Hydrogen Combustion using IRC, AIMD and normal modes

The generation of reference data for deep learning models is challenging for reactive systems, and more so for combustion reactions due to the extreme conditions that create radical species and alternative spin states during the combustion process. Here, we extend intrinsic reaction coordinate (IRC) calculations with <em>ab initio </em>Molecular dynamics (AIMD) simulations and normal mode displacement calculations to more extensively cover the potential energy surface for 19 key reaction channels involved in hydrogen combustion. A total of ~290,000 potential energies and ~1,260,000 nuclear force vectors are evaluated with a high quality range-separated hybrid density functional, ωB97X-V, to construct the reference data set, including transition state ensembles, for the deep learning models to study hydrogen combustion reaction.

对于反应体系而言，为深度学习模型生成参考数据本就颇具挑战；而燃烧反应因过程中存在极端条件，会催生自由基与多种自旋态，其参考数据的生成难度更甚。本研究将从头算（ab initio）分子动力学（AIMD）模拟与简正模位移计算拓展应用于内禀反应坐标（intrinsic reaction coordinate, IRC）计算流程，以更全面地覆盖氢燃烧相关19个关键反应通道的势能面。本次研究采用高精度范围分离杂化密度泛函（range-separated hybrid density functional）ωB97X-V，共计评估了约29万个势能点与约126万个核力矢量，以此构建包含过渡态系综的参考数据集，以供深度学习模型开展氢燃烧反应相关研究。