Ferroelectric Domains and Dynamics: Insights Enabled by Machine Learning

发布日期:2025-11-24

报告人:Prof. Liangzhi Kou

单位:Queensland University of Technology

报告时间:20251128日星期五 9:00

报告地点:知新楼C7楼量子报告厅

邀请人:马衍东


摘要:Polar domains and their dynamic manipulation have attracted considerable interest due to their rich physics and potential applications in next-generation digital memory devices. However, theoretical investigations are often limited by the lack of methods capable of accurately capturing the complex underlying phenomena. Recent advances in machine learning interatomic potentials, particularly those based on neural network frameworks, offer a powerful alternative. By constructing large-scale first-principles training datasets and directly fitting potential energy surfaces, these approaches effectively bypass the computational bottleneck of solving the Schrödinger equation.In this talk, I will present two recent studies that integrate density functional theory (DFT), machine learning potentials, and deep learning molecular dynamics (DLMD) simulations to investigate ferroelectric domain formation and dynamics. Specifically, we demonstrate the creation and manipulation of polar domains in ferroelectric In₂Se₃ through localized deformation, and in twisted bilayers of CuInP₂S₆. Unlike the polar vortices and skyrmions reported in (PbTiO₃)n/(SrTiO₃)n superlattices, the mechanisms here arise from stacking- or strain-dependent energy barriers for ferroelectric switching, as well as variations in switching speeds under thermal fluctuations. Importantly, the thermal stability and polarization lifetimes are found to be highly sensitive to curvature, twist angle, and temperature, and can be further tuned by external electric fields and strain.


报告人简介Prof. Kou received his PhD from Nanjing University of Aeronautics and Astronautics, and has held research and academic positions at Bremen University (Germany), the University of New South Wales, and Queensland University of Technology (Australia). His research focuses on first-principles simulations of low-dimensional nanomaterials, with particular interest in multi-physical coupling phenomena—mechanical, electrical, and magnetic—and their applications in energy conversion and storage, nanoelectronic devices, and catalysis.

He has received numerous prestigious awards and recognitions, including a Humboldt Research Fellowship (2012–2014), an ARC Discovery Early Career Researcher Award (DECRA) (2018–2021), inclusion among the world's top 2% of scientists (2021–2025), and the Friedrich Wilhelm Bessel Research Award (2025).

To date, Professor Kou has published over 200 peer-reviewed articles in leading journals such as Nature Communications, Journal of the American Chemical Society, Nano Letters, ACS Nano, Advanced Science, and Advanced Functional Materials. His work has garnered around 15,000 citations, with an h-index of 65.