This paper examines the role of molecular diffusion in the mobilization of waterflood residual oil. A numerical model that simulates the swelling of residual oil blobs by CO, diffusion through a blocking water phase is developed. The diffusion times calculated with the model are compared with those calculated in laboratory micromodel and coreflood experiments. The comparison shows that diffusion plays a major role in the mobilization and recovery of waterflood residual oil and that high unit or local displacement efficiencies are achieved when there is sufficient time for diffusion to swell the oil significantly. For the length scales normally associated with laboratory corefloods, diffusion is sufficiently rapid to reduce effectively the adverse effects of bypassing on overall recovery efficiency. This is not so for field floods, where bypassing of oil by injected CO2, may be expected to occur on a much larger scale.
The displacement of oil by CO2 has been the subject of numerous studies, and a considerable volume of experimental data now exists for the process. In these studies, CO2, usually displaces oil from a fully saturated slim tube or test core in a secondary displacement under conditions where phase behavior favors the development of miscibility. For these conditions, recoveries are observed to be high and rate-insensitive, provided that the effects of gravity segregation, viscous fingering, and bypassing are minimized. In direct contrast, for tertiary recovery experiments where CO2, is injected into a previously watered-out test core, recoveries of residual oil are observed to be considerably lower, dependent on both flood rate and core length. and different for water-wet and oil-wet systems. The differences in performance between secondary and tertiary flood experiments are usually explained in terms of the high water saturations present in the tertiary flood experiments and their effect on the microscopic displacement efficiency. As a consequence of the highly unfavorable mobility ratio for the immiscible CO2, water displacement, injected CO2, bypasses considerable volumes of water, leaving high water saturations behind the displacement front in water-wet rock. The water blocks or shields the residual oil from direct contact with the injected CO2. This prevents development of miscibility and results in a considerable reduction in microscopic displacement efficiency. Molecular diffusion of CO2, through the water blocking phase has been suggested as an important mechanism in the mobilization and recovery of the residual oil in water-wet rock. Swelling of the oil phase causes a breakdown of the original capillary equilibrium, resulting in a pore-scale redistribution of the phases. For this process to be fully effective in recovering residual oil, sufficient time must be available for diffusion of CO2, to swell the oil significantly. Recent flow visualization studies in two-dimensional (2D) micromodels qualitatively confirmed the importance of diffusion and oil swelling in tertiary CO2 displacement experiments. These studies have demonstrated clearly that the bulk of the waterflood residual oil is left behind the CO2, displacement front, trapped by water and therefore inaccessible to the flowing CO2, channels. It was further observed that CO, diffusion through the blocking water phase caused isolated oil droplets in water-wet rock to swell. When sufficient time (overnight) was allowed for diffusion, the swelling eventually was enough to break the water barrier that blocked the oil from the flowing CO2, channel and therefore to allow high oil recoveries. Although the available experimental data suggest that diffusion is important in the recovery of waterflood residual oil, the role of diffusion has not yet been quantified. It is therefore not possible to scale the diffusion times and associated nonequilibrium effects responsible for oil recovery in tertiary CO2, floods correctly on the laboratory scale or in the field. The purpose of this paper is to examine the role of molecular diffusion and to determine the time scales necessary for diffusion to be an effective recovery mechanism both in laboratory floods and in the field. An idealized one-dimensional (1D) pore-scale model is developed that simulates CO2 diffusion through a water barrier to a trapped oil phase and the subsequent swelling of the oil phase. The model clearly demonstrates the important role of diffusion in the recovery process. Computed diffusion times are compared with available experimental data both to validate the proposed model and to relate correctly the time scales over which this process is effective both in laboratory corefloods and in the process is effective both in laboratory corefloods and in the field.
The actual mechanism by which waterflood residual oil is mobilized during a tertiary CO2, flood is not yet fully understood.