[Pw_forum] the problem of scf convergence

Eyvaz Isaev eyvaz_isaev at yahoo.com
Sat Sep 15 14:11:41 CEST 2007


Hi,

Axel just suggested using a "non-ak", i.e. not
generated by Axel Kohlmeyer, pseudopotential. 

It was quite interesting to test my own
pseudopotential for Li which gave a=4.64Ang (Exp.
4.60Ang), and B=83GPa (Exp. 81GPa) for Li20. Phonons
at Gamma point look good (without splitting of the
T_1u mode):  T_1u - 431cm^-1, and T_2g - 524cm^-1. 

Bests,
Eyvaz.

--- hyello25 Qgrui <hyello25 at gmail.com> wrote:

> hi,
>   I performed other USPPs tests (blyp-van_ak,
> pw91-van_ak, pz-van_ak) and
> found that all these calculations with USPPs were
> not convergence both in
> press and Etot.For example, with blyp-van_ak( result
> is at the bottom of the
> mail).All these USPPs are downloaded from the
> homepage of
> PWSCF/pseudopotential.
>   However, this unconvergence situation disappeared
> with NCPP (PBE-TM-GGA,
> converted from fhi format PP using fhi2upf code).
> The result is at the
> bottom as well.
>   As Axel Kohlmeyer mentioned:
>   "please try a different Li pseudopotential (i.e. a
> non-ak one).
> the one you are using seems to have some not yet
> understood
> problems for higher-order properties: lattice
> constant from
> EOS fit, MD and geometry optimizations were fine,
> phonons
> were bad."
>   And can anyone tell me what does the 'ak'(in
> blyp-van_ak) mean? What kind
> of systems are these 'ak' USPPs suitable for
> calculation?
> 
> WITH blyp-van_ak
>
USPP:------------------------------------------------------
>     ---!   convergence scf caculation for Li2O  
> !---
>  ecutwfc = 100 Ry,  ecutrho = 320 Ry :
> !    total energy              =   -61.82021122 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 19.03
>  ecutwfc = 100 Ry,  ecutrho = 360 Ry :
> !    total energy              =   -61.82317722 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=   -
> 6.75
>  ecutwfc = 100 Ry,  ecutrho = 400 Ry :
> !    total energy              =   -61.82694596 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 4.67
>  ecutwfc = 100 Ry,  ecutrho = 600 Ry :
> !    total energy              =   -61.82089409 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 4.29
>  ecutwfc = 100 Ry,  ecutrho = 800 Ry :
> !    total energy              =   -61.81983678 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 19.40
>  ecutwfc = 100 Ry,  ecutrho = 1000 Ry :
> !    total energy              =   -61.82031614 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 22.33
>  ecutwfc = 200 Ry,  ecutrho = 800 Ry :
> !    total energy              =   -61.82432402 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 9.13
>  ecutwfc = 200 Ry,  ecutrho = 1000 Ry :
> !    total energy              =   -61.82480674 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 12.09
>  ecutwfc = 200 Ry,  ecutrho = 1600 Ry :
> !    total energy              =   -61.82437317 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 6.83
>  ecutwfc = 200 Ry,  ecutrho = 1800 Ry :
> !    total energy              =   -61.82423710 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=
> 5.72
>
-------------------------------------------------------------------------------
> 
> WITH PBE-MT-GGA
>
NCPP:-----------------------------------------------------------
>     ---!   convergence scf caculation for Li2O  
> !---
>  ecutwfc = ecut = 50 Ry :
> !    total energy              =   -33.53719310 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P= -
> 329.31
>  ecutwfc = ecut = 80 Ry :
> !    total energy              =   -33.60362083 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.81
>  ecutwfc = ecut = 100 Ry :
> !    total energy              =   -33.60420031 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 48.44
>  ecutwfc = ecut = 140 Ry :
> !    total energy              =   -33.60518304 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.36
>  ecutwfc = ecut = 180 Ry :
> !    total energy              =   -33.60525632 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 45.97
>  ecutwfc = ecut = 200 Ry :
> !    total energy              =   -33.60552604 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.30
>  ecutwfc = ecut = 240 Ry :
> !    total energy              =   -33.60553853 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.24
>  ecutwfc = ecut = 280 Ry :
> !    total energy              =   -33.60553917 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.23
>  ecutwfc = ecut = 300 Ry :
> !    total energy              =   -33.60553970 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.24
>  ecutwfc = ecut = 340 Ry :
> !    total energy              =   -33.60553987 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.23
>  ecutwfc = ecut = 400 Ry :
> !    total energy              =   -33.60554013 Ry
>           total   stress  (Ry/bohr**3)              
>     (kbar)     P=  -
> 46.23
>
---------------------------------------------------------------------------------
> > _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
> 


-------------------------------------------------------------------
Prof. Eyvaz Isaev, 
Theoretical Physics Department, Moscow State Institute of Steel & Alloys, Russia, and 
Condensed Matter Theory Group, Uppsala University, Sweden Eyvaz.Isaev at fysik.uu.se, eyvaz_isaev at yahoo.com


       
____________________________________________________________________________________
Boardwalk for $500? In 2007? Ha! Play Monopoly Here and Now (it's updated for today's economy) at Yahoo! Games.
http://get.games.yahoo.com/proddesc?gamekey=monopolyherenow  


More information about the Pw_forum mailing list