Random search as high throughput computation
Recorded 30 May 2017 in Lausanne, Vaud, Switzerland
The first principles computation of materials properties was once restricted to the investigation of a few structures at a time. The dramatic growth in the availability of computational resources now allows many thousands of computations to be performed independently and at the same time. One use of this power is to screen for novel or extreme materials, taking existing databases of crystal structures, or modifications of them, and computing their properties. While making use of existing knowledge of materials is an excellent starting point, in many cases, important structures are missing, not known, or even readily accessible. This is particularly the case for materials under extreme conditions, non-crystalline material structure, such defects – point defects, interfaces, or surfaces – and nano-structures.
To complement what we already know, it is essential to be able to make good, diverse and realistic, suggestions of possible structures at the atomic level. I will describe a straightforward approach to first principles structure prediction - Ab Initio Random Structure Searching (AIRSS) , and focus on the technical, high throughput, aspects of the method, from data acquisition to analysis and management.
 C. J. Pickard, and R. J. Needs, Journal of Physics-Condensed Matter, 23(5), 053201 (2011)
Watched 345 times.Watch