Diffusion coefficients of model contaminants in dense CO2

Dense CO2 extraction is an emerging technology in cleaning and remediation. The successful design and implementation of processes based on this technology require the accurate determination of solute diffusivities in the systems. In this work, a Taylor dispersion apparatus was constructed for measuring binary diffusion coefficients of benzoic acid, biphenyl, and p-dichlorobenzene in dense CO2 at temperatures of 293.15, 298.15, 308.15, 318.15, and 323.15 K and pressures from approximately 71 to 171 bar. For correlation and prediction of binary diffusivity, the Enskog kinetic approach with density correction, the Liu-Silva-Macedo (LSM) model, and the Eaton-Akgerman (EA) model were evaluated. The Enskog approach gave satisfactory predictions in liquid CO2, but it may be inappropriate to extrapolate this correlation to the supercritical region. The LSM model based on the Lennard-Jones fluid theory gave poor diffusivity predictions in liquid CO2 but its accuracy was improved in the supercritical region. The EA correlation based on the rough hard sphere theory gave accurate predictions in both the liquid and supercritical CO2 regions. (C) 2000 Elsevier Science B.V. All rights reserved.