[Pw_forum] About LDA+U method on ZnO structural optimiztion

Thu May 22 12:07:12 CEST 2014

Dear J.J.
I do not know how the projection of Kohn-Sham orbitals on the local atomic basis set (one of the possible choices...) is performed in VASP. It may be 
the source of discrepancies between VASP and QE results. Moreover, as you use a single angular momentum channel to perform the projection, and you 
must add the on site correction to the GGA energy "purged" from the self interaction term, I suspect that all of this mechanism is also affected by the 
shape of the pseudopotential projectors, and that different PPs may provide different results. Extensive benchmarks are difficult because VASP is quite 
expensive, and because it uses only PAW PPs, while most of the QE results have been obtained by using NC and US PPs.
Anyway, thank you for having put across the UHF method for closed shell systems, which I did not know.
YS
Giuseppe

On Thursday 22 May 2014 06:12:35 jijun gong wrote:
> To Giuseppe,
> 
>     I really appreciate your help! Your precious experiences of LDA+U
> for ZnO are of great help to me.
>     Here, I still have some confusion about the LDA+U method. In the
> J. Chem. Phys. 140, 121105 (2014), they used the Hartree-Fock theory
> (UHF) and got a U value of 4.5 eV for Zn d electrons in ZnO. With this
> U value, they performed calculations using VASP code and found a
> decreasing trend for ZnO cell parameters, the result fairly agrees
> with the experimental data.
>     However, to my surprise, by using the PWSCF code with the same U
> value, I got a slightly increasing trend and the error becomes mildly
> larger than DFT's. Obviously, there was an opposite tendency between
> the LDA+U results calculated by these two codes in this issue. So,
> with respect to the DFT+U method, I confused if there was a dependence
> on codes.
> 
> Thanks again!
> 
> J. J. Gong
> 
> > Date: Tue, 20 May 2014 21:34:44 +0200
> > From: Giuseppe Mattioli <giuseppe.mattioli at ism.cnr.it>
> > Subject: Re: [Pw_forum] About LDA+U method on ZnO structural
> > 
> > 	optimiztion
> > 
> > To: pw_forum at pwscf.org
> > 
> > Message-ID:
> > 	<20140520213444.Horde.9n2sgIz_PqySRXQdrNtORg1 at webmail.sic.rm.cnr.it>
> > 
> > Content-Type: text/plain; charset=UTF-8; format=flowed; DelSp=Yes
> > 
> > 
> > Dear J.J.
> > I've performed several ZnO-related calculations, including DFT+U. In
> > particular, you may be interested in the supporting information of
> > Mattioli, et al.; J. Phys. Chem. C (2012) 116, 15439-15448, where the
> > properties of ZnO bulk obtained at different levels of theory are
> > discussed. In my experience, DFT+U alone does not guarantee good
> > structural results in metal oxides, if you do not use it with care. In
> > the above paper (and in a couple of following contributions), I've
> > tested a double (and substantially semiempirical...) DFT+U(Zn,O)
> > correction, applied to both Zn 3d (7.0 eV) and O 2p (5.9 eV) shells. I
> > was more interested in electronic than structural properties, and the
> > former were good. Neverheless I've obtained quite reasonable DFT+U
> > lattice parameters (a=3.248, c=5.223). At any rate, your DFT+U results
> > seem not so bad: +1.3% on a and +1.9 on c should be considered as
> > acceptable overestimates.
> > HTH
> > Giuseppe
> > 
> > Giuseppe Mattioli
> > ISM-CNR
> > Italy
> > 
> > Quoting jijun gong <jijungong at gmail.com>:
> >> Dear all:
> >>     I am working on ZnO structural optimization using the LDA+U method.
> >> 
> >> As
> >> in the journal of chemical physics 140, 121105 (2014), the U value on Zn
> >> d
> >> electrons was set to be 4.5 eV, calculated by vasp code, there was a
> >> good agreement with experimaental data in cell parameters by DFT+U than
> >> DFT method did.
> >> 
> >>     the a, c and u in experimental values are 3.250, 5.210, 0.382,
> >> 
> >> respectively.
> >> 
> >>     Calculated by VASP code,
> >>     In DFT,the a, c and u values are 3.286, 5.295, 0.3807, respectively.
> >>     In DFT+U,the a, c and u values are 3.243, 5.218, 0.3809,
> >> 
> >> respectively.
> >> 
> >>     It seems that the DFT+U correction works pretty well with the vasp
> >> 
> >> code.
> >> 
> >>     But when using pwscf in vc-relax mode, I got the fellowing results:
> >>     In DFT,the a, c and u values are 3.290, 5.305, 0.379, respectively.
> >>     In DFT+U,the a, c and u values are 3.292, 5.309, 0.379,
> >>     respectively. It shows that in pwscf, the DFT+U correction does not
> >>     much affect the
> >> 
> >> results. also the datas show that there is significant difference in DFT
> >> +U
> >> results between the two codes.
> >> 
> >>     My input file with pwscf attached to the following. I want to know
> >> 
> >> where I went wrong in the pwscf input file.
> >> 
> >>     Any help will be appreciated!
> >>     
> >>     J.J. Gong
> >>     Lanzhou University of technology, China
> >> 
> >> the pwscf input file
> >> 
> >>  &CONTROL
> >> 
> >> calculation = "vc-relax",
> >> restart_mode = "from_scratch",
> >> prefix = "ZnO-opt",
> >> outdir = "/tmp/",
> >> pseudo_dir = "~/pp",
> >> !wf_collect=.TRUE.
> >> !etot_conv_thr = 1d-5
> >> !forc_conv_thr = 1d-3
> >> tstress = .true.
> >> tprnfor = .true.
> >> /
> >> &SYSTEM
> >> ibrav=0,
> >> celldm(1)=1.88972613,
> >> ecutwfc=55, ecutrho=550,
> >> nat=4,ntyp=2,
> >> lda_plus_U=.true., lda_plus_u_kind = 0, Hubbard_U(1)=4.5, Hubbard_U(2)=
> >> 0,
> >> /
> >> &ELECTRONS
> >> electron_maxstep = 80,
> >> mixing_beta=0.5,
> >> conv_thr=1D-8,
> >> /
> >> &ions
> >> /
> >> &cell
> >> /
> >> 
> >> 
> >> ATOMIC_SPECIES
> >> 
> >>   Zn  65.3900  Zn.pbe-van.UPF
> >>   
> >>    O  15.9994   o_pbe_v1.2.uspp.F.UPF
> >> 
> >> CELL_PARAMETERS (alat=  1.88972613)
> >> 
> >>    3.292518870   0.000394478   0.000000646
> >>   
> >>   -1.640667267   2.850447665  -0.000000288
> >>   
> >>    0.000001154   0.000000426   5.310381081
> >> 
> >> ATOMIC_POSITIONS (crystal)
> >> O        0.332685749   0.665993276   0.379582752
> >> O        0.667312132   0.334006494   0.879563839
> >> Zn       0.333421586   0.666985007   0.000318706
> >> Zn       0.666579130   0.333015022   0.500232704
> >> K_POINTS (automatic)
> >> 4 4 3 1 1 0
> >> 
> >> 
> >> --
> >> JiJun-Gong
> >> 
> >> Department of Applied Physics
> >> Lanzhou University of Technology
> >> Lanzhou
> >> China
> > 
> > --
> > ********************************************************
> > - Article premier - Les hommes naissent et demeurent
> > libres et ?gaux en droits. Les distinctions sociales
> > ne peuvent ?tre fond?es que sur l'utilit? commune
> > - Article 2 - Le but de toute association politique
> > est la conservation des droits naturels et
> > imprescriptibles de l'homme. Ces droits sont la libert?,
> > la propri?t?, la s?ret? et la r?sistance ? l'oppression.
> > ********************************************************
> > 
> >     Giuseppe Mattioli
> >     CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
> >     v. Salaria Km 29,300 - C.P. 10
> >     I 00015 - Monterotondo Stazione (RM)
> >     Tel + 39 06 90672836 - Fax +39 06 90672316
> >     E-mail: <giuseppe.mattioli at ism.cnr.it>

********************************************************
- Article premier - Les hommes naissent et demeurent
libres et ègaux en droits. Les distinctions sociales
ne peuvent être fondèes que sur l'utilitè commune
- Article 2 - Le but de toute association politique
est la conservation des droits naturels et 
imprescriptibles de l'homme. Ces droits sont la libertè,
la propriètè, la sùretè et la rèsistance à l'oppression.
********************************************************

   Giuseppe Mattioli                            
   CNR - ISTITUTO DI STRUTTURA DELLA MATERIA   
   v. Salaria Km 29,300 - C.P. 10                
   I 00015 - Monterotondo Stazione (RM)          
   Tel + 39 06 90672836 - Fax +39 06 90672316    
   E-mail: <giuseppe.mattioli at ism.cnr.it>
   http://www.ism.cnr.it/english/staff/mattiolig
   ResearcherID: F-6308-2012