GAPW - only soft AO for one kind

20 views
Skip to first unread message

Arthur France-Lanord

unread,
Nov 5, 2025, 3:34:13 AM (2 days ago) Nov 5
to cp2k
Hi all, 

I'm using GAPW with pseudos for a system with antimony as I need a high precision on the atomic forces - everything works as intended, but with the default parameters, Sb has no hard part: 

&KIND Sb
  BASIS_SET ORB TZVP-MOLOPT-PBE-GTH-q5
  POTENTIAL GTH-GGA-q5
&END KIND

From the output: 

 QS| GAPW|      At least one kind is NOT PAW, i.e. it has only soft AO
 QS| GAPW|      The NOT PAW atoms are treated fully GPW

This might not be a problem (since I get fairly well-converged forces), but the quantities I usually check and which should converge to zero are very much non-zero here: 

  Total Rho_soft + Rho1_hard - Rho1_soft    -1440.0000494132
  Total charge density (r-space):           -1290.0000494132
  Total Rho_soft + Rho0_soft (g-space):        -0.0000531123

I know it's due to the hard/soft separation, since when I use a smaller pseudo for Sb (such as q23), there is a hard part, and these quantities converge to zero, but I wanted to ask you if this is intended or important. 

Otherwise, would it make sense to tune HARD_EXP_RADIUS in KIND? 

Also, is there a reason why there is a q15 pseudo available for Sb (in POTENTIAL_UZH), but no associated basis set? 

thanks, 
Arthur

Jürg Hutter

unread,
Nov 5, 2025, 4:10:13 AM (2 days ago) Nov 5
to cp...@googlegroups.com
Hi

You can force the PAW treatment of all atoms with
FORCE_EVAL%DFT%QS%FORCE_PAW

Changing HARD_EXP_RADIUS in the KIND section, as well as EPS_FIT
in the QS section can help to increase accuracy of the total charge integration.

I would also use
FORCE_EVAL%DFT%QS%GAPW_1C_BASIS EXT_SMALL
to improve the PAW projection.

For elements beyond the third row it is a good idea to also increase the
atomic grid
LEBEDEV_GRID and RADIAL_GRID in the KIND section.

best regards
JH

PS Not all basis sets have been generated for the UZH PP. There was no request for
the missing ones up to now.

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Arthur France-Lanord <arthur....@gmail.com>
Sent: Wednesday, November 5, 2025 9:34 AM
To: cp2k
Subject: [CP2K:21946] GAPW - only soft AO for one kind
--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k+uns...@googlegroups.com>.
To view this discussion visit https://groups.google.com/d/msgid/cp2k/f819c101-df6e-4316-a967-96df20b6b9a9n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/f819c101-df6e-4316-a967-96df20b6b9a9n%40googlegroups.com?utm_medium=email&utm_source=footer>.
Reply all
Reply to author
Forward
0 new messages