The use of gas hydrocarbon solvents in the recovery of heavy oil has been increased because of the advantages they have over the thermal methods under some reservoir conditions. The injection of a miscible solvent in the reservoir implies a mass transfer process which is governed by a diffusion coefficient. Consequently the measurement of the diffusion coefficient is extremely important. This, however, presents a significant amount of challenges in the laboratory and in the data analysis.
In spite of the importance of the diffusion processes not much effort has been put in the understanding and calculation of the gas-liquid diffusion coefficient. In a recent work Guerrero-Aconcha and Kantzas (2008) used the "Slopes and Intercepts" analytical technique to successfully obtain the diffusion coefficient of liquid hydrocarbons in heavy oil. However the technique involves the previous knowledge of the composition density relationship. Here a non-iterative finite volume method was used to obtain the diffusion coefficient dependent on concentration without previous knowledge of relations for the diffusion coefficient and density with concentration. Computed Assisted Tomography (CAT) was used to obtain the density profiles and back calculate the concentration-dependent diffusion coefficients.
The results agree very well with the theory of diffusion in binary mixtures. An empirical model was used to perform predictions on the studied systems. The results are the vital importance for VAPEX and cyclic solvent injection recovery processes.