The effect of silica structure on resilience.

The effect of silica structure on resilience

Previous investigations[1-3] and commercial application have shown that partial replacement of HAF or ISAF carbon black by precipitated silica can produce a significant increase in resilience with a related reduction in tan delta. This effect is illustrated by data (figure 1) taken from recent tread compounding studies of SBR and natural rubber compounds, where it is seen that the maximum change in 100 [degrees] C rebound occurs in a replacement range of 15 to 30 phr. Since silica has long been associated with stiffening effects, it was not immediately apparent why resilience should increase. The purpose of this article is to describe an investigation of the cause and probable mechanism of resilience enhancement by precipitated silica.

Procedures and materials

The experimental program was based on successive laboratory studies in which various materials known to alter silica behavior were added to natural rubber formulas (table 1). Separate studies, each of which included 6 to 8 compounds. with controls, were made over the course of a year.

The bulk of the work was based on a peroxide cure system to avoid the crosslinking interference normally produced by silica in sulfur-zinc oxide systems. Since previous programs had established an optimum range between 15 and 25 phr, silica content was maintained at 15 phr in the peroxide series and at 20 phr in the sulfur cured series.

Resilience was evaluated with the Goodyear-Healey pendulum rebound tester (ASTM D1054). Values obtained at 80-100 [degreees] C generally correlate with tan delta (higher rebound % equals lower tan delta). Dynamic modulus was determined on the Du 9900 thermal analyzer 983 DMA module at 1 Hz frequency with a temperature scan from 0 [degrees] C to 100 [degrees] C. As work progressed, the importance of low strain (20%) extension modules became apparent. This property, together with other stress strain data, was obtained on the United Elastomatic tensile tester. Crosslink state of cure represented by the difference between minimum and maximum rheometer torque values. The gradual rise of the peroxide cure curve allowed a selection of cure times among compounds to make comparisons at equal crosslink density. The abrupt form of the sulfur cure rheometer curves precluded this approach to equal crosslink comaprisons among the sulfur cured compounds. A concern for equal crosslink evaluations arises from the need to separate crosslink from filler contributions to resilience. As subsequent data will show, most silica modifiers have a substantial effect on crosslinking, negative with peroxide and positive with sulfur cure systems.

The precipitateed silica used in this work (designated HS200) was Hi-Sil 210, a non-dusting pellet form with primary particle size of 19nm and surface area of 150 m2/g. Mercapto silane was A189 and methyl silane A163. Zinc octoate was the commercial product, Octoate Z.

Results

Peroxide cured NR: Modifiers The initial comparison of silica and N330 (table 2) shows a decrease, rather than the expected increase, in resilience for a 15 phr replacement of N330 by silica. Part of this loss can be related to slightly higher hardness and stiffness (M20) with silica, but this is off-set by greater crosslinking in the silica compound.

Although the difference in rebound is statistically significant, it is confirmed by the equal tan delta values. The complete replacement of black by an equal volume of silica produces changes in hardness and stiffness too great to allow any dynamic property comparison.

These results indicate that a simple substitution of silica for black is not sufficient to produce the demonstrated increase in resilience, and that additional compound modification is necessary to achieve this effect.

Customary industrial compounding practice for resilients silica-black tread compounds employs a silane coupling agent, sulfur donor, ultra accelerator a combination of these. When the silane and donor modifiers in general use were added to the 15 phr silica contro formula, substantial improvement in resilience occurs (table 3). With the same modifiers, the resilience of the black control remains unchanged or, with DTDM, decreases. This difference between black and silica compounds suggests that increased resilience is a result, not of the modifier acting alone, but of its intentions with silica.

There are, of course, other modifying or activating materials which are known to react with silica to produce profound changes in rheological and vulcanizate properties. Soluble forms of zinc, glycols and amine derivatives as well as various ultra accelerators could be expected to influence resilience. Single point evaluations of representatives from these groups (table 4) in the peroxide natural rubber formual showed a generally low level of effectiveness. However, since resilience is strongly influenced by crosslinking, and since the extent of peroxide crosslinking is reduced by most modifiers, comparable rebound comparisons should be made at equal crosslink levels.

To meet this requirement, cure times for the control and several other compounds were reduced from the original 40 minutes to 20 or 40 minutes to afford a comparison at an equal crosslink state of 50 dN-m. These adjusted values (table 5) revealed that mercapto silane produced the largest increase in rebound. Significant improvement was found for both morpholino dithio benzothiazole (MDB) and dithio dimorpholine (DTDM).

Subsequent experiments confirmed the effectiveness of mercapto silane and DTDM. No significant activity took place with zinc dimethyl dithiocarbamate (ZMDC), morpholino thio benzothiazole (MBS), tetrabutyl thiuran disulfide (TBTD). zinc octoate, diethylene glycol (DEG), or hexamethylene tetramine (HMT).

Silica water content

Any examination of silica's influence on compound properties must consider the effect of lightly bound water, normally present at from 5 to 7%, depending on the relative humidityl. The effect of water can be evaluated by its removal, either by drying the silica before use of by mixing at high mixer temperatures. The results of both these procedures (table 6) are a substantial increase in resilience with no change in ODR crosslinks. Hardness and 20% modulus are correspondingly lower and 300% modulus rises slightly.

Silica acidity

The strength of the filler network or agglomeration is also affected by the intrinsic acidity (pKa) of the silica surface. The source of the Lewis acid sites is generally aluminum, and the degree of acidity depends on the hydration form and the position of the aluminum ion in the silica lattice[7]. The less acidic (pKa - 3.0) of two silicas which are equal in primary particle size provides lower M20 modulus and higher resilience (table 7).

Dynamic properties

Dynamic modulus data (table 8) reveal the difference in effectiveness between DTDM and mercapto silane. The previously noted decrease in M20 and hardness, which accompanies higher resilience, is reflected by lower storage and loss modulus.

Although both modifiers reduce E' and E", the drop in E" is greater for the silane. Thus, higher resilience is the result of lower loss modulus (E") rather than any increase in storage modulus (E').

Sulfur cured natural rubber

Further exploration of silica modifier effects on resilience was carried out in sulfur cured compounds to confirm the peroxide cure findings and to reveal changes in activity related to the cure system.

The test formula (table 1) is based on a normal sulfur cure system with silica content of 20 phr rather than the 15 phr evaluated with the peroxide cured series. Unmodified replacement of HAF by HS200 silica produces no significant change in resilience. In contrast to peroxide curing where silica provides higher crosslinking than HAF, here crosslinking and M300 both are reduced (table 9). It appears that any substantial increase in resilience will have to be obtained from silica modifications similar to those described above.

The four modifiers evaluated included methyl silane as well as mercapto silane, DTDM and zinc octoate. A summary of curing, viscosity and cured properties (table 10) for three concentrations of modifier demonstrated the expected increase in rebound (figure 2). In all cases, including DTDM and zinc octoate, the resilience versus concentration curve reaches a plateau at about 1 phr of modifier, or 5% of the silica content.

The two silanes afford a comparison of coupling (mercapto) and noncoupling (methyl) activity. Here the previously found correlation of high resilience with low M20 obtains, partially, only for mercapto functionality, notwithstanding a moderate increase in ODR crosslinks. This behavior, together with a slight decrease in Mooney viscosity, indicates that a substantial reduction in silica structure or agglomeration has occurred. It appears that the silica - polymer attachment provided by the silane sulfur link (confirmed by increased M300) leads to a more effective structure dissipation than is available from silica surface modification alone by methyl silane.

Although both DTDM and zinc octoate provide equal resilience enhancement, there are marked differences in their effect on cure, viscosity and physical properties. Increasing modulus, both M20 and M300, and hardness in the DTDM compounds correspond to the higher crosslink values contributed by the donor sulfur and obliterate any evidence of reduced structure. Zinc octoate presents opposite trends. Reductions in viscosity, hardness and M20 all offer evidence of reduced structure.

Further insight of zinc octoate behavior is gained by observing its activity in a non-silica formula (table 11). The lack of any discernible change in properties of the all-black compounds offers conclusive evidence that zinc octoate reacts exclusively with silica to produce the structure related changes noted.

Dynamic modulus data (table 12) indicate that increased rebound is most closely related to loss modulus (E") for DTDM, zinc octoate and mercapto silane. Significant reductions in E" correspond to higher rebound (figure 3). With methyl silane, E" was unchanged and storage modulus (E') rose. A major deviation from these patterns occurs with concentrations of zinc octoate higher than 0.3 phr where sharp reductions in E' and E" correspond to the previously noted changes in viscosity, hardness and low strain modulus.

These results generally confirm the peroxide cure findings that silica structure dissipation and polymer coupling act through a reduction in loss modulus, rather than an increase in storage modulus, to enhance resilience.

Discussion

To explore possible mechanisms by which silica modifiers and water removal increase resilience, it is helpful to examine the relationships between resilience and other vulcanizate properties.

It is not necessary to perform rigorous regression analyses to see that higher resilience is generally accompanied by lower 20% modulus and hardness. A plot of M20 vs. 100 [degrees] C rebound, which includes many of the modified and control compounds prepared for this project, confirms that a real correlation exists (figure 4). When these properties are plotted for compounds cured to equal crosslink levels (table 5), a linear relationship is obtained (figure 5).

Variation of low strain modulus in the absence of any compounding change in filler particle size and content (or plasticization) is generally explained in terms of filler structure[4]. Silica structure, unlike that of carbon black, is not a fixed material property but varies with surface modification during mixing and curing[5]. It can best be described as a filler to filler network held together by hydrogen bonded water.

Electron microscopy of HS200 silica reinforced NR illustrates this silica structure and the changes produced by mercapto silane modification (figure 7). The change from many large to fewer small agglomerates corresponds to increased rebound and decreased M20 modulus.

When the extent and strength of this network or agglomeration are reduced by removing the normal water content, a more energy efficient filler state is obtained. Less energy is irrevocably consumed in breaking down silica structure during deformation. The effectiveness of mercapto silane, MDB and DTDM can be related to their ability to replace water at the silanol interface and provide reaction sites more compatible with polymer than with silica.

Although the bonding ability of silane alkoxy groups to silica is well known, a similar explanation for the morpholine derivatives is not so obvious. The literature contains a reference[6] which may offer some support for morpholine-silica bonding. In this case, it was demonstrated that the carbonyl group of dimethyl acetamide reacts strongly with the surface silanols of partly dehydrated silica. More evidence of morpholine reactivity was seen in DSC analyses of silica-DTDM mixtures where a 10 [degrees] C shift in the melting endotherm of the mixture - which did not occur in an HAF control - indicated a probable reaction (figure 6).

The reduction in hardness and low strain modulus which accompany increased resilience will quite likely raise the question of a direct relationship between softening in general and resilience. Will plasticization by any means produce higher resilience? A series of peroxide cured silica-HAF compounds which includes additional increments of paraffinic oil sufficient to produce a significant reduction in hardness and 20% modulus provides at least a partial answer:

Oil, Phr     100 [degrees] rebound  M20      Hardness  Tan D
5 (Control)  84.1%                  .71 MPa  57        .086
8            84.4                   .62      55        .082
11           84.7                   .57      53        .084

Although the reduction in M20 is similar, the slight increase in rebound is well below that obtained from silica modification.

Summary and conclusions

Simple substitution of silica for carbon black does not increase resilience. Silica water removal, by drying or dissipation through the addition of silane or morpholine modifiers, is necessary to break down the characteristics silica agglomerates and thereby provide a filler structure of lower hysteresis.

Changes found in filler structure parameters such as low strain modulus and in electron microscopy provide a basis for the structure theory of resilience enhancement by silica. [Figures 1 to 7 Omitted] [Table 1 to 12 Omitted]

References

[1]N.L. Hewitt, Rubber World, November 1984. [2]N.L. Hewitt, Rubber & Plastics News, July 28, 1986. [3]N.L. Hewitt, paper presented at Rubbercon '88, Sydney, Australia, October 1988. [4]M.P. Wagner, H.J. Wartman, J.W. Sellers, Kautsch. Gummi Kunsist., 2D, 407 (1967). [5]N.L. Hewitt, Elastomerics, March 1981. [6]R.K. Ilcr, The chemistry of silica, Wiley, New York, 1979, p. 658. [7]M.J. Clark, paper presented at ASTM symposium on pesticide application, October 1982, Ft. Mitchell, KY.