Yuxiang Chen 陈宇翔

Ph.D. candidate in Chemical Process in East China University of Science and Technology

Update: On September 14, 2023, I am designing a universal model for predicting the theoretical value of organic sulfide absorption amount by solvents under any environmental conditions.

I am currently in the final year of my M.S./Ph.D. program at East China University of Science and Technology, working with Prof. Hui Sun in the Department of Petroleum Processing on Intelligent Molecular Design. My current research focuses on developing novel theoretical and computational frameworks for modeling molecular properties, with applications to desulfurization, purification, and solvent development.

Prior to Ph.D., I studied software engineering on my own, where I immersed myself in the fields of mathematical modeling and machine learning. Since 2020, I have been contributing to computational chemistry, data analysis, and mathematical modeling for Prof. Fahai Cao's research group.

Outside of research, I am actively involved in volunteer work both within the department and on campus. During the COVID-19 pandemic in 2022, I worked with the Graduate Student Council at the College of Chemical Engineering at East China University of Science and Technology to support my fellow graduate students.

Email: chenyuxiang@mail.ecust.edu.cn

CV  /  Google Scholar  /  Github /  ZhiHu /  WeChat

I'm interested in molecular desgin, machine learning, optimization, and mathematical modeling. Much of my research is about determining the relationship between the chemical structure and its properties. Representative papers are highlighted.

We obtained the reaction rate constants of determining step via computation and constructed a machine learning model to predict reaction kinetics of carbonyl sulfide with solvents.
Two custom-defined descriptors representing the steric hindrance of amine groups has improved kinetics predication.
We provided a progressive and in-depth explanation on how chemical structures affect reaction rates from a big data perspective.

We provided an approach to the synthesis of AgY@Cu-BTC hybrid composite for enhanced sulfides capture and moisture resistance.

We revealed the relationship between the structural characteristics of the solvents and their interactions with MeSH and proposed rules for designing molecules to generate a potential solvent with enhanced dissolving affinity named 1-(2-(diethylamino)ethoxy)butan-2-amine.

We revealed the influence of the compositions and structures of the supports on the catalyst performances through experiments together with density functional theory (DFT) calculations.

We proposed an intelligent molecule design approach to encode and decode discrete molecules into multidimensional continuous variables by variational autoencoder (VAE), based on Bayesian optimization (BO) to define the next molecule with a potentially high reaction rate.

We constructed a computational framework by integrating molecular similarity search and active learning methods, namely, molecular active selection machine learning (MASML).
We are proud that the final developed molecule has already reached the pilot stage of industry.

We synthesized a family of copper-1,3,5-benzenetricarboxylic acid (Cu-BTC) sorbents for organosulfur compound capture.