1% does not destroy a significant portion of the elastic quality (G), then the tan [delta] value at low strain is not significantly affected.
However, by rerunning this frequency sweep at 7% strain instead of 1%, one can see that this moderately higher strain is beginning to break up the filler network and is causing a crossover in the tan [delta] data points at a frequency of 70 radians/second, where there is an inversion in the ordinal relationships. For example, the tan [delta] value for 75 phr N234 was the lowest at lower frequencies, but suddenly becomes the highest tan [delta] at the higher frequencies. In other words, for uncured tan [delta] values, what was the lowest at low frequency becomes the highest at high frequency. This is shown in figure 15.
[FIGURE 14 OMITTED]
Further work has shown that this uncured tan [delta] inversion is even more noticeable when the frequency sweep is performed at 50% strain amplitude.
This means that in order to achieve the best prediction of die swell with the RPA, one must make adjustments in the applied strain during frequency sweeps in order to simulate more effectively the destruction of the filler network which is also happening in the factory during an extrusion operation.
This "fine tuning" of strain in RPA testing is probably more critical for fully reinforcing carbon blacks than it is with semi-reinforcing carbon blacks (carcass grades) such as N660.
Additional work using 15 to 75 phr loading of N660 instead of N234 has shown that increasing the applied strain amplitude during RPA frequency sweep testing does not affect the RPA results for N660 nearly as much as it did for the N234.
By applying a series of frequency sweeps with progressively higher and higher strains to both the silica compounds (without TESPT vs. the silica compounds with the TESPT), some very interesting facts were observed which may be helpful in future RPA studies to determine the degree of silanization that is achieved.
Figures 16 and 17 compare RPA frequency sweeps at 1% strain for silica without TESPT vs. with TESPT.
Figure 16 shows that without the TESPT silanization, the silica filler network is very dense, as witnessed by the wide spread of uncured tan [delta] values with no uncured tan 5 inversion crossover at 1% strain. However, figure 17 shows that with the TESPT silanization, the uncured tan 8 values overall are clustered much closer together and there is a well defined uncured tan delta inversion crossover point at 15 radians/second. This is a signature that the silica filler network is very much destroyed from the silanization reaction with the TESPT. Also, this inversion crossover point could be useful in the future for quantifying the degree of silanization that has actually occurred.
[FIGURE 15 OMITTED]
[FIGURE 16 OMITTED]
[FIGURE 17 OMITTED]
Figure 18 shows that the RPA uncured tan 5 inversely correlates to the capillary rheometer running die well, if it is on one side of this inversion crossover point, but could actually have a positive correlation with die swell if the tan [delta] point is on the other side of this crossover point.
Further work has shown this same phenomenon at 7% applied strain.
Repeatability comparisons of capillary rheometer die swell measurements vs. RPA rheological measurements
From analysis of 285 capillary rheometer die swell measurements vs. 1500 RPA rheological measurements, a repeatability value was calculated for each instrument.
[FIGURE 18 OMITTED]
These calculations give the capillary rheometer a coefficient of variation of 3.68% vs. the RPA rheological measurements of 1.02%.
Repeatability is three times better with the RPA to measure rheological properties relating to die swell than with the capillary rheometer.
RPA rheological properties such as tan [delta] and G' can correlate to capillary rheometer die swell.
Capillary rheometer running die swell has an R squared correlation of 72% explained variation with relaxed die swell.
The RPA is three times faster and more productive than the capillary rheometer.
The Alpha Technologies extended dynamic range (EDR) for the RPA 2000 exhibits much better measurement of the Payne Effect for commercial compounds.
The ARC 2020 capillary rheometer effectively measured increases in die swell from higher applied shear rates, the use of lower filler concentrations, or the use of larger particle size fillers (instead of smaller particle size fillers).
Silanization of precipitated hydrated silica significantly increased the compound's die swell.
The specific strain amplitude applied during the performance of an RPA frequency sweep of a filled rubber compound is very important. Increasing the applied strain in a sequence of multiple frequency sweeps can measure an uncured tan [delta] inversion. Also, in a single frequency sweep, at a properly set strain, a specific frequency may be recorded where the tan [delta] inversion crossover occurs.
The uncured tan [delta] inversion crossover may prove useful in quantifying the degree of silanization that has occurred from a factory silica rubber mixing process.
Even though increasing strain amplitude and/or frequency applied during RPA testing of filled rubber compounds can give additional useful rheological information, running at ultra low strain with ultra low frequency gives no useful information, just noise.
This article is based on a paper presented at the 188th Technical Meeting of the Rubber Division, ACS, October 2015.
(1.) Kejian Wang, Chapter 4, "Die swell of complex polymeric systems," Viscoelasticity--From Theory to Biological Applications, Intech 2012.
(2.) J. Dick, E. Norton and T. Xue, "Effects of variation in strain measurements with the extended dynamic range (EDR) of the RPA on the accuracy of shear thinning measurements for rubber compounds, " Paper No. 88 presented at the Rubber Division, ACS, Technical Meeting, October 14, 2014.
(3.) J. Dick, "Overview of capillary rheometry, part 1," Rubber World, January, 2007, p. 30.
(4.) J. Dick, "Overview of capillary rheometry, part 2," Rubber World, February, 2007, p. 36.
(5.) Christopher Macosko, Chapter 6, "Shear rheometry pressure driven flows, " Rheology, Principles, Meaurements and Applications, Wiley-VCH, 1994.
(6.) Nicholas P. Cheremisinoff, "Product design and testing of polymeric materials, " J. of Polymer Science, Part A, Vol. 29, Issue 9, 1,365-1,366, August 1991.
(7.) J. Dick and H. Pawlowski, "Applications for stress relaxation from the rubber process analyzer in the characterization and quality control of rubber, " October, 1995, Rubber World, January 1997.
(8.) John Dick, How to Improve Rubber Compounds, 2nd Edition, Hanser, 2014.
(9.) John Sezna, MPT Applications Manual, Monsanto Chemical Corp., 1987.
(10.) Jean Leblanc, Chapter 5, "Polymers and carbon black, filled polymers, science and industrial applications," CRC Press, 2010.
(11.) J.S. Dick and E. Norton, "Applications for the RPA enhanced dynamic range in measuring silanization reactions from the mixing of silica-loaded stocks, " Paper No. 69 presented at the Fall Meeting of the Rubber Division, ACS, 2013.
(12.) J. Dick, E. Norton and T. Xue,"Effects of variation in strain measurements with the extended dynamic range (EDR) of the RPA on the accuracy of shear thinning measurements for rubber compounds," Paper No. 88 presented at the Fall Meeting of the Rubber Division, ACS, Nashville, TN, October 14, 2014.
(13.) J. Dick, H. Pawlowski and J. Moore, "Viscous heating and reinforcement effects of different fillers using the rubber process analyzer, " Rubber World, January, 2000.
(14.) Alpha Technologies, ARC 2020 Brochure, 2006.
(15.) J. Dick, M. Ferraco, K Immel, T. Mlinar, M. Senskey and J. Sezna, "Utilization of the rubber process analyzer in Six Sigma programs," Rubber World, January, 2003, p. 32.
(16.) J. Dick, "Overview of capillary rheometry, part 1," Rubber World, January, 2007, p. 30.
(17.) J. Dick, "Overview of capillary rheometry, part 2," Rubber World, February, 2007, p. 36.
John S. Dick, Edward Norton and Tianxiang Xue, Alpha Technologies
Table 1--variations in subject filler Compound 1 2 3 4 5 phr phr phr phr phr SBR 1500 100 100 100 100 100 Subject filler 15 30 45 60 75 Naphthenic oil 15 15 15 15 15 Stearic acid 2 2 2 2 2 6PPD 2 2 2 2 2 Zinc oxide 3 3 3 3 3 Total 137 152 167 182 197 Table 2--RPA test configuration Sequence of frequency sweeps 0.1 0.2 0.5 1.0 2.0 5.0 10 20 50 Hz at 0.07% strain, 100[degrees]C 0.1 0.2 0.5 1.0 2.0 5.0 10 20 50 Hz at 1.0% strain, 100[degrees]C 0.1 0.2 0.5 1.0 2.0 5.0 10 20 50 Hz at 7.0% strain, 100[degrees]C 0.1 0.2 0.5 1.0 2.0 5.0 10 20 50 Hz at 14.0% strain, 100[degrees]C 0.1 0.2 0.5 1.0 2.0 5.0 10 20 50 Hz at 50.0% strain, 100[degrees]C 0.1 0.2 0.5 1.0 2.0 5.0 10 20 50 Hz at 300% strain, 100[degrees]C
|Printer friendly Cite/link Email Feedback|
|Author:||Dick, John S.; Norton, Edward; Xue, Tianxiang|
|Date:||May 1, 2016|
|Previous Article:||NBR/PVC blends for optimized physical properties in extrusion compounds.|
|Next Article:||Rubber Division offers education.|