[Pw_forum] ld1.x generate Sb PP

[Yongsheng Zhang]

Dear PP experts, 


I spend long long time to generate the Sb PBE pseudopotential. The "best" one I get is still not very successful to get the good experimental or VASP-PAW lattice constants. 

exp    :            4.308420000000000         4.308419999850220         11.27460000000000
PAW-PBE:     4.387963702059904         4.387963701901805         11.42545713239960
PP-PBE:        4.44711025                      4.44858214                     11.81877927
The c-direction is significantly different to the experimental data and PAW results. And I notice in the PP database, there has a PW91 Sb PP. Using the PW91 PP, the theoretical lattice constants are quite good. (4.35479533, 4.35557221,  11.32102582). 


The my Sb PP input file is: 

&input
    title='Sb'
    zed=51.
    rel=1,

    rlderiv=2.2,
    eminld=-2.0,
    emaxld=2.0,
    deld=0.01,
    nld=3,

    config='[Kr] 4d10.0 5s2.0 5p3.0'
    iswitch=3,
    dft='PBE'
 /
 &inputp
   pseudotype=3,
   lloc=1,
   rho0=0.001,
   file_pseudopw='Sb.pbe.UPF',
   nlcc = .true.,
   author='YS-Zhang',
 /
4
4D  3  2 10.00  0.00  1.60  2.00  1
4D  3  2  0.00 -1.70  1.60  2.00  1
5S  1  0  2.00  0.00  2.10  2.10  1
5P  2  1  3.00  0.00  2.30  2.30  1

During the tests, I didn't find any ghost state and the largest difference between PP and AE is 

     3 2     4D   1(10.00)       -6.59082       -6.55589       -0.03493
     1 0     5S   1( 0.00)       -4.49315       -4.49962        0.00647
     2 1     5P   1( 0.00)       -3.63377       -3.64128        0.00751
     dEtot_ae =      10.908350 Ry
     dEtot_ps =      10.913263 Ry,   Delta E=      -0.004913 Ry

Please help me improve the Sb PP input file. Thank you very much. 


Yongsheng Zhang

Northwestern University
Dept. of Materials Science and Eng.
2220 Campus Dr.
Evanston, IL 60208