[Pw_forum] problem with vc-relax
Masoud Nahali
masoudnahali at live.com
Wed Jul 28 16:54:42 CEST 2010
Dear Olga Sedelnikova
At first please use the C.pbe-rrkjus.UPF as the pseudopotential. This is
well tested by Nicola Marzari , Phys. Rev. B 71, 205214 (2005), and also I
have tested it. The paper of Nicola is very useful. please be careful that
it is an ultra soft pseudopotential and you should select ecutrho 10X
greater than ecutwfc. Don't be worry about the cell size in the Z-direction
since it is due to london interactions and may not be predicted well by DFT.
But you can use “london=.true.” to improve the results. For both
ion_dynamics and cell_dynamics I have used “bfgs”. I am sure that you will
get better results. Please examine and don't hesitates to inform me about
the problem.
Best wishes
Masoud Nahali
SUT
> The problem is not with wrong charge. The error "charge is wrong" appeared
> only for current options (smearing='mv', dynamics='damp', cg
> diagonalization). If I use other physically appropriate options I will
> obtain other errors: problem with davidson diagonalization, too small
> trust_radius for bfgs method and sometimes error from efirmig.f90 "internal
> error, cannot braket Ef". I have checked the dynamics of unit cell during
> vc-relaxation by XCrySDen and found that interlayer distance and bond
> length
> in graphite decreased to ~ 2 and 1.2 Angstroms correspondingly. For
> graphene calculations the unit cell was compressed along the z-axis and C-C
> bond length decreased. However the "relax" calculation is fine.
>
> Can anybody recommend the right options for "vc-relax" calculation of
> graphite?
>
> Thank you,
>
> Olga Sedelnikova
> Nikolaev Institute of Inorganic Chemistry of SB RAS
>
>
>
> > Please visit this achieve
> > http://www.democritos.it/pipermail/pw_forum/2006-August/004713.html, and
> > http://www.democritos.it/pipermail/pw_forum/2008-December/010853.html,
> > you may find the solution.
> > Best Regards
> > Sanjeev
> >
> > Dept. of Physics
> > Bhavnagar University
> > Gujarat
> >
> >
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> http://www.democritos.it/pipermail/pw_forum/attachments/20100728/5efc99a5/attachment-0001.htm
>
> ------------------------------
>
> Message: 2
> Date: Wed, 28 Jul 2010 11:44:27 +0200
> From: Merlin Meheut <merlin.meheut at lmtg.obs-mip.fr>
> Subject: [Pw_forum] Ni.pbe-sp-mt_gipaw.UPF
> To: pw_forum at pwscf.org
> Message-ID: <4C4FFBFB.6030206 at lmtg.obs-mip.fr>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Dear PWSCF users,
>
> I would like to compute fcc Nickel metal using the pseudopotential
> Ni.pbe-sp-mt_gipaw.UPF, available on your website.
> I do not need the GIPAW part, but I have been said that this should not
> be an issue. However, using an input very
> similar to example02, the calculation fails by "not recognizing the
> pseudo":
>
> -------------------------------------------------------
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>
> from read_pseudo_gipaw : error # 1
> UPF/GIPAW in unknown format
>
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
>
> stopping ...
> --------------------------------------------------------
>
> I use Quantum Espresso version 4.1.2.
>
> Below is my input file (which works with pseudo Ni.pbe-nd-rrkjus.UPF
> <http://www.quantum-espresso.org/pseudo/1.3/UPF/Ni.pbe-nd-rrkjus.UPF>) :
>
> &control
> disk_io = 'default' ,
> pseudo_dir = '$PBS_O_WORKDIR',
> outdir = '/tmpdir/$PBS_O_LOGNAME/',
> tprnfor = .true.,
> tstress = .true.,
> calculation = 'scf'
> prefix = 'ni',
> verbosity = 'high',
> /
> &system
> ibrav=2, celldm(1) =6.65, nat= 1, ntyp= 1,
> nspin=2,
> starting_magnetization(1)=0.5,
> degauss=0.02,
> smearing='mp',
> occupations='smearing',
> ecutwfc =27.0
> ecutrho =300.0
> /
> &electrons
> conv_thr = 1.0d-8
> mixing_beta = 0.7
> /
> ATOMIC_SPECIES
> Ni 58.6934 Ni.pbe-sp-mt_gipaw.UPF
> ATOMIC_POSITIONS
> Ni 0.00 0.00 0.00
> K_POINTS AUTOMATIC
> 8 8 8 1 1 1
>
> Thanks for any help,
>
> Merlin Meheut
>
> --
> Merlin M?heut
> Maitre de conf?rences
> LMTG- Universit? Paul Sabatier Toulouse 3
> 14 avenue Edoaurd Belin
> 31400 Toulouse
> tel: (+33)5 61 33 26 21
> Fax: (+33) 5 61 33 25 60
> Mobile: (+33) 6 34 67 57 02
>
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> http://www.democritos.it/pipermail/pw_forum/attachments/20100728/a31abd1f/attachment-0001.htm
>
> ------------------------------
>
> Message: 3
> Date: Wed, 28 Jul 2010 12:16:27 +0200 (CEST)
> From: Ari P Seitsonen <Ari.P.Seitsonen at iki.fi>
> Subject: Re: [Pw_forum] Ni.pbe-sp-mt_gipaw.UPF
> To: PWSCF Forum <pw_forum at pwscf.org>
> Message-ID: <alpine.LNX.2.00.1007281209050.28489 at pcihpc19.uzh.ch>
> Content-Type: text/plain; charset="iso-8859-15"
>
>
> Dear Merlin,
>
> You can either
>
> 1) remove the whole GIPAW part from the pseudo potential,
> 2) modify the pseudo potential file:
>
> <PP_PAW_FORMAT_VERSION>
> 0.10000000000E+00
> </PP_PAW_FORMAT_VERSION>
>
> and
>
> <PP_GIPAW_FORMAT_VERSION>
> 0.1
> </PP_GIPAW_FORMAT_VERSION>
>
> The version 4.1.2 of the code wants an integer in the latter case, so
> you can modify the "0.1" into "1",
>
> 3) or you modify the source code in Modules/read_upf_v1.f90 so that it
> accepts this syntax also.
>
> Sorry for the confusion (I am partially guilty for it). Please tell if
> this doesn't solve your problem.
>
> Greetings from Zurich,
>
> apsi
>
>
> -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
> Ari P Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/<http://www.iki.fi/%7Eapsi/>
> Physikalisch-Chemisches Institut der Universitaet Zuerich
> Tel: +41 44 63 54 497 / Mobile: +41 79 71 90 935
>
> On Wed, 28 Jul 2010, Merlin Meheut wrote:
>
> > Dear PWSCF users,
> >
> > I would like to compute fcc Nickel metal using the pseudopotential
> > Ni.pbe-sp-mt_gipaw.UPF, available on your website.
> > I do not need the GIPAW part, but I have been said that this should not
> be
> > an issue. However, using an input very
> > similar to example02, the calculation fails by "not recognizing the
> pseudo":
> >
> > -------------------------------------------------------
> >
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> > %%%
> > ???? from read_pseudo_gipaw : error #???????? 1
> > ???? UPF/GIPAW in unknown format
> >
> ?%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> > %%%%
> >
> > ???? stopping ...
> > --------------------------------------------------------
> >
> > I use Quantum Espresso version 4.1.2.
> >
> > Below is my input file (which works with pseudo Ni.pbe-nd-rrkjus.UPF) :
> >
> > &control
> > ?????????? disk_io = 'default' ,
> > ??? pseudo_dir???? = '$PBS_O_WORKDIR',
> > ??? outdir???????? = '/tmpdir/$PBS_O_LOGNAME/',
> > ??? tprnfor??????? = .true.,
> > ??? tstress??????? = .true.,
> > ??? calculation??? = 'scf'
> > ??? prefix???????? = 'ni',
> > ??? verbosity????? = 'high',
> > ?/
> > &system
> > ??? ibrav=2, celldm(1) =6.65, nat=? 1, ntyp= 1,
> > ??? nspin=2,
> > ??? starting_magnetization(1)=0.5,
> > ??? degauss=0.02,
> > ??? smearing='mp',
> > ??? occupations='smearing',
> > ??? ecutwfc =27.0
> > ??? ecutrho =300.0
> > ?/
> > &electrons
> > ??? conv_thr =? 1.0d-8
> > ??? mixing_beta = 0.7
> > ?/
> > ATOMIC_SPECIES
> > ?Ni? 58.6934? Ni.pbe-sp-mt_gipaw.UPF
> > ATOMIC_POSITIONS
> > ?Ni 0.00 0.00 0.00
> > K_POINTS AUTOMATIC
> > 8 8 8 1 1 1
> >
> > Thanks for any help,
> >
> > Merlin Meheut
> >
> > --
> > Merlin M?heut
> > Maitre de conf?rences
> > LMTG- Universit? Paul Sabatier Toulouse 3
> > 14 avenue Edoaurd Belin
> > 31400 Toulouse
> > tel: (+33)5 61 33 26 21
> > Fax: (+33) 5 61 33 25 60
> > Mobile: (+33) 6 34 67 57 02
> >
> >
> >
>
> ------------------------------
>
> Message: 4
> Date: Wed, 28 Jul 2010 06:36:36 -0700 (PDT)
> From: Eyvaz Isaev <eyvaz_isaev at yahoo.com>
> Subject: Re: [Pw_forum] problem with vc-relax
> To: PWSCF Forum <pw_forum at pwscf.org>
> Message-ID: <846523.52444.qm at web65713.mail.ac4.yahoo.com>
> Content-Type: text/plain; charset="us-ascii"
>
> Dear Olga,
>
> I could not reproduce the error you report using different pw.x versions.
> They finished by the error message (QE 4.2, compiler gfortran on Ubuntu
> 9.4 on
> my laptop, serial):
>
> from scale_h : error # 1
> Not enough space allocated for radial FFT: try restarting with a larger
> cell_factor.
>
> Usually, that means unreasonable lattice parameters. Say, one of them is
> much
> larger than expected. In fact, this turned out to be dependent on ecutwfc
> and
> pseudopotential type. The error persists for norm-conserving type pseudo
> you
> used. But using US pseudo for C, and the cutoff energy 30Ry (360 for
> ecutrho) ,
> no problem arises.
>
> Besides, for Norm-conserving type pseudo for C 18Ry seems to be extremely
> low.
> You can play around ecutwfc, upscale and cell_factor parameters.
>
> If you like, you can fix c-axis, do relax, and using a set of c-axis
> parameter
> find the minimum.
>
>
> Last, if you study a graphene sheet why you used 6 6 6? The correct one
> seems to
> be 6 6 1.
>
> Bests,
> Eyvaz.
> -------------------------------------------------------------------
> Prof. Eyvaz Isaev,
> Department of Physics, Chemistry, and Biology (IFM), Linkoping University,
> Sweden
>
> Theoretical Physics Department, Moscow State Institute of Steel & Alloys,
> Russia,
>
> isaev at ifm.liu.se, eyvaz_isaev at yahoo.com
>
>
>
>
> ________________________________
> From: Olga Sedelnikova <o.sedelnikova at gmail.com>
> To: pw_forum at pwscf.org
> Sent: Tue, July 27, 2010 12:42:53 PM
> Subject: [Pw_forum] problem with vc-relax
>
>
> Dear All,
> I have a problem with 'vc-relax' option for graphene. I have used david and
> cg
> diagonalization proceedures, different types of smearing and ion and cell
> dynamics but all calculations were crashed.
> One of my input file:
>
> &CONTROL
> calculation = 'vc-relax',
> prefix = 'graphene',
> restart_mode='from_scratch'
> tstress=.true.,
> tprnfor=.true.,
> nstep=45,
> etot_conv_thr = 1.0E-4 ,
> forc_conv_thr = 1.0D-3 ,
> max_seconds = 36000 ,
> dt = 100,
> /
> &SYSTEM
> ibrav = 4,
> celldm(1)= 4.64833
> celldm(3)=5.447621
> nat = 2,
> ntyp = 1,
> ecutwfc = 18.0,
> nbnd =8,
> occupations='smearing'
> smearing = 'mv'
> degauss = 0.005
> /
> &ELECTRONS
> conv_thr = 1.0d-7
> electron_maxstep = 70
> diagonalization='cg'
> /
> &IONS
> ion_dynamics='damp'
> /
> &CELL
> cell_dynamics = 'damp-w' ,
> press=0.00 ,
> wmass=0.0015 ,
> /
>
> ATOMIC_SPECIES
> C 12.01100 C.pz-vbc.UPF
> ATOMIC_POSITIONS crystal
> C 0.00000000 0.00000000 0.50000000
> C 0.33333333 0.66666666 0.50000000
> K_POINTS {automatic}
> 6 6 6 0 0 0
>
> Inthis cause:
>
> from electrons : error # 1
> charge is wrong
>
> Maybe somebody has already dealed with this problem and can advice the
> correct
> options for graphene.
> Any suggestion will be appreciated.
>
> --
> Best wishes,
> Olga Sedelnikova
> Nikolaev Institute of Inorganic Chemistry of SB RAS
>
>
>
>
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> http://www.democritos.it/pipermail/pw_forum/attachments/20100728/46ff1114/attachment.htm
>
> ------------------------------
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>
> End of Pw_forum Digest, Vol 37, Issue 60
> ****************************************
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20100728/0ea9aea0/attachment-0001.htm
More information about the Pw_forum
mailing list