Molecular Balances: Computation meets Experiment

-该项目试图解决的研究问题/该项目的目标；分子平衡是相对较小的系统，旨在通过实验隔离给定类型的相互作用，以及它在多大程度上对整个系统的行为负责。虽然关于分子平衡的实验文献稳步增长，但其理论和计算对应的文献却严重不发达。总体目标是从根本上改变这种不可接受的局面。更具体的目标是：-使用带有外部色散校正的DFT将相互作用量子原子(IQA)的能量分配方法应用于研究良好的命名天平(例如Cockroft、Gung、Shimizu)(D3，D4)。-通过将两个粒子密度矩阵(2PDM)电子关联能量与一个粒子密度矩阵(1PDM)电子关联能量联系起来，能够将非DFT后Hartree-Fock波函数结合到与IQA兼容的色散能量评估中。-应用内部方法相对能量梯度(REG)，根据能量贡献在解释系统整体行为方面的重要性对它们进行排名。-将内部力场FFLUX应用于水溶液中选定的分子平衡上，并使用REG-IQA监测分配的能量轨迹。研究相应的多极静电学。-使用开发的方法和获得的知识来获得对褶皱的化学洞察力，并雄心勃勃地开始指导其合成设计。-将采取的方法来回答这些问题(学生将实际做什么)；学生将受益于三种主要方法的精确组合：IQA、REG和FFLUX。这种“交叉受精”组合带来的现实主义和复杂性将为分子平衡提供前所未有的洞察力。因此，这项工作将为上述文献提供急需的计算对应。学生将有效地使用计算机程序，但需要编写特定的Python脚本来处理本地计算机将生成的大量数据。他还将使用机器学习方法高斯过程回归来创建模型，从中可以快速计算原子能量。作为另一种类型的活动，学生将在2pdm(参考或“校准器”)和1pdm之间找到一种模式，这是获得与IQA兼容的非DFT电子相关能从而获得色散能量的唯一计算可行的方法。最后，学生将在Fortran 90中进行一些编程，以修改DL-FFLUX，使其能够处理基于IQA的色散能量。-研究的新的工程和/或物理科学内容(将其置于EPSRC职权范围内的科学)。总体而言，该项目符合物理科学主题的两个分项目，即“计算和理论”和“凝聚态物质：电子结构”，与超分子化学有最终联系。该项目也是在定向组装具有靶向性质的扩展结构的化学科学大挑战(DAESTP)下进行的。最后，这个项目有一个强大的机器学习组件，因此与人工智能重叠。

- the research questions the project is trying to address/the objectives of the project; Molecular balances are relatively small systems designed to experimentally isolate a given type of interaction, and to what extent it is responsible for the behaviour of the total system. While the experimental literature on molecular balances steadily grows, its theoretical and computational counterpart is woefully underdeveloped. The overall objective is to fundamentally change this unacceptable situation. More specific objectives are:- apply the energy partitioning method Interacting Quantum Atoms (IQA) on well-studied named balances (e.g. Cockroft, Gung, Shimizu) using DFT with external dispersion correction (D3, D4).- enable the incorporation of non-DFT post-Hartree-Fock wavefunctions in the IQA-compatible evaluation of dispersion energy by linking the two-particle-density-matrix (2PDM) electron correlation energies to one-particle-density-matrix (1PDM) ones.- apply the in-house method Relative Energy Gradient (REG) to rank energy contributions by their importance in explaining the system's overall behaviour.- apply the in-house force field FFLUX on a selection of molecular balances in aqueous solution and monitor partitioned energy trajectories with REG-IQA. Study the corresponding multipolar electrostatics. - Use the developed methodology and acquired knowledge to obtain chemical insight in foldamers and, ambitiously, start guiding their synthetic design.- the approach that will be taken to answer these questions (what the student will actually be doing); The student will benefit from the precise combination of three main methodologies: IQA, REG and FFLUX. The realism and sophistication following from this "cross-fertilising" combination will offer unprecedented insight into molecular balances. This work will thereby deliver the much needed computational counterpart to the literature mentioned above. The student will make efficient use of the computer programs but will need to write specific Python scripts to handle the wealth of data that the local computers will generate. He will also use the machine learning method Gaussian Process Regression to create models from which the atomic energies can be quickly calculated. As a further type of activity, the student will find a pattern between the 2PDM (the reference or "calibrator") and the 1PDM, which is the only computationally feasible approach to obtain IQA-compatible non-DFT electron correlation energies and thus dispersion energy. Finally, the student will carry out some programming in FORTRAN90, in order to modify DL-FFLUX so that it can handle IQA-based dispersion energies. - the novel engineering and/or physical sciences content of the research (the science that places it within EPSRC's remit). Overall the project fits under two sub-portfolios of the Theme Physical Sciences, namely "Computational and Theoretical", and "Condensed Matter: Electronic Structure" with an ultimate connection to Supramolecular Chemistry. This project also resorts under the Chemical Sciences Grand Challenge of Directed Assembly of Extended Structures with Targeted Properties (DAESTP). Finally, there is a strong Machine Learning component to this project, and thus overlap with Artificial Intelligence.