Maybe you can check if different starting magnetizations converge into one.
On Tue, 17 Jun 2014 02:02:41 +0000, "Kurth, Michael R." <kurth.17 at buckeyemail.osu.edu> wrote: > Dear Juanjo, > > Thank you for the advice. We did try a DFT+U calculation only on the > neodymium atoms, but it was also with the default number of bands. We saw > similar results in terms of total magnetization as the other calculations > we ran without the +U correction. > > Thanks, > Michael > > Graduate Student in Nuclear Engineering > Department of Mechanical and Aerospace Engineering > The Ohio State University > > > From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On > Behalf Of Juanjo Mel?ndez > Sent: Monday, June 16, 2014 1:37 PM > To: PWSCF Forum > Subject: Re: [Pw_forum] Nd2Fe14B Magnetization > > Dear Michael > > I think that the problem is the presence of Nd in your lattice. As far as > I know, f-electrons are not only tricky to handle within a DFT formalism, > but they constitute the border of its validity. I would not expect a great > degree of accuracy in your calculations for any PP you may use. > > Anyway, you could try with DFT+U with the proper correction for correlated > f-electrons (and maybe d-electrons from Fe as well). Sometimes it is hard > to converge, but it could help. Anyway, I would not expect any miracle. > > BTW. Do not waste your time trying to change ecutrho. You are using a > norm-conserving pseudopotential, so that ecutrho is meaningless. > > Good luck > > Juanjo > > Juan J. Mel?ndez > Associate Professor > Department of Physics ? University of Extremadura > Avda. de Elvas, s/n 06006 Badajoz (Spain) > Phone: +34 924 28 96 55 > Fax: +34 924 28 96 51 > Email: melendez at unex.es > Web: http://materiales.unex.es/miembros/personal/jj-melendez/Index.html > > From: Kurth, Michael R.<mailto:kurth.17 at buckeyemail.osu.edu> > Sent: Monday, June 16, 2014 5:52 PM > To: PWSCF Forum<mailto:pw_forum at pwscf.org> > Subject: Re: [Pw_forum] Nd2Fe14B Magnetization > > Dear Dr. Rebaza, > > Thank you for the reply. We did increase ecutrho to approximately 6 times > the ecutwfc value and did not see a notable difference in the calculation. > As far as the value of ecutwfc=260, we arrived at this value after running > calculations at ecutwfc=300 and not observing a 1 mRy/atom difference in > the total energy. We also have tested PBE and PW pseudopotentials but have > not observed a difference in the total magnetization value. > > Sincerely, > Michael Kurth > > Graduate Student in Nuclear Engineering > Department of Mechanical and Aerospace Engineering > The Ohio State University > From: pw_forum-bounces at pwscf.org [mailto:pw_forum-bounces at pwscf.org] On > Behalf Of Arles V. Gil Rebaza > Sent: Monday, June 16, 2014 11:28 AM > To: PWSCF Forum > Subject: Re: [Pw_forum] Nd2Fe14B Magnetization > > Dear Michael, are you sure that "ecutwfc = 260.0" is a enough value for > your system, may be you should increse, and you're using the default value > of "ecutrho", try to converge with this tag.!!! > > Best > > PhD. Arles V. GIl Rebaza > Instituto de F?sica La Plata > La Plata - Argentina > > 2014-06-16 12:16 GMT-03:00 Kurth, Michael R. > <kurth.17 at buckeyemail.osu.edu<mailto:kurth.17 at buckeyemail.osu.edu>>: > > > Hello, > > > > I am trying to calculate the magnetization of NdFeB which is a > ferromagnetic material and a metal. This calculation has been done before > using different methods than plane waves and pseudopotentials. Moreover, > the magnetization per formula unit has been experimentally measured. The > previous calculations and experimental values agree and report a > magnetization per formula unit of around 37 Bohr Magneton. The previous > calculations used the Generalized Gradient Approximation for the exchange > correlation functional. > > > > We first tried the calculation using the default number of bands and the > input file is attached below: > > > > &control > > calculation='scf' > > restart_mode='from_scratch', > > wf_collect=.TRUE., > > prefix='NdFeB', > > pseudo_dir = '/usr/local/espresso/espresso-5.0.3/pseudo/', > > outdir = '/fs/lustre/osu7834/NdFeB/' > > / > > &system > > ibrav=0, nat=68, ntyp=3, > > nspin=2, starting_magnetization(1)=0.6, starting_magnetization(2)=0.6, > starting_magnetization(3)=0.6, > > ecutwfc = 260.0 > > occupations='smearing', smearing='mv', degauss=0.002 > > / > > &electrons > > conv_thr=1.0e-6 > > mixing_beta=0.02 > > mixing_mode = 'local-TF' > > / > > CELL_PARAMETERS {angstrom} > > 8.8 0.0 0.0 > > 0.0 12.19 0.0 > > 0.0 0.0 8.8 > > ATOMIC_SPECIES > > Nd 144.24 Nd.pbe-mt_fhi.UPF > > Fe 55.845 Fe.pbe-mt_fhi.UPF > > B 10.811 B.pbe-mt_fhi.UPF > > ATOMIC_POSITIONS {angstrom} > > B 5.54400 0.00000 3.25600 > > B 3.25600 0.00000 5.54400 > > B 7.65600 6.09500 7.65600 > > B 1.14400 6.09500 1.14400 > > Fe 4.40000 0.00000 0.00000 > > Fe 4.40000 6.09500 0.00000 > > Fe 0.00000 6.09500 4.40000 > > Fe 0.00000 0.00000 4.40000 > > Fe 4.40000 1.37747 4.40000 > > Fe 0.00000 7.47247 0.00000 > > Fe 0.00000 4.71753 0.00000 > > Fe 4.40000 10.81253 4.40000 > > Fe 0.85360 2.49895 0.85360 > > Fe 7.94640 2.49895 7.94640 > > Fe 5.25360 8.59395 3.54640 > > Fe 3.54640 8.59395 5.25360 > > Fe 5.25360 3.59605 3.54640 > > Fe 3.54640 3.59605 5.25360 > > Fe 0.85360 9.69105 0.85360 > > Fe 7.94640 9.69105 7.94640 > > Fe 2.79840 3.01093 2.79840 > > Fe 6.00160 3.01093 6.00160 > > Fe 7.19840 9.10593 1.60160 > > Fe 1.60160 9.10593 7.19840 > > Fe 7.19840 3.08407 1.60160 > > Fe 1.60160 3.08407 7.19840 > > Fe 2.79840 9.17907 2.79840 > > Fe 6.00160 9.17907 6.00160 > > Fe 4.99840 1.56032 1.97120 > > Fe 3.80160 1.56032 6.82880 > > Fe 6.37120 7.65532 8.20160 > > Fe 2.42880 7.65532 0.59840 > > Fe 0.59840 4.53468 2.42880 > > Fe 8.20160 4.53468 6.37120 > > Fe 1.97120 10.62968 4.99840 > > Fe 6.82880 1.56032 3.80160 > > Fe 3.80160 10.62968 6.82880 > > Fe 4.99840 10.62968 1.97120 > > Fe 2.42880 4.53468 0.59840 > > Fe 6.37120 4.53468 8.20160 > > Fe 8.20160 7.65532 6.37120 > > Fe 0.59840 7.65532 2.42880 > > Fe 6.82880 10.62968 3.80160 > > Fe 1.97120 1.56032 4.99840 > > Fe 3.15920 2.14544 0.34320 > > Fe 5.64080 2.14544 8.45680 > > Fe 4.74320 8.24044 1.24080 > > Fe 4.05680 8.24044 7.55920 > > Fe 7.55920 3.94956 4.05680 > > Fe 1.24080 3.94956 4.74320 > > Fe 0.34320 10.04456 3.15920 > > Fe 8.45680 2.14544 5.64080 > > Fe 5.64080 10.04456 8.45680 > > Fe 3.15920 10.04456 0.34320 > > Fe 4.05680 3.94956 7.55920 > > Fe 4.74320 3.94956 1.24080 > > Fe 1.24080 8.24044 4.74320 > > Fe 7.55920 8.24044 4.05680 > > Fe 8.45680 10.04456 5.64080 > > Fe 0.34320 2.14544 3.15920 > > Nd 2.34080 0.00000 2.34080 > > Nd 6.45920 0.00000 6.45920 > > Nd 6.74080 6.09500 2.05920 > > Nd 2.05920 6.09500 6.74080 > > Nd 7.56800 0.00000 1.23200 > > Nd 1.23200 0.00000 7.56800 > > Nd 5.63200 6.09500 5.63200 > > Nd 3.16800 6.09500 3.16800 > > > > K_POINTS {automatic} > > 4 4 4 0 0 0 > > > > > > This input file was prepared after checking convergence with respect to > the energy cutoff and the number of points and did converge after 2 restart > runs. Unfortunately, we obtained a magnetization of 25 Bohr Magneton/f.u. > (which is too small). > > > > Furthermore, and after checking for convergence with respect to the number > of bands (by adding more bands) we found that we end up with a > magnetization of 50 Bohr Magneton/f.u. (which is too big). > > I also think that the degauss value used for smearing is small enough and > can't be the problem. > > Are there any recommendations for what other parameters I could examine to > help me get the right answer? > > > > Your help is greatly appreciated! > > > > Sincerely, > Michael Kurth > > > Graduate Student in Nuclear Engineering > > Department of Mechanical and Aerospace Engineering > > The Ohio State University > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org<mailto:Pw_forum at pwscf.org> > http://pwscf.org/mailman/listinfo/pw_forum > > > > -- > ###---------> Arles V. <---------### > ________________________________ > No virus found in this message. > Checked by AVG - www.avg.com<http://www.avg.com> > Version: 2014.0.4592 / Virus Database: 3964/7686 - Release Date: 06/16/14 > ________________________________ > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > > ________________________________ > [Image removed by sender.]<http://www.avast.com/> > > > Este mensaje no contiene virus ni malware porque la protecci?n de avast! > Antivirus<http://www.avast.com/> est? activa. > > > ________________________________ > No virus found in this message. > Checked by AVG - www.avg.com<http://www.avg.com> > Version: 2014.0.4592 / Virus Database: 3964/7686 - Release Date: 06/16/14 -- PostDoc Department of Physics, South University of Science and Technology of China
