Title:	Rational and data-driven computational research on modern molecular materials
DNr:	SNIC 2022/3-34
Project Type:	SNIC Large Compute
Principal Investigator:	Hans Ågren <hagren@kth.se>
Affiliation:	Uppsala universitet
Duration:	2023-01-01 – 2024-01-01
Classification:	10407 10603 30105
Homepage:	https://www.katalog.uu.se/profile/?id=N18-41
Keywords:

Abstract

The research in this project involves, as before, method development, programming and applications, where experimental - theoretical collaboration almost always takes place. The main focus is on rational approaches, thus implementing the laws of quantum and classical physics on a computer in order to derive new molecular materials. This has recently been complemented with the use of in-house data-driven approaches in the form of machine learning, something that will be expanded with deeper AI in the present project. Still, and emphasized here, our experience and original expertise is in the rational modelling, something that is also reflected by our application of computer resources. Our “rational” computing activities are divided into three main areas: 1) Molecular Properties, Spectroscopy, Structures and Reactivity; 2) Multiscale modelling and method development; 3) Nano- and Bio-Photonics and Electronics. In collaboration with experimentalists we build up, step by step, know-how and understanding which makes us better suited to tackle complicated systems and processes in biology, chemistry and in the life and materials sciences. We aim to interpret modelling results in terms of chemical structure, properties and dynamics, where we deal with real problems by using models that join the accuracy of quantum mechanics and the applicability of classical physics. Our portfolio of modelling tools to do this is wide, our computational resources have increased immensely, and so our knowledge, leading to a situation in our research fields which is more promising and inspiring than ever before. Our groups are also involved in computer science, making our codes perform optimally on modern computer architectures as provided by SNIC and with wide ramifications with respect to e-Science projects, both nationally and internationally. We have generated a considerable breakthrough in multiscale modelling and presented new models within photophysics, nanoparticle technology, low-dimensional materials and biomedical markers for diseases. New and unprecedented results have been introduced in these areas with considerable academic as well as societal value. Our groups have earlier quite extensively used Kebnekaise and Tetralith. This year we would like to request increased time allocations as our modelling is very resource demanding.