Transition structure calculations via quantum chemistry methods have become a staple in modern chemical reaction research.Yet,success rates in optimizing transition structures rely heavily on rational initial guesses ...Transition structure calculations via quantum chemistry methods have become a staple in modern chemical reaction research.Yet,success rates in optimizing transition structures rely heavily on rational initial guesses and expert supervision.We develop a machine learning approach that utilizes a bitmap representation of chemical structures to generate high-quality initial guesses for modeling transition states of chemical reactions.The core of the approach comprises a convolutional neural network methodology with a genetic algorithm.An extensive dataset derived from quantumchemistry computations is built,providing sufficient data on which the model can be trained,validated and tested.By applying the method to typical bi-molecular hydrogen abstraction reactions involving hydrofluorocarbons,hydrofluoroethers,and hydroxyl radicals—reactions critical in atmospheric fluoride degradation and global warming potential evaluation,yet extremely challenging to model,we achieve transition state optimizations with an impressive,verified success rate of 81.8%for hydrofluorocarbons and 80.9%for hydrofluoroethers.The reported work demonstrates the effectiveness of employing visual representation in chemical space exploration tasks and opens new avenues for the transition structure modeling.展开更多
The bimolecular single collision reaction potential energy surface of an isocyanate NCO radical with a ketene CH2CO molecule was investigated by means of B3LYP and QCISD（T） methods. The computed results indicate tha...The bimolecular single collision reaction potential energy surface of an isocyanate NCO radical with a ketene CH2CO molecule was investigated by means of B3LYP and QCISD（T） methods. The computed results indicate that two possible reaction channels exist on the surface. One is an addition-elimination reaction process, in which the CH2CO molecule is attacked by the nitrogen atom at its methylene carbon atom to lead to the formation of the intermediate OCNCH2CO followed by a C-C rupture channel to the products CH2NCO＋CO. The other is a direct hydrogen abstraction channel from CHzCO by the NCO radical to afford the products HCCO＋HNCO. Because of a higher barrier in the hydrogen abstraction reaction than in the addition-elimination reaction, the direct hydrogen abstraction pathway can only be considered as a secondary reaction channel in the reaction kinetics of NCO＋ CH2CO. The predicted results are in good agreement with previous experimental and theoretical investigations.展开更多
基金support from the National Natural Science Foundation of China(No.52488201,No.52276212)National Key Research and Development Program of China(No.2022YFB3803600)+3 种基金the Suzhou Science and Technology Program(SYG202101)the Natural Science Foundation of Jiangsu Province(No.BK20231211)the Key Research and Development Program in Shaanxi Province of China(No.2023-YBGY-300)the China Fundamental Research Funds for the Central Universities.O.V.P.acknowledges support of the USA National Science Foundation(CHE-2154367).
文摘Transition structure calculations via quantum chemistry methods have become a staple in modern chemical reaction research.Yet,success rates in optimizing transition structures rely heavily on rational initial guesses and expert supervision.We develop a machine learning approach that utilizes a bitmap representation of chemical structures to generate high-quality initial guesses for modeling transition states of chemical reactions.The core of the approach comprises a convolutional neural network methodology with a genetic algorithm.An extensive dataset derived from quantumchemistry computations is built,providing sufficient data on which the model can be trained,validated and tested.By applying the method to typical bi-molecular hydrogen abstraction reactions involving hydrofluorocarbons,hydrofluoroethers,and hydroxyl radicals—reactions critical in atmospheric fluoride degradation and global warming potential evaluation,yet extremely challenging to model,we achieve transition state optimizations with an impressive,verified success rate of 81.8%for hydrofluorocarbons and 80.9%for hydrofluoroethers.The reported work demonstrates the effectiveness of employing visual representation in chemical space exploration tasks and opens new avenues for the transition structure modeling.
文摘The bimolecular single collision reaction potential energy surface of an isocyanate NCO radical with a ketene CH2CO molecule was investigated by means of B3LYP and QCISD（T） methods. The computed results indicate that two possible reaction channels exist on the surface. One is an addition-elimination reaction process, in which the CH2CO molecule is attacked by the nitrogen atom at its methylene carbon atom to lead to the formation of the intermediate OCNCH2CO followed by a C-C rupture channel to the products CH2NCO＋CO. The other is a direct hydrogen abstraction channel from CHzCO by the NCO radical to afford the products HCCO＋HNCO. Because of a higher barrier in the hydrogen abstraction reaction than in the addition-elimination reaction, the direct hydrogen abstraction pathway can only be considered as a secondary reaction channel in the reaction kinetics of NCO＋ CH2CO. The predicted results are in good agreement with previous experimental and theoretical investigations.
