Mechanical properties of rubbers reinforced with porcelainite-derived precipitated silica.

Silica has been used in recent years as an alternative to carbon black for the reinforcement of rubbers, specifically for tire applications (refs. 1-3). Silica surface characterization and silane chemistry advances have improved the design of silica-filled elastomers (refs. 4-6), although the fundamental mechanisms of silica reinforcement are still not well understood (refs. 7 and 8). This article examines the mechanical properties of rubbers formulated with silica of differing structure. Ultra small-angle x-ray scattering is demonstrated as a technique to determine the primary particle and aggregate structure of the powders.

The materials studied were synthesized at the Dimona Silica Industries pilot plant located at Ben Gurion University of the Negev, in Beer-Sheva, Israel (ref. 9). The samples are a series of developmental precipitated silicas derived from porcelainite, a mineral rich in amorphous silica. The synthesis scheme follows the conventional synthesis of precipitated silica from water-glass solution, although new technologies have been developed that take advantage of using porcelainite as the starting mineral. The new process has reduced energy requirements and also yields precipitated calcium carbonate (PCC) at a second production line. The lowered energy consumption results from the low-temperature dissolution process employed to convert porcelainite ore to water-glass (refs. 10 and 11).

The synthesis process (figure 1) starts with the pulverizing of porcelainite mined from the Negev desert in southern Israel. The pulverization is followed by the formation of sodium silicate:

[Na.sub.2]O x [H.sub.2]O + nSi[O.sub.2] [right arrow] ([Na.sub.2]O)[(Si[O.sub.2]).sub.n] + [H.sub.2]O (1)

[FIGURE 1 OMITTED]

The sodium silicate or water-glass solution then goes through a carbonization process in which C[O.sub.2] is introduced with a bicarbonate-carbonate buffer. The result is a basic solution of Si[O.sub.2], calcium carbonate and sodium hydroxide:

(NaO)(Si[O.sub.2])n + C[O.sub.2] [right arrow] [Na.sub.2]C[O.sub.3] +n Si[O.sub.2] (2)

[Na.sub.2]C[O.sub.3] + Ca[(OH).sub.2] [right arrow] [Na.sub.2]O x [H.sub.2]O + CaC[O.sub.3] (3)

In the remaining steps of the process, the silica precipitate is separated, washed, filtered, neutralized and dried. The pH and processing conditions can be varied to modify the properties of the silica. The process is capable of making precipitated silica with surface areas that range from 60 [m.sup.2]/g to over 300 [m.sup.2]/g.

Production and marketing

Silica produced by the above process is marketed under the trademark Dimosil by Dimona Silica Industries (DSI). DSI's target date for commercial production is the second quarter of 2008. The design capacity of the plant is 40,000 metric tons/ year.

The flexibility of the synthetic process, which we demonstrate in this article, allows DSI to tailor silica to specific customer requirements. The initial product will be highly dispersing silica (Dimosil GT series), which is often associated with high performance tires and so-called "green" tires. DSI is planning to develop silica specifically tailored for winter tires with improved wet/ice skid traction. The DSI product portfolio will also include lower-surface-area powders designed for truck tires, having good reinforcing properties and improved abrasion resistance.

DSI recently developed a new product with enhanced surface reactivity that requires less coupling agent and reduces the compounding time in tire manufacturing. These reactive fillers will become the first products in the Dimosil GT-R line of cost-effective fillers.

Highly dispersing silica

Previously, we studied the properties of Dimosil developmental grade 288 and compared this powder with commercial highly dispersing silicas (ref. 10). Similarities were observed in the morphology of Dimosil 288, Zeosil 1165 (Rhodia) and Ultrasil 7005 (Degussa) as measured by small-angle light and x-ray scattering. Mechanical tests (figure 2) showed that Dimosil 288 is similar to Ultrasil 7005 in rubber compounds. The data shown in figure 2a were obtained for unvulcanized rubber. The stress-strain curves in 2b indicate that in vulcanized rubber, Dimosil produces higher modulus at low strain and lower ultimate elongation and stress compared to the Ultrasil sample. Figures 2c and 2d show the dynamic mechanical data for a temperature scan. The Dimosil compound had slightly lower hysteresis at 60[degrees]C and slightly higher hysteresis at -20[degrees]C. The storage modulus for the Dimosil sample is also lower at 60[degrees]C and higher at -20[degrees]C.

[FIGURE 2 OMITTED]

The current study was undertaken to determine the range of silica particle sizes achievable by the DSI synthetic scheme, with the goal of producing silica powders having morphologies tailored to specific applications. The synthetic protocol was adjusted to control the size of the primary particle, as well as the size and morphology of the aggregate. A series of samples with varying morphological characteristics was then tested in rubber compounds.

A series of powders having a range of surface areas and DBP numbers was chosen for this study. Small-angle x-ray scattering was used to characterize the morphology of the powders such as the primary particle size, the aggregate size, the aggregate fractal dimension and the aggregation number. The developmental powders were compounded into rubbers using model tire formulations. Mechanical and rheological properties were then measured.

Small-angle scattering 

Figure 3 shows a combined plot of small-angle scattering data obtained using light and x-ray instruments for one of the developmental samples. Three distinct structural levels are present, corresponding to the primary particle, aggregate and agglomerate features. The figure represents the variation in the scattered intensity with the scattering vector, q, which is defined by: q = 4[pi] sin ([theta]/2)/[lambda], where [lambda] is the wavelength of the incident radiation (x-ray, light) and [theta] is the scattering angle. By combining the scattering data obtained from light and x-rays it is possible to cover a q-range from [10.sup.-6] < q < 1.0 ([Angstrom.sup.-1]), which corresponds to a size range (~[q.sup.-1]) of ~10 Angstrom, to 100 [micro]m. The scattering data in figure 3 are fit using the unified approach developed by Beacuage (ref. 12) to obtain specific information about each structural level. Figure 4 shows such a fit to the ultra small-angle x-ray scattering data portion shown in figure 3.

[FIGURES 3-4 OMITTED]

Figure 4 identifies local power-law and Guinier regimes by the labels P (power-law) and Rg (Guinier). At high q (0.03 - 0.2 [Angstrom.sup.-1]), the scattered intensity decays as a power-law described by:

I(q) = B[q.sup.-4] with B = 2[pi]N[([DELTA][rho]).sup.2]S (4)

where N is the number density of primary particles in the scattering volume, [DELTA][rho] is the contrast (the difference in the electron density of the scattering material and the background air) and S is the surface area of a primary particle. The power-law exponent of -4 is called Porod's law (ref. 13) for scattering from a smooth surface. The data, therefore, confirm that the primary particle surface is, in fact, smooth.

Scattering in the Guinier region is governed by Guinier's law (ref. 14):

I(q) = [G.sub.i] exp (- [q.sup.2][R.sup.2]gi/3) with [G.sub.i] = [N.sub.i] [([DELTA][rho]).sup.2][V.sub.i.sup.2] (5)

where [N.sub.i] is the number density and [V.sub.i] is the volume for the object giving rise to level-i scattering (i = 1 for primaries and i = 2 for aggregates) and [R.sub.gi] is the radius of gyration of the objects giving rise to the scattering at level i. These Guinier regions show up on a log-log plot as transition regions or "knees" between power-law regions. Level 1 refers to the primary particles that appear at high-q (q [congruent to] 0.01 [Angstrom.sup.-1]). A second Guinier region is also present, indicated by [R.sub.g2], corresponding to the size of the silica aggregates. Between the two Guinier regimes is a power-law feature, as indicated by [P.sub.2], the power-law exponent obtained from the slope on a log-log plot. In this regime, the scattering follows a fractal dependence (ref. 15):

[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII.] (6)

where [d.sub.f] is the mass-fractal dimension ([P.sub.2] = -[d.sub.f]) and [B.sub.2] is the power-law prefactor for the aggregates (level 2). For a fractal cluster, the power law prefactor, [B.sub.2], is approximately:

[B.sub.2] = [G.sub.2][d.sub.f] / [R.sub.g2][d.sub.f] [GAMMA]([d.sub.f]/2]) (7)

[G.sub.2] is the Guinier prefactor defined in (5) and [GAMMA] is the gamma function (ref. 16).

The radius of gyration, [R.sub.g], which is obtained for the primary particles, can be correlated to a particle radius, when a spherical geometry is assumed, by:

[R.sub.Sphere] = [square root of 5/3] [R.sub.g] (8)

Beyond describing the morphology of the aggregates in terms of the mass-fractal dimension, [d.sub.f], the silica aggregates can be described by the degree of aggregation, z, or the number of primary particles that make up the aggregate. The degree of aggregation is calculated from the unified fit parameters by:

z = [G.sub.2]/[G.sub.1] (9)

where [G.sub.1] and [G.sub.2] are the Guinier prefactors for level 1 and level 2.

Materials and methods

Materials

The eight DSI silica samples chosen for this study had a range of specific surface areas and were believed to have different large-scale aggregate and agglomerate morphologies as well. Two commercial precipitated silica samples were also evaluated: Zeosil 1165 from Rhodia and Ultrasil 7005 from Degussa.

Methods

Ultra small-angle scattering (USAXS) measurements on the silica powders were performed using a Bonse-Hart camera covering a q-range of 0.0002 - 0.4 [Angstrom.sup.-1] (Beam line ID-33, Advanced Photon Source, Argonne National Laboratory). The x-ray beam was 0.4 mm high and 2-mm wide. Powder samples about 0.1 mm thick were held between two strips of Scotch Magic Tape (3M). The data were corrected for sample transmission and for background scattering arising from the tape and air. The slit-smeared data were desmeared with Indra software (available on-line at www.uni.aps.anl.gov) and analyzed with Irena software, also available from Argonne National Laboratory.

The tire recipe used for formulating the filled elastomer compounds is shown in table 1, where phr is grams per hundred grams of rubber.

Results and discussion

A series of samples was produced using the same reactor, but different synthesis conditions. The aim was to decrease the size of the primary particle while maintaining similar aggregate dimensions. Figure 5 shows the scattering curves for five different Dimosil samples, highlighting the variation in particle size and aggregate size achieved by changing the synthesis conditions. Table 2 lists the unified parameters for the eight Dimosil samples. The parameter z is the mean number primary particles in the aggregate. The level 2 fit values for sample B8 were obtained from a combined small-angle x-ray and light scattering curve, since the aggregate size fell in the light scattering regime.

[FIGURE 5 OMITTED]

One of the goals for this series of samples was to synthesize silica containing various particle sizes. Sample B1 had the smallest primary particle size followed by sample B3, with [R.sub.g] values of 96.6 Angstrom for sample B1 and 108 Angstrom for B3. Both of these samples are close in size to the commercial Zeosil 1165 ([R.sub.g] = 96.5 Angstrom) and Ultrasil 7005 ([R.sub.g] = 85.2 Angstrom) (ref. 10). The Dimosil samples showed a range of sizes, with B8 being much larger, while the rest of the samples had [R.sub.g] values that were between 90 and 460 Angstrom. The aggregate sizes for the B1 and B3 samples varied slightly from the aggregate size of the commercial samples. The aggregate size for Dimosil B1 was [R.sub.g] = 1,100 Angstrom, and for Dimosil B3 the aggregate size was [R.sub.g] = 1,590 Angstrom. There is also similarity in the degree of aggregation for Dimosil B1, B3, Ultrasil (z = 914) and Zeosil (z = 571) (ref. 10). No trend in the degree of aggregation with particle size is observed. Similar primary particle sizes have different aggregation numbers.

The series of silica samples was compounded using a model tire formulation. The uncured gums were then tested according to ISO 5794-2 1998. Table 3 shows the specific surface area (SSA) and DBP numbers for all of the samples. The B8 sample shows a specific surface area of 217 [m.sup.2]/g, but also shows a very large primary particle size that does not correspond to such a high surface area. This discrepancy seems to indicate that the B8 sample is porous at small size scales, which leads to an underestimation of particle size when calculated from gas adsorption data.

The uncured samples were tested using a Mooney viscometer and a Rubber Process Analyzer (RPA), which measures the dynamic mechanical properties as the samples cure. Figure 6 shows the results of these tests for the Mooney viscosity at 100[degrees]C, storage modulus, loss modulus and tan delta. The Mooney viscosity at 100[degrees]C increases as the number of particles in the aggregates increases. The same compounds were cured and tested, measuring tensile properties, tear resistance, hardness and wear resistance. The values from these tests are shown in figure 7.

[FIGURES 6-7 OMITTED]

Dimosil B1, B3 and B6 show similar tensile strength and elongation as the Zeosil and Ultrasil samples. The tear resistance values were also similar for the B1 and B3 samples compared to Zeosil and Ultrasil samples.

The stress-strain curves for the two commercial samples and the Dimosil samples are plotted in figure 8. The samples B1, B3 and B6 are plotted with the commercial silica samples, showing the similarity in the stress-strain behavior (figure 8a). As the primary particle size and degree of aggregation decreases for the Dimosil samples (table 1), the tensile elongation and ultimate stress decrease as well.

[FIGURE 8 OMITTED]

The Dimosil samples showed a wide variation in properties in both the uncured and cured states, which followed trends with particle size and degree of aggregation. Samples with the largest primary particle size showed the lowest mechanical properties, specifically samples B8 and B4. The trends with degree of aggregation are of interest, since these observations represent the first time that mechanical properties of silica-filled rubber compounds showed a correlation with the degree of aggregation. For carbon black and silica fillers, the DBP number has historically been used to correlate the "structure" of fillers with mechanical properties. The degree of aggregation, however, is measured directly from small angle scattering. Presumably, DBP reflects both the degree of aggregation and the fractal dimension.

Conclusion

The ability to make precipitated silica using a new synthesis protocol is demonstrated as well as the ability to vary the structure of the silica during synthesis. The developmental silica samples show differing structure at the primary particle, aggregate and agglomerate size scales, with samples B1 and B3 having the highest aggregation number. The similarities between the B1 and B3 samples were also observed in the testing of the compounded samples in both the uncured and cured state. While one of the goals was to produce silica with similar rubber reinforcement properties to that of commercial highly dispersible silica, this study also demonstrates the ability to control the structure through a synthetic protocol designed to control particle sizes.

Small-angle x-ray scattering was used in this study to characterize the structure of the silica samples at three different length scales. The degree of aggregation showed some correlation with the mechanical properties and appears to be a potential feature that correlates the structure of silica to the reinforcing behavior in elastomers.

This article is based on a paper presented at a meeting of the Rubber Division, ACS (www.rubber.org).

References

(1.) Y. Bomal, Y. Chevallier and P. Cochet (to Rhodia Chimie), U.S. Patent No. 6,702,888 (2004).

(2.) J.B. Donnet, Rubber Chemistry and Technology, 71, 323 (1998).

(3.) H. Esch, U. Gorl, R. Kuhlmann and R. Rausch (to Degussa AG), U.S. Patent No. 6, 977,065, (2005).

(4.) J. Frohlich, W. Niedermeier and H.D. Luginsland, Composites Part A--Applied Science and Manufacturing 36, 449 (2005).

(5.) H.D. Luginsland, J. Frohlich and A. Wehmeier, Rubber Chemistry And Technology 75 (4), 563 (2002).

(6.) J.E. Mark, B. Erman and R.E. Frederick, Science and Technology of Rubber, 3rd Ed., p. 743, Elsevier Academic Press, Burlington, MA, 2005.

(7.) M. Kluppel, Adv. Polym. Sci. 164, 1 (2003).

(8.) D.W. Schaefer, T. Rieker, M. Agamalian, J.S. Lin, D. Fischer, S. Sukumaran, C.Y. Chen, G. Beaucage, C. Herd and J. Ivie, Journal Of Applied Crystallography 33, 587 (2000).

(9.) E. Feinblum (to R&D Silicate Products Ltd.), U.S. Patent No. 5,302,364 (1992).

(10.) D.W. Schaefer, D.J. Kohls, R. Kosso and A. Vorobiev, Paper 69, "New generation precipitated silica for elastomer reinforcement," presented at the 107th Technical Meeting of the Rubber Division, American Chemical Society, Cincinnati, OH, October, 2006.

(11.) I. Tsyganova, Chemical Week, Sept. 27, 2006, p. 20.

(12.) G. Beaucage and D. W. Schaefer, Journal Of Non-Crystalline Solids 172, 797 (1994).

(13.) G. Porod, Small-Angle X-Ray Scattering; Academic Press: London, 1982.

(14.) A. Guinier and G. Fournet, Small-Angle Scattering of X-Rays; Wiley: New York, 1955.

(15.) D.W. Schaefer, J.E. Martin and K.D. Keefer, Journal De Physique 46, 127 (1985).

(16.) G. Beaucage, Journal Of Applied Crystallography 29, 134 (1996).

Doug J. Kohls and Dale W. Schaefer (email: [e-mail]), University of Cincinnati, and Raissa Kosso and Irina Tsyganova, Dimona Silica Industries (www.dsilica.com)

Table 1--rubber formulation

Ingredient                             PHR

Buna VSL 5025-1 HM (oil extended)    120
Buna CB 25                            25
Silica filler                         80
Si 75 (TESPD)                          6.3
Edenor C 18 98-100 (stearic acid)      2.5
Enerthene 1849-1 (mineral oil)         3.5
Zinc oxide active                      2.5
Vulkacit D/C (DPG)                     2
Vulkacit CZ/C (CBS)                    1.7
Sulfur 90/95 Chancel                   2.1

Table 2--unified fit parameters obtained for eight Dimosils

Level 1                                 B1                  B2

[R.sub.g] (Angstrom)                  96.6                 166
P                                        4                   4
G                                      701               4,290
B                         7.28x[10.sup.-5]    3.77x[10.sup.-5]

Level 2
[R.sub.g] (Angstrom)                 1,100                 958
P                                     2.40                2.85
G                          4.25x[10.sup.5]     6.47x[10.sup.5]
B                         8.68x[10.sup.-2]    9.50x[10.sup.-3]
z                                      606                 151

Level 1                                 B3                  B4

[R.sub.g] (Angstrom)                   108                 464
P                                        4                   4
G                                    1,720              42,900
B                         1.14x[10.sup.-4]    1.06x[10.sup.-5]

Level 2
[R.sub.g] (Angstrom)                 1,590               2,340
P                                     2.11                3.37
G                          1.23x[10.sup.6]     4.63x[10.sup.6]
B                         5.23x[10.sup.-1]    1.62x[10.sup.-4]
z                                      716                 108

Level 1                                 B5                  B6

[R.sub.g] (Angstrom)                   214                 143
P                                        4                   4
G                                    7,760               2,230
B                         2.43x[10.sup.-5]    3.32x[10.sup.-5]

Level 2
[R.sub.g] (Angstrom)                 1,030                 992
P                                     3.20                2.84
G                          9.71x[10.sup.5]     6.52x[10.sup.5]
B                         1.37x[10.sup.-3]    8.60x[10.sup.-3]
z                                      125                 293

Level 1                                 B7                B8 *

[R.sub.g] (Angstrom)                   240               3,500
P                                        4                3.73
G                                    7,740     1.77x[10.sup.8]
B                         1.94x[10.sup.-5]    1.71x[10.sup.-5]

Level 2
[R.sub.g] (Angstrom)                 1,840              38,400
P                                     2.51                2.87
G                          1.76x[10.sup.6]    1.63x[10.sup.10]
B                         4.54x[10.sup.-2]    1.71x[10.sup.-5]
z                                      227                  92

* Level 2 fit values for sample B8 were obtained from light scattering
data.

Table 3--specific surface area and DBP
measurement results

Sample                B1     B2     B3     B4     B5     B6     B7

SSA ([m.sup.2]/g)      1    114    122    178    209    134    170
DBP (g/100g)         294    294    252    244    265    233    216

Sample                B8    Ultrasil    Zeasil
                                7005      1165

SSA ([m.sup.2]/g)    217         178       164
DBP (g/100g)         160         238       232