Carbon black's effects on vibration isolation.

Elastomeric products reinforced with carbon black are widely used in many dynamic applications where controlled damping of mechanical vibrations is essential for performance. In the elastomeric compounds used in disease applications, carbon black serves both to modify the dynamic properties and to improve the strength properties of the rubber end-product. Proper selection of the type of carbon black for optimum performance in vibration isolation applications relies on knowledge of the properties that the carbon black provides. To aid in this selection, this article describes the effects that carbon black characteristics have on dynamic and failure properties in representative elastomeric formulations -- i.e., natural rubber (NR) and butyl rubber (IIR) compounds typical of those found in engineered vibration isolators for engine mounts. Information presented in this article would also have utility in other dynamic applications involving these types of elastomers.

Comparisons of dynamic properties of rubber compounds normally are evaluated over ranges of strains, frequencies and temperatures. Analysis of the data has sometimes relied on plots of G' (or tan [Delta]) vs. strain at a fixed frequency or of G' (or tan [Delta]) vs. frequency at a fixed strain. (In this context, G' is the dynamic elastic modulus and 8 is the phase angle between the elastic modulus and the viscous modulus of a viscoelastic material.) Alternative analysis of data has been based on plots of tan [Delta] vs. G' at constant strain and frequency. Although these kinds of analyses provide useful information about the dynamic performance of rubber compounds, it is difficult to relate results from these dynamic analyses to the performance of the rubber components expected under the operating conditions experienced in engine mounts. Causes for this difficulty are related to the Payne effect found in filled rubber compounds (ref. 1) and have been described previously (ref. 2).

Dynamic performance of engine mounts depends on their characteristics under both static and dynamic conditions. In this article, dynamic properties are evaluated in terms of springrate ratio ([k.sub.d]/[k.sub.s]) vs. tan [Delta] which have been shown to be representative of dynamic performance important in the design of a vibration isolator (ref. 2). In this representation, [k.sub.d] is the dynamic elastic springrate of a vibration mount under low strain and high frequency which characterizes the dynamic operating environment of the mount. In terms of laboratory test variables, [k.sub.d] is equal to G' ([Omega, Gamma]) which approximates compound elastic modulus at the frequency ([Omega]) and strain ([Gamma]) where an engine mount normally provides isolation. The parameter [k.sub.s], is the static springrate of a vibration mount under high strain and low deformation rates characteristic of a mount under static conditions. In terms of laboratory test variables, [k.sub.s] is equal to [G.sub.s] ([Gamma]) at strain, [Gamma], and represents the elastic modulus at the static loading of the mount. The hysteretic parameter, tan [Delta], is the loss factor of the material evaluated at high strain and relatively low frequency intended to simulate conditions at the resonant frequency of an engine mount.

In terms of engine mount performance, the parameter [k.sub.d] is a measure of the ability of a rubber compound to influence transmissibility at normal (high) operating frequencies; i.e., rubber compounds with low values of [k.sub.d] give improved isolation performance at their normal operating frequencies. The parameter [k.sub.d] is not used by itself but rather as part of the [k.sub.d]/[k.sub.s] ratio which normalizes the dynamic stiffness with the static modulus. This facilitates comparison of results from compounds with different levels of hardness. In the low frequency range near the resonance of the isolation system, the ability to reduce transmissibility can be estimated from the tan [Delta] of the rubber compound; i.e., rubber compounds with high tan [Delta] give improved isolation performance near the resonance frequency of the vibration isolator. Hence, for best performance of an isolation system at operating (high) frequencies and at resonant (low) frequencies, the rubber compound should have the lowest value of [k.sub.d/k.sub.s] and the highest value of tan [Delta], both measured as described here. Derivations and more details of these relationships and their relevance to engine mount performance are given in (ref. 2) where it is also shown that the parameters [k.sub.d/k.sub.s] and tan [Delta] are directly related, and both are measures of the Payne effect in filled rubber systems (ref. 1). Dynamic parameters described in this article are defined as follows:

[k.sub.d] = G'(100 Hz, 0.2% single-strain amplitude)

[k.sub.s] = G(t) (30 sec., 20% single-strain amplitude)

tan [Delta] = tan [Delta] (10 Hz, 2% single-strain amplitude, interpolated from strain sweeps)

These conditions were selected as those which approximate performance of an engine mount in its normal operating environment. The strain in [k.sub.d] was deliberately chosen to be one-tenth the strain in tan [Delta] because the tan [Delta] of natural rubber was expected to be 0.1 which, in turn, gives an expected strain magnification of 10.

Experimental

The surface areas, structures and aggregate size distributions of the carbon black samples were evaluated. Surface area in m2/g was determined with cetyltrimethylammonium bromide (CTAB) according to ASTM procedure D-3765. Structure as measured in ml/100g of dibutyl phthalate absorption was determined after crushing (CDBP) according to ASTM procedure D-3493. Parameters for aggregate size distribution (ASD) were measured by using a disc centrifuge photosedimentometer. Each carbon black sample for ASD determination was sonified as a 10 mg per 25 ml water slurry to which a nonionic surfactant had been added. The D-ratio used later in this article was calculated from the ASD curve as the ratio of the still width at half maximum height of the curve divided by its mode.

The formulation for the NR compounds is given below. Except where noted in the results given in table 1, the oil loading was 10.0 phr (parts per 100 parts of rubber, by weight) in all experiments. The mixing procedure for NR compounds in the internal mixer (77 rpm, 32 [degrees]C water) is given in table 2. The formulation for the IIR compounds is given in table 3. The mixing procedure for IIR compounds in the mixer (77 rpm, 32 [degrees]C water) is given in table 4.

  Table 1 - NR compound formulation

Ingredient                  phr
NR                          100.0
Carbon black                Variable
Zinc oxide                  5.0
Stearic acid                2.0
Naphthenic oil              Variable
Wax                         1.5
DPPD                        1.5
TBBS                        1.5
Sulfur                      2.0
Total                       113.5
                     + carbon black
                            and oil
  Table 2 - NR mixing procedure

Time (min.) Action
      0     Add polymer
    0.5     Add oil, wax, zinc oxide, stearic acid, DPPD
    2.5     Add carbon black, turn water on
    6.0     Sweep and add sulfur, and TBBS
    7.5     Dump at 104 [degrees] C, turn off cooling water
  Table 3 - IIR compound formulation

  Ingredient                phr

  IIR                       100.0
  Carbon black              Variable
  Zinc oxide                5.0
  Sulfur                    1.5
  Paraffinic oil            8.0
  TMTD                      2.0
  MBT                       1.0
  Total                     117.5 +
                            carbon black
   Table 4 - IIR mixing procedure

Time (min.)    Action
      0       Add polymer
    0.5       Add 1/2 carbon black, zinc oxide, sulfur
    2.5       Add rest of carbon black and oil
    6.0       Add TMTD and MBT
   7.25       Dump

Dynamic tests were performed on a Rheometrics RDS II (10-10,000 g-cm transducer) at 23 [degrees] C with parallel plate mode and a specimen size of 8 mm diameter by 3.25 mm thick. Specimens were cured and subsequently bonded to the endplates with cyanoacrylate adhesive. In some of the tests involving swollen rubber samples, the disc specimen was not satisfactory and a torsion rectangular specimen 20 mm long with rectangular cross section of 12.5 mm by 6.25 mm was used. This specimen was gripped across the 6.25 mm width and oscillated about the axis parallel to its 20 mm length.

The following conditions were used for the dynamic tests because of the sensitivity of low strain properties to previous, strain history:

* Three-cycle conditioning step to [+ or -] 25% strain at 0.1 Hz;

* stress relaxation test at 20% shear strain for 30 sec.;

* strain sweep at 10 Hz from 0.1 to 20% strain;

* frequency sweep at 0.2% strain from 0.08 to 80 Hz.

Tear strengths of the rubber compounds were evaluated according to ASTM D-624 (Die C).

Results and discussion

The selection of carbon blacks evaluated in these experiments was made to cover the range of carbon black surface area and structure that are of major interest in engine mounts (figure 1). This includes several standard ASTM-grade carbon blacks designated by their appropriate N-numbers and several non-ASTM-grade carbon blacks that are designated by arbitrary CB-numbers. The CB-numbered carbon blacks include some that are commercially available and others that are experimental.

Eight of the blacks shown in figure 1 were evaluated in natural rubber formulations at five loadings of each carbon black; e.g., from 10 to 120 phr. The carbon blacks that were evaluated and the dynamic properties measured on samples cured from each formulation are shown in figure 2, where the data have been plotted as [k.sub.d]/[k.sub.s] vs. tan [Delta]. It is readily apparent in figure 2 that the dynamic performance of all of the compounds is well represented by the single exponential curve drawn through the data points. It is possible to move along the curve by changing loading of the carbon blacks or by changing the type of carbon black in the formulation, but moving off the curve is not easily done. It should be noted that increasing the loading or selecting a more reinforcing carbon black in the rubber compound shifts the dynamic performance to the right along the curve; i.e., to the more hysteretic end.

Data from the dynamic tests of natural rubber compounds are also displayed as functions of the analytical properties of the carbon blacks. The hysteresis of the compounds as measured by tan [Delta] is shown to depend significantly on the CTAB of the carbon black (figure 3) and to a lesser extent on the CDBP of the carbon black (figure 4). This is expected. The values of the [k.sub.d]/[k.sub.s] ratio were found to depend on the CTAB and CDBP in ways similar to those shown for tan [Delta]. The similarity between the dependencies of tan [Delta] and [k..sub.d]/[k.sub.s] of the rubber compound on the carbon black analytical properties was also expected because of the close relationship between tan [Delta] and [k.sub.d]/[k.sub.s] shown in figure 2 and described previously (ref. 2).

Although desired levels of tan [Delta] and [k.sub.d]/[k.sub.s] can be obtained with many different combinations of carbon black type and loading as seen in figure 2, other properties influenced by the selection and loading of carbon black must also be optimized to achieve acceptable performance in engine mounts. Tear strength is one example. In the present experiments, tear strengths of all the natural rubber compounds were evaluated at ambient temperature and at 100 [degrees] C. Since there was a direct correlation observed between tear values measured at the two test temperatures, only data from the 100 [degrees] C tests are considered here. The relationships between the tear results and the [k.sub.d]/[k.sub.s] results of natural rubber compounds containing three loadings (40, 60 and 80 phr) of selected carbon blacks are shown in figure 5 (with the more highly reinforcing carbon blacks at each loading giving both higher [k.sub.d]/[k.sub.s] and higher tear strength values). It is seen that the best tear values are achieved by selecting the lowest loading of carbon black in the rubber compound which also gives the required dynamic performance; i.e., [k.sub.d]/[k.sub.s] ratio. For example, at [k.sub.d]/[k.sub.s] ratios of 2.0 to 2.5, the best tear properties are achieved by using a rubber compound containing 40 phr of a more highly reinforcing carbon black. In the intermediate range of [k.sub.d]/[k.sub.s] ratios (e.g., 3.0 to 5.0), 60 phr of a more highly reinforcing carbon black in the rubber compound gives higher tear values. At high [k.sub.d]/[k.sub.s] ratios ([is less than] 5.0), the choices are more limited because a high loading of a highly reinforcing carbon black is needed to obtain acceptable dynamic performance.

Another analytical property for selecting carbon blacks for engine mounts is the width of the ASD of the carbon black. It has been reported (ref. 3) that increasing the width of the ASD of a carbon black reduces the hysteresis of rubber compounds. That is, carbon blacks with similar surface areas and structures but with broader ASDs provide lower hysteresis.

The application of this concept to engine mounts was evaluated by selecting carbon blacks with similar structures and surface areas but with significantly different ASDs. As described earlier, the width of the ASD of each carbon black was quantified by its D-ratio; i.e., the ratio of the width of the ASD curve at half-maximum height divided by its mode (peak value). Carbon blacks selected for evaluation of the effects of ASD on dynamic performance are shown in figure 6. These include several commercially available ASTM N-762 type 762 carbon blacks as well as experimental and non-ASTM-grade carbon blacks.

Dynamic properties of the carbon blacks with different ASDs were evaluated in the natural rubber compound at loadings of 25, 62.5 and 100 phr. Dynamic data shown as [k.sub.d]/[k.sub.s] vs. tan [Delta] in figure 7 are well represented by the single exponential curve fitted to all of the data. Furthermore, data points from the curve shown in figure 7 can be integrated over the lower (left) portion of the curve given in figure 2.

The effect of using carbon blacks with different ASDs is to move along the exponential curve but not off of it; this is the same as was observed for the broader range of carbon blacks shown in figure 2. At constant carbon black loading, the values of [k.sub.d]/[k.sub.s] are essentially unaffected by using carbon blacks with different D-ratios.

Another consideration for performance of engine mounts is the amounts of oil present in the rubber compound. Oil is often included in the formulation to function as a plasticizer in the rubber compound. In addition, oil can become swollen into a rubber isolator during service in an oily underhood environment. In both cases, the presence of the oil would influence the dynamic properties of the engine mount and, hence, its ability to give acceptable isolation performance.

The effects on dynamic properties of increasing oil content in natural rubber compounds were evaluated using N-330 and 330 CB-7 carbon blacks. Compounds containing these blacks, selected because they represent large differences in carbon black structure and surface area, were found to behave similarly with similar changes in oil concentration; i.e., both changes in oil concentration and in method of oil addition. Hence, only the data from the evaluations of the compounds containing CB-7 carbon black are given here.

The data for compounds in which the oil was added during mixing are shown in figure 8. Here, the dynamic data are plotted as tan [Delta] vs. [G.sub.s] in order to highlight separately the effects of adding oil on the hysteresis and on the static spring rate. Data for compounds containing 0 to 80 phr carbon black and no oil were fitted with a straight line. The data point for the compound containing 50 phr carbon black and no oil was interpolated from the experimental data shown. It is seen in figure 8 that adding oil to the compound causes essentially a horizontal shift of the data points to the left, which indicates that the effect of the oil is to lower the value of [G.sub.s] but not to appreciably influence tan [Delta] of the rubber compound. It follows from the observed effect of oil on [G.sub.s] that the main contribution of oil in the rubber compound is to dilute the polymer phase which, in turn, decreases the density of chemical and entanglement crosslinks. The Payne effect, which is mainly responsible for hysteresis (tan [Delta]) in filled compounds, is expected to depend only on the small state of micro-dispersion of the carbon black. Other than a somewhat lower viscosity of the polymer phase due to the oil addition in the compound, no changes in the state of micro-dispersion of carbon black (Payne effect) would be expected and, hence, little effect on hysteresis from addition of oil to the compound would be anticipated.

A second set of experiments was performed in which a comparison of the mechanisms for addition of oil to the rubber compounds was made. The same natural rubber compound, the same two carbon blacks, and similar levels of oil (25 phr) were added to the rubber compound. The samples in which all the oil was added during the mixing of the compound were compared with samples which contained the same amount of oil but which were. prepared by adding a portion of the oil during mixing of the compound (10 phr) and by swelling the remainder of the oil (15 phr) into previously vulcanized specimens. A different Rheometrics specimen was required because of swelling the oil into some of the samples, but this did not appreciably influence the values of the properties.

The characteristics of the dynamic properties of samples made by adding all the oil during compound mixing were similar to those described earlier even though different sample geometry was used for dynamic testing in the two sets of evaluations (compare figures 8 and 9). As seen earlier, the addition of all the oil during the mixing of the compound reduced significantly the values of [G.sub.s] without appreciably influencing tan [Delta] of the natural rubber samples. The effects of swelling a portion of the oil into the vulcanized compound are also shown in figure 9 (designated as "swollen") where the values of both tan [Delta] and [G.sub.s] are significantly reduced compared to the compound containing no oil. It is seen, for example, that the tan [Delta] and the [G.sub.s] values of the compound containing 80 phr carbon black and 25 phr oil, 15 phr of which had been swollen into the samples, are similar to the properties of the compound containing 50 phr carbon black but which had been prepared by adding all 25 phr of the oil during the compound mixing. The lower hysteresis observed in the oil swollen samples appears to result from an increase of the inter-aggregate distance of adjacent carbon black aggregates by the addition of the oil. The larger distance between aggregates in the swollen samples would reduce the Payne effect which, in turn, would reduce the hysteresis. The change in both [G.sub.s] and tan [Delta] caused by oil swelling into rubber vulcanizates adds complexity to engineering the long-term dynamic performance of mounts which are required to function in oily environments.

Dynamic performance of butyl compounds containing the carbon blacks shown in figure 1 was also determined. The specific carbon blacks and dynamic data from the butyl compounds plotted as [k.sub.d]/[k.sub.s] vs. tan [Delta] are shown in figure 10; as with natural rubber, the dynamic properties for carbon black loaded butyl rubber in figure 10 are well represented by a single exponential function. A comparison of the dynamic behaviors of the natural rubber and the butyl rubber compounds is shown in figure 11.

It is seen that the relationships for the two elastomeric compounds are similar but that the values of [k.sub.d]/[k.sub.s] and tan [Delta] are substantially higher for the butyl compounds. This behavior is expected because of the inherently higher hysteresis of butyl rubber compared to natural rubber. The difference in dynamic behavior between the two elastomeric compounds, shown in figure 11, demonstrates an important feature about rubber performance in engine mounts. In both the natural rubber and butyl rubber compounds, the choice of carbon black type or carbon black loading allows the design engineer to move up or down a single exponential curve but not off of it. To move off a [k.sub.d]/[k.sub.s] vs. tan [Delta] curve requires a change in the type of elastomer used in the engine mount (or other vibration isolator). In the present example, changing from natural rubber to butyl rubber allows movement off the curve in the direction of higher hysteresis (or higher [k.sub.d]/[k.sub.s]); and the converse is also observed. It is also important to note that the same blacks give significantly different values of [k.sub.d]/[k.sub.s] (or tan [Delta]) in the different elastomers. This provides the design engineer flexibility through the selection of an appropriate combination of carbon black type, carbon black loading and elastomer type to achieve the required value for [k.sub.d]/[k.sub.s].

The dependencies of tan [Delta] and [k.sub.d]/[k.sub.s] for butyl rubber compounds on the surface areas and structures of the carbon blacks were similar to those seen in natural rubber compounds except for the higher level of hysteresis (and [k.sub.d]/k.sub.s]) observed in the butyl compounds. In addition, the relationship between tear and dynamic properties for the butyl compounds was similar to that observed in natural rubber compounds; i.e., at the same [k.sub.d]/[k.sub.s], the tear strength is highest for compounds containing low loadings of highly reinforcing carbon blacks. The characteristics of the curves are similar to those seen for natural rubber (figure 5) except the overall range of tear strength for the butyl compounds was smaller.

Summary and conclusions

Dynamic properties of compounds containing wide ranges of carbon black types and loadings were shown to be well represented on a [k.sub.d]/[k.sub.s] vs. tan [Delta] curve by a single exponential curve. In compounds with a single kind of elastomer, changes in carbon black type or loading facilitate movement up or down the [k.sub.d]/[k.sub.s] vs. tan [Delta] curve but not off of it. Moving off the curve could be accomplished by changing the elastomer used in the rubber formulation to one of higher or lower hysteresis depending on the desired direction. Carbon blacks with aggregate size distributions of differing widths were found to have little effect on the [k.sub.d]/[k.sub.s] ratio (or on tan [Delta]) in a natural rubber compound. Carbon blacks with higher surface areas were found to-provide higher tear strengths of rubber compounds as expected. It was further shown that improved tear properties with equivalent dynamic performance could be achieved by using lower loadings of a more highly reinforcing carbon black in both natural rubber and butyl rubber compounds. Adding oil to a carbon black loaded natural rubber compound, either during mixing or by swelling it into the vulcanizate, reduced the [G.sub.s] of the compound. When the oil is added by swelling it into the vulcanizate, however, a reduction in the tan [Delta] of the compound is also observed.

References

[1.] A.R. Payne and R.E. Whittaker, Rubber Chem. & Tech., 44, 440 1971).

[2.] R.L. Warley, Rubber World, 213, 33 (March 1996).

[3.] G. Kraus, Angew. Makromol., Chem. 60161, 215 (1977).