Are results trustworthy?

Skip to first unread message

Sep 20, 2022, 8:31:31 AMSep 20
to cp2k

I have been trying to optimize a bulk CuO using PBE and PBEsol with D3.  I've tried GPW and GAPW with various combinations of basis sets, kpoints and broken symmetry and have not been able to match the optimized geometry from VASP.

The goal is to study adsorption onto the CuO surface.  I'm wondering, if I constrain the CuO slab in the experimentally determined coordinates and cell length/angles, if the adsorption results would be trustworthy?

Overall, I must admit, that I'm somewhat surprised and disappointed in CP2K's lack of ability to match both experimentally and calculated results.  So again, my real question is, can the results be trusted, given that I have been unable to recreate the first step of the process?

Best regards,

Matt Watkins

Sep 23, 2022, 10:23:56 AMSep 23
to cp2k
I ran your system with a few tweaks (went back to pseudos and molopt basis) and it seemed reasonably well behaved.
I only performed a fixed shape cell_opt not full relaxation. A cell_ref would be better but not sure how it plays with multiple_unit_cell flag.
Cell parameters came out as:
ut.out: CELL| Vector a [angstrom]:      14.111     0.000     0.000   |a| =    14.110920
out.out: CELL| Vector b [angstrom]:       0.000     9.973     0.000   |b| =     9.972731
out.out: CELL| Vector c [angstrom]:      -2.516    -0.000    15.068   |c| =    15.276187
which seem reasonable. No idea if your +U value is reasonable or not.
Reply all
Reply to author
0 new messages