From electrons to phase diagrams (8-10 June 2022)

Today large numbers of DFT calculations are a routine task, thanks to efficient DFT codes, automatized workflow management, and powerful high-performance supercomputers. The DFT data can then be used to parameterize accurate interatomic potentials, in particular potentials based on machine learning (ML) approaches that promise near-DFT accuracy even for complex multicomponent materials. When combined with efficient sampling methods for the computation of free energies, these ML models enable to estimate phase diagrams directly from DFT data and help to supplement and assess experimental input.

At this three-day workshop, we will provide tutorials and hands-on sessions that cover the complete chain from high-throughput electronic structure calculations to the computation of phase diagrams. Day 1 will focus on automated workflows for the generation of DFT data based on the pyiron IDE. On day 2 we will present how to parameterize and validate different classes of interatomic potentials from DFT reference data using dedicated software packages (Atomicrex, Pacemaker, RuNNer). Day 3 will then introduce methods and tools for the computation of thermodynamic properties and phase diagrams as implemeted in the package Calphy.

Onsite and online: the workshop will be run in hybrid mode, onsite and online participation is possible.

Onsite location: Ruhr-University Bochum, Bochum, Germany

Registration: Participation is free of charge. Please register by 15 May 2022 here: http://potentials.rub.de/2022/registration.php

Website: http://potentials.rub.de/2022/index.php