A LOCAL COMPOSITION MODEL FOR LIQUID MIXTURE SHEAR VISCOSITY
Doctor of Philosophy
A predictive local composition model for multicomponent liquid mixture shear viscosity has been developed based on the Eyring and NRTL local composition models. Only binary equilibrium thermodynamic information is used in this model in addition to pure component data. The equilibrium thermodynamic data are used to compute the local compositions of the mixture. No mixture shear viscosities and no adjustable parameters are required. Binary interaction terms which appear in the expression can be obtained from a proposed correlation yielding a final equation devoid of any adjustable viscosity parameters. Liquid mixture viscosities and densities have been measured for 28 binary and 20 ternary systems as a function of composition between 15 and 50(DEGREES)C at ambient pressure. Systems were chosen in order to provide a range of constituent types to test the presented model. In terms of the property (xi) (TBOND) ln((eta)V), where (eta) is dynamic viscosity and V is molar volume, upon which the model presented herein is based, calculated values were compared to experimental data obtained from the literature and in this laboratory, yielding an average absolute deviation (AAD) of 1.0 % for 44 binary and 1.6 % for 20 ternary nonaqueous systems, the predictive equations must be modified for aqueous mixtures by addition of a purely empirical factor. While the theoretical basis for this modification is unknown, it does provide a predictive equation for aqueous systems which represents experimental data better than existing nonparametric correlations. For 14 binary and 4 ternary aqueous systems, the agreement between experimental and predicted (xi)-value was 2.8 % and 2.6 %, respectively, using this modified equation.