An Approach for Verifying Filler Component In Rubber Compounds

Material engineers rely on various fillers and filler blends to define the physical, physiochemical and performance characteristics of rubber compounds. It is therefore sometimes necessary to do verification characterization of the filler component in production mixes and end products. The current paper explores the use of Thermogravimetric analysis (TGA) along with other known analytical and statistical tools for such purposes. TGA of rubbers usually involve thermal degradation of the polymer component in an inert environment to an equilibrium state followed by degradation in an oxidative environment. The later allows for determination of carbon black /graphite and mineral filler content.

The oxidative profile for N339, N550, N990, graphite, ground coal and blends, were investigated. These fillers were studied as free powders and as components of SKF’s experimental elastomeric recipes.  One early observation was that the polymer environment affected the degradation rates of the fillers. Hence the verification process should first involve generation of relevant baseline data. Studies done under isothermal conditions (640°C to 750°C) showed that for the carbon blacks (N339, N550, N990 and ground coal) the rate law da/dt = k was relevant to most of the reaction range (a represents the amount of conversion/degradation that took place). The respective rate constants responded uniquely to temperature changes, hence respective

Activation energies and Pre-exponential factor formed the basis for typing the different grades

Mineral filler verification involved XRF analysis of residues that were retrieved from TGA studies (oxidative conditions). The test samples included Calcium Silicate, Magnesium Silicate, (crystalline, amorphous, diatomaceous) silica, Alumino- Silicates, Calcium Carbonates, and Barium Sulfate. XRF data were further evaluated via Principal Component Analysis (PCA) to give “score” and “loading” projections based on mineral composition. XRF data on rubber recipes that contained known levels of various mineral fillers/blends were then incorporated into the PCA matrix. The coordinates of these observations within the PC1 /PC2 plane were then compared with the theoretical expectations. A significant correlation was seen and a predictor model is reported. More work needs to be done with other mineral blends to provide a broader application of the predictor model.