Electronic Structure Community

Introduction

The research within the Electronic Structure community aims at understanding and predicting the properties of materials and molecules, by computing the quantum-mechanical state of the many-electron system. The computations are ideally based on first principles (often called "ab initio"), i.e., without any parameters from experiments entering the models. This feature is fundamentally important since it puts electronic structure calculations apart as an independent research method with respect to experiment and analytical theory.

            The researchers in this field come from both chemistry (quantum chemistry) and physics (materials science/condensed matter), and historically, their approaches to the many-electron problem have differed. For a long time, Hartree-Fock-Roothan theory was the method of choice for describing molecules, whereas Density Functional Theory (DFT) in the local density approximation (LDA) did a reasonable job at describing metals. Further, methods with a more sophisticated description of electron-electron correlations (so-called explicitly correlating methods), such as multiconfiguration and coupled cluster methods, have led to highly predictive, hierarchical, tools for molecular properties. These latter methods, however, are still too computationally demanding to be used on large systems.  With the invention of hybrid functionals such as B3LYP, generalized gradient corrected approximations to the exchange-correlation functional, and several other developments in the field, DFT has become vastly better at describing a wide range of systems and is now widely applied by both chemists and physicists. Today, the quantum chemistry and condensed matter research fields are in many aspects very similar, but for historical reasons, the methods and research emphasis still differ and the research groups still live partly parallel lives. One goal of creating the SeRC Electronic Structure community is to help merge these two research traditions, both in terms of interaction and in terms of method development.

            Although the development has been very rapid the last 20 years, there are still a number of outstanding challenges that contemporary electronic structure is facing, in particular in

consistency (DFT), scalability (explicitly correlating methods) and applicability (both). A strategic research initiative within SeRC would be to join forces with other communities like Numerical Analysis, Molecular Simulation, FLOW and Visualization, to solve some of these bottlenecks. In particular the collaboration with Molecular Simulation may lead to substantial advances for the applicability problem in terms of multiscale approaches that include long time and length scales, temperature and pressure. This can lead to seamless calculations moving from atoms, molecules, grains, grids, and bulk, with outstanding applications in materials science, chemistry, biology, and nanotechnology.

 

Research Environment

The SeRC Electronic Structure community consists of four distinct research groups. Two are located at KTH (Hans Ågren at BIO and Anna Delin at ICT) and two at LiU (Igor Abrikosov at  Theoretical Physics and Patrick Norman at Computational Physics). In addition, we have close collaboration with electronic structure groups in Uppsala (Olle Eriksson and Kersti Hermansson).

            At ICT KTH the tenured staff is Anna Delin (Professor) and Lars Bergqvist (Associate Professor, SeRC Faculty). Relevant international collaborations include participation in the EU-project NEXTEC on thermoelectric properties and method development together with researchers in Los Alamos, Uppsala, Linköping and Ireland. KTH together with Uppsala are the main sites for the development of the software UppASD for atomistic spin dynamics simulations which has users world-wide.

            At BIO, KTH the research group is led by Hans Ågren (Professor) with Zilvinas Rinkevicius (Associate Professor, SeRC Faculty) as additional SeRC tenured staff. The group has strong international collaboration implemented in a number of EU projects, bilateral programs and research centers. In particular, the collaboration with China is strong.

            At LiU Theoretical Physics , the research group is led by Igor Abrikosov (Professor) with Marcus Ekholm (Assistant Professor, SeRC Faculty) as additional SeRC tenured staff.

            At LiU Computatonal Physics, the research group is led by Patrick Norman (Professor and also Director of NSC) with Mathieu Linares (Assistant Professor, SeRC Faculty) as additional SeRC tenured staff.

Relevant collaborations (for both the groups at LiU) include the Strategic Centre in Materials Science for Nanoscale Surface MSE2, Linköping Linnaeus Initiative for Novel Functional Materials, SSF Strategic Research Center and Multifilms, Uppsala (Olle Eriksson), KTH (Börje Johansson), Antoine Georges (Professor), Ecole Polytechnique, Paris, France; M. I. Katsnelson (Professor, University of Nijmegen, The Netherlands), L. Dubrovinsky (Professor,Universität Bayreuth, Germany) and Yu. Kh Vekilov (Professor, Moscow Institute of Steel and Alloys, Russia).

 

Overall Research Goals

The long-term research goal is to contribute to solving important outstanding problems in the area of electronic structure calculations such as developing and structuring software for electronic structure calculations that can be used jointly by our communities. In particular, the prioritized aims are to:

  • Developing all-purpose multiscale software with a DFT/TDDFT kernel.
  • Developing interfaces of electronic structure softwares to visualization and haptics.
  • Increase the efficiency and exploring novel algorithms of existing computational methods dealing with charge and spin transport in materials, strong correlations and spin dynamics in order to be able to address orders of magnitude larger and more complex systems.
  • Bridge different length and time scales to be able to perform true multiscale modeling without loss in accuracy.