Prediction of binary diffusion coefficients in supercritical CO2 with improved behavior near the critical point

In this work, a predictive model for binary diffusivities at infinite dilution (D-12) in SC-CO2 is proposed. It combines two terms - background and singular - with the objective to represent D-12 accurately not only far but also near the critical point, where critical enhancement is always observed. The model provides an average error of 6.20% for a large database including 149 systems and 4469 data points over wide ranges of temperature and pressure. The models selected for comparison (Wilke-Chang, Scheibel, Lusis-Ratcliff, Lai-Tan, Tyn-Calus and Reddy-Doraiswamy) achieve scattered and biased results, with average errors from 11.62% to 75.17%. In the whole, the new model exhibits an excellent performance for any kind of molecules in terms of size, molecular weight, polarity and sphericity, in all critical region. In order to help interested readers, a spreadsheet for the calculation of D-12 is given in Supplementary data. The input data is: temperature, pressure, CO2 viscosity, and solute properties (acentric factor, critical constants, molar volume at normal boiling point, and molecular weight - given in this paper for the systems studied). (C) 2014 Elsevier B.V. All rights reserved.