Phase Behavior of Mixtures of Bitumen, Solvent, and Water

Solvent assisted thermal processes for in situ bitumen recovery are a less energy and greenhouse gas intensive alternative to conventional thermal processes. The understanding and modeling of these processes is partly limited by a lack of phase behavior data for pseudo-ternary mixtures of bitumen, solvent, and water at reservoir conditions. This thesis is among the first comprehensive pseudo-ternary phase behavior studies at conditions relevant to potential SAGD operations. It focuses on the impact of water on the phase behavior of mixtures of bitumen with: 1) n-pentane at temperatures from 180 to 280°C; 2) dimethyl ether (DME) from 90 to 180°C and pressures up to 17 MPa. Saturation pressures (liquid/liquid-vapor and liquid-liquid/liquid-liquid-vapor boundaries), water solubility limits (liquid/aqueous-liquid boundary), and onset conditions (liquid/liquid-liquid boundary) were measured. Saturation pressure measurements were performed using constant composition pressure-step expansion method. Onsets were measured with three different methods: 1) from the intersection of liquid/liquid-vapor region with the constant pressure of the liquid-liquid/liquid-liquid-vapor region on a pressure-composition diagram; 2) from the change in the asphaltene content of the light liquid phase above the onset; 3) from visual observation of a solvent titration performed in a high-pressure microscope coupled with a PVT cell. The water solubility limits were also determined using the intersection method. The phase boundaries were first modeled with the Advanced Peng-Robinson equation of state (APR EoS). The APR EoS model with tuned binary interaction parameters matched saturation pressures, water solubility limits, and heavy phase onsets for all the systems with n-pentane generally to within the measurement uncertainty, except near and within the supercritical region. For DME systems, the model could be tuned to match the saturation pressures but not both the saturation pressures and onsets with the same interaction parameters. Therefore, the Modified Regular Solution (MRS) model for modeling onsets was adapted to accommodate polar solvents. New property correlations were developed for DME and water. The MRS model with tuned binary interaction parameters matched the onsets for all the systems generally within the experimental error.