基于基本不变量神经网络势能面的H+C_(2)D_(2)反应漫游机理研究

Roaming Dynamics of H+C_(2)D_(2) Reaction on Fundamental-Invariant Neural Network Potential Energy Surface

基于最近发展的基本不变量神经网络方法构建的精确全维势能面,本文对H+C_(2)D_(2)→HD+C_(2)D/D_(2)+C_(2)H反应开展了广泛的准经典轨线计算.发现了直接抽取反应途径对整体反应性的影响很小,与漫游路径相比可以忽略不计.乙炔促进的漫游路径在总体反应性中占主导地位,而亚乙烯基促进的漫游路径贡献很小.结果表明,虽然漫游路径可以通过短寿命或长寿命的复合物形成过程进行,但是产物HD与D_(2)的分支比与2:1很接近,这意味着漫游动力学主要是山长寿命的复合物形成过程促成的.并且,这两个产物通道的角分布也完全不同.这些计算结果为深入了解同位素效应在双分子反应漫游动力学中的重要性提供了有价值的见解.

We performed extensive quasiclassical trajectory calculations for the H+C_(2)D_(2)→HD+C_(2)D/D_(2)+C_(2)H reaction based on a recently developed,global and accurate potential energy surface by the fundamental-invariant neural network method.The direct abstraction pathway plays a minor role in the overall reactivity,which can be negligible as compared with the roaming pathways.The acetylenefacilitated roaming pathway dominates the reactivity,with very small contributions from the vinylidene-facilitated roaming.Although the roaming pathways proceed via the short-lived or long-lived complex forming process,the computed branching ratio of product HD to D_(2) is not far away from 2:1,implying roaming dynamics for this reaction is mainly contributed from the long-lived complex-forming process.The resulting angular distributions for the two product channels are also quite different.These computational results give valuable insights into the significance and isotope effects of roaming dynamics in the biomolecular reactions.