The use of rheological testing for predicting processability.
If an uncured piece of rubber were to be tested by pouting it from one glass to the next, it would appear not to flow. If, however, it were pressed through a syringe, as in extrusion, it would possess flow characteristics. Elastomers present new challenges because of the viscoelastic nature they exhibit, as well as the chemical process of crosslinking. When an uncured sample is subjected to heat, the physical properties of that rubber change drastically, making the measurement of flow necessary but incomplete. Other tests associated with vulcanization, such as scorch and the measurement of modulus during curing, help the rubber manufacturer to learn valuable information about the processing and performance of a product.
To begin with, it is useful to understand the unique aspects of an elastomer, and how those characteristics apply to the processing of a rubber compound. Polymers vary in many crucial ways. Polymers differ not only in their chemical composition, but also in their respective molecular weight (mw). Of all available polymers, rubber displays the highest molecular weight. The higher the molecular weight, the higher the viscosity. Due to the nature of the polymerization process, not all chain molecules are 'born equal', i.e., they vary in the number of repetitive units and therefore in weight. Figure 1 shows the statistical distribution of the chain lengths, which usually has the familiar shape of a bell. The average molecular weight is represented by the maximum of this curve. The variation in weight can be widespread (for instance for emulsion polymers) or rather narrow (often for solution polymers). Processes are usually quite sensitive to changes in molecular weight distribution (mwd). Wider distributed polymers are easier to process (due to the softening action of the higher number of short chains) and show higher elongation (due to the higher number of very long chains). Narrowly distributed polymers, however, have their advantage: They form much more homogeneous networks with superior tear and abrasion resistance. To make the issue even more complicated, we have to be aware that during the polymerization process, depending on the conditions, polymers do not always grow linearly, but instead may form branches, as can be seen in figure 2. In other words, chains with the same mw may not have the same length, and this has consequences for the processing behavior. Branched molecules behave like already entangled ones, i.e., more elastic. Figure 3 shows that branching is more evident at low shear rates.
[FIGURES 1-3 OMITTED]
Even though the polymeric structure plays an important, sometimes dominant, role in the prediction of the processing behavior of a compound, the influence of the filler(s) has to be taken into account, as well. They differ in particle size, particle size distribution and particle shape, as well as the chemical affinity of their surface to the polymer chains. Thus, all these properties influence the theological behavior of the compound.
Elastomers are viscoelastic in that they have a component that behaves like a fluid and one that behaves like a metallic spring. A Maxwell Element, as shown in figure 4, represents this behavior as a viscoelastic fluid (ref. 2). Here, the fluid, or viscous component is represented by a dashpot, Kv, and the elastic component is represented by a spring, Ke. When a force is applied to a Maxwell Element, both the dashpot and the spring elongate. When the force is released, the energy stored by the spring is used to return the spring to its original shape. However, when the force is released, the dashpot does not return to its original shape and the energy is dissipated as heat. This behavior is represented in figure 5. The spring is the storage or elastic component of a material, and the dashpot is the viscous or loss component. Although other models describe viscoelasticity, Maxwell's element presents a simple interpretation.
[FIGURES 4-5 OMITTED]
Processability of a rubber compound is often considered to be related to its viscoelastic nature. Wise and Barker explained this by stating: "The relative performance of a compound in these steps (the forming steps) is dependent upon its plasticity/ elasticity relationship, which in turn depends upon the polymer and its temperature, the force applied, the rate at which the force is applied and the degree of deformation (strain)" (ref. 3). ASTM further defines processability as the "relative ease at which raw or compounded rubber can be handled in rubber machinery" (ref. 4). The types of rubber machinery in mixing are typically open roll mills or internal mixers. The shaping stages of the process can include extruders, calenders or any type of molding: compression, injection or transfer. Finally, after the material is mixed and shaped, vulcanization must occur. This may occur in the same equipment as the shaping. For instance, in molding, the material is shaped and cured in the same equipment. On the other hand, a profile may be shaped by an extruder and then cured in a salt bath. Though processability can encompass all of the phases in the making of a product, for the purposes defined here, the focus will be on a mixed compound as it is being formed until the time when vulcanization begins, scorch.
The great depression presented American manufacturing companies with a challenge not seen since the Industrial Revolution. Between the years of 1930 and 1936, the number of automobile manufacturers went from 97 to 10, yet many of those that survived still exist to this day. In order to survive, corporations needed to create products that were more innovative and processes that were more efficient. The viable market had shrunk, but the value of the remaining enterprises increased out of necessity. The testing of rubber in the world's major rubber companies experienced a similar renaissance. In addition to the economic problems of the time, the rubber industry was experiencing a shortage of natural rubber, leading researchers from around the world to improve synthetic polymers to reduce the dependency on natural rubber. Melvin Mooney, from U.S. Rubber, developed a rheological testing instrument, which was called the Mooney viscometer. Mooney's aim was to develop a quantitative method that measures not only plasticity, but also other properties such as viscosity and recovery (ref. 5). The purpose, as he described it, was to give manufacturers a tool that will help them to predict the behavior of the rubber in the manufacturing process. The Mooney viscometer became a key quality control test, as well as an important tool for research and development of new products from tires to aircraft parts, which were soon to play a role in World War II. Though improved methods were introduced through the remainder of the 20th century, the Mooney viscometer remains one of the most common processabilty testers in the polymer and rubber products industries. But as so often happens, a simple method gives only a simple answer. The biggest disadvantage is that polymers with different mwd deliver the same Mooney viscosity. The Mooney viscosity test unfortunately reflects only the average molecular weight, as can be seen in figure 1. It cannot differentiate between polymers with narrow and wide mwd. The second issue is that, due to the relatively low shear rates, the method does not reflect the conditions in actual processing procedures. The die design for the Mooney viscometer has been consistent over the years, and can be seen in figure 6.
[FIGURE 6 OMITTED]
The original values that the Mooney viscometer looked at were viscosity, scorch and relaxation. Viscosity is simply the torque as the rubber flows. Scorch is the point when crosslinking begins. Knowing the scorch time is important, since it is that point that marks the end of potential shaping. Once a material has scorched, it can no longer flow freely. Relaxation is a measurement that will predict viscoelasticity without a shear force. All of the three basic Mooney tests can be seen in figure 7.
[FIGURE 7 OMITTED]
Molecules are entangled and form a 'physical' network, but align during shearing. Once the shear force is stopped, the deformed molecules try to recover into their original shape, which is done by slipping. The time it takes for that rearranging of molecules to be completed is an excellent prediction of viscoelasticity, and can be measured with Mooney relaxation. Branched molecules behave like overly entangled ones, which is evident by a slower stress relaxation. Because the Mooney stress relaxation (MSR) can detect this easily, it is a simple and efficient way for the characterization of a rubber polymer. If viscoelasticity can be predicted by doing a Mooney stress relaxation test, processing phenomena such as die swell can be anticipated. Figure 8 shows the parameters that are most commonly used for Mooney stress relaxation testing. These are also described in ASTM D1646 (ref. 6). Figure 9 shows an extreme, though unrealistic, example of how perfectly elastic and perfectly inelastic material will behave. Figures 9 also shows how the parameters being evaluated in MSR relate to elastic and inelastic behavior.
[FIGURES 8-9 OMITTED]
Another method to measure the viscoelastic components of a polymer or rubber compound is to subject the material to oscillatory shear. The use of oscillatory shear rheometry in the rubber industry began in the mid-1960s with the advent of the oscillating disk rheometer. In 1962, J.R. Beatty, et al., invented an instrument that used many of Mooney's ideas, while focusing beyond the point of scorch (ref. 7). The result was the combination of something similar to Mooney viscosity with tensile modulus, which was until this time used for determination of vulcanization. The instrument was initially called viscurometer, but has since been referred to as curemeter or oscillating disk rheometer (ODR).
The curemeter avoided the two drawbacks of the Mooney viscometer by creating a bi-conical disk to replace the rotor, and by oscillating rather than completing full rotations. The bi-conical disk made constant shear rates possible. The material closest to the center of the rotor will shear at the same rate as the material on the outside. Torque could be measured in inch-pounds or dNm rather than the arbitrary Mooney units. The choice of 1, 3 or 5 degrees of oscillation allowed the test to expand beyond scorch, since the sample would no longer tear in the dies. The ODR was run at higher temperatures than the Mooney viscometer in order to facilitate vulcanization, but still made it possible to measure viscosity, scorch, cure rate and reversion characteristics in one test. What remained the same was the basic die design of the Mooney that included dies surrounded by heated platens. The die design can be seen in figure 6.
The current oscillating disk rheometer requires no preheat. According to ASTM D2084, the lowest point in the curve was taken, the minimum torque, and designated as ML. Scorch was recorded as the time when a 1 or 2 dNm rise above ML occurred. This was called ts1 (or ts2), the 's' being scorch (ref. 8). The vulcanization process from this point of scorch is then plotted until vulcanization is complete. Though the shaping of the compound is finished by this point in the process, the amount of heat applied during the cure remains important to the success of the final product. A rheometer can help make these predictions. Once the torque curve has leveled off, it is assumed that the sample has completed vulcanization. The torque value at this highest point is called maximum torque, and designated as MH. Cure time is calculated by:
tc90 = the time at torque: 0.9 (MH-ML) + ML
A cure rate can be found by using the following formula:
CRI, cure rate index = 100/(tc90 - ts1)
A standard rheometer curve is shown in figure 10.
[FIGURE 10 OMITTED]
The moving die rheometer improved on this by simplifying the test method and adding the capabilities of measuring the viscoelastic components of a rubber compound. This method is described in ASTM 5289 (ref. 9). These instruments, however, were limited in that they were operated at conditions of constant temperature, frequency and strain. To fully characterize a polymer or rubber compound, a range of these conditions needed to be available. This occurred with the introduction of versatile dynamic mechanical rheological testers (DMRTs) introduced in the 1990s. DMRTs were designed both for sophisticated research and development, as well as for routine process and quality control testing. The DMRT allows full characterization of a material's viscoelastic properties. ASTM 6204 and ASTM 6601 describe several applications of DMRTs (refs. 10 and 11). The die configuration of a moving die rheometer die is shown in figure 6.
As in previous rheometers, the DMRT measures the torque required to oscillate a die. While varying the physical conditions of frequency, strain and temperature, the DMRT can then produce a series of results that relate to Maxwell's Element. The elastic modulus (G') is the spring, the loss modulus (G") is the dashpot, and the dynamic modulus (G*) represents the total applied force. Figure 11 shows a simple vector diagram relating these values. Rather than modulus, represented as G, the DMRT measures torque directly, represented as S. A simple mathematical conversion can be performed to convert torque values to modulus. Figure 12 shows this relationship in a common cure curve.
[FIGURES 11-12 OMITTED]
The shaping process
Figure 13 represents a simplified relationship seen in a rubber shaping process. Although this uses a rubber extruder to represent the shaping stage, all rubber processes possess the similar components encompassing a force, the rubber compound, a forming stage and a final shape. For example, as we see in the extruder, the force required to rotate a screw or push a ram is applied to the rubber, which is shaped by an orifice. In calendering the rubber is forced by the calender rolls to a desired thickness of sheeted material. In molding, the rubber is forced into the shape of the mold generally by ram pressure. In all cases temperature is an added critical component, since the material's properties also change as a function of temperature. Also, in all cases, a force is applied to the robber in order to obtain a desired shape.
[FIGURE 13 OMITTED]
Figure 14 provides an example of a perfectly elastic compound after it is shaped. Though perfect elasticity in rubber is impossible, this theoretical figure shows a material that maintains its previous shape exactly. This characteristic is represented by elastic modulus (G') only. Since there is no viscous response, loss modulus (G") would not exist. Tangent [delta], the ratio of G" to G', will also be 0. The figure also shows a perfectly viscous example, which is also impossible for rubber. In this example, the shaped material does not hold any shape at all, instead experiencing only flow. In this case, only G" would exist, while G' would be 0. Here, tan [delta] is undefined. By these examples it is clear that tan [delta] is inversely related to elasticity. The goal of producing the shape of the orifice is not easily achieved. The figure also represents a more realistic elastomeric response, where shape is neither perfectly held, nor lost. By understanding the viscoelastic characteristic of the material, processability may be better defined.
[FIGURE 14 OMITTED]
Providing real world examples allows us to see the advancement from these new technologies, and how they can easily be interpreted. These are only four simple examples, as a very large range of test possibilities exists and has been well documented.
As was previously discussed, Mooney viscosity has been the most common way to test processability, mainly because of the simplicity to interpret results and the fact that it has been used for so many years. Because of the extremely high strains involved (full rotation), it is known that it sacrifices to distinguish filler networks, and intricacies in the physical network are lost. Several situations often present themselves when looking at a rubber compound and trying to differentiate quality compounds or batches. Viscoelasticity, as was explained, is a key material property that relates to processability. Scorch is another issue which presents itself for a number of reasons, including cure system variations, shelf life and mixer temperature variations. To analyze these particular issues, four examples were evaluated (ref. 12):
1. Comparison of viscoelastic measurements of four compounds to the measurement of Mooney viscosity;
2. the use of a DMRT variable temperature test to better define a material's scorch characteristics;
3. the use of a DMRT frequency sweep to differentiate between batches of differently aged materials; and
4. the use of a DMRT test for quality control optimization.
The results from the first series of tests are shown in figure 15. Here, the Mooney viscosity results show compounds A, B and C to be approximately the same, and compound D to be only 4% lower. The complex modulus ([G.sup.*]) shows the same trend, but compound D is nearly 25% lower than the other compounds. The Mooney viscosity and complex modulus show similar trends, but complex modulus shows significant differences. When tan [delta] is analyzed, additional differences are seen. The tan [delta] values show a range of nearly 15%, with compound C being the most elastic, and compound A being the least elastic. It would be expected, therefore, that these materials may process differently, for example, in an extrusion process where die swell is important. The differences in dynamic modulus and tan delta, however, may not be critical to a compression molding process. It is important to always consider the conditions of processing when analyzing the material's properties and relating them to processability.
[FIGURE 15 OMITTED]
In this series, three different batches of the same compound were tested using a standard moving die rheometer, isothermal test at 177[degrees]C. The test conditions were then changed to set the test temperature at 121[degrees]C for the first 25 minutes of the test and then the temperature was automatically raised to 177[degrees]C. The results are shown in figure 16. It can be seen that the isothermal test did not show any difference between the three batches. When tested with a variable temperature profile, however, the scorch time was seen to vary significantly. Scorch differences must be considered in regard not only to the process, but also the material flow characteristics. Scorch may therefore be more critical in injection molding, where scorch times are typically shorter and the material is subject to additional work, which may result in viscous heating.
[FIGURE 16 OMITTED]
This next series uses three batches of the same compound that had been aged for different times. To evaluate this, a variable frequency test (oscillating frequencies of 0.1, 2.0 and 20 Hz) was made at 100[degrees]C and 7% strain. The tan 5 results are shown in figure 17. These results showed significant differences at the lowest frequency, indicating that the highest tan [delta] was the least-aged and the lowest tan [delta] was the most-aged. This is logical, since small amounts of crosslinking will occur as a compound ages. Therefore, a compound will become more elastic as it ages.
[FIGURE 17 OMITTED]
The moving die rheometer has been an excellent tool for providing a fast test that keeps up with the mixing process. Often, a moving die rheometer test on a compound takes less than five minutes. There are several limitations to the standard moving die rheometer test, which can be seen in figure 10. The most obvious is the relatively small fraction of the test that occurs before ts1 (scorch), and how that section is near the bottom of the overall vulcanization curve. It is this section that most completely represents processability; therefore, the common quality control test does not fully provide a picture of processability. For this reason, the DMRT has been used to design a test that magnifies the shaping section of the vulcanization curve, which is the section before scorch occurs, while still completing the test in the same length of time.
An FST test (frequency, strain, temperature) was designed in order to take advantage of those variable capabilities of a DMRT. In addition to varying conditions, a unique timing approach was employed in order to provide for the desired short test time. This uses the software system to change test conditions based upon achieving scorch time. The goal was then to select FST conditions to complete a test in four minutes or less, so that the test turn-around time could be minimized (ref. 13).
Frequency and strain were set initially 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. When a 1 dNm rise above the minimum torque occurred, the temperature was raised to 185[degrees]C, the frequency was reduced to 1.67 Hz and the strain was reduced to 7%. A comparison of the standard isothermal moving die and the FST test is shown in figure 18.
[FIGURE 18 OMITTED]
Sixty batches were then run consecutively on a DMRT using the FST parameter, and separately run on a DMRT at constant conditions of 175[degrees]C, 7% strain, 1.67 Hz. The cure curves are shown in figure 19. The standard moving die rheometer tests showed very few differences in viscosity and scorch times, while the FST showed a much larger range in both, suggesting that it is a more discriminating quality control test for processability.
[FIGURE 19 OMITTED]
Many years of rubber mixing and processing experience help us to develop and use compounds fulfilling the desired requirements. Among other things, however, the more we know why compounds behave the way they do and not just how, the faster and more economically we are enabled to serve our markets.
One thing that has helped us tremendously is the understanding of structure-property relationships of the ingredients that make up a rubber compound. Starting from the polymer chains' repetition unit, up to the macroscopic structure, we are able to predict the rhecal behavior, the crosslinking step and the ultimate properties of the finished rubber part. Of course, the interaction with fillers, plasticizers and, above all, the crosslinking system has to be taken into account as well.
Without the instruments, however, which probe the rheol ogy, the curing and the properties we are interested in, we are rather blind. For many years, the practical and simple solutions like the Mooney viscometer, the ODR and the durometer hardness tester seemed to be sufficient. Needless to say, they left more questions then answers.
We cannot claim to have instruments and methods at hand today which give us all the answers, but we are much closer after the introduction of modem DMRTs. Not only do they help us to lift the veil further and give us a fuller picture of a compound under development in the lab, they also find their way into the mill room or the work floor, helping to monitor and maintain the desired uniformity.
This article is based on a paper presented at a meeting of the Rubber Division, ACS, www.mbber.org.
(1.) J. White, in "Science and technology of rubber," editors Mark, Erman and Eirich, 1994, p. 257.
(2.) ASTM D 6048-96, "Stress relaxation of testing of raw rubber, unvulcanized rubber compounds and thermoplastic elastomers," 09.01, American Society for Testing and Materials, West Conshohocken, PA, p. 921 (2001).
(3.) Barker, Sullivan and Wise, in "Basic elastomer technology," K.C. Baranwal, H.L. Stephens, Eds., The Rubber Division, American Chemical Society, 2001, p. 145.
(4.) ASTM D1566-00b, "Standard terminology relating to rubber," 09.01, American Society for Testing and Materials, West Conshohocken, PA, p. 295 (2001).
(5.) Melvin Mooney, U.S. Patent 2,037,529, "Plastometer," April 14, 1936, Assigned to United States Rubber Company.
(6.) ASTM 1646--03, "Standard test method for measurement of viscosity, stress relaxation and pre-vulcanization characteristics," American Society for Testing and Materials, West Conshohocken, PA, p. 343. (2003).
(7.) J.R. Beatty, A.E. Juve and P.W. Carper, U.S. Patent 3,182,494, "Viscurometer," May 11, 1965, assigned to the BF Goodrich Company.
(8.) ASTM 2084--01, "Standard test method for vulcanization using oscillating disk curemeter," American Society for Testing and Materials, West Conshohocken, PA, p. 384 (2003).
(9.) ASTM 5289--95, "Standard test method for vulcanization using rotorless cure meters," American Society for Testing and Materials, West Conshohocken, PA, p. 851 (2003).
(10.) ASTM 6204--01, "Standard test method of unvulcanized rheological properties using rotorless shear rheometers," American Society for Testing and Materials, West Conshohocken, PA, p. 1,027 (2003).
(11.) ASTM D6601--02, "Measurement of cure and after-cure dynamic properties using a rotorless shear rheometer," American Society for Testing and Materials, West Conshohocken, PA, p. 1,058 (2003).
(12.) M.C. Putman and J.B. Putman, "Processability for viscoelastic compounds," Tire Technology International (2003).
(13.) J.B. Putman and M.C. Putman, "A simplified approach to testing for quality control and process control," paper 26, ACS Rubber Division, October 2003.
Matthew C. Putman and Peter Richter, Tech Pro, email@example.com
|Printer friendly Cite/link Email Feedback|
|Date:||May 1, 2007|
|Previous Article:||A new nitrosamine-free curing agent.|
|Next Article:||Impact of inflation pressure retention on tire aging and durability.|