Narrow Escape Brownian Dynamics Modeling in the Three-Dimensional Unit Sphere (2107.01233v2)

Abstract: The narrow escape problem is a first-passage problem concerned with randomly moving particles in a physical domain, being trapped by absorbing surface traps (windows), such that the measure of traps is small compared to the domain size. The expected value of time required for a particle to escape is defined as mean first passage time (MFPT), which satisfies the Poisson partial differential equation subject to a mixed Dirichlet-Neumann boundary condition. The primary objective of this work is a direct numerical simulation of multiple particles undergoing Brownian motion in a three-dimensional sphere with boundary traps, compute MFPT values by averaging Brownian escape times, and compare the results with asymptotic results obtained by solving the Poisson PDE problem. A comprehensive study of results obtained from the simulations shows that the difference between Brownian and asymptotic results for the escape times mostly not exceed $1\%$ accuracy. This comparison in some sense validates the narrow escape PDE problem itself as an approximation (averaging) of the multiple physical Brownian motion runs. This work also predicted that how many single-particle simulations are required to match the predicted asymptotic averaged MFPT values. The next objective of this work is to study dynamics of Brownian particles near the boundary by estimating the average percentage of time spent by Brownian particle near the domain boundary for both the anisotropic and isotropic diffusion. It is shown that the Brownian particles spend more in the boundary layer than predicted by the boundary layer relative volume, with the effect being more pronounced in a narrow layer near the spherical wall. It is also shown that taking into account anisotropic diffusion yields larger times a particle spends near the boundary, and smaller escape times than those predicted by the isotropic diffusion model.