Abstract

Polar materials have a dipole moment or polarization. Ferroelectrics are a special class of polar materials wherein the polarization can be switched with an external electric field. Because of their characteristics, ferroelectrics are especially useful in adjustable capacitors, non-volatile memories, and sensors. Materials databases like Materials Project contain a large number of materials, but identifying new metastable polar, and more specifically ferroelectric materials can be time consuming. In this project, we train a machine learning model to distinguish between binary compounds having a wurtzite structure—which have a permanent polarization—and their nonpolar zincblende and rock salt polymorphs. We use this model to predict a large number of ternary materials that we find to have high chances to be most stable as a wurtzite. These compounds serve as a smaller dataset that is tractable for high-throughput DFT calculations to calculate their stability in the wurtzite phase, their polarization, and switching barriers with higher accuracy. In the second chapter, we first attempt to identify factors or descriptors that can screen ferroelectric materials from a materials database. Using these descriptors, we predict new ternary oxides that can be potential ferroelectrics and evaluate these predictions using first-principles density-functional theory (DFT) calculations.

Committee Chair

Rohan Mishra

Committee Members

Elijah Thimsen Sanghoon Bae

Degree

Master of Science (MS)

Author's Department

Mechanical Engineering & Materials Science

Author's School

McKelvey School of Engineering

Document Type

Thesis

Date of Award

Spring 5-19-2022

Language

English (en)

Author's ORCID

https://orcid.org/0000-0002-3961-1656

Download

Included in

Engineering Commons

Share

COinS