Describe the bug
(From AiiDA-QE) When running core-hole SCF calculations in the XpsWorkChain and XspectraCrystalWorkChain with default settings in combination with the Grimme D3 vdW correction, the SCF calculation returns NaN for the total energy. This is because the DFT-D3 module of QE sets D3 parameters based on the ATOMIC SPECIES name, but if the name is not a valid element symbol then the program continues to the SCF procedure rather than crashing due to the unidentified element symbol.
This will happen using the default settings for both WorkChains because the default abs_atom_marker is set to X:
Proposed Solution
Changing the absorbing atom to <element_symbol><number> easily fixes the issue, as numbers are automatically trimmed when the DFT-D3 module reads the species name:
For AiiDA-QE, this is possible through overrides, however there is currently no way to set the abs_atom_marker in AiiDALab-QE and thus the WorkChains will fail unless the absorbing atom marker is changed.
Describe the bug (From AiiDA-QE) When running core-hole SCF calculations in the
XpsWorkChainandXspectraCrystalWorkChainwith default settings in combination with the Grimme D3 vdW correction, the SCF calculation returnsNaNfor the total energy. This is because the DFT-D3 module of QE sets D3 parameters based on theATOMIC SPECIESname, but if the name is not a valid element symbol then the program continues to the SCF procedure rather than crashing due to the unidentified element symbol.This will happen using the default settings for both
WorkChains because the defaultabs_atom_markeris set toX:Proposed Solution Changing the absorbing atom to
<element_symbol><number>easily fixes the issue, as numbers are automatically trimmed when the DFT-D3 module reads the species name:For
AiiDA-QE, this is possible through overrides, however there is currently no way to set theabs_atom_markerinAiiDALab-QEand thus theWorkChains will fail unless the absorbing atom marker is changed.