Surface morphology control of Transparent Conducting Oxide layers for improved light trapping using wafer grating and feedback control

Abstract

This work first introduces a kinetic Monte-Carlo simulation model for a two-species thin film deposition process on a grated (patterned) wafer, and then utilizes a model predictive controller, which manipulates a spatially distributed deposition rate, to produce thin films whose surface morphology has a structure that improves light trapping. This approach to thin film surface morphology control can be applied to many deposition processes, and in particular, in the context of Transparent Conducting Oxide (TCO) thin film manufacturing processes used in thin film solar cells where it is desirable to produce thin films with precisely tailored surface morphology. Specifically, a two species thin film deposition process involving atom adsorption, surface relaxation and surface migration is considered and is modeled on a large-scale lattice (lattice size=40,000) via kinetic Monte-Carlo methods and aggregate surface roughness and slope are used to describe the surface morphology. Subsequently, multiple sets of simulations are carried out to understand the process dynamics dependence on wafer grating parameters, such as magnitude and period of grating, and other process parameters, such as temperature and deposition rate. From these simulations, it is concluded that a spatially distributed deposition rate profile is necessary to be utilized across the lattice in order to induce desirable surface morphology at light wavelength spatial scales that lead to desired thin film solar cell performance. Then, an Edwards–Wilkinson-type equation is utilized to predict the surface evolution and forms the basis for the design of a predictive feedback controller. The model parameters of the Edwards–Wilkinson equation are identified from kinetic Monte-Carlo open-loop simulations. Analytical solutions of the expected surface roughness and surface slope at the visible light wavelength spatial scale are obtained by solving the Edwards–Wilkinson equation and are used in the predictive controller formulation and in the control action calculation. The controller is applied to the kinetic Monte-Carlo simulation of the deposition process taking place on a sinusoidal grated wafer. Simulation results demonstrate that the proposed controller, wafer grating and patterned actuator design successfully regulate aggregate surface roughness and slope to desirable set-point values at the end of the deposition.