DFT modelling of the temperature dependent elastic constants and planar defects in zirconium hydrides

Dnr:

SNIC 2016/1-419

Type:

SNAC Medium

Principal Investigator:

Pär Olsson

Affiliation:

Malmö högskola

Start Date:

2016-11-01

End Date:

2017-11-01

Primary Classification:

10304: Den kondenserade materiens fysik

Webpage:

Allocation

Abstract

Because of their low thermal neutron capture cross section and good corrosion resistance, zirconium-based alloys are commonly used as fuel cladding material in the core of nuclear power reactors. While in service, the fuel cladding is in contact with water, which promotes the oxidation of Zr. This process releases free hydrogen, of which a portion enters the alloy and gives rise to the formation of hydrides once the solid solubility limit has been exceeded. This can have a detrimental effect on the integrity and longevity of the material, as it can lead to phenomena such as embrittlement and delayed hydride cracking. Thus, to make reasonable lifetime assessments and ensure strutural integrity of fuel claddings subjected to operating conditions, it is necessary to estimate the mechanical properties of the material whilst subjected to such conditions. For these reasons it is important to investigate (i) the temperature dependent elastic constants (TDEC) and isotropic averages, (ii) surface energy of densely packed and low index surfaces and (iii) generalized stacking fault energy curves of the two most common zirconium hydrides: the delta- and epsilon-hydrides. In this project, these properties will be modelled using quantum mechanical density functional theory (DFT), where in particular we will use VASP and the open source Quantum-Espresso (QE) package. Our intent is to perform phonon calculations at triolith and self-consistent-field calculations at the Abisko system. To supplement the modelling, the project collaborators at Studsvik Nuclear Corporation will perform nanoindentation experiments on zirconium hydrides. At this stage of the project we are in a very computationally intensive phase. In particular workpackage (i) requires substantial computational resources because of the phonon calculations that need to be performed. This means that I estimate the need for about 50 000 core-hours/month at triolith. The Abisko system is somewhat slower than the triolith system which is why we will use it for relaxation and self-consistent-field calculations for decohesion and generalized stacking fault energy calculations. Those considered systems are quite large containing up to about 100 atoms.