Theoretical Studies of Spectroscopy and Dynamics of Molecular Systems

新墨西哥州大学的Hua Guo得到了理论和计算化学计划的支持，以开发准确有效的方法来研究小分子系统的高激发光谱和反应动力学。 采用的方法是基于递归生成的Krylov子空间中的特征值问题部分解决。 由于主要的计算任务涉及矩阵向量乘法，这样的方法具有良好的缩放律。 通过谨慎选择离散化方案，递归方法（如Lanczos算法或Chebyshev传播）可以有效计算高激发振转能级以及受长寿命共振影响的系统的反应概率、截面和速率。 原型系统亚甲基，由于存在的伦纳-泰勒耦合和几个锥形交叉点的兴趣，将使用高水平的从头计算，包括一个新的策略进行研究。另一个要研究的体系是乙炔和它的高能亚乙烯异构体。 为了在全维中表征这种基本异构化，将应用一种新的方案来收敛异构化势垒以下的所有振动能级，这项研究处于理论化学物理的前沿，其结果有望为重要分子系统和相关化学过程的实验观察提供有价值的见解。 此外，还将普遍提高计算方法，并将研究产生的计算机代码免费提供给感兴趣的研究人员。 由于该项目涉及大型本征或线性方程求解器，因此其结果可能有利于出现类似问题的其他科学和工程领域。 实例包括但不限于高分辨率成像、纳米材料、光子晶体和生物物理学。

Hua Guo of the University of New Mexico is supported by the Theoretical and Computational Chemistry Program to develop accurate and efficient methods to study highly excited spectroscopy and reaction dynamics of small molecular systems. Methods to be employed are based on recursive generation of the Krylov subspace in which the eigenproblem is partially solved. Since the major computational task involves matrix-vector multiplications, such methods have favorable scaling laws. With prudently chosen discretization schemes, recursive methods such as the Lanczos algorithm or the Chebyshev propagation allow efficient computation of highly excited ro-vibrational levels as well as reaction probabilities, cross sections, and rates of systems affected by long-lived resonances. The prototypical system methylene, of interest due to the presence of Renner-Teller coupling and several conical intersections, will be studied using high level at initio calculations including a new strategy. The other system to be investigated is acetylene and its higher energy vinylidene isomer. To characterize this fundamental isomerization in full dimension, a new scheme will be applied to converge all vibrational levels below the isomerization barrier.This research is at the cutting edge of theoretical chemical physics, with outcomes expected to provide valuable insights into experimental observations of important molecular systems and associated chemical processes. As well, computational methodologies will be advanced in general, with computer codes resulting from the research to be made freely accessible to interested researchers. Since this project deals with large eigen or linear equation solvers, outcomes may benefit other fields in science and engineering where similar problems emerge. Examples include but are not limited to high-resolution imaging, nanomaterials, photonic crystals, and biophysics.