"Design of Functional Perovskite Oxides from Atomistic Simulations, Machine Learning and Causal Analysis"
Thursday, November 16, 2023 3 PM to 4 PM
About this Event
6548 Forest Park Pkwy, St. Louis, MO 63112, USA
https://imse.wustl.edu/Dr. Saurabh Ghosh, Dept of Nuclear Engineering, University of Tennessee
The family of ABO3 perovskite oxides form a large class of compounds with diverse physical and electronic properties. This set of materials shows a wide range of functional properties, including ferroelectricity, ferromagnetism, multiferroicity, catalysis, radiation tolerance, etc. Single or multiple structural modes in bulk crystal materials can drive high symmetry to low symmetry phase transition. Coupling between structural modes can go two or more functional properties into the low symmetry phase. The celebrated Hybrid Improper Ferroelectricity (HIF) falls into this category of multimode-driven phase transition [1]. It has been reported to drive weak ferromagnetism, linear magnetoelectricity, and ferroelectricity into cation-ordered A/A systems with Pnma-ABO3 as building blocks [2].
In this talk [2-6], I will focus on how ABO3 oxides can be functionalized as (a) hybrid improper multiferroic and (b) nuclear energy material. I will discuss the emergent multiferroic properties by structural mode coupling considering (i) ABO3 / A’BO3 oxide superlattices and (ii) AA’BB’O6 double perovskites, where A and A’ are rare-earth or alkali metals and B, B’ are transition metals. The aim is to establish meaningful cross-coupling between functional properties such as ferroelectricity, magnetism and metal-to-insulator transition. Understandings developed from first principles calculations, ab initio molecular dynamics (AIMD), and hypothesis-driven active machine learning (ML) will be presented. I will also discuss how we can go beyond the traditional correlative nature of ML models to establish a causal relationship between the descriptor space constructed from structural modes plus other geometry-driven features and target property, i.e., cation ordering in double perovskites For nuclear energy materials, the critical performance matrix is tolerance against radiation damage. I will briefly discuss how AIMD simulations and ML can help design potential ABO3 complex oxides for applying radiation shields and nuclear waste forms.
Faculty, students, and the general public are invited.
Hosted by: Rohan Mishra, MEMS