DMA of rubber in immersion environments.

Dynamic mechanical analysis (DMA) simultaneously gives a rapid and unique insight into the thermal and mechanical properties from a single experiment. It also offers exceptional sensitivity to glass transitions and provides information about the modulus and damping behaviors of the material with respect to temperature. Secondary relaxations are also frequently evident (ref. 1).

In DMA, the sample is deformed and released in a smooth sinusoidal fashion. Modulus is calculated from the amplitudes of motion and force, while damping is measured from the time delay (phase lag) from the applied force to the resulting motion. The usual thermal scan can profile the properties of a material from as low as -150[degrees]C to temperatures in excess of 300[degrees]C in as little as an hour. Additional variables include deformation mode, frequency, and amplitude of stress or motion applied to the sample. Additional information that can be collected from typical DMA instruments includes average stress and motion for expansion and contraction studies in concert with modulus measurements. Furthermore, many modern instruments allow non-sinusoidal force experiments for creep or ramped force stress-strain evaluations.

One clear drawback of most DMA measurements is that most evaluations are conducted in conditions far from the end-use conditions for rubber (such as air and low, 1 Hz, frequencies). Some work has been conducted on samples that have been immersed in liquids, or a variety of atmospheric conditions (refs. 2-4), with adapted DMA instruments. These procedures and equipment suffer from various drawbacks, as they were not specifically designed for these tasks. The physical instruments modified to accept fluid, humid or other gaseous environments lack fully integrated temperature control or suffer from limited specimen geometries or both. Furthermore, experiments where the sample has been preconditioned by immersion in a liquid or exposure to a particular gaseous environment and then analyzed in a conventional instrument are difficult to interpret. These specimens often lose the solvent, moisture or gas to which they were exposed during the experiment. This loss of material alters the specimen behavior, beyond the changes due to the increasing temperature, and thus complicates interpretation.

This article describes a new instrumentation system designed specifically for dynamic mechanical analysis of materials in air, fluids and a variety of other environmental conditions. The instrumentation is described in detail, and a series of applications illustrate the role this equipment can play in resolving issues where the environment affects the mechanical properties of materials.

Instrumentation

The immersion data described in this paper have all been obtained using a T2000 DMA fitted with the environmental furnace, which is manufactured by Triton Technology and available in North America through Mettler-Toledo (figure 1).

[FIGURE 1 OMITTED]

The samples can be mounted using any of the standard deformation modes (clamps) used for traditional DMA experiments including cantilevered bending, shear, tension, compression or three-point bending. The most common modes, cantilevered bending and tension, are shown in figure 2. The patent-pending rotational head mechanism allows the mech-anical section of the instrument (the middle tower in figure 1) to be positioned optimally for each type of experiment. For immersion studies, the specimens are initially mounted with the head in the horizontal position. The head is then rotated to the vertical down position, ready to accept the environmental furnace. The furnace slides forward for filling with liquids. It is then slid backwards and raised to close and immerse the sample. The liquids are contained within a liner that is constructed of either borosilicate glass or anodized aluminum.

[FIGURE 2 OMITTED]

The environmental chamber can also be used to envelope the sample in a known gaseous atmosphere. This can allow, for example, experiments in certain humidity conditions. The service collar that meets the bath chamber has purge inlet and outlet ports. These are in addition to the standard purge ports that are provided on the standard instrument system. A humidity generator can be attached to the inlet with the outlet being left open or attached to a withdrawal system. A port in the service collar accepts a humidity measurement device for checking the actual humidity. This service port can also be used for other sensors, such as pH meters, or for injecting materials into the system.

Although primarily designed for good isothermal control of liquids, the system is capable of slow, controlled thermal ramps with liquid or gaseous environments. No additional power supply or software is needed for temperature control, since the environmental chamber plugs into the instrument's furnace power port and uses the standard temperature programmer. The temperature programmer's PID control parameters are readily calibrated for each particular fluid. Three sets of PID values are stored in the instrument at one time, but unlimited sets can be stored in a spreadsheet and pushed into the instrument memory as needed. The environmental chamber allows a number of different types of temperature sensors, in a number of different locations, to be used for temperature monitoring and control. Sub-ambient experiments are possible because of the closed loop cooling system on the environmental furnace. Experiments can be started at ambient temperatures followed by heating or cooling ramps. Alternatively, the fluid can be raised to the start temperature prior to immersing the sample.

The temperature range of the system fitted with the bath is -100[degrees]C to 150[degrees]C, although this depends more on the nature of fluids used than on the instrumentation itself. The temperature performance is highly dependent on the physical properties of the fluid. Care should be taken to follow all safety precautions for the fluids being used; most importantly flash points and boiling temperatures. The analytical head can be purged of fumes if odorous or toxic vapors are evolved from the analytical fluids. These can be vented into an appropriate fume extraction system. The T2000B DMA was used in all the immersion studies presented in this article.

Another new DMA instrument, the DMA/SDTA[861.sup.e] was used to measure the modulus of polychloroprene rubber at high frequencies. This DMA has a lower instrument compliance so samples can be measured at much higher frequencies than with the Triton DMA. There are six major design differences between this instrument (figure 3) and others. First, a force sensor attached to the fixed clamp measures the force, rather than simply measuring the drive motor force. Second, the displacement sensor (LVDT core) is attached to the moving clamp on the opposite side of the drive shaft near the sample. Third, the frame that connects the motor to the measurement section is extremely stiff. Fourth, the clamping mechanism is very stiff, yet it allows very accurate mounting of shear and tension mode samples and wide ranges of operating lengths in bending modes. These qualities lead to the capability for a very wide range of sample stiffness measurements. Fifth is the addition of a four-axis alignment device to ensure the drive system and the measurement system are colinear. Sixth, a separate temperature sensor accurately measures the sample temperature, particularly in shear mode.

[FIGURE 3 OMITTED]

Moving the displacement sensor off the drive shaft and adding the force sensor simplify the operating model of the instrument. In this system, the force supplied by the motor is important for control of the applied amplitude, but is irrelevant to the measured force. In this arrangement, there is no need to account for the inertia of the drive system or suspension stiffness. The system moves according to the specimen stiffness. This motion is detected directly, since the displacement sensor is connected to the moving clamp and not to the drive shaft. The motion generates internal forces in the sample, which are detected directly by the force sensor attached to the fixed portion of the clamping assembly. This system measures specimen stiffness accurately to within 1% for frequencies up to about 10% of resonance frequency (erf. 5), even at very small displacement amplitudes. Since the theoretical resonance frequency is approximately 3 kHz (shear clamps) and the damping frequency is around 60 kHz, the system is ideal at frequencies up to 300 Hz. Beyond this frequency and for very stiff samples, the stiffness and mass of the fixed clamps, along with the stiffness and damping of the force sensor, must be taken into account. The corrections extend the usable frequencies to 50% of the resonance (1,000 Hz). The full derivation of this model can be found in Wrana (ref. 5).

The clamps are arranged into two families according to the modes of deformation. The small clamping assembly (figure 4) is for shear and small tension samples, and the large assembly is for bending, compression and larger tension samples. These can be seen in figure 4. The clamping assembly (frame) mechanism allows shear and tension samples to be mounted into the clamps outside the instrument, which provides reproducible sample preparation and accurate sample thickness measurements. The clamp/sample combinations are then inserted into the clamping assembly, which maintains system alignment. The clamping assembly arrangement also keeps the system aligned when changing modes within a single family.

[FIGURE 4 OMITTED]

The four-axis alignment device ensures colinear motion of the drive and measurements systems with proper sample orientation. With a misaligned system, force can be applied either off-axis (rotated) or off-center relative to the specimen. The four-axis device repositions the fixed clamps relative to the moving clamp. Alignment is checked whenever the clamping assembly is changed, e.g., when changing from shear to bending. When working within a single clamping assembly, four-axis realignment is not necessary.

The measurement system has been optimized so that parasitic instrument resonances occur above the maximum measurement frequency of 1,000 Hz. This large frequency range can be used for the shear clamps with proper sample conditions. The DMA/SDTA 861 e was used to measure the modulus of polychloroprene rubber samples at frequencies up to 1,000 Hz.

Results and discussion

Nitrile rubber seals

Three grades of nitrile rubber were tested for suitability in seal applications utilizing the T2000B DMA. In the first test, the sample was heated from -100 to +100[degrees]C at 4[degrees]C/min. in single cantilever bending. The results of these experiments, bending storage modulus (E') and damping (tan delta), as functions of temperature, are shown in figure 5. At normal use temperatures these materials have a bending modulus around 20 MPa (2E+07 Pa). At lower temperatures, they become glassy with a modulus that is approximately 100 times stiffer than in the rubbery region. In the transition zone, the damping (tan delta) exhibits a peak due to molecular rearrangements. This peak temperature is often used as the glass transition temperature. The standard grade rubber, marked nitrile in figure 5, shows a glass transition at around -15[degrees]C. Low ACN rubber, which is used to enhance the low temperature properties of the seal compounds, has its transition around -40[degrees]C. The third rubber is an injection moldable grade (the other materials are processed by compression molding) that has an even lower glass transition temperature (Tg). Its modulus above the glass transition temperature is slightly higher than the low ACN grade.

[FIGURE 5 OMITTED]

A number of standard tests had been performed on rubber materials. These included hardness, tensile strength, ultimate elongation, stress at 100% elongation and volumetric swell. Of these properties, ultimate elongation correlated to glass transition temperature (Tg) and tan delta. For example, the nitrile rubber had the greatest elongation and the highest Tg. More experimentation is needed to definitively characterize these correlations.

Comparison of durometer and modulus for polyurethane

The purpose of this test was to correlate the modulus values of various polyurethanes with their durometer (ref. 6) values. The T2000B DMA was used in all these tests. The first step in analyzing a new type of sample is to evaluate its response to frequency and strain. The parameter of frequency is very important to the analysis of elastomers. The modulus can change significantly with frequency, as seen in figure 6. This frequency sweep experiment was performed on one specimen of polyurethane at ambient temperature in the single cantilever mode with a fixed displacement amplitude of 40 gm. The modulus was reduced at frequencies below 1.5 Hz and above 3.5 Hz. Ideally, we would like to operate the DMA in the region where the modulus is invariant to frequency, in this case 1.5 to 3.3 Hz. The DMA, however, is more sensitive to sample differences at higher frequencies, so we chose 5 Hz as a compromise.

[FIGURE 6 OMITTED]

The next step was to determine the polyurethane's response to strain. The purpose of this experiment was to determine the sample's strain region for linear viscoelastic behavior. The sample can permanently deform if it receives too much strain, and the measured modulus would drift with time. The results would be too variable if there was insufficient strain. In this test, the frequency was held constant at 5 Hz, and the strain was changed from low to high by increasing the displacement amplitude applied to the sample. The results of this experiment are shown in figure 7. The modulus be-comes noisy below about 0.015 mm amplitude, and the modulus decreases rapidly above 0.06 mm. Ideally, the region from 0.015 to 0.03 mm is optimum, because the sampie's modulus is invariant. The instrument is more sensitive at higher amplitudes, so we picked 0.04 mm amplitude as a compromise. This amplitude represents a maximum strain of 0.16% for the sample.

[FIGURE 7 OMITTED]

The next step was to check the range of samples being tested, in this case different durometer values, to determine if the frequency and strain, selected above, are in the linear viscoelastic region for the extreme samples. Figure 8 shows the amplitude sweep results for 40 and 75D durometer polyurethane samples. Both samples were at ambient temperature with a frequency of 5 Hz. Notice how the two samples behave differently versus displacement. The durometer 40 specimen has a constant modulus above 0.04 mm displacement. The durometer 75D specimen has a large amplitude dependence above 0.04 mm. In this case, we are correct in selecting 0.04 mm dynamic displacement for these experiments. The frequency sweep for both samples was also performed, but not reported here. The five-hertz frequency was a good frequency for both polyurethane samples.

[FIGURE 8 OMITTED]

Now we can test evaluate the correlation between modulus and durometer. A range of polyurethanes with differing durometer values was run in the T2000B DMA (figure 9). As can be seen, the durometer value for these polyurethanes correlates with the log of the modulus of the specimen. Caution must be taken in extrapolating this correlation to all classes of elastomers. This test must be performed for each class of elastomers being investigated.

[FIGURE 9 OMITTED]

Modulus of polychloroprene immersed in hot steering fluid

The T2000B DMA can be used to go beyond the durometer test and test the modulus of a sample submersed in a fluid. This experiment studied the effect of hot steering fluid on the modulus of polychloroprene rubbers of several durometer values. We selected 80[degrees]C for these tests to simulate the temperature environment in an automotive engine compartment. These tests were all done at 5 Hz frequency and 0.04 mm dynamic amplitude. The samples all had a nominal thickness of 3 mm. The results are shown in figure 10. The top two lines, durometer 80 and 60, show a drop in the modulus over time as the hot steering fluid penetrated the polychloroprene matrix. The lower durometer specimens, 30 and 40, showed the opposite effect; the modulus increased with immersion time. These results highlight the danger in extrapolating correlations too far. In this case, analyzing only the high, or low, durometer samples and extrapolating to the others would have given wrong conclusions. We postulate that the modulus increase of the lower durometer samples is due to the fluid entering the matrix and displacing the air. In this case, we would expect the specimen modulus to increase as the air is displaced with the higher modulus fluid.

[FIGURE 10 OMITTED]

Modulus of polychloroprene immersed in hot automotive fluids

A rubber selected for automotive use may come in contact with many different fluids. In this experiment, we examined the comparative effects of different fluids on the modulus of polychloroprene (durometer 60) at 80[degrees]C. The results of these tests are shown in figure 11. A most interesting result is the effect of air on the polychloroprene--the modulus increases over time. We suppose this is due to the polychloroprene curing further during the exposure to 80[degrees]C heat. This serves as a glaring reminder that one must always run controls in order to properly interpret the results. This polychloroprene would become stiffer with time if it were used in the engine compartment of a car.

[FIGURE 11 OMITTED]

Ethylene glycol (figure 11) reduces the modulus of the polychloroprene over time, relative to the (air) control. Water reduces the modulus even more. We suspect that the anhydrous ethylene glycol is removing water from the native polychloroprene at the same time the ethylene glycol is penetrating the rubber. Further work needs to be done varying the concentration of water in ethylene glycol. The steering and transmission fluids have a similar softening effect on the polychloroprene. The steering fluid decreases the modulus faster, presumably due to its lower viscosity. The T2000B DMA with the fluid bath is ideally suited to testing the effect of fluids on elastomers.

Modulus of polychloroprene as a function of frequency

The selection of an elastomer for use in any application must consider the strain and frequency it will be subjected to when in use. A rubber may have a low enough durometer, but that is a measurement at approximately zero frequency. The end-use conditions will require the rubber to absorb energy at a non-zero frequency. Auto engine mounts must absorb energy at many frequencies, such as 13 Hz at idle, 33 Hz at normal speeds and 83 Hz under rapid acceleration. The DMA/SDTA 861e was used to test polychloroprene from 0.01 to 1,000 Hz as a function of durometer. These tests were done at ambient temperature, shear deformation and 1% strain. The results are shown in figure 12. Normal automotive engine vibrations would be between the 1 and 2 in figure 12. The high durometer polychloroprene may last longer due to its high modulus, but at high frequencies it becomes stiffer. The higher modulus means the durometer 80 polychloroprene would be more likely to transmit engine vibration to the passenger compartment. These test results are needed to assist design engineers in the proper selection of elastomers for use where vibration damping is desirable. Further work would include repeating these experiments at different temperatures to simulate real world conditions.

[FIGURE 12 OMITTED]

Conclusions

Many different DMA experiments have been presented in this article. A summary of the important results includes:

* Nitrile rubber with the greatest elongation has the highest glass transition temperature;

* polyurethane modulus changes with strain and frequency;

* at ambient temperature, durometer values of polyurethanes correlate with the log of modulus;

* modulus change for polychloroprene in hot steering fluid does not correlate with the durometer values;

* elastomers must be selected for their actual performance in the hot contact fluid; and

* high frequency moduli are needed to properly select an elastomer for vibration damping.

The new Triton DMA instrument system simplifies the study of material behavior, not only in fluids, but also in defined environmental atmospheres. The system provides a safe and easily adaptable arrangement, allowing users in a wide variety of disciplines the ability to extract accurate mechanical and thermal information. The immersion experiments can mimic behavior in end-use conditions and can be used to accelerate changes in the materials. Measurements in controlled environments can prevent loss of solvent or additive during heating, thus yielding better insight into the effect of conditioning on the properties of materials.

The DMA/SDTA[861.sup.e] is ideally suited to testing the modulus of elastomers at high frequency. This enables better elastomer selection for end-use applications. The results shown here are only a start. These new capabilities offer many new tests for faster basic research and application development for elastomers.

References

(1.) R.E. Wetton, R.D.L. Marsh and J.G. Van de Velde, Thermochimica Acta 175, 1-11 (1991).

(2.) "Dynamic mechanical response of adhesively bonded beams: Effect of environmental exposure and interfacial zone properties," C. Li, R.A. Dickie and K.N. Norman, Ford Motor publication,.

(3.) "Thermal analysis of food components," D.M.R. Georget and A.C. Smith, Inst. of Food Research, Norwich Laboratory, Loughborough University.

(4.) M. Odlyha, T.Y.A. Chan and O. Pages, Thermochimica Acta, 263, 7-21 (1995).

(5.) "Dynamisch-mechanische anayse von gefullten elastomersystemen," Wrana, C., Doctoral Dissertation from Fakutat fur Naturwissenschaften der Universitat Ulm, Ulm, Germany, 1996.

(6.) ASTM International Standard D2240-02b, "Standard Test Method for Rubber Property--Durometer Hardness."