|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Computational Materials Discovery and Optimization
||A Combined Experimental-computational Approach to Determining Nanoscale Structures
||Spencer Hills, Alper Kinaci, Fatih Sen, Maria Chan
|On-Site Speaker (Planned)
Nanoscale structures are difficult to determine due to reduced symmetry compared to bulk. Computational predictions of the lowest energy structure require sampling of incredibly large parameter space of all possible atomistic configurations, and suffer from uncertainties in both energy evaluations and global optimizer effectiveness. Experimentally, the pair distribution function (PDF) provides information on the structure of a nanocluster, but the inversion of the PDF signals is non-trivial and sometimes non-unique.
In this talk, we will discuss an approach to combine computational and experimental information into one optimization framework to reduce these difficulties. We use a multi-objective genetic algorithm to minimize energy calculated by density functional theory (DFT) and PDF residual. We benchmark this framework on Au nanoclusters, and compare them to using only DFT or PDF. We find that the combined approach accurately determines the target structure more often than single data type optimization.
||Planned: Supplemental Proceedings volume