Dear Dan, well, this was long ago (2008) when I was testing the Siesta Basis sets. But I will try to dig out some more details:
At the time I was using siesta-2.0. If my recollection is correct the defaults regarding the basis set generation of hydrogens changed with siesta-3.0, which improved the situation. Some data using PBE and a basis described by PAO.BasisSize DZP PAO.EnergyShift 100 meV PAO.SplitNorm 0.13 I found a dipol moment of 1.4938 debye and Mulliken Populations of 1.074 on H, 5.852 on O (negative H, positive O) After playing with many parameters to minimize the DFT total energy (without increasing the cutoff radii too much) I ended up using a basis defined by: %block PAO.Basis H 1 split 0.00 n=1 0 2 P 1 E 10. 5.088 6.088 2.519 1.000 1.000 O 2 split 0.00 n=2 0 2 E 10. 4.220 5.220 2.774 1.000 1.000 n=2 1 2 P 1 E 10. 5.298 6.298 2.630 1.000 1.000 %endblock PAO.Basis dipole moment 1.720 debye Mulliken Population 0.788 on H 6.423 on O Thus, the hydrogen is positive and the oxygen negative. The dipol moment is reasonably close to a benchmark calculation at PBE/aug-cc-pV5Z which gives 1.8028 debye. I hope this illustrates my comment and does not confuse even more. Regards, Ulrich. From: Dan Gil <[email protected]> To: <[email protected]> Sent: 11.10.2017 17:26 Subject: Re: [SIESTA-L] Mulliken Charges on Water Thank you all for your responses. Dr. Camps - Thank you for the paper. The article further makes me more worried about the results I obtained from SIESTA - I had a positive partial charge on Oxygen from the Mulliken population. None of the basis set/functional/population method combinations in the paper have this result. Dr. Biedermann - I ran more calculations with the split-valence basis set and I've played around with the orbital-confining cutoff radii and the hydrogen split radius. I've noticed the dependence of the Mulliken population on these parameters but I still don't get the correct dipole moment. Dan On Mon, Oct 9, 2017 at 5:08 PM, I. Camps <[email protected]> wrote: Take a look at the following paper. They discus the effect of various methods in the determinations of water atomic charges: http://onlinelibrary.wiley.com/doi/10.1002/jcc.20157/full F. Martin, H. Zipse, Charge distribution in the water molecule. A comparison of methods, J. Comput. Chem. 26 (2005) 97-105. DOI: 10.1002/jcc.20157 []'s, Camps On Mon, Oct 9, 2017 at 4:44 AM, Ulrich Biedermann <[email protected]> wrote: Dear Dan, I have noticed similar problems. Mulliken charges depend on the atomic basis set used. In Siesta the cutoff radius (of H) seems to be the critical parameter. I suggest that you check the dipol moment of an isolated water molecule (10 Angstrom box). It should have the correct magnitude and sign for the basis set settings you use. Regards, Ulrich. From: Dan Gil <[email protected]> To: <[email protected]> Sent: 06.10.2017 16:13 Subject: [SIESTA-L] Mulliken Charges on Water Hi, I am using Siesta 3.2 and ran into a strange problem. I simulate a water molecule and the Mulliken charges are wrong. It shows that the oxygen atom is positively charged and the two hydrogens are negatively charged. The Hirshfeld and Voronoi populations are fine. Did I run into a bug? input.fdf NumberOfAtoms 3 NumberOfSpecies 4 %block ChemicalSpeciesLabel 1 13 Al 2 14 Si 3 8 O 4 1 H %endblock ChemicalSpeciesLabel PAO.BasisType nodes %block PAO.BasisSizes Al DZP Si DZP O DZP H DZP %endblock PAO.BasisSizes XC.functional GGA XC.authors PBE # SCF options MaxSCFIterations 200 DM.MixingWeight 0.25 # New DM amount for next SCF cycle DM.Tolerance 1.d-6 # Tolerance in maximum difference SolutionMethod diagon # OrderN or Diagon MeshCutoff 400 Ry DM.Require.Energy.Convergence .true. DM.Energy.Tolerance 1.d-6 eV MD.TypeOfRun CG MD.NumCGsteps 100 MD.InitialTemperature 300.0 K MD.TargetTemperature 300.0 K WriteHirshfeldPop .true. WriteVoronoiPop .true. WriteMullikenPop 1 LatticeConstant 1 Ang %block LatticeVectors 9.5160000000 0.0000000000 0.0000000000 -4.7580000000 8.2411000000 0.0000000000 0.0000000000 0.0000000000 48.2792733630 %endblock LatticeVectors WarningMinimumAtiomicDistance 1.0 Ang AtomicCoordinatesFormat NotScaledCartesianAng %block AtomicCoordinatesAndAtomicSpecies < Coord.fdf Coord.fdf 5.265981 -1.943725 -6.378528 3 4.222545 -1.943725 -6.378528 4 5.612381 -0.997652 -6.378528 4 Output (excerpts) mulliken: Mulliken Atomic and Orbital Populations: Species: O Atom Qatom Qorb 2s 2s 2py 2pz 2px 2py 2pz 2px 2Pdxy 2Pdyz 2Pdz2 2Pdxz 2Pdx2-y2 1 5.764 1.436 -0.027 1.363 1.793 1.235 -0.023 -0.003 -0.034 0.003 0.002 0.005 0.001 0.014 Species: H Atom Qatom Qorb 1s 1s 1Ppy 1Ppz 1Ppx 2 1.118 0.845 0.036 0.063 0.104 0.070 3 1.118 0.845 0.036 0.091 0.104 0.042 Hirshfeld Net Atomic Populations: Atom # Qatom Species 1 -0.234 O 2 0.117 H 3 0.117 H Voronoi Net Atomic Populations: Atom # Qatom Species 1 -0.171 O 2 0.086 H 3 0.085 H Best Regards, Dan ---------------- ------------------------------------------------- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf Handelsregister B 2533 Amtsgericht Düsseldorf Geschäftsführung Prof. Dr. Gerhard Dehm Prof. Dr. Jörg Neugebauer Prof. Dr. Dierk Raabe Dr. Kai de Weldige Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 Please consider that invitations and e-mails of our institute are only valid if they end with …@mpie.de. If you are not sure of the validity please contact [email protected] Bitte beachten Sie, dass Einladungen zu Veranstaltungen und E-Mails aus unserem Haus nur mit der Endung …@mpie.de gültig sind. In Zweifelsfällen wenden Sie sich bitte an [email protected] ------------------------------------------------- ------------------------------------------------- Max-Planck-Institut für Eisenforschung GmbH Max-Planck-Straße 1 D-40237 Düsseldorf Handelsregister B 2533 Amtsgericht Düsseldorf Geschäftsführung Prof. Dr. Gerhard Dehm Prof. Dr. Jörg Neugebauer Prof. Dr. Dierk Raabe Dr. Kai de Weldige Ust.-Id.-Nr.: DE 11 93 58 514 Steuernummer: 105 5891 1000 Please consider that invitations and e-mails of our institute are only valid if they end with …@mpie.de. If you are not sure of the validity please contact [email protected] Bitte beachten Sie, dass Einladungen zu Veranstaltungen und E-Mails aus unserem Haus nur mit der Endung …@mpie.de gültig sind. In Zweifelsfällen wenden Sie sich bitte an [email protected] -------------------------------------------------
