[G] Hydrogen and Battery Related Materials(2)

9:30 AM - 9:45 AM JST(12:30 AM - 12:45 AM UTC)

[85] Hybrid Sampling Machine-Learning Potentials for Accurate Hydrogen Adsorption in MOF-303

○Kartik Sau Sau¹, Ikutaro Hamada², Tamio Ikeshoji³, Yiming Lu¹, Susmita Roy⁴, Shohichi Furukawa¹, Linda Zhang¹, Hung Ba Tran¹, Takahiro Kondo^4,5,1,6, Hao Li¹, Shin-ichi Orimo^7,1 (1. Advanced Institute for Materials Research (WPI-AIMR) Tohoku University, Aoba-ku, Sendai 980-8577, Japan, 2. Department of Precision Engineering, Graduate School of Engineering, Osaka University, Suita, Osaka, 3. Mathematics for Advanced Materials Open Innovation Laboratory (MathAM-OIL), AIST, c/o WPI-AIMR, Tohoku University, Sendai, 4. Department of Materials Science, Institute of Pure and Applied Sciences, University of Tsukuba, Tsukuba, Ibaraki, 5. Hydrogen Boride Research Center, TIAR, University of Tsukuba, Tsukuba, Ibaraki, 6. Tsukuba Research Center for Energy Materials Science, University of Tsukuba, Tsukuba, Ibaraki, 7. Institute for Materials Research, Tohoku University, Sendai, Miyagi)

Keywords:

MOF-303,Adsorption,molecular dynamics simulations,machine learning,machine learning potential,hydrogen adsorption,grand canonical Monte Carlo,density functional theory,isosteric heat

We developed a hybrid sampling workflow combining force-field GCMC and ab initio MD to train robust machine-learning potentials for gas adsorption. Applied to MOF-303, our model accurately predicts H2 uptake, isosteric heat, and diffusion, revealing temperature-dependent adsorption mechanisms.