Effect of CO2 Pre-Extraction on Water Flooding in Nanopores: Insights from Molecular Dynamics Simulations

Shale oil is predominantly stored in nanoscale pores with ultra-low porosity and permeability, where conventional waterflooding commonly delivers poor recovery. While CO₂-enhanced oil recovery (CO₂-EOR) can improve production by inducing oil swelling, reducing viscosity, and promoting desorption, many existing evaluations still rely on bulk-phase properties and thus inadequately capture nano-confinement and mineral-specific surface effects, obscuring quantitative relationships among CO₂ fraction, desorption efficiency, and mobility. In this study, equilibrium and non-equilibrium molecular dynamics simulations are performed to quantify density layering, competitive adsorption, and rheological/slip behavior of shale oil–CO₂ mixtures confined in quartz and kaolinite nanopores. The simulations show that CO₂ preferentially enriches near pore walls, displaces adsorbed oil, and weakens oil–rock interactions, facilitating the release of interfacial hydrocarbons. Compared with bulk behavior, confinement increases apparent viscosity by about two- to threefold, and kaolinite exhibits pronounced boundary resistance manifested as adverse (negative) slip. As the CO₂ fraction increases to ~20–40%, viscosity decreases markedly and interfacial transport improves, shifting the displacement from unstable fingering toward a more coherent piston-like front. Building on these pore-scale insights, a multiscale coupling framework is developed by embedding MD-derived transport and interfacial parameters into reservoir numerical simulations to conduct 3D field-scale forecasts for the Gulong Sag. The resulting recovery factors that account for nano-confinement (~8–20%) better match field behavior, whereas bulk-parameter simulations substantially overestimate performance. Sensitivity analyses further indicate mineral-dependent economically favorable CO₂ windows (>20% for quartz-dominated pores and ~30–40% for kaolinite-rich pores), highlighting the need for differentiated injection strategies; overall, the proposed multiscale approach bridges microscopic interfacial physics and macroscopic development prediction, providing quantitative support for optimizing CO₂-EOR and enhancing CO₂ utilization and storage in unconventional reservoirs.

Keywords: Shale oil; Nano-confinement effects; Molecular dynamics simulations; Unconventional reservoirs