Efficient and Accurate Quantum Simulation for Large Periodic Systems

描述（由申请人提供）：Q-Chem是一个最先进的商业量子化学程序，用于在广泛的学科中模拟分子过程，包括生物学，化学和材料科学。在所有的量子模型中，密度泛函理论（DFT）是应用最广泛的。在这个SBIR应用中，我们寻求将分子DFT计算的最新进展应用于周期系统。目前能够处理PBC的最先进的程序在准确性和效率方面存在重大缺陷，并且缺乏对非常大的单元的能力。近年来，我们开发了一些新的分子系统的DFT方法，在不损失精度的情况下，将DFT计算的效率提高了几倍。在这个第一阶段的建议中，我们计划将这些方法扩展到周期系统。我们将把我们的多分辨率交换相关（mrXC）方法应用于周期系统。该方法将两种不同类型的数值网格分别与周期系统和分子量子化学的传统无缝结合。mrXC在不牺牲精度的情况下利用了这两种网格。对于库仑问题，我们在应用傅里叶变换库仑方法的基础上，提出了一种有效处理四中心积分的新方法。哈密顿量的对角化成为包含数百个或更多原子的单元胞的计算瓶颈。我们将应用我们最近建立的绝对定域分子轨道模型，该模型对分子簇的研究非常有效。在这个项目结束时，Q-Chem将成为一个独特的高效和精确的大型周期系统的电子结构包。

DESCRIPTION (provided by applicant): Q-Chem is a state-of-the-art commercial quantum chemistry program that is used to model molecular processes over a wide range of disciplines, including biology, chemistry, and materials science. Among all the quantum models, density-functional theory (DFT) is the most-widely applied. In this SBIR application, we seek to apply the most recent advances in the molecular DFT calculation to the periodic systems. The current state-of-art programs capable of handling the PBC has major deficiencies in accuracy and efficiency, and lacks the capability for very large unit cells. In recent years, we have developed some new DFT methods for molecular systems that have improved the efficiency of the DFT calculation by several folds without loss of accuracy. In this Phase I proposal, we plan to extend those methods to the periodic systems. We will apply our multiresolution exchange-correlation (mrXC) method to the periodic systems. This method combines seamlessly the two different types of numerical grid with traditions in periodic systems and the molecular quantum chemistry, respectively. mrXC takes advantage of both grids without sacrifice of accuracy. For the Coulomb problem, we will develop a new method to treat the four-center integrals efficiently, in addition to the applying the Fourier-transform Coulomb method. The diagonalization of the Hamiltonian becomes a computational bottleneck for unit cells containing hundreds of atoms or more. We will apply our recently developed absolutely localized molecular orbitals model, which has been very effective for the study of molecular clusters. At the end of this project, Q-Chem will become a uniquely efficient and accurate electronic structure package for large periodic systems. PUBLIC HEALTH RELEVANCE: This project aims to develop and implement efficient and accurate DFT methods for periodic systems. DFT is at the core of molecular modeling and is applied widely in biological research/development and in drug discovery. Periodic systems are routinely used in biological simulations. There is a lack of efficient and accurate DFT methods for this type of systems. The efficient and accurate application of DFT to the periodic systems will significantly increase researchers' quality of work and extend the application scope of quantum simulation.