报告题目：Accurate Prediction of Ice Nucleation on Graphene Oxide Surfaces Using Supervised Machine Learning
报告时间：2024年05月28日（周二） 16：00
报告地点：3号楼309会议室
报告人：张志森 副教授
邀请人：周昕 研究员
简介： Dr. Zhisen Zhang, Associate Professor from Physical Department of Xiamen University, studied in Chemistry Department at Zhejiang University since 2004 as an undergraduate student, and received his doctoral degree in Physical Chemistry from Zhejiang University in 2014. Then he started his research in Physical Department of Xiamen University. His research interests focus on nucleation of ice and gas hydrate. He has published over 40 papers as first author and corresponding author, as well as more than 80 papers as a co-author. The majority of these publications focus on the research of crystal nucleation and growth mechanisms in systems related to ice, methane hydrates, and biominerals.
摘要： Ice formation is a common natural phenomenon that not only impacts Earth's climate but also plays a crucial role in scientific research. Although a large number of materials have been studied as nucleation agents, most of them are uniform surfaces. Ice nucleation on the complicated non-uniform surfaces is not clear. Herein, the graphite oxide surface is used as a complicated non-uniform model surface. We employed molecular dynamics simulations to investigate ice nucleation on two types of GO models: i) orderly distributed oxidation groups on the GO surface, and ii) randomly distributed oxidation groups on the GO surface. For the ii type of GO models, we accurately predicted the ice nucleation sites on the GO surface with randomly distributed oxidation groups by utilizing machine learning techniques. For the complicated non-uniform surfaces, we design a machine learning scheme to study ice nucleation on these surfaces. Our findings indicate that the ice nucleating efficiency of the complicated non-uniform surface is determined by the structural compatibility between the atomic configuration of the surface and the ice crystal.
