Development And Application Of MOF-808 Encapsulated Platinum Cluster Artificial Neural Network Potentia

Metal organic framing（MOFs）materials are considered one of the ideal loading anti-sintering catalyst materials for their high specific surface area and unique nanoporous structure to effectively regulate the size and shape of encapsulating metal clusters.In order to explore the influence of metal organic skeleton on metal clusters during catalysis and design more effective catalysts,it is very necessary to theoretically study the binding site,stability and migration mechanism of metal clusters in MOFs materials.With the current computational power,the traditional density functional theory DFT consumes a lot of resources to perform computational simulations of structurally complex materials such as MOFs.However,with the continuous development of artificial neural networks,developing the potential energy that accurately describe the interaction between metal clusters and MOFs materials can greatly compensate for the deficiency of density functional theory,which can increase the simulation scale to nanoscale or nanosecond levels at lower energy consumption.In this paper,we developed the artificial neural network potential（G-NN）encapsulating platinum（Pt）clusters using a strategy combining global neural network technology of machine learning and DFT in MOF-808.A series of potential energy validations,including structural optimization,adsorption energy,and migration energy barriers,show that the artificial neural network potential energy G-NN we have developed not only has DFT accuracy for test systems,but is several orders of magnitude faster than density functional theory.This neural network potential energy was used to further study the adsorption and migration behavior of Pt_n clusters（n=1-13）in the thousands-atomic model system MOF-808,and we found that the most stable adsorption sites of Pt_n clusters on MOF-808 change with the size of Pt_n clusters.Furthermore,we find that the level of the migration energy barrier is positively correlated with the size of the adsorption energy during Pt_n cluster migration.Meanwhile,Pt₄ isomerizes during the migration of MOF-808.Finally,we also show that after the adsorption of carbon monoxide by Pt atoms effectively promotes the aggregation of Pt_n clusters through the Ostwald ripening mechanism.