报告题目：Computational Insights into Degradation and Phase Segregation in Hybrid Perovskites via Machine Learning Potentials
报告人：殷骏，助理教授，香港理工大学物理与材料学系
主持人（邀请人）：张立军教授，韩丹教授
报告时间：2026年8月15日10:00-12:00
报告地点：中心校区唐敖庆楼D区429会议室
主办单位：汽车材料教育部重点实验室 百家乐在线游戏
报告人简介：殷骏，香港理工大学物理与材料学系助理教授，校长青年学者。2024年获国家优秀青年科学基金资助。他于新加坡南洋理工大学获物理学博士学位，随后赴沙特阿卜杜拉国王科技大学开展博士后研究并担任研究科学家。长期致力于新型光电和能源材料的理论计算与模拟研究，在光物理过程微观机制解析、材料构效关系理论建模以及性能智能调控等前沿领域取得了多项原创性成果，为新型功能材料的理性设计与开发提供了重要的理论基础和技术支撑。迄今已发表SCI论文280余篇（总被引超过25,000次，h因子87），多项成果发表于Science、Nature、Nature Energy、Nature Photonics、Nature Synthesis等国际顶尖期刊。2022年至2025年连续入选斯坦福大学发布的全球前2%高被引科学家榜单。
报告摘要：The evolution of computational materials science has been propelled by the integration of diverse computational approaches, including density functional theory (DFT), molecular dynamics (MD), and machine learning potentials (MLPs). In this talk, I will highlight our recent efforts using MLP-assisted MD simulations to study hybrid perovskites, with particular emphasis on how surface orientation and structural heterogeneity govern surface and phase stability. I will also discuss our recent ML-driven studies on the design of organic spacers for achieving desirable band alignment and enhanced band splitting in two-dimensional hybrid perovskites, as well as self-assembled molecules for improving interfacial binding and surface coverage on metal oxides toward efficient perovskite/silicon tandem solar cells. By developing an ML-driven workflow together with a set of structural and electronic descriptors, we aim to uncover the fundamental structure-stability relationships in hybrid perovskites and establish a predictive framework for the rational design of more stable hybrid perovskite materials for photovoltaic and spintronic applications.
