Effect of solvent on the adsorption behavior of asphaltene on silica surface: A molecular dynamic simulation study

In recent years, the hybrid thermal-solvent process has been widely applied to improve the recovery performance of steam injection processes in heavy oil reservoirs. In this paper, the method of Molecular Dynamics (MD) simulation is employed to illuminate the asphaltenes adsorption behavior in the thermal-solvent recovery process. Three different solvent molecules (CO2, C3H8, and nC4H10) and SARA (Saturates, Aromatics, Resins, Asphaltenes) simulated heavy oil model are constructed as the basic simulation model. A series of MD simulations at different temperature conditions are performed. Results show that for the SARA model, the asphaltene molecules can interact with the silica by a T-shape stacking, finally forming the asphaltene dense aggregates as a basic heavy oil occurrence state. The steric hindrance effect of other SARA components can also contribute to this configuration. Temperature significantly affects the adsorption configuration of asphaltenes by disassembling the dense core and loosening the structure of the aggregates. For the SARA model in three solvent atmospheres, the increasing temperature can benefit the extraction of light components. CO2 can only extract saturates, while nC4H10 and C3H8 can simultaneously extract the saturates and aromatics. Besides, asphaltenes re-precipitation behavior can be observed in the 393 K CO2 atmosphere. Both nC4H10 and C3H8 have mutual solubility with the heavy oil system. No apparent precipitation of asphaltenes occurs in the above two atmospheres. Comparing the performance of extraction capability and diffusion capability in all MD simulations, the nC4H10 can both extract light oil components and control the asphaltenes precipitation. It further reveals that nC4H10 can recover heavy oil more efficiently at a microcosm level. Among the three different solvents, nC4H10 is the optimal solvent for hybrid thermal-solvent processes in heavy oil reservoirs. © 2022 Elsevier B.V.