Hello,
I am trying to reconstruct the non-local part of the pseudopotential to compute some matrix elements. I am using the PSPS.nc output file (generated with iomode = 3 and prtpsps = 1) and I have some trouble with the radial form factors specifically. The way I understand it, Abinit stores the cubic spline data (ffspl) and the grid on which they are stored (qgrid_ff). My issue is that even with a stringent pseudopotential (e.g. the NC SR (ONCVPSP v0.5 PBE stringent psp8 for Carbon from Pseudodojo), the qgrid given by the PSPS.nc output file stops at q=3.1416 … I have to evaluate the radial form factors at some |k+G| vectors that are much larger than that, in fact, around 90% of my |k+G| vectors lie outside of qgrid, for an ecut of 40 Ha. I suspect this making my non-part not very good. Am I missing something or is it normal that Abinit tabulates qgrid until 3.14? If that’s the case, do I have to compute the form factors myself on a custom qgrid from the input .psp8 file? I have attached my input file.
Thank you.
Tony Vuu, on behalf of a colleague
graphene_p_uc_5x5x1.abi (1.9 KB)