Lee Hsun Lecture Series                                                             

Topic: Ab initio quo vadis: first principles calculations now and in the future 

Speaker: Prof. Georg Kresse 

Faculty of Physics and Centre for computational materials sciences, University of Vienna, 1090 Vienna, Austria 

Time: 10:00-12:00, (Thur.) Sept.26, 2019 

Venue: Room 403，Shi Changxu Building, IMR CAS 

Abstract: 

First principles calculations are now common practice in many areas of materials research, condensed matter physics, and chemistry. This great success is mostly linked to the now ubiquitous density functional theory, for which Walter Kohn jointly with John A Pople received the Nobel Prize in 1998. A little bit more than 20 years have passed, and the materials modelling landscape has changed drastically, with millions of calculations performed every year solving problems that many computational scientists had not dreamed off few decades ago. The Vienna ab initio package (VASP) was an important stepping stone towards these developments, and VASP remains the most popular code to perform density functional theory calculations for condensed matter systems. I will briefly recapitulate the history of VASP and give a small glimpse of what can be done nowadays. The main thrust of this talk is, however, on what the future will bring. Specifically, how can we make materials modelling truly predictive for real materials taking into account all their intricate complexities. The first issue to address is that density functional theory, although in principle exact, needs to rely on approximate functionals, and importantly, there is no systematic way to improve these functionals. Many body perturbation theory is a means to resolve this dilemma, since it is, at least in principle, systematically improvable.  The second issue we need to solve is that, even though density functional theory calculations are relatively fast on today's high performance computers, large system sizes and long simulation times relevant to many processes are not accessible. Using machine learning to construct accurate interatomic force fields can lessen this issue.  I show for the example of a novel organic perovskite, how we can combine insight from many body perturbation theory with the computational speed of machine learning to obtain finite temperature phase diagrams in beautiful agreement with experiment. This example shows that accurate predictions for finite temperature properties, phase diagrams, chemical potentials, and catalytic properties of ordered and disordered materials will soon become routine. An exciting future lies ahead for computational materials modelling. 

CV 

Georg Kresse was born on 21 July 1967 in Vienna. He completed his doctoral thesis at the Institute for Theoretical Physics of the Vienna University of Technology in 1993 under the supervision of Jürgen Hafner. Kresse then worked as a scientific assistant in Vienna and held a postdoctoral position at Staffordshire University with Mike Gillan. After his habilitation at the Vienna University of Technology in 2001, he was offered a full professorship from Oxford University as well as from the University of Vienna.  He accepted the chair for Computational Quantum Mechanics in Vienna in 2007. Since 2011 Kresse is a full member of the Austrian Academy of Sciences and since 2012 of the International Academy of Quantum Molecular Sciences. He is the recipient of several awards, including the 2003 "START Grant" of the Austrian Science Fund, the "Hellmann Preis" of the International Working group for Theoretical Chemistry, and the 2016 Kardinal-Innitzer-Preis. 

His main scientific focus lies in the fields of Theoretical Solid State Physics, Surface Sciences and Computational Materials Physics. His work on ab initio density functional theory for solids, liquid and amorphous systems and surfaces has contributed significantly to basic and applied research and has shaped the application of density functional theory worldwide. Kresse is the main author of the computer code "VASP" (Vienna ab initio simulation package) which his research group develops. VASP is the internationally most widely used program for quantum mechanical simulations of condensed matter. The publications on which this code is based received between 30.000 and 50.000 citations each and are amongst the 100 most cited research articles worldwide. 

Until 2019 Kresse directed the Special Research Area "Vienna Computational Materials Laboratory" funded by the Austrian Science Fund. The main goal of this large collaborative project was the precise description of electron interactions in solids and real materials. His current research focus is on the accurate prediction of mechanical, electronic and optical properties in condensed matter beyond simple mean field methods, a research area to which he has already made significant contributions. Georg Kresse is the author of about 400 research articles and has an h-index of over 100.