Machine Learning Guided Polymorph Selection in Molecular Beam Epitaxy of In2Se3

TL;DR

This paper demonstrates that Bayesian Optimization can efficiently guide the molecular beam epitaxy process to selectively grow high-purity In2Se3 polymorphs, significantly reducing experimental trials and overcoming growth challenges.

Contribution

It introduces a Bayesian Optimization framework for phase-selective growth of In2Se3, enabling efficient exploration of growth parameters and achieving high phase purity.

Findings

Achieved 91% phase purity of γ-In2Se3 after fewer than ten trials.

Limited success in isolating α-In2Se3 due to amorphous film formation at low temperatures.

Validated Bayesian Optimization as an effective tool for complex materials growth.

Abstract

Indium selenide (In2Se3), a layered chalcogenide with multiple polymorphs, is a promising material for optoelectronic and ferroelectric applications. However, achieving polymorph-pure thin films remains a major challenge due to the complex growth space. In this work, Bayesian Optimization (BO) is successfully leveraged to guide the molecular beam epitaxy (MBE) growth of In2Se3 on Al2O3 substrates. By training a predictive Gaussian Process Regressor with sequential learning, we efficiently explored substrate temperature, indium flux, selenium flux, and cracker temperature, reducing experimental trials required for successful synthesis. A {\gamma}-In2Se3 film with 91% phase purity was achieved after fewer than ten trials. Attempts to isolate {\alpha}-In2Se3 were limited by amorphous film formation at low temperatures, indicating that single-step co-deposition is unsuitable for crystalline {\alpha}-In2Se3 on Al2O3. Overall, this study validates Bayesian Optimization as a powerful approach for phase-selective growth in complex materials systems.