Estimation of the diffusion coefficient of a model food dye (malvidin 3,5-diglucoside) in a high pressure CO2+methanol system

Experimental data for the diffusion coefficient of malvidin 3,5-diglucoside in carbon dioxide and methanol as cosolvent at high pressure are compared with the values estimated by predictive equations. Those equations studied are based on the Stokes–Einstein theory and the rough-hard-sphere theory. The values predicted using these equations are, in general, higher than experimental values and these differences become more marked as the percentage of methanol in the solvent system is increased. The best fit was obtained with the Wilke–Chang, Scheibel, Hayduck–Minhas, and Dymond equations when the cosolvent is present at 5%. Correlation of the experimental data with the Stokes–Einstein equation allows to estimate the diffusing solute radii. The increase in these radii with the percentage of cosolvent is related to the formation of solute/methanol clusters. Finally, modifications of the different equations from the Stokes–Einstein model and the rough-hard-sphere theory are proposed.