HCl在Ag和Pt金属表面解离吸附的精确量子动力学模拟研究

The dissociative chemisorption of molecular species on transition-metal surfaces is the essential step in heterogeneous catalytic reactions. Due to its significant importance, a quantitative understanding is greatly desired. Although many significant progresses have been achieved in the past few decades in this gas-surface area, only for very limited reactions, the theoretical calculations can reproduce the experimental measured reaction probability with chemical accuracy (within 1 kcal/mol). The large discrepancies between the theoretical calculations and experiments in gas-surface reactions can be attributed to two mainly reasons: the first one is the error of the density function theory (DFT) with the approximate exchange-correlation functionals which were used to construct the potential energy surfaces (PESs); the second one is the complexity of the surface lattice motions which were very difficult to simulate accurately. In this project, the full-dimensional global PESs for the dissociative chemisorption of HCl on rigid Ag(111) and Pt(111) will be constructed using the neural networks (NN) method based on the DFT with the PBE, PW91, and RPBE functionals, respectively. An explicit comparison between the three GGA functionals will be given through the time dependent wave packet (TDWP) method. We will also use the NN method to fit the new PESs in which the lattice motion will be incorporated and do the quantum calculations to investigate the effect of the lattice motion. Our aim is to quantitatively reproduce the experimental measured reaction probability for this two title reaction systems. Furthermore，we hope that we can obtain some general rules for the two directions mentioned above to give some promising theoretical support for the chemically accurate simulation of other gas phrase molecules especially for the polyatomic molecules dissociating reactions through this project.

气相分子在金属表面解离吸附是工业异相催化过程中很重要的步骤。尽管这些年该领域的理论模拟工作取得了很大的进步，但在定量模拟分子束实验结果方面仍存在非常大的缺陷。只有非常少的反应体系理论能大体重复实验结果，误差约为1kcal/mol。原因主要有两方面：一是目前用来计算构型能量的密度泛函理论（DFT）采用的交换关联函数的误差，二是由于金属表面晶格运动的复杂性导致其很难被精确的模拟。本项目以HCl在Ag和Pt表面解离吸附为研究对象，针对上面提到的两个问题进行深入探究，拟通过神经网络方法基于DFT下不同的泛函来分别构建全维高精度势能面，通过含时波包法研究不同泛函对反应机制的影响，另外拟采用同样方法来拟合晶格运动结果，从量子层面对其进行研究。最终希望能定量上精确模拟这两个反应体系的分子束实验结果，并通过本项目为以后的气相-金属反应体系特别是多原子分子解离体系的精确性模拟提供良好的理论支持。

气相分子在金属表面解离吸附是工业异相催化过程中很重要的步骤。理论工作在定量模拟分子束实验结果方面仍存在非常大的缺陷。只在非常少的反应体系理论能大体重复实验结果。本项目以HCl在Ag和Pt表面解离吸附为研究对象，拟通过神经网络方法基于DFT下不同的泛函来分别构建全维高精度势能面，通过含时波包法研究不同泛函对反应机制的影响。希望能定量上精确模拟这两个体系的实验结果。我们构建了HCl/Ag(111)基于PBE、PW91、PBE泛函的六维势能面，基态反应几率显示PW91最大，RPBE最小。研究发现这三种泛函在定性上对反应的振动激发和转动方向性效应影响不大，只是定量上有很大的差别。我们又构建了HCl+Ag(100)/Ag(110)体系的六维势能面，发现反应势垒按Ag(110)<Ag(100)<Ag(111)的顺序变大，动力学计算发现振动激发等效应在这三个晶面上也基本一致。接下来我们拟合了HCl+Pt(111)基于PBE和RPBE的六维势能面，发现该体系top位置的反应几率和其他位置结果相差不大，而且第一激发态的反应几率在总能条件下却比基态结果小，这和之前HCl解离体系结果有很大的差异，主要缘于其过渡态的位置相比其他HCl解离体系低。另外我们构建了H2O+Ni(100)体系的九维势能面并利用全维量子动力学来验证位置平均近似的适用性。研究还发现该体系模式选择效应很强烈。 通过对HCl+Ag体系研究发现不同DFT泛函以及不同金属晶面对反应性质的影响，只在定量上有很大差别，在HCl+Pt体系中发现了反应过渡态的位置和势能面的拓扑结构都对反应活性有很大的影响，这为以后的气相-金属反应体系特别是多原子分子解离体系的精确性模拟提供了良好的理论支持。