Development of Fast and Accurate Computational Chemistry Methods based on Atom-Centred Potentials and their Application to Crystal Structure Prediction

尽管现代计算基础设施的能力大大增加，基于量子力学的新方法的发展，仍然需要开发模拟技术，使之能够高精度地预测大系统的广泛化学性质。开发基于量子力学(QM)方法的世界领先小组主要致力于创建：1)半经验方法(SEM)，旨在预测包含10,000个原子的体系(例如蛋白质)的结构；1和2)改进的基于密度泛函理论(DFT)的方法，可用于以合理的精度预测最多包含100个原子的分子的各种性质。我的研究计划旨在开发易于使用的基于量子的方法，计算成本介于扫描电子显微镜和离散傅立叶变换之间，适用于多达1000个原子的分子。我们建议开发具有比传统DFT方法更高的目标精度的方法。因此，我的项目中出现的创新将为化学、物理、制药科学和其他领域的广大用户带来有价值的新模拟工具。

Despite enormous increases in the power of modern computing infrastructure and the development of new quantum mechanical-based methods, there remains a need to develop simulation techniques that enable the prediction of a wide range of chemical properties of large systems with high accuracy. World-leading groups developing quantum mechanics (QM) based methods have focused primarily on creating: 1) semi-empirical methods (SEMs) aimed at structure prediction in systems containing >10,000 atoms (e.g. proteins),1 and 2) improved density-functional theory (DFT) based methods that can be used to predict a wide range of properties for molecules containing up to 100 atoms with reasonable accuracy. My research program aims to develop easy-to-use quantum-based methods with computational costs that are intermediate to SEMs and DFTs, and are applicable to molecules of up to 1000 atoms. We propose to develop methods with target accuracies that are much higher than conventional DFT methods. Therefore, the innovations that emerge from my program will bring valuable new simulation tools to a broad community of users in chemistry, physics, pharmaceutical sciences, and other areas.