1 sujet IRFU/DPhN

Dernière mise à jour : 21-01-2021


««

• Nuclear Physics

 

Pushing ab initio calculations of atomic nuclei to higher precision

SL-DRF-21-0293

Research field : Nuclear Physics
Location :

Service de Physique Nucléaire (DPhN)

Laboratoire études du noyau atomique (LENA) (LENA)

Saclay

Contact :

Thomas DUGUET

Vittorio SOMA

Starting date : 01-10-2021

Contact :

Thomas DUGUET
CEA - DRF/IRFU/DPhN/LENA

0169082338

Thesis supervisor :

Vittorio SOMA
CEA - DRF/IRFU/DPhN/LENA

0169083236

Laboratory link : http://irfu.cea.fr/Phocea/Vie_des_labos/Ast/ast_visu.php?id_ast=4070

Pushing ab initio calculations of atomic nuclei to higher precision



The theoretical description of atomic nuclei from first principles, or in a so-called ab initio fashion, has become possible only recently thanks to crucial advances in many-body theory and the availability of increasingly powerful high-performance computers. Such ab initio techniques are being successfully applied to study the structure of nuclei starting from the lighter isotopes. Still, extensions to heavy elements and nuclear reactions are posing considerable difficulties. The objective of the thesis is to contribute to this on-going progress in nuclear many-body theory. The project will focus on a developing ab initio technique (the so-called Gorkov-Green function approach, devised at CEA Saclay) designed to describe open-shell or superfluid systems (the majority of atomic nuclei). After the first promising applications to light and medium-mass nuclei, the method faces crucial upgrades to reach the precision and competitiveness of state-of-the-art approaches. The proposed work will aim to put in place the necessary tools towards this direction.



In particular, the Gorkov-Green function approach will be extended to the next level of precision. After some formal work, this will require a careful numerical implementation on top of the existing code. Given the increased cost of the corresponding numerical calculations, expected to go from moderately (100 proc.) to massively parallel (1000 proc.), special attention will have to be paid to the code optimisation and the use of pre-processing techniques like importance truncation or tensor factorisation.



Overall, the thesis work will exploit the latest advances in nuclear theory, including the use of nuclear interactions from chiral effective field theory and renormalisation group techniques, as well as high-performance computing codes and resources. The work will consist in formal developments, computational tasks and application of the new technology to cases of experimental interest. International collaborations are envisaged.

 

Retour en haut