Application of experimental design for optimization of physicochemical properties of the inorganic pigment, iron(III) silicate

Abstract

Optimization of physicochemical properties (bulk density, D n , capacity to absorb water, H w , capacity to absorb paraffin oil, H o ) of highly dispersed precipitated iron(III) silicate was performed, as related to temperature ( T), flow rate of sodium metasilicate solution ( v) and concentration of metal salts ( c). The analysis was performed in three consecutive stages: (1) testing of silicate properties as affected by individual variables of the process, (2) empirical optimization, (3) experimental corroboration of the optimization. At each of the stages, appropriate experimental design was set up, mathematical model was established to describe the studied properties of silicate and statistical analysis of results was performed. The studies permitted to define optimum variables of precipitation of highly dispersed iron(III) silicate from sodium metasilicate solution using salt solution. Increase in precipitation temperature was found to bring about a decrease in bulk density and increases in capacities to absorb water and oil by the formed iron(III) silicate. Increase in the flow rate of sodium metasilicate solution induced a decrease in capacities to absorb water and oil and an increase in iron(III) silicate bulk density. On the other hand, increased concentration of iron(III) sulphate solution resulted in increased bulk density and capacity to absorb water but a decreased capacity to absorb oil by iron(III) silicate. This showed that individual properties of the formed silicate are related in various, sometimes opposite, ways to conditions of silicate precipitation. In the study the potential was demonstrated of applying factorial and compound designs for optimization of processes. The experimental optimization involved in parallel three different physicochemical properties: bulk density, capacity to absorb water and capacity to absorb paraffin oil. The properly planned and conducted experiment was shown to determine optimum conditions for conducting the process in cases when models describing the tested properties exhibit no extremum while variables of conducting the process exert opposite effects on the studied properties.