Efficient and Adaptive Methods for Computing Accurate Potential Surfaces for Quantum Nuclear Effects: Applications to Hydrogen-Transfer Reactions
Efficient and Adaptive Methods for Computing Accurate Potential Surfaces for Quantum Nuclear Effects: Applications to Hydrogen-Transfer Reactions
复制标题
计算量子核效应精确势面的高效自适应方法：在氢转移反应中的应用
DOI：
10.1021/acs.jctc.7b00927
复制
发表时间：
2017
影响因子：
5.5
通讯作者：
Iyengar, Srinivasan S.
中科院分区：
文献类型：
--
作者：
DeGregorio, Nicole;Iyengar, Srinivasan S.
我们提出了两种采样方法来测量势能表面的关键区域。这些采样测量采用 (a) 瞬时量子波包密度、(b) 势面的近似值、其 (c) 梯度，以及 (d) 基于香农信息论的表达式，该表达式估计与量子波包相关的局部熵。这四个标准一起实现了对势表面的定向采样，这似乎正确地描述了势表面的本地振荡频率或本地奈奎斯特频率。然后利用采样函数导出曲面细分方案，该方案离散化多维空间以实现对潜在表面的有效采样。然后将采样的势面与四种不同的插值程序相结合，即（a）局部埃尔米特曲线插值，（b）低通滤波拉格朗日插值，（c）鲍曼及其同事开发的单项式对称化近似（MSA），以及（d）修改的谢泼德算法。采样程序和拟合方案用于计算（a）高度非谐氢键系统中的电势表面，以及（b）研究生物挥发性有机化合物（异戊二烯）中的氢转移反应，其中发现转移氢原子表现出关键的量子核效应。就异戊二烯而言，此处讨论的算法用于沿氢转移反应路径导出多维势面，以测量量子核自由度对氢转移过程的影响。基于减少的计算量，通过使用此处讨论的采样函数对势表面进行最佳采样以及相关势表面的准确性，我们相信该方法将在量子核动力学问题的研究中找到巨大的用途，此处演示了其在氢转移反应和氢键系统中的应用。
We present two sampling measures to gauge critical regions of potential energy surfaces. These sampling measures employ (a) the instantaneous quantum wavepacket density, an approximation to the (b) potential surface, its (c) gradients, and (d) a Shannon information theory based expression that estimates the local entropy associated with the quantum wavepacket. These four criteria together enable a directed sampling of potential surfaces that appears to correctly describe the local oscillation frequencies, or the local Nyquist frequency, of a potential surface. The sampling functions are then utilized to derive a tessellation scheme that discretizes the multidimensional space to enable efficient sampling of potential surfaces. The sampled potential surface is then combined with four different interpolation procedures, namely, (a) local Hermite curve interpolation, (b) low-pass filtered Lagrange interpolation, (c) the monomial symmetrization approximation (MSA) developed by Bowman and co-workers, and (d) a modified Shepard algorithm. The sampling procedure and the fitting schemes are used to compute (a) potential surfaces in highly anharmonic hydrogen-bonded systems and (b) study hydrogen-transfer reactions in biogenic volatile organic compounds (isoprene) where the transferring hydrogen atom is found to demonstrate critical quantum nuclear effects. In the case of isoprene, the algorithm discussed here is used to derive multidimensional potential surfaces along a hydrogen-transfer reaction path to gauge the effect of quantum-nuclear degrees of freedom on the hydrogen-transfer process. Based on the decreased computational effort, facilitated by the optimal sampling of the potential surfaces through the use of sampling functions discussed here, and the accuracy of the associated potential surfaces, we believe the method will find great utility in the study of quantum nuclear dynamics problems, of which application to hydrogen-transfer reactions and hydrogen-bonded systems is demonstrated here.