DFTB-based global optimization approach to large scale interfacial systems

该项目的目的是开发一种面向大规模界面系统结构测定的通用计算协议。这样的系统通常具有一定程度的复杂性，其决定了利用全局优化技术的无偏几何确定。同时，内在目标是需要捕获量子效应的维度，因此能量学需要电子结构方法。这两个约束一起导致目前在很大程度上棘手的计算成本。在这个项目中，我的目标是克服这一瓶颈，通过引入几何预筛选在一个更具成本效益的理论水平，并建立一个专门面向界面系统的采样方法，往往是未知的组成和化学计量。对于后者，我将依赖于巨正则采样技术，而对于前者，我将依赖于自洽电荷密度泛函紧束缚（SCC-DFTB）。将作出具体努力，为SCC-DFTB制定一个一致的参数化方案，使这种方法适用于周期系统的所有元素。开发的协议将被应用到不同的问题，共享广泛的结构筛选的必要性，即Si和RuO 2的表面重建和氧化诱导的再分散的表面支持的TanOm集群的预测。

Aim of the project is the development of a versatile computational protocol geared towards the structure determination of large-scale interfacial systems. Such systems are generally of a level of complexity that dictates unbiased geometry determination with global optimization techniques. Simultaneously, intrinsically targeted are dimensions where quantum effects need to be captured and thus electronic structure methods are required for the energetics. Both constraints together lead presently to largely intractable computational costs. In this project I aim to overcome this bottleneck by introducing a geometric pre-screening at a more cost efficient level of theory and establishing a sampling approach specifically geared towards interfacial systems with often unknown composition and stoichiometry. For the latter I will rely on grand-canonical sampling techniques, while for the prior I will rely on Self-Consistent Charge Density Functional Tight Binding (SCC-DFTB). Specific efforts will be placed into a consistent parametrization scheme for SCC-DFTB that makes this approach available for all elements of the periodic systems. The developed protocol will be applied to diverse problems sharing the necessity for extensive structural screening, namely the prediction of Si and RuO2 surface reconstructions and the oxidation-induced redispersion of surface-supported TanOm clusters.