[QE-users] Wrong prediction for Pd-Pd bond length

[pboulet]

As a supplement to Giuseppe’s and Lorenzo's comments:

1- forc_conv_thr should be smaller: about 1e-4 or 1e-5 (why using 1.94469e-03 with that many digits?)
2- add: etot_conv_thr= 1e-7
2- I would use ibrav=0 together with CELL-PARAMETERS (see online manual)
3- occupation=fixed, no smearing, and nbnd=number of occupied states+ a few virtual ones. It should work as you have a dimer.
4- why using vdW corrections? not sure they are significant. Furthermore, as a remember Grimme’s corrections are not that good for metals. I got too large bond distances for gold (but it was for a solid)… 
5- mixing_beta=0.2 should do the job and you eventually avoid oscillations in the SCF convergence if you use a too large value (or even no convergence at all).
6- &CELL … / section is not necessary as you only optimise atomic positions.
7- K_POINTS {Gamma} is all right as you have a molecule (k-points are only meaningful for slabs and bulk structures)

HTH,
Best


Pascal Boulet
—
Professor in computational chemistry - DEPARTMENT OF CHEMISTRY
Director of the Madirel laboratory
Aix-Marseille University - Avenue Escadrille Normandie Niemen - F-13013 Marseille - FRANCE
Tél: +33(0)4 13 55 18 19 - Fax : +33(0)4 13 55 18 50
Email : pascal.boulet at univ-amu.fr <mailto:pascal.boulet at univ-amu.fr>




> Le 28 janv. 2019 à 16:44, Paolo Costa <paolo.costa85 at gmail.com> a écrit :
> 
> Dear QE users,
> 
> I am trying to optimize (relax) Pd dimer (Pd2) as cluster placed on cubic lattice.
> However I got a value of Pd-Pd bond length which is far too long (3.36 A) compare to the experimental one (about 2.6 A).
> Am I doing something wrong? Here below the input file:
> 
> &CONTROL
>      calculation   = "relax"
> prefix ='Pd2'
>     forc_conv_thr =  1.94469e-03
>     max_seconds   =  1.34369e+14
>     nstep         = 800
>     pseudo_dir    = "/home/pcosta/pseudo"
> 	verbosity     ='high'
> disk_io ='none'
>     tprnfor       = .TRUE.
>     tstress       = .TRUE.
> /
> 
> &SYSTEM
>     a     =  4.50000e+01
>     degauss                   =  1.00000e-02
>     ecutrho                   =  1.83747e+01
>     ecutwfc                   =  1.46997e+00
> 	lda_plus_u                = .FALSE.
>     ibrav                     = 1
>     nat                       = 2
>     nspin                     = 2
>     ntyp                      = 1
>     occupations               = "smearing"
>     smearing                  = "gaussian"
>     starting_magnetization(1) =  2.00000e-01
>     assume_isolated = 'makov-payne'
> 	vdw_corr                  = 'grimme-d3'
> /
> 
> &ELECTRONS
>     conv_thr         =  1.00000e-08
>     diagonalization  = "david"
>     electron_maxstep = 800
>     mixing_beta      =  4.00000e-01
>     startingpot      = "atomic"
>     startingwfc      = "atomic+random"
> /
> 
> &IONS
>     ion_dynamics = "bfgs"
> 	trust_radius_min=1e-5
> /
> 
> &CELL
>     cell_dofree    = "all"
>     cell_dynamics  = "bfgs"
>     press_conv_thr =  5.00000e-01
> /
> 
> K_POINTS {automatic}
>  3  3  1  0 0 0
>   
> ATOMIC_SPECIES
> Pd    106.42000  Pd.pbe-n-kjpaw_psl.1.0.0.UPF
> 
> ATOMIC_POSITIONS {angstrom}
> Pd     16.194860  19.259990  18.659720
> Pd     18.251640  18.665930  18.594650
> 
> Thanks.
> -- 
> Paolo Costa, Ph.D.
> Postdoctoral Researcher
> Department of Chemistry and Biomolecular Sciences
> University of Ottawa
> 10 Marie Curie, Ottawa, ON K1N 6N5, Canada
> Room number: DRO 326 (D'Iorio Hall)
> _______________________________________________
> users mailing list
> users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190128/fc93eefe/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 1585 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20190128/fc93eefe/attachment.p7s>