Summary Dimethyl ether (DME) was investigated as a potential additive to steam to improve steam-assisted gravity drainage (SAGD) in a previous simulation study. The main objective of this research is to compare DME with n-hexane in terms of the capability of viscosity reduction for Athabasca bitumen. In addition, new experimental data are presented for bubblepoint pressures, densities, and viscosities of Athabasca bitumen and its mixtures with DME and n-hexane. Results show that DME results in slightly higher viscosity than n-hexane when they are mixed with the same Athabasca bitumen at a given pressure, temperature, and molar concentration. For example, the equimolar mixture of DME with Athabasca bitumen is 79 cp, and that of n-hexane with the same bitumen is 49 cp at 328 K and 60 bar. However, the two solvents are equivalent as diluent at temperatures higher than 380 K. For example, the difference is approximately 1 cp at 382 K and 35 bar between the equimolar mixture of Athabasca bitumen with DME and that with n-hexane. The viscosity data measured for bitumen/n-hexane mixtures and bitumen/DME mixtures in this research were correlated with three different viscosity models: a modified Arrhenius model, the power-law model, and the Walther (1931) model. The viscosity data were well-correlated with the modified Arrhenius model, but not with the original Arrhenius (log-linear mixing) rule. The modified Arrhenius model can be used directly with a commercial simulator. Liquid/liquid separation for solvent/bitumen mixtures, which occurred for n-butane/Athabasca bitumen in Gao et al. (2017), was not observed for any of the DME/bitumen and n-hexane/bitumen mixtures in this research. The highest solvent concentration in this study was 80 mol% DME for the DME/bitumen system and 92 mol% n-hexane for the n-hexane/bitumen system.
Read full abstract