[Pw_forum] wrong result from converted abinit LDA FHI PP

[Axel Kohlmeyer]

On Thu, Sep 20, 2012 at 11:12 PM, James Mao <xjamesmao at gmail.com> wrote:
> Dear all,
>
> I learned how to generate and test PPs in the past several days, and
> generated NC-LDA PPs for oxygen and carbon using PSLibrary code for
> practice. They were tested using ld1.x first, then a CO2 vibrational
> run to compare with results from known pbe-rrkjus PPs, as well as all
> electrons results. Everything looks fine.

> I also converted fhi PP from Abinit and tested them. The results of
> ld1 test looks fine: e(eigen) of AE-PS and \delta E are small,  wfc
> and logarithmic derivative match AE results. However, the results of
> CO2 vibrational calculations using these converted PPs are completely
> wrong. My question is: what could cause a completely wrong result of
> photon calculation for PPs which look fine in ld1 code test?

the problem is your complete neglect of the differences between
ultra-soft and norm-conserving pseudopotentials in your input.

your input is for ultra-softs, for NC you need a higher wavefunction cutoff.
and usually the "hardest" potential, in your case oxygen determines
the cutoff needed. try using 70 ry and then crank it up in steps to, say,
120ry or even 150ry. also you should not set the density cutoff, as it
should be 4x the wfc cutoff (which is the default). you should observe
some convergence to consistent values this way, and i would bet
they are not as wrong as your current one.

cheers,
     axel.

>
> I don't think I messed up the converting. But if anyone would like to
> help, just let me know and I will send you the converted PPs for your
> testing.
>
> Another thing is that although pbe-rrkjus PPs produced best results,
> they also produced negative frequencies, is it normal? In all cases I
> ran a "relax" calculation first to make sure the geometry was optimized.
>
> I am learning everything in solid states calculations, so any comments
> are welcome and will be greatly appreciated!
>
> James Mao
> ------------
> Postdoc, Department of Chemistry
> University of Pittsburgh
> ------------
>
> Here is the all electron results:
>    [cm-1]
>      641
>     1373
>     2438
>
> results from pbe-rrkjus PPs:
>
> # mode   [cm-1]    [THz]      IR
>     1   -188.02   -5.6368    0.0000
>     2   -188.02   -5.6368    0.0000
>     3      0.00    0.0000    0.0000
>     4      0.00    0.0000    0.0000
>     5      0.00    0.0000    0.0000
>     6    658.64   19.7456    0.4938
>     7    658.64   19.7456    0.4938
>     8   1304.71   39.1142    0.0000
>     9   2340.86   70.1772   12.1795
>
> results from NC LDA PPs I generated using PSLibrary code:
>
> # mode   [cm-1]    [THz]      IR
>     1      0.00    0.0000    0.0000
>     2      0.00    0.0000    0.0000
>     3      0.00    0.0000    0.0000
>     4    218.30    6.5444    0.0000
>     5    218.30    6.5444    0.0000
>     6    577.10   17.3010    0.0521
>     7    577.10   17.3010    0.0521
>     8   1303.15   39.0676    0.0000
>     9   2386.77   71.5536   11.0101
>
> results (WRONG!) from LDA PPs converted from Abinit fhi files:
>
> # mode   [cm-1]    [THz]      IR
>     1   -219.57   -6.5825    0.0000
>     2   -219.57   -6.5825    0.0000
>     3      0.00    0.0000    0.0000
>     4      0.00    0.0000    0.0000
>     5      0.00    0.0000    0.0000
>     6    246.28    7.3833    0.1700
>     7    246.28    7.3833    0.1700
>     8    499.18   14.9652    0.0000
>     9    675.29   20.2448    9.8485
>
> rlx input:
>
>  &control
>     calculation='relax',
>     restart_mode='from_scratch',
>     prefix='co2'
>     pseudo_dir='./pseudos'
>     outdir='./'
>     tprnfor = .true.
>  /
>  &system
>     ibrav = 1, celldm(1) =20.0, nat=  3, ntyp= 2,
>     ecutwfc = 30.0, ecutrho = 300.0
>  /
>  &electrons
>     mixing_beta = 0.7
>     conv_thr =  1.0d-8
>  /
>  &ions
>     ion_dynamics="bfgs",
>  /
> ATOMIC_SPECIES
>  C  12.01 06-C.LDA.fhi.UPF
>  O  16.01 08-O.LDA.fhi.UPF
> ATOMIC_POSITIONS (angstrom)
>  C     0.000000     0.000000     0.000000
>  O     0.000000     0.000000     1.220000
>  O     0.000000     0.000000    -1.220000
> K_POINTS (automatic)
> 1 1 1 0 0 0
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum



-- 
Dr. Axel Kohlmeyer  akohlmey at gmail.com  http://goo.gl/1wk0
International Centre for Theoretical Physics, Trieste. Italy.