[交流] 【求助完毕】怎么用第一性原理分子动力学来做复杂结构的粗优化？

凝聚态和固体表面催化反应相关计算都有建模型和对模型的结构优化的步骤，复杂的初始结构的优化往往是非常痛苦和耗时的过程。如果能够用先第一性原理分子动力学来做复杂结构的粗优化，再按照正常的DFT计算来做结构优化，可能会极大的缩短时间。

下面一篇文章在TiO2面上放两层Au原子条带作为催化剂模型，初始结构就是先用分子动力学来优化的。

Zhi-Pan Liu etc. PRL. Catalytic Role of Metal Oxides in Gold-Based Catalysts: A First Principles Study of CO Oxidation on TiO2 Supported Au.

To obtain a physically meaningful structure for the Au/TiO2 composite system, Nose thermostat molecular dynamics simulation (T = 200 K, 5 fs/step, 500 steps) was performed. Five low energy structures were selected and each one was further optimized by a conjugate gradients procedure. The lowest energy structure was then used for all subsequent studies. In all calculations, the Au atoms and the top three layers of TiO2(110) surface were relaxed, keeping the rest of the TiO2 atoms
fixed at their bulk-truncated positions.

Nose thermostat molecular dynamics simulation 看起来是针对恒温系综的。对于一般的supercell体系，是不是就是NVT系综呢？

Vasp 做分子动力学模拟设置 SMASS>0(Nose mass parameter.)(此时默认IBRION=0),就会用Nose算法模拟正则系综。但是，如何如上面一段英文那样得到Five low energy structures呢？

[ Last edited by cenwanglai on 2011-5-15 at 22:10 ]

回复此楼