Electronic Structure; Simulation; Structure Prediction; Materials Design
Stephan Mohr Marc Eixarch Maximilian Amsler Mervi J.Mantsinen Luigi Genovese (2018), Linear scaling DFT calculations for large tungsten systems using an optimized local basis, in Nuclear Materials and Energy
Maximilian Amsler Zhenpeng Yao and Chris Wolverton (2017), Cubine, a Quasi Two-Dimensional Copper–Bismuth Nanosheet, in Chem. Mater.
, 29(22), 9819.
Kim S Hegde VI Yao Z Lu Z Amsler M He J Hao S Croy JR Lee E Thackeray M Wolvert C, First-principles Study of Lithium Cobalt Spinel Oxides: Correlating Structure and Electrochemistry, in ACS Appl. Mater. Interfaces
Materials design has become one of the most rapidly increasing fields of condensed matter science during the last years. Discovery of novel materials are called for with potential applications in various technologically relevant fields such as energy production and storage, novel electronic devices, data storage, medical application and catalysis. Instead of synthesizing and characterizing novel materials in time-consuming, expensive experiments, theoretical material scientists have started to create databases obtained from all materials known to date and to systematically analyze them with high-throughput computation methods, trying to gain insight on how chemistry and structures are linked to material properties and to extract patterns from the underlying data. New, innovative computational methods are needed to efficiently characterize the continuously growing amount of materials data, and techniques need to be developed to extrapolate materials properties to design novel, undiscovered compounds with improved properties. The presented project is aimed at solving two key limitations in materials design. First, most structure prediction methods do not access the knowledge stored in the large structural databases. By combining the Minima Hopping structure prediction scheme and machine-learning techniques a sophisticated method will be developed to explore new chemistries and to discover compounds with improved or new properties in energy applications. The new method will be applied to real-life challenges in materials design to find thermoelectric and hydrogen storage materials, or materials for use in photovoltaic applications. Second, Quantum Monte Carlo methods will be used to refine the energetic ordering of different phases in various compounds to significantly improve the predictive power by going beyond the accuracy of conventional density functional theory calculations used in current approaches. The outcome of this project will be of great value for many material scientists and will considerably accelerate the theoretical discovery of novel materials.