RPA-studies into the silica/silane system.

The addition of fillers to rubber compounds has a strong impact on the static and dynamic behavior of rubber samples. Besides the strain-independent contribution of the hydrodynamic effect, the filler-to-rubber interaction and the crosslinkage of the matrix to the modulus, the dynamic modulus G* shows a strong strain dependency at low strains. This stress softening at small deformations, also known as the Payne effect (ref. 1), plays an important role in the understanding of reinforcement mechanism of filled rubber samples and can be attributed to the breakdown of the filler-filler network. In figure 1, the different contributions to the complex modulus G* of filled rubber Samples are depicted (ref. 2). The theory of the dynamic behavior of filled rubber samples is reviewed by Medalia (ref. 3).

[ILLUSTRATION OMITTED]

The aim of this article is to give an insight into the reinforcement of the silica/silane filler system by analysis of the strain dependency of the complex modulus G* and the loss factor tan [Delta]. The Rubber Process Analyzer (refs. 4 and 5) allows a very reliable and detailed investigation of this dynamic behavior in a much easier and faster way than the common dynamic analysis of vulcanizates. The main focus of this work lies on the influence of the silane modification of silica on the dynamic properties. In the case of carbon black, the filler polymer interaction is mainly of a physical nature (physisorption), whereas in the case of bifunctional silanes as coupling agents, a chemical bond between the silica and the rubber is established.

RPA measurement

The RPA is a vulcameter that is able to measure the complex modulus of rubber compounds under shear deformation. It can be operated with strain and frequency sweeps in wide ranges of strain amplitude and temperature. Compounds can be studied in the uncured (green compound), as well as in the cured state. Vulcanization can be performed with the same sample used for the green compound analysis, and therefore a good insight into several characteristics of the rubber compound is obtained. Furthermore, it is possible to measure samples several times in a row to determine the stress softening and relaxation processes. The possible settings for the measurement are given in table 1. Due to the biconical geometry of the die, the shear rate in the sample is nearly constant over the whole test specimen. The most appropriate strain sweeps for filled vulcanizates are at 1.6 Hz in a range of 0.28 to 42% or 1.0 Hz in the range of 0.28 to 100%.

Table 1 - possible settings for the RPA
measurement

Range of measurement
Temperature range:          50 [degrees] C to 230 [degrees] C
Frequency:                            0.002 - 33 Hz
Strain (SSA):               [+ or -] 0.28% to [+ or -] 1,250%
(Single strain amplitude)
Shear speed:                         <=[30s.sup.-1]
=>
Theoretical:
Frequency [Hz]                        max. SSA (%)
0.33                                      1,250
0.83                                        630
1.66                                        140
8.33                                         42
33.3                                         14
Practical
1.6                                          42
1.0                                         100

Experimental

All investigated rubber compounds are based on a solution SBR/high cis-1,4 BR formulation according to table 2. The mixing procedure is given in table 3. The high dispersible Ultrasil 7000 GR was used for the silica; this should eliminate the influence of the degree of dispersion on the Payne effect.

Table 2 - basic formulation

Stage 1
S-SBR             Styrene 25%                           96
                  Vinyl 50%
                  Arom. oil 37.5 pbw.
BR                cis-1, 4B>96%                         30
Silica            N2 surface [175m.sup.2]/g             80
Silane                                               Variable
Other chemicals   ZnO 3; stearic acid 2; arom. oil
                  10; 6PPD 1.5; wax 1
Stage 2
Batch stage 1

Stage 3
Batch stage 2
DPG                                                      2
CBS                                                      1.5
Sulfur                                               Variable
Table 3 - basic mixing procedure

Stage 1   Intermix GK 1.5E
0 - 1'    Polymers
1 - 3'    1/2 silica, ZnO, stearic acid
          arom. oil, silane
3 - 4'    1/2 silica, 6PPD, wax
4'        Sweep
4 - 5'    Mix
5'        Sweep
5 - 6'    Mix and dump

Stage 2
0 - 2'    Batch stage 1
2 - 5'    Remill at 150 [degrees] C

Stage 3
0 - 2'    Batch stage 2, accelerators
          sulfur
2'        Dump and homogenize on an open mill

The different silanes used in this study are shown in table 4. The monofunctional silanes can only react with the silanol groups on the silica surface and make the silica hydrophobic, whereas the bifunctional silanes can react with the silica and the rubber. The reaction and processing of these silanes has been reported (refs. 6-10). The disulfane silane TESPD with [S.sub.2] [is greater than] 80% reacts similar to VP Si 75 and was used for this study. Without limitations, all results achieved with TESPD are also relevant for VP Si 75.

Table 4 - types of silanes and their molecular
weight

Monofunctional silanes                                           MW
                                                               (g/mol)
PTES                     Propyltriethoxysilane                   206
OTES                     Octyltriethoxysilane                    276
HDTES                    Hexadecyltriethoxysilane                388
TESPS                    Bis-(triethoxysilylpropyl)sulfide       442

Bifunctional silanes
TESPT                    Bis-(triethoxysilylpropyl)              539
                           tetrasulfane, Si 69
TESPD                    Bis-(triethoxysilylpropyl)disulfane     478

The RPA strain sweep analysis was performed at a temperature of 60 [degrees] C, a frequency of 1.6 Hz and a strain amplitude in a range of 0.28-42% (SSA). For vulcanizates, strain amplitudes higher than 42% at 1.6 Hz are not used in this work, due to a possible slippage of the sample in the die. For each strain, 13 cyclic strains were applied and the monitored data are the the average values of the last 10 cycles.

Results

The silica-silica network

It is well known that silica as a filler itself shows a very strong filler networking due to its polar character and the ability to form hydrogen bonds (ref. 11). In order to study this strong network, two silicas with different surface areas were investigated in a SBR/BR formulation (70 phr loading, without silane) and compared to the carbon blacks N 234 and N 356. The CTAB and DBP values of the silicas and the carbon black are given in table 5.

Table 5 - CTAB and DBP of the investigated silicas
and carbon black

                      CTAB                  DBP
                   [m.sup.2]/g           [ml/100g]

Ultrasil VN 2 GR       125       [approximately equals] 185
Ultrasil 7000 GR       160       [approximately equals] 235
N 234                  119                              125
N 356                   95                              145

The degree of the filler network corresponds to the G* value at low strains. As can be seen in figure 2, the filler-filler interaction of the 7000 GR with the high surface area of 160 [m.sup.2]/g is significantly stronger than the one of VN 2 GR. In the case of carbon black N 356 with the lower CTAB surface compared to N 234 has a lower filler network. Furthermore it can be seen that the silica network starts to break down at higher strain amplitudes, and this breakdown takes place over a much wider range of amplitudes than in the case of carbon black. This can also be seen in the course of the tan [Delta] vs. strain. In the case of the carbon blacks, the maximum energy dissipation, due to the breakdown of the filler network, is reached at about 7% strain, whereas even at 42% strain the maximum is not reached in the case of the silicas. The breakdown of the interaggregate bonding of the silica over a broad range of strain amplitudes indicates that this interaction differs strongly from the van der Waals bonding of carbon black aggregates (ref. 12).

[GRAPH OMITTED]

For carbon blacks, the G* value at high strains is predominantly affected by the in-rubber structure corresponding to the DBP value. In the case of 7000 GR, the slightly higher G* at 42% strain compared to the VN 2 GR has to be attributed to the stronger filler networking which has not diminished as much as for VN 2 GR. The influence of the silanol group density and surface activity of the silica is not referred to here.

Silanization of silica

The modification of the polar silica surface with the apolar silane, the so called silanization reaction, leads to hydrophobation of the surface. This hydrophobation of the silica surface with a monofunctional alkylsilane reduces the filler network strongly and makes the silica compatible with the rubber matrix. This reduction depends not only on the degree of hydrophobation (e.g. amount of silane [ref. 13]), but also on the kind of silane. The decrease of G* under strain can be correlated to the processability of the compound under shear forces, e.g., an extruder (ref. 14).

The three alkyl silanes PTES, OTES, HDTES and the thioether silane TESPS were added equimolarly (6.5, 8.7, 12.2, 7 phr with 2 phr [S.sub.8]). From the point of sterical hindrance and polarity, TESPS is very similar to Si 69, but does not undergo a reaction with the polymer. As can be seen in figure 3, an increase in the alkyl chain length leads to a lower Payne effect - the hydrophobation improves.

[GRAPH OMITTED]

As can be seen, the addition of HDTES to a silica compound results in an outstanding low Payne effect, which makes this alkylsilane useful as a very efficient processing aid. Two explanations for the loss of silica-to-silica contacts may be stated: a) the long alkyl chains cover the silica surface very efficiently, or b) the long alkyl chains form self-orientated layers that shield the surface. All the other silanes show a significantly higher filler networking at this dosage. Compared to PTES, with the highest G* values at low strains, the use of TESPS, which is similar to Si 69, gives a quite good degree of hydrophobation.

The sequence of the Payne effect of these samples is principally not changed after curing, but the absolute values differ (figure 4). At 7% strain, which corresponds to an angle of 0.5 [degrees], the filler network has not broken down completely (except for the compound with HDTES); only high strains can achieve the almost complete destruction of the filler network. But vulcameter analyses, e.g., with the MDR, are normally performed at 0.5 [degrees], which means that not only the crosslinking density, but also the filler-to-filler interaction are measured. In order to correlate the crosslinking density with the [Delta] torque-value, it can be recommended to measure the vulcanization behavior of silica-filled rubber samples at 3 [degrees] (42% strain) - in the case of carbon black, 0.5 [degrees] is much better.

[GRAPH OMITTED]

As demonstrated in figure 5, heating of an unaccelerated compound (30 minutes at 165 [degrees] C) leads to a strong increase of the filler network (a [right arrow] b), but the stress softening effect of the applied strain sweep leads to much lower G* values, when the measurement is repeated twice (b [right arrow] c). A second heat treatment results in a reformation of the filler network, which can be seen in the increase of the G* (c [right arrow] d), but the effect is less strong than in the first heat treatment. Again, the stress softening effect leads to a considerable decrease of the Payne-effect (d [right arrow] e) on the level of the first stress softening cycle. This means that the history of heat and shear of the rubber sample influences the Payne-effect and therefore also the results of the RPA measurement. The hysteresis effect is well known in literature (ref. 14) and has to be taken into consideration when the RPA curves are interpreted. It is advisable to let the sample relax and be conditioned before the RPA analysis is started, e.g., 3' at 60 [degrees] C in the closed die, and to apply two strain sweeps.

[GRAPH OMITTED]

The degree of the stress softening and heating effects depends on the quality of hydrophobation. In the case of HDTES, these effects are nearly not observable, because the formation of a silica network is suppressed very efficiently.

Rubber-to-rubber and filler-to-rubber crosslinking

The crosslinking of the matrix results in a constant increase of the modulus G* at all strains (see figure 1), and the tan [Delta] decreases. The degree of the matrix crosslinkage can also be seen in the RPA. The delta G* at higher strain amplitudes of the green compound and the vulcanizate (figure 6; c-a) correlates with the matrix crosslink - the formation of filler-to-rubber crosslinks is ruled out by the use of TESPS. As expected, the tan [Delta] is lowered significantly by the matrix crosslink.

[GRAPH OMITTED]

The filler-to-rubber coupling can be observed by comparing the course of G* vs. strain of compounds with mono- and bifunctional silanes, e.g., PTES and Si 69. In figure 7, the RPA curves of SBR/BR compounds with silica, silica modified with PTES and Si 69, respectively (6.4 phr, 1.5 phr [S.sub.8]), as well as the hydrophobic silica Sipernat D17, are shown.

[GRAPH OMITTED]

The G* curves of 7000 GR + PTES and Sipemat D17 show at high strains the same G* value, whereas the G* curve of 7000 GR + Si 69 is significantly higher at high strains. This higher G* value can be attributed to the filler-to-rubber coupling, assuming a constant crosslink density. The effect of the filler-to-rubber coupling with Si 69 can also be seen in the lowered tan [Delta] curves of figure 7. The silica network of the 7000 GR is not completely broken down at 42% strain.

Pre-scorch during mixing

It is known that mixing of silica/Si 69 batches at excessive temperatures ([is greater than] 155 [degrees] C) leads to a significant pre-crosslinking in the non-productive mixing stages (ref. 10). This pre-scorch results in a deteriorated compound processing and significantly changed vulcanizate data (refs. 8 and 9). The pre-crosslinking can easily be observed by RPA measurements of the tan [Delta] at low strains in the green compound.

To demonstrate this effect, S-SBR/BR compounds with Si 69 and the disulfane silane TESPD ([S.sub.2] [is greater than] 80%) were mixed at different dump temperatures. As can be seen in figure 8, the tan [Delta] of the Si 69 compounds decreases strongly at the high mixing temperature of 166 [degrees] C (the modulus at high strains increases). This is due to the formation of crosslinks that makes the viscoelastic state of the green compound more elastic; a decision between a filler-to-rubber coupling or a matrix crosslink due to a sulfur donation of the Si 69 can not be made in this measurement. In the case of the disulfane silane TESPD, this decrease of tan [Delta] with increasing mixing temperature is much less. This means that the use of a disulfane silane reduces strongly the tendency to a pre-scorch during mixing.

[GRAPH OMITTED]

Silanization vs. mixing time

The progress of the silanization reaction, respective to the decrease of the Payne effect, with preceding mixing time has been followed by RPA. These studies were performed in an S-SBR/BR formulation using a 4 L mixer, and TESPD was chosen as silane to suppress the influence of a possible pre-scorch on the modulus. At fixed times, samples were taken from the batch and analyzed by RPA. The mixing procedure is: 0-1 min. polymers; 1-2 min. one-half silica, ZnO, stearic acid, oil; 2-3 min. one-half silica, TESPD, wax, 6PPD; 3-15 min. mixing and taking samples. These measurements were done at three different dump temperatures (120, 145 and 170 [degrees] C); the results are visualized in figures 9-11.

[GRAPHS OMITTED]

The comparison of the three figures demonstrates that the rate of the silanization reaction rises strongly with increasing mixing temperature. Even after 15 minutes of mixing at 120 [degrees] C, the silanization is not completed (figure 9), while at 145 [degrees] C and 7-8 minutes, no further significant decrease of the Payne effect is observed (figure 10). Mixing at high temperatures (165 [degrees] C) leads to a very fast silanization reaction. In correlation with earlier findings, at too high mixing temperatures the tan [Delta] is slightly lowered, i.e., the quality of hydrophobation is somewhat deteriorated (valid for green compounds: high tan [Delta] [equivalence] low viscosity).

A similar experiment was carried out with Si 69 at 170 [degrees] C. As described earlier, this polysulfane silane causes a prescorch at high temperatures. As can be seen in figure 12, the tan [Delta] is, as expected, lower than in the case of the TESPD (figure 11). Mixing Si 69 at lower temperatures leads to tan [Delta] values similar to the corresponding TESPD experiments described above.

[GRAPH OMITTED]

Conclusion

In the characterization of silica/silane reinforcement systems, the RPA measurement can be used:

* To study the influence of the silanization of the silica on the reinforcement;

* to estimate the compound processability;

* to determine the pre-scorch and scorch problems during mixing; and

* to optimize the mixing procedure of silica filled compounds.

References

(1.) A.R. Payne and R.E. Whittaker, Rubber Chem. Technol. 44, (1971) 440.

(2.) A.R, Payne, Rubber Plast. Age 42, (1961) 963.

(3.) A.I. Medalia, Rubber Chem. Technol. 51, (1978) 437.

(4.) H. Pawlowski and J. Dick, Rubber World 6/1992, 35.

(5.) J. Frohlich, H.-D. Luginsland and W. Niedermeier, "Reinforcement mechanism in the rubber matrix by active fillers," ACS Rubber Division, April 2000.

(6.) A. Hunsche, U. Gorl, A. Muller, M. Knaack and Th. Gobel, Kautsch. Gummi Kunstst., 50 (1997) 881.

(7.) U. Gorl, A. Hunsche, H.G. Koban and Th. Lehmann, Kautsch. Gummi Kunstst., 51 (1998) 525.

(8.) U. Gorl, J. Munzenberg, D. Luginsland and A. Muller, Kautsch. Gummi Kunstst., 52 (1999) 588.

(9.) H.-D. Luginsland and A. Hasse, Tire Technology, March 2000, 52.

(10.) H.-D. Luginsland, Kautsch. Gummi Kunstst., 53 (2000).

(11.) J.H. Bachmann, J.W. Sellers, M.P. Wagner and R.F. Wolf, Rubber Chem. Technol. 32, (1959) 1,286.

(12.) S. Wolff, M.-J. Wang and E.-H. Tan, Kautsch. Gummi Kunstst., 47 (1994) 102.

(13.) S. Wolff, Kautsch. Gummi Kunstst., 34 (1981) 280.

(14.) J.S. Dick, Rubber World 1/1994, 19.