[Pw_forum] Wrong effective charges in AFM insulator using "zue"

Andrea Dal Corso dalcorso at sissa.it
Wed Jan 19 09:41:54 CET 2011


On Wed, 2011-01-19 at 16:27 +0800, xirainbow wrote:
> In 4.2.1 manual, it is said that:If .zue=true.is an alternative
> algorithm, different from the default one (if trans .and. epsil ) The
> results should be the same within numerical noise.
> 
> 
> However, I get wrong "effective charges" in zue=true in AFM insulator
> CaMnO3 with 4.2.1
> These two Mn atoms are the same except antiparallel spin.
> In "epsil+trans", effective charges of Mn1=7.79, Mn2=7.78;
> In "zue"             , effective charges of Mn1=3.1,   Mn2=51.2;
> 
> 
> I will describe my calculation procedure using 4.2.1 in short.
> 1:  I do scf calculation in scf.in
> restart_mode  = 'from_scratch'
> occupations   = 'smearing'
> smearing      = 'gauss'
> degauss       = 0.002
> starting_magnetization(1) = 1
> starting_magnetization(2) = -1
> starting_magnetization(3) = 0
> starting_magnetization(4) = 0
> 
> 
> 2: I do a scf calculation  in scf.in
> restart_mode  = 'restart'
> occupations               = 'fixed'
> nspin                     = 2
> tot_magnetization         = 0
> 
> 
> 3: I do ph calculation with 
> zue         = .TRUE. !the same with epsil and trans = TRUE
> (In this condition, effective charges of Mn1=3.1,   Mn2=51.2;)
> or with:
> epsil       = .TRUE. , trans       = .TRUE. 
> (In this condition , effective charges of Mn1=7.79, Mn2=7.78;)
> 
> 
> (The reason is why I restart scf is below: 
> "starting_magnetization" must be used with "smearing". However,
> "smearing" will give a metal result. Effective charges can not be
> calculated for metal. But, CaMnO3 is really a insulator AFM.
> "occupations=fixed" can get a insulator result. However it must used
> with "tot_magnetization". If I use "tot_magnetization=0", I will get a
> PM no AFM state.
> In the end, I use "smearing"+"starting_magnetization" to get a AFM
> metal. Then I use "fixed"+"tot_magnetization" to get a AFM insulator.
> Is there any other good method to get a AFM insulator?)

> 
> 
> 
> The following my script for "zue= .TRUE "
> And I have attached output file of "zue= .TRUE " in accessary.
> Thanks:)

The phonon with LSDA + constraint is not programmed. So it will not
work. If the AFM insulator is the ground state then you should obtain it
with noncolin=.true. without constraint. In this case ph.x at gamma
should work and you should get the correct effective charges. If not
please provide the complete input.

HTH,

Andrea



> 
> #*********************************************
> # scf-antimag-smearing
> cat>scf.in<<EOF
> for vc-relax
> &CONTROL
> calculation   = 'scf'
> title         = 'G-10-scf'
> verbosity     = 'high'
> restart_mode  = 'from_scratch'
> wf_collect    = .FALSE.
> tstress       = .TRUE.
> tprnfor       = .TRUE.
> prefix        = 'G-10-scf'
> etot_conv_thr = 1.0e-4
> forc_conv_thr = 1.0e-3
> disk_io       = 'low'
> pseudo_dir    = './'
> /
> 
> 
> &SYSTEM
> ibrav         = 2
> celldm        = 14.20419765
> nat           = 10
> ntyp          = 4
> nbnd          = 50
> ecutwfc       = 70
> ecutrho       = 560
> nosym         = .FALSE.
> !nosym_evc     = .TRUE.
> !noinv         = .TRUE.
> occupations   = 'smearing'
> smearing      = 'gauss'
> degauss       = 0.002
> starting_magnetization(1) = 1
> starting_magnetization(2) = -1
> starting_magnetization(3) = 0
> starting_magnetization(4) = 0
> !occupations               = 'fixed'
> nspin                     = 2
> !tot_magnetization         = 0
> /
> 
> 
> &ELECTRONS
> electron_maxstep  = 100
> conv_thr          = 1.0e-7
> mixing_mode       = 'plain'
> mixing_beta       = 0.7
> mixing_ndim       = 8
> diagonalization   = 'david'
> diago_david_ndim  = 4
> /
> 
> 
> ATOMIC_SPECIES
>  Mn1 55  Mn.pbe-sp-van.UPF
>  Mn2 55  Mn.pbe-sp-van.UPF
>  Ca  40  Ca.pbe-nsp-van.UPF
>  O  16  O.pbe-rrkjus.UPF
> 
> 
> ATOMIC_POSITIONS {crystal}
>  Mn1 0.00    0.00    0.00
>  Mn2 0.50    0.50    0.50
>  Ca  0.25    0.25    0.25
>  Ca  0.75    0.75    0.75
>  O   0.75    0.25    0.75
>  O   0.25    0.75    0.25
>  O   0.75    0.75    0.25
>  O   0.25    0.25    0.75
>  O   0.25    0.75    0.75
>  O   0.75    0.25    0.25
>  
> 
> 
> K_POINTS {automatic}
> 6 6 6 0 0 0
> EOF
> ./dopw <scf.in >G-scf-10-smearing.out
> 
> 
> # scf-antimag-insulator
> cat>scf.in<<EOF
> for vc-relax
> &CONTROL
> calculation   = 'scf'
> title         = 'G-10-scf'
> verbosity     = 'high'
> restart_mode  = 'restart'
> wf_collect    = .FALSE.
> tstress       = .TRUE.
> tprnfor       = .TRUE.
> prefix        = 'G-10-scf'
> etot_conv_thr = 1.0e-4
> forc_conv_thr = 1.0e-3
> disk_io       = 'low'
> pseudo_dir    = './'
> /
> 
> 
> &SYSTEM
> ibrav         = 2
> celldm        = ${lattic_parameter} !bohr
> nat           = 10
> ntyp          = 4
> nbnd          = 50
> ecutwfc       = 70
> ecutrho       = 560
> nosym         = .FALSE.
> !nosym_evc     = .TRUE.
> !noinv         = .TRUE.
> !occupations   = 'smearing'
> !smearing      = 'gauss'
> !degauss       = 0.002
> !starting_magnetization(1) = 1
> !starting_magnetization(2) = -1
> !starting_magnetization(3) = 0
> !starting_magnetization(4) = 0
> occupations               = 'fixed'
> nspin                     = 2
> tot_magnetization         = 0
> /
> 
> 
> &ELECTRONS
> electron_maxstep  = 100
> conv_thr          = 1.0e-7
> mixing_mode       = 'plain'
> mixing_beta       = 0.7
> mixing_ndim       = 8
> diagonalization   = 'david'
> diago_david_ndim  = 4
> /
> 
> 
> ATOMIC_SPECIES
>  Mn1 55  Mn.pbe-sp-van.UPF
>  Mn2 55  Mn.pbe-sp-van.UPF
>  Ca  40  Ca.pbe-nsp-van.UPF
>  O  16  O.pbe-rrkjus.UPF
> 
> 
> ATOMIC_POSITIONS {crystal}
>  Mn1 0.00    0.00    0.00
>  Mn2 0.50    0.50    0.50
>  Ca  0.25    0.25    0.25
>  Ca  0.75    0.75    0.75
>  O   0.75    0.25    0.75
>  O   0.25    0.75    0.25
>  O   0.75    0.75    0.25
>  O   0.25    0.25    0.75
>  O   0.25    0.75    0.75
>  O   0.75    0.25    0.25
>  
> 
> 
> K_POINTS {automatic}
> 6 6 6 0 0 0
> EOF
> ./dopw <scf.in >G-scf-10-insulator.out
> 
> 
> #for phonon-antimag
> cat>ph.in <<EOF
> for phonon
> &INPUTPH
> amass(1)    = 55
> amass(2)    = 55
> amass(3)    = 40
> amass(4)    = 16
> outdir      = "./"
> prefix      = 'G-10-scf' !must be the same with scf
> ldisp       = .FALSE.
> niter_ph    = 100
> tr2_ph      = 1.0e-12
> alpha_mix(1)= 0.7
> nmix_ph     = 4
> iverbosity  = 1
> fildyn      = 'matdyn'
> !epsil       = .TRUE. !Do not set epsil to .true. for metallic system
> or q/=0
> !trans       = .TRUE. !if trans .and. epsil effective charges are
> calculated
> zue         = .TRUE. !the same with epsil and trans = TRUE
> /
> 0.0 0.0 0.0
> EOF
> 
> 
> ./doph <ph.in >G-ph-10.out
> done
> #*********************************************
> 
> -- 
> ____________________________________
> Hui Wang
> School of physics, Fudan University, Shanghai, China
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
-- 
Andrea Dal Corso                    Tel. 0039-040-3787428
SISSA, Via Bonomea 265              Fax. 0039-040-3787249
I-34136 Trieste (Italy)             e-mail: dalcorso at sissa.it





More information about the users mailing list