Monte Carlo study of the molecular beam epitaxy process for manufacturing magnesium oxide nano-scale films

• Published in: IIE transactions Vol. 47; no. 2; pp. 125 - 140
• Main Authors: Uddin, Ghulam Moeen, Ziemer, Katherine S., Zeid, Abe, Kamarthi, Sagar
• Format: Journal Article
• Language: English
• Published: Norcross Taylor & Francis 01.02.2015; Taylor & Francis Ltd
• Subjects: Bonding; Chemical bonds; factor-wise sensitivity analysis; functional oxide heterostructures; interface engineering; Magnesium; Magnesium oxide; magnesium oxide thin films; Mathematical models; Molecular beam epitaxy; Monte Carlo experiments; Monte Carlo methods; Monte Carlo simulation; Nano manufacturing; nano-scale metal oxide films; neural network; non-parametric regression; Process control; Process controls; Production management; Sensitivity analysis; Silicon carbide; Spectrum analysis; Studies; Variables
• ISSN: 0740-817X; 2472-5854; 1545-8830; 2472-5862
• DOI: 10.1080/0740817X.2014.905732

This article presents a Monte Carlo-based factor-wise sensitivity analysis conducted on the performance variables of a Molecular Beam Epitaxy (MBE) process. Using lab-scale MBE equipment, magnesium oxide (MgO 111) films are grown on a hexagonal silicon carbide 6H-SiC (0001) substrate. The thin film surface chemistry in terms of O‒Mg and OH‒Mg bonding states is examined using X-ray photoelectron spectroscopy. A multi-layer perceptron is used to model the process. Monte Carlo experiments are conducted on the process model to study the causal relationship between the critical process control variables and the key performance indicators. The sensitivity of O‒Mg and OH‒Mg bonding states in MgO films to each of the four process control variables (growth time, substrate temperature, magnesium source temperature, and percentage starting oxygen) is examined. Each control variable is varied individually while keeping other control variables constant at their mid values in one case and randomly varying in another case. The sensitivity of the performance variables to the interaction between a select set of control variable pairs is also examined. The interaction between substrate temperature and oxygen on the starting surface is found to significantly affect the dynamics of OH‒Mg bonding state.