【大学频道】中国科学院大学卡弗里理论科学研究所呈献 | Patrick Rinke

更多精彩视频登陆网站

www.koushare.com

图 | Patrick Rinke教授

题   目：GW and beyond: core excitations, strong correlation and artificial intelligence

报告人：Patrick Rinke

单   位：Aalto University

时   间：2019-05-30

地   点：中国科学院大学

扫码观看精彩报告视频

报告摘要

The GW approximation to many-body perturbation theory (MBPT) has been incredibly successful in theoretical spectroscopy. Yet, challenges abound and new frontiers await. In this presentation, I will introduce our first exploration of perturbative GW (G0W0) for core excitations. We have implemented the real-frequency contour deformation technique into theall-electron numerical atomic orbital code FHI-aims. The real-frequency treatment proves to be necessary to treat core excitations accurately. The computed core level binding energies deviate by generally less than 0.5 eV from experiment (on an absolute scale) outperforming the popular density-functional theory based Delta Self-consistent Field (△SCF) method. To tackle strongly correlated systems, we have developed a new quantum embedding theory. It captures strong (static) correlation in a subspace by configuration interaction (CI) theory and high-energy dynamic correlation with MBPT in the GW and Bethe-salpeter equation (BSE) approximation. For the challenging multi-reference problems of H2 and N2 dissociation, we obtain good agreement with benchmark results. Our theory treats ground and excited states on equal footing, and we compute vertical excitation energies of N2 and free-base porphyrin in excellent agreement with high level quantum chemistry methods. Despite the successes of theoretical spectroscopy, calculations and experiments are costly and time consuming. For this reason, we recently enhanced theoretical spectroscopy with artificial intelligence (AI). Once trained, the Al can make predictions of spectra instantly and at no further cost. I will present artificial neural networks that can learn excitation spectra of molecules with an accuracy of 97%.

个人简介

In the Department of Applied Physics at Aalto, I lead the Computational Electronic Structure Theory (CEST) group. My group and I are developing advanced quantum mechanical methods and apply them to pertinent problems in material science, surface science, physics, chemistry and the nano sciences. Quantum mechanics gives us access to the behavior of atoms and electrons and allows us to gain an atomistic view of matter. Examples are the atomic structure of surfaces or hybrid organic-inorganic interfaces. Perturbing the atoms and electrons, as done in spectroscopies, reveals more information about matter. We use theoretical spectroscopy to probe the properties of molecules, molecules on surfaces, nanostructures, semiconductors including d- and f-electron oxides and materials for photovoltaic applications.

—— ——往期精彩回顾—— ——

END

扫描二维码，关注微信公众号

戳这里，观看精彩视频哟！