混合金属团簇几何结构的优化算法及其应用研究

For the importance in nano property-modification, catalysis, and magnetism etc, the mixed metal clusters have been attracted many interests. The first step to understand the physical and chemical properties of clusters is to determine the cluster’s geometries (especially the ground-state structures). As it is still difficult to directly measure the geometry of free clusters containing more than three atoms experimentally, thus the combination with theoretical calculations is particularly important for acquiring information of the ground-state structure of a cluster. ..In this project, the constant-temperature molecular dynamics method is adopted based on the Gupta-type interatomic potential, a combination with the ideal quenching technique and the constant-temperature molecular dynamics method (as called the molecular dynamics quenching algorithm) is proposed as a new global optimization method to search the ground-state geometry of the mixed metal cluster. Meanwhile, by combining the molecular dynamics quenching algorithm with the traditional genetic algorithm, a global optimization algorithm of hybrid genetic algorithm is proposed. With focusing on some mixed clusters containing transition metal element, the effectiveness as well as the high-efficiency to acquire the ground-state geometry of the mixed metal cluster through the two global optimization algorithms will be investigated in detail. And the inherent relationship between the efficiency for acquiring the ground-state structure and the different type, composition and size of the mixed cluster will also be investigated in detail. Meanwhile, the reliability of results from the two optimization algorithms will be explored in further through the combination of first-principles calculations, and the electronic structure of the mixed-metal clusters will also be investigated in this project.

由于在纳米颗粒改性、催化及磁性等方面的重要性，混合金属团簇已引起了人们的极大研究兴趣。考虑到团簇结构（尤其是基态结构）的确定是了解团簇物理、化学性质的基础，而实验上仍难以直接测量多原子自由团簇的几何结构，因而结合理论计算以获取团簇基态结构信息就显得尤为重要。.本项目基于Gupta原子间相互作用势函数采用分子动力学方法，提出理想淬火技术与恒温分子动力学方法的结合（分子动力学淬火算法）作为寻找混合团簇基态几何结构的新的全局优化算法。同时，将分子动力学淬火算法与传统的遗传算法相结合提出了混合遗传算法的全局优化算法。具体以若干含过渡金属混合团簇为研究对象，详细研究这两种全局优化算法在搜寻混合金属团簇基态几何结构方面的有效性（高效性）、及搜寻基态结构的效率与混合团簇不同种类、组分及尺寸的内在联系。同时，结合第一性原理计算方法进一步探讨优化算法所得结果的可靠性，并进一步研究混合金属团簇的电子结构特征。

本项目中，将第一性原理高温分子动力学几何结构搜索方法推广至全温度区间的全局几何结构搜索优化方法，并详细研究了该方法运用于（半）经验势时的结果与量子力学结果的内在关联，说明了该优化算法的可靠性。同时，系统研究若干混合金属团簇（主要为催化混合金属团簇）的电子结构特征。从微观机制上揭示与理解混合金属团簇体系的物理性质（如混合金属团簇的基态几何结构、电子结构与其组分、尺寸的内在联系），为解释实验现象和进一步的理论研究提供理论依据和物理图象。.此外、超额完成项目申请（任务）书中科学研究成果及人才培养目标。