Direct Coupling of Free Diffusion Models to Microscopic Models of Confined Crystal Growth and Dissolution

TL;DR

This paper introduces a novel method for coupling free diffusion models with microscopic crystal growth models, demonstrating improved accuracy in confined environments and reducing computational costs by ignoring line of sight in certain conditions.

Contribution

The study develops a direct coupling approach between diffusion and surface growth models, incorporating mean first passage time for deposition, and compares variants with and without line of sight considerations.

Findings

Models reproduce macroscopic surface dynamics accurately.

Free diffusion models yield lower equilibrium roughness in confinement.

Lattice diffusion performs poorly in tight confinements.

Abstract

We couple a free solute diffusion model to a model of crystal surface growth represented by, but not limited to, a (2 + 1)-dimensional solid-on-solid (SOS) model confined by a flat surface. We use kinetic Monte Carlo (KMC) with dissolution rates based on nearest-neighbor interactions to solve the Master equation for the surface dynamics, and we use an offlattice random walk to model the Fickian diffusion of the solute particles. The two solvers are coupled directly through deposition rates of the free particles calculated using the mean first passage time (MFPT) of deposition that is found to scale as $r^{-4}$. Two variants are studied: ignoring (radial) and not ignoring the line of sight (pathfinding). Reference models such as uniform concentration (random deposition) and lattice diffusion (crystal lattice extended into the liquid) are used for comparison. We find that the macroscopic limit of the surface dynamics is reproduced by all models. The free diffusion models produce a lower equilibrium roughness and a smaller height autocorrelation length than the reference models, and are found to behave very well in tight confinements. It is also demonstrated that lattice diffusion does not work well in tight confinements. The two MFPT models behave very similarly close to equilibrium and for dissolution, but becomes increasingly different with increasing surface growth speed. The model is put to use by simulating a cavity with a flux boundary condition at one side. The conclusion is that the new model excels in confinement, and line of sight can in practice be ignored since the dominant deposition sites likely are in line of sight, which minimizes the CPU-time needed in the coupling.