[Pw_forum] discrepancy in molecular binding energy between pwscf and localized basis set program

[Axel Kohlmeyer]

On Mon, 11 Jun 2007, Garold Murdachaew wrote:

GM> Dear pwscf users,
GM> 
GM> I am doing some calculations for relatively weakly interacting species and 
GM> decided to compare to what I obtained with localized wavefunction based 
GM> programs such as molpro.  For example, I did calculations of the binding 
GM> energy of Xenon dimer (at the minimum separation of 4.42 Angstroms).
GM> I obtained (in meV):
GM> 
GM>  	molpro	pwscf3.2
GM> LDA	-34	-34
GM> PBE	 -4	-11

garold,

one stupid question: how is your basis set convergence 
in molpro and more importantly, did you correct for BSSE?

i found that BSSE can be very large for weakly binding
systems, even with large basis sets.

axel.

GM> 
GM> As can be seen, the LDA calculations agree perfectly but the PBE pwscf 
GM> calculation gives an interaction energy almost three times more negative. 
GM> (I realize that for most pwscf users who are usually only concerned about 
GM> energies on the eV rather than on the meV scale these differences may seem 
GM> insignificant.)
GM> 
GM> Here is my PBE pwscf input file for the Xe dimer (the monomer and the LDA 
GM> inputs are similar):
GM> 
GM> &CONTROL
GM>    calculation   = 'scf',
GM>    prefix        = '$HEAD',
GM>    title         = '$HEAD',
GM>    restart_mode  = 'from_scratch',
GM>    outdir        = '$TMPDIR/',
GM>    PSEUDO_DIR    = '$PSEUDODIR/',
GM> /
GM> &SYSTEM
GM>      ibrav     = 0,
GM>      celldm(1) = 25.d0,
GM>      nat       = 2,
GM>      ntyp      = 1,
GM>      ecutwfc   = 30.0d0,
GM>      ecutrho   = 300.0d0,
GM>      occupations = 'smearing',
GM>      degauss     = 0.001D0,
GM>      smearing    = 'mp',
GM>      nbnd        = 12,
GM> /
GM> &ELECTRONS
GM>      conv_thr    = 1.0d-8,
GM>      mixing_mode = 'plain',
GM>      mixing_beta = 0.3d0,
GM> /
GM> ATOMIC_SPECIES
GM> Xe 131.30   Xet2tPBE.RRKJ3
GM> ATOMIC_POSITIONS { angstrom }
GM> Xe     .000000000    .000000000    .000000000  0  0  0
GM> Xe     .000000000    .000000000   4.420000000  0  0  0
GM> K_POINTS { gamma }
GM> CELL_PARAMETERS { cubic }
GM>     1.00000    0.00000    0.00000
GM>     0.00000    1.00000    0.00000
GM>     0.00000    0.00000    1.00000
GM> 
GM> As can be seen, the molecule is in a cubic box of edge length 25 bohr and 
GM> thus there is sufficient vacuum to minimize the spurious interactions with 
GM> its periodic images.  I played with the convergence setting conv_thr but 
GM> setting it to a factor of 100 tighter had very small effect.  Changing the 
GM> degauss setting had no effect on energy (only Fermi level was changed.) 
GM> However, when I increased  ecutwfc and ecutrho to 70 and 700 respectively, 
GM> I obtained a slight change (-10 meV rather than -11 meV).  Can somebody 
GM> please tell me how I can reproduce my PBE  localized basis sets 
GM> calculation.  Do I need to make some changes to the pseudos?  Or is this 
GM> what I can expect from gradient-corrected calculations (such as PBE) due 
GM> to some sort of errors in estimating the gradients (which are not needed 
GM> in the LDA calculation)?
GM> 
GM> Thank you for you assistance.
GM> 
GM> Best regards,
GM> Garold Murdachaew
GM> 
GM> ps:  Here are the details of the pseudo I used in pwscf:
GM> 
GM> LDA pseudo: XeLDA.RRKJ3.UPF: nonrel, nc, nlcc=t, valence=8, maxl=2
GM> PBE pseudo: Xet2tPBE.RRKJ3 : nonrel, nc, nlcc=t, valence=8, maxl=2
GM> 
GM> --
GM> Garold Murdachaew
GM> Postdoctoral Fellow - Condensed Matter Sector
GM> SISSA-ISAS (International School for Advanced Studies)
GM> New Building Room 208
GM> Via Beirut 2-4, 34014 Trieste, Italy
GM> 
GM> tel   : +39 040 3787 477
GM> fax   : +39 040 3787 528
GM> email : murda at sissa.it
GM> 
GM> 
GM> _______________________________________________
GM> Pw_forum mailing list
GM> Pw_forum at pwscf.org
GM> http://www.democritos.it/mailman/listinfo/pw_forum
GM> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.