UTA team exploring artificial ferroelectric materials

Researchers working with artificial ferroelectric materials to induce mechanical behaviors

Tuesday, Oct 12, 2021 • Herb Booth : Contact

 

Engineering research

A pair of University of Texas at Arlington researchers is trying to understand how “smart” mechanical behaviors can be engineered in artificial ferroelectric materials.

Ye Cao, an assistant professor in the Materials Science and Engineering Department, and Joseph Ngai, an associate professor in the Physics Department, were awarded a $597,856 grant from the National Science Foundation for the project. They are working with ferroelectric materials, which have the ability to have a spontaneous electric polarization that can be manipulated to create mechanical motion by applying electric fields.

The researchers hope to use thin films made from varied materials and with varied thicknesses to create a free energy landscape that is tunable by light. Free energy is the energy that can be organized and used to do work. The free energy landscape is like a topographic map, with peaks and valleys, and the material characteristics of the ferroelectric materials are determined at the valleys.

Cao and Ngai hope to tune the mechanical behaviors of these materials with light, and by doing so, determine how they will react. They will then use machine learning and computational modeling to create different landscapes that could be rationally used in various applications.

“We’re dealing with a class of materials where we can subtly change the landscape and change behaviors,” Ngai said. “When we combine thin layers of various chemical compositions, we can engineer the free energy landscape and tune its characteristics to our needs.”

Cao will use machine learning and phase-field modeling to predict the materials’ natural properties. He and Ngai will take samples and create data points, stitch the data points together using machine learning, then connect the points with computational models that can create 3D or greater models that are impossible to replicate in a lab.

“We want to do high-throughput simulations to generate sufficient data points, then use machine learning to find hidden connections between them that we can exploit for future uses,” Cao said. “The number and thickness of the layers can be well controlled in the computational model to tune the mechanical behaviors and eventually be validated by experiments.”

- Written by Jeremy Agor, College of Engineerng