Multiscale computational fluid dynamics modelling of spatial atomic layer deposition processes: Application to chamber design and process control

초록

In this work, we present a multiscale modelling study that integrates density functional theory (DFT) calculations and computational fluid dynamics (CFD) simulations to investigate spatial atomic layer deposition (SALD) of Al2O3 films using trimethylaluminum (TMA) and H2O precursors. DFT was used to elucidate the reaction energetics on hydroxyl-terminated alpha-Al2O3(0001) surfaces, revealing mechanistic details, such as adsorption pathways, activation barriers, and rate-determining steps in both half-reactions. These atomic-level insights were incorporated into a transient CFD model of the in-line SALD reactor to explore the coupled phenomena of precursor transport, flow hydrodynamics, and surface chemistry under various processing conditions. Special attention was paid to the comparison of the top-and bottom-purging configurations, where the exhaust positions clearly influenced flow patterns and vortex formation, thereby affecting precursor uniformity and mixing. Simulations captured the self-limiting half-cycles of TMA and H2O, demonstrating that once the substrate surface sites were saturated, no further reactions occurred, regardless of the precursor excess. Parametric analyses further showed how substrate velocity and temperature can be optimized to balance film growth rate, precursor utilization, and overall process efficiency. Taken together, these results offer practical guidelines for the design of the SALD reactor by revealing how geometric parameters and operating conditions govern uniform film deposition. By combining DFT-derived chemical kinetics with reactor-scale CFD simulations, this study provides a robust framework for predicting and fine-tuning thin film growth in efficient, cost-effective SALD systems.

키워드