Ab initio molecular dynamics with quantum nuclear effects: potential surfaces and gradients from on-the-fly fragment based electronic structure methods

斯里尼瓦桑湾印第安纳州大学的艾扬格获得了化学理论、模型和计算方法项目的奖项支持。 能源、环境和生物研究前沿的许多问题都需要对电子和原子核进行量子力学处理。但是，即使在今天，对这些问题的详细量子描述也太复杂了。的高性能计算环境。这是因为这些问题的计算复杂度随着系统规模呈指数级增长，这使得它们难以处理。 艾扬格和他的研究小组正在开发新的计算方法来解决这些问题。这些方法有望对生物酶、大气化学和材料科学中的广泛问题的研究产生重大影响，包括聚合物电解质燃料电池中的氢转移研究。 这些方法处于现代计算量子化学和化学物理学的前沿。因此，PI组的学生有机会学习和发展新的理论方法，并将这些方法应用于重要的实际问题。研究结果，包括计算机代码以及新的科学思想，被广泛传播到科学界。具体来说，计算机程序完成后将作为NSF资助的SEAGrid科学网关的一部分出现。此外，PI继续发起化学系新生的参与和指导。PI进一步设想通过开发学生指导的基于网络的学习方法来连接化学的子学科，并实现连续而不是离散的本科生教育。精确计算复杂化学体系的经典轨线和势能面，使之与高水平的电子结构相一致，是计算化学和化学物理中的一个重大挑战。 艾扬格和他的同事们正在这一领域做出贡献，包括发展动态分子动力学方法， 允许在MP2和耦合团簇水平上精确处理电子结构，使用类似DFT的方法。此外，他们正在计算量子力学核自由度的势面，与更高层次的电子结构方法（如MP2和耦合簇）非常一致，计算成本与DFT相当。虽然从头算分子动力学（AIMD）方法很有吸引力，因为它们不需要先验拟合的势面，但这种优势通常被排除，因为对势和力的动态评估是一个很强的计算约束。因此，大多数AIMD的实现仅限于电子结构的密度泛函理论（DFT）处理。 本项目的目标是缓解这种计算限制，并使用允许更高水平电子结构、后Hartree-Fock精度的有效方法，使AIMD更常规和更广泛地适用。具体而言，该提案涉及（a）开发新方法，以有效计算AIMD和Car-Parrinello类方法的动态电子结构，与MP2、耦合簇和其他后Hartree-Fock方法一致，（B）使用这些方法计算势面，以促进动态量子动力学。研究的化学问题包括：（i）质子化水分子团簇的构象动力学和振动性质的研究，以及（ii）反应路径的精确势能面的确定和羟基异戊二烯氧化途径中氢转移反应的量子波包动力学研究，羟基异戊二烯是一种生物挥发性有机化合物。

Srinivasan S. Iyengar of Indiana University is supported by an award from the Chemical Theory, Models and Computational Methods program. Many problems at the forefront of energy, environmental and biological research demand the quantum mechanical treatment of electrons and nuclei. But the detailed quantum description of such problems is much too complex even in today?s high performance computing environment. This is because the computational complexity in these problems grows exponentially with system size, which makes them intractable. Iyengar and his research group are developing new computational methods to address these issues. These methods are poised to have major impact on the study of a wide class of problems in biological enzymes, atmospheric chemistry and materials science including study of hydrogen transfer in polymer electrolyte fuel cells. The methods are at the forefront of modern computational quantum chemistry and chemical physics. Hence students in the PI's group have the opportunity to learn and develop new theoretical methods and apply these methods to important practical problems. The results, involving computer codes as well as novel scientific ideas are widely disseminated to the scientific community. Specifically, the computer programs when completed will appear as part of the NSF-funded SEAGrid science gateway. In addition, the PI continues to initiate participation and mentoring of freshman by chemistry faculty. The PI further envisions homogenizing the chemistry curriculum by developing student directed web-based learning methods to connect the sub-disciplines of chemistry and achieve a continuous rather than discrete education of undergraduates. Computing accurate classical trajectories and potential surfaces in agreement with high levels of electronic structure for complex chemical systems is a major challenge in computational chemistry and chemical physics. Iyengar and co-workers are making contributions in this area, including the development of on-the-fly molecular dynamics methods that allow accurate treatment of electronic structure at the MP2 and Coupled Cluster level, using methods that scale like DFT. Furthermore, they are computing potential surfaces for quantum-mechanical nuclear degrees of freedom that are in good agreement with higher levels of electronic structure methods such as MP2 and Coupled Cluster, at computational costs comparable to DFT. While ab initio molecular dynamics (AIMD) methods are appealing because they do not need an a priori fitted potential surface, this advantage is normally precluded since on-the-fly evaluation of potential and forces is a strong computational constraint. Hence most implementations of AIMD are limited to density functional theoretic (DFT) treatment of electronic structure. The goal of this project is to alleviate this computational constraint and make AIMD more routinely and broadly applicable using efficient methods that allow higher levels of electronic structure, post-Hartree-Fock accuracy. Specifically, the proposal deals with (a) the development of new methods to efficiently compute on-the-fly electronic structure for AIMD and Car-Parrinello-like methods in agreement with MP2, Coupled Cluster and other post-Hartree-Fock methods, (b) the computation of potential surfaces using these approaches to facilitate on-the-fly quantum dynamics. Chemical problems to be studied include, (i) the study conformation dynamics and vibrational properties of protonated water clusters, and (ii) the determination of accurate potential surfaces for reaction paths and quantum wave packet dynamics studies on hydrogen transfer reactions involved in the oxidation pathways of hydroxyl-isoprene, which is a biogenic volatile organic compound.