摘要：AbstractThis paper presents a comparison between worst-case and distributional uncertainty analysis in a thin film deposition process. The key idea is to evaluate the effect of model parameter uncertainties on thin film properties employing power series expansion (PSE). The worst-case deviation in the film properties is obtained under bounded parameter uncertainties while the probabilistic bounds are estimated under distributional uncertainties. To describe the growth process on the surface of a substrate, a multiscale approach that integrates kinetic Monte Carlo (KMC) simulations with continuum modelling is employed. The uncertainty analysis in this work is applied to estimate the optimal substrate temperature profile for robust optimization of the thin film end-point properties.
关键词：KeywordsThin film depositionMultiscale modellingUncertainty analysisPower series expansionRobust optimization
