Gas–surface interaction effects on Rayleigh–Bénard convection in rarefied flows

Non-continuum Rayleigh–Bénard convection of a monatomic gas is investigated using the Direct Simulation Monte Carlo method for Knudsen numbers in the range 0.01≤Kn≤0.04. The influence of gas–surface interaction (GSI) on flow stability and heat transfer is examined using fully diffuse, partially diffuse, and Cercignani–Lampis–Lord (CLL) kernels applied at isothermal hot and cold boundaries. The CLL formulation permits independent variation of the tangential momentum accommodation coefficient, σt, and the normal energy accommodation coefficient, αn, enabling their individual contributions to be assessed. The onset and development of convection exhibit pronounced sensitivity to the choice of GSI model and accommodation coefficients. Lower values of σt increase velocity slip and reduce viscous damping, thereby promoting instability, whereas larger values of αn reduce temperature jumps and enhance thermal coupling between the gas and the walls. The combined influence of these accommodation processes alters the predicted stability boundaries, with a low σt, high αn combination extending the admissible rarefaction limit for convection onset from Kn=0.029 for the fully diffuse kernel to Kn=0.037 for the CLL kernel. The results indicate that the critical Rayleigh number is a function of accommodation coefficients, Knudsen and Froude numbers. Since these parameters are functions of the gas species, surface properties, and thermodynamic conditions, a single critical Rayleigh number cannot adequately characterize convection in the non-continuum regime. Heat transfer analysis shows broadly similar trends across different GSI kernels, with enhanced transport observed for larger αn and smaller σt, highlighting the sensitivity of quantitative transport predictions to gas–surface interaction modeling in rarefied convection.