Multiscale modeling of gallium nitride atomic layer etching in chlorinated plasmas: A combined dynamic global model, <i>ab initio</i> , and kinetic Monte Carlo approach

GaN is a key material in microelectronics that requires precise etching to meet technological demands. Plasma Atomic layer etching (ALE) is a promising process to achieve a defectless surface. However, the plasma/surface interaction mechanisms involved remain complex. In this work, we propose a multiscale model of GaN plasma ALE in chlorinated plasmas that combines a plasma module using a dynamic global model to describe Cl2/Ar plasmas, a sheath module described by a Monte Carlo model, and a surface module considering a kinetic Monte Carlo model. In addition, density functional theory calculations of Cl adsorption on GaN surfaces are used for a better description of the adsorption mechanisms. The surface module allows simulations of the GaN surface during ALE with an atomic-scale resolution. This multiscale approach enables comprehensive evolution of surface morphology, material composition, and etching products over time. Results reveal a self-limited etching behavior consistent with ALE principles and show that residual chlorine penetrating in shallow layers significantly impacts the etch per cycle. Our simulator provides a currently offline predictive framework that could contribute as a step toward future digital twin development to optimize and control GaN ALE processes.