2022 Annual Meeting
Design Rules for MOFs As Material for Hydrogen Storage Focusing on the Adsorption Energy Predicted By Machine Learning
The combustion of energy resources is known to be a major contributor to climate change due to greenhouse gas such as carbon dioxide and methane released by consumption of fossil fuel. Hydrogen gas has been suggested as an adequate substitution of fossil fuels since it's only byproduct of combustion is water. However, commercialization of hydrogen gas as energy resource requires efficient storage due to its high volatility with potential of explosion. Metal organic framework (MOF) is a promising candidate for efficient hydrogen storage because it is easy to synthesize and provide a large surface area due to their porous nature. The objective of this research is to identify features of MOFs that lead to efficient hydrogen storage and to use machine learning to predict the adsorption energy of hydrogen molecule on different MOFs. The adsorption energy of hydrogen was calculated via the density functional theory. Results showed that the metal center, functional group and organic linker had a significant effect on the adsorption energy. The mean absolute error(MAE) of the neural network model was 0.209eV for the whole data set and 0.162eV without outliers.