[QE-users] SCF calculations on WN supercell not converging
Nicola Marzari
nicola.marzari at epfl.ch
Sun Jun 21 20:16:17 CEST 2020
Thanks Ari!
and I know I've said it 1,000 times, but it would be ideal to alway
double check against this
https://www.materialscloud.org/work/tools/qeinputgenerator
and then understand why your own choices are different.
Not that the above is the truth, but at least it has careful guesses of
parameters and pseudopotentials. BTW, the instructions mention
explicitly the k-point sampling of slabs.
nicola
On 21/06/2020 20:09, Ari P Seitsonen wrote:
>
> Dear Coralie,
>
> Besides the concern of the possibly errorneous pseudo potential, I
> would recommend that you begin with a simpler system at the beginning if
> you are not yet familiar with the (periodic) DFT calculations, in
> particular the broadening, empty bands, choice of k points and so on.
>
> In particular here, it seems that you have been trying to use quite
> many k points in the directions corresponding to the surface plane
> (earlier 2x2, now 4x4 will consume plenty of computing time), and in
> particular in the direction corresponding to the surface normal, where
> you have (in principle) no periodicity there is no need to specify more
> than "1" in the k points. I would also recommend to try if you can make
> the laterally 1x1 unit cell, without the adsorbate to converge, before
> going to such large a slab. Are you sure about the values of the
> 'starting_magnetization()' on the nitrogen and tungsten atoms - 60 %
> polarisation on the latter for example (please notice that this is in
> _relative_ units, so the tungsten atoms have a magnetic moment of 0.6 *
> 14 valence electrons in the initial electron density - the magnetisation
> is larger than the "true valence" of four electrons of W)? The initial
> electronic structure might be very far from the final, self-consistent
> one that you are trying to reach. What is the magnetic ordering in the
> bulk, is it ferro-magnetic?
>
> Greetings from Paris,
>
> apsi
>
> -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
>
> Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
>
>
> On Fri, 19 Jun 2020, Coralie Khabbaz wrote:
>
>> Hello,
>> Thank you so much for your response. Do you think it will converge if
>> I use the PAW pseudopotential instead of
>> the ultrasoft?
>>
>> On Fri, 19 Jun 2020 at 04:15, Andrea Urru <aurru at sissa.it> wrote:
>> Dear Coralie,
>> the issue you are facing is most likely due to the W pseudopotential,
>> which gives convergence issues if
>> used in slab systems
>> with vacuum space, as reported
>> here: https://dalcorso.github.io/pslibrary/PP_list.html
>>
>> The behavior you describe might be due to a ghost state, but I am not
>> deeply expert in pseudopotentials
>> and I cannot swear.
>> I would suggest you to try a different pseudopotential. If you wish to
>> use Ultrasoft pseudopotentials you
>> may try the one from
>> Pslibrary 0.3.1.
>>
>> Best regards,
>>
>> Andrea Urru
>>
>> Ph. D. Student in Condensed Matter
>> SISSA - Trieste (Italy)
>>
>> Il giorno 17 giu 2020, alle ore 23:31, Coralie Khabbaz
>> <khabbaz.coralie at gmail.com> ha
>> scritto:
>>
>> Hello,
>> I am doing an scf calculation on a tungsten nitride (WN) slab
>> (catalyst), with a methane molecule 5
>> Angstrom away from the surface. The energy values are not converging,
>> even after 533 iterations.
>> The energy values are very negative, and then they increase to a
>> positive value then decrease a lot
>> again. Before building the supercell, I had a WN unit cell with 433
>> K-points. Then, I built the
>> super cell by using a scaling of 333 and miller indices of 100. I
>> tried using K points of 111 and
>> 222 for my super cell, but the calculations didn't converge for both.
>>
>> This is the input file I am using:
>>
>> &CONTROL
>> calculation = "scf"
>> max_seconds = 1.72800e+05
>> pseudo_dir =
>> "/home/coralee/projects/def-jkopysci/coralee/.pseudopot"
>> restart_mode = "from_scratch"
>> /
>>
>> &SYSTEM
>> a = 1.24907e+01
>> angle1(1) = 0.00000e+00
>> angle1(2) = 0.00000e+00
>> angle2(1) = 0.00000e+00
>> angle2(2) = 0.00000e+00
>> b = 1.34803e+01
>> c = 2.52767e+01
>> cosab = 6.12323e-17
>> cosac = 6.12323e-17
>> cosbc = -1.85547e-01
>> degauss = 1.00000e-02
>> ecutrho = 4.50000e+02
>> ecutwfc = 5.00000e+01
>> ibrav = 14
>> nat = 113
>> nspin = 2
>> ntyp = 4
>> occupations = "smearing"
>> smearing = "gaussian"
>> starting_magnetization(1) = 2.00000e-01
>> starting_magnetization(2) = 6.00000e-01
>> starting_magnetization(3) = 0.00000e+00
>> starting_magnetization(4) = 0.00000e+00
>> /
>>
>> &ELECTRONS
>> conv_thr = 1.00000e-06
>> diagonalization = "david"
>> electron_maxstep = 528
>> mixing_beta = 4.00000e-01
>> startingpot = "atomic"
>> startingwfc = "atomic+random"
>> /
>>
>> K_POINTS {automatic}
>> 4 4 2 0 0 0
>>
>> ATOMIC_SPECIES
>> N 14.00674 N.pbe-n-rrkjus_psl.1.0.0.UPF
>> W 183.84000 W.pbe-spn-rrkjus_psl.1.0.0.UPF
>> C 12.01070 C.pbe-n-rrkjus_psl.1.0.0.UPF
>> H 1.00794 H.pbe-rrkjus_psl.1.0.0.UPF
>>
>> ATOMIC_POSITIONS {angstrom}
>> N 2.481849 0.418242 14.837784
>> N 6.645429 0.418242 14.837784
>> N 10.809008 0.418242 14.837784
>> N 2.481849 4.911686 14.837784
>> N 6.645429 4.911686 14.837784
>> N 10.809008 4.911686 14.837784
>> N 2.481849 9.405131 14.837784
>> N 6.645429 9.405131 14.837784
>> N 10.809008 9.405131 14.837784
>> W 2.481849 2.628097 14.748648
>> W 6.645429 2.628097 14.748648
>> W 10.809008 2.628097 14.748648
>> W 2.481849 7.121541 14.748648
>> W 6.645429 7.121541 14.748648
>> W 10.809008 7.121541 14.748648
>> W 2.481849 -1.865347 14.748648
>> W 6.645429 -1.865347 14.748648
>> W 10.809008 -1.865347 14.748648
>> N 0.400059 3.110895 14.468179
>> N 4.563639 3.110895 14.468179
>> N 8.727218 3.110895 14.468179
>> N 0.400059 7.604339 14.468179
>> N 4.563639 7.604339 14.468179
>> N 8.727218 7.604339 14.468179
>> N 0.400059 -1.382549 14.468179
>> N 4.563639 -1.382549 14.468179
>> N 8.727218 -1.382549 14.468179
>> W 0.400059 0.783425 14.415413
>> W 4.563639 0.783425 14.415413
>> W 8.727218 0.783425 14.415413
>> W 0.400059 5.276869 14.415413
>> W 4.563639 5.276869 14.415413
>> W 8.727218 5.276869 14.415413
>> W 0.400059 9.770314 14.415413
>> W 4.563639 9.770314 14.415413
>> W 8.727218 9.770314 14.415413
>> N 2.481849 3.081849 12.630078 0 0 0
>> N 6.645429 3.081849 12.630078 0 0 0
>> N 10.809008 3.081849 12.630078 0 0 0
>> N 2.481849 7.575293 12.630078 0 0 0
>> N 6.645429 7.575293 12.630078 0 0 0
>> N 10.809008 7.575293 12.630078 0 0 0
>> N 2.481849 -1.411595 12.630078 0 0 0
>> N 6.645429 -1.411595 12.630078 0 0 0
>> N 10.809008 -1.411595 12.630078 0 0 0
>> W 2.481849 0.798260 12.540942 0 0 0
>> W 6.645429 0.798260 12.540942 0 0 0
>> W 10.809008 0.798260 12.540942 0 0 0
>> W 2.481849 5.291704 12.540942 0 0 0
>> W 6.645429 5.291704 12.540942 0 0 0
>> W 10.809008 5.291704 12.540942 0 0 0
>> W 2.481849 9.785148 12.540942 0 0 0
>> W 6.645429 9.785148 12.540942 0 0 0
>> W 10.809008 9.785148 12.540942 0 0 0
>> N 0.400059 1.281057 12.260473 0 0 0
>> N 4.563639 1.281057 12.260473 0 0 0
>> N 8.727218 1.281057 12.260473 0 0 0
>> N 0.400059 5.774502 12.260473 0 0 0
>> N 4.563639 5.774502 12.260473 0 0 0
>> N 8.727218 5.774502 12.260473 0 0 0
>> N 0.400059 10.267946 12.260473 0 0 0
>> N 4.563639 10.267946 12.260473 0 0 0
>> N 8.727218 10.267946 12.260473 0 0 0
>> W 0.400059 3.447032 12.207706 0 0 0
>> W 4.563639 3.447032 12.207706 0 0 0
>> W 8.727218 3.447032 12.207706 0 0 0
>> W 0.400059 7.940476 12.207706 0 0 0
>> W 4.563639 7.940476 12.207706 0 0 0
>> W 8.727218 7.940476 12.207706 0 0 0
>> W 0.400059 -1.046412 12.207706 0 0 0
>> W 4.563639 -1.046412 12.207706 0 0 0
>> W 8.727218 -1.046412 12.207706 0 0 0
>> N 2.481849 1.252012 10.422372 0 0 0
>> N 6.645429 1.252012 10.422372 0 0 0
>> N 10.809008 1.252012 10.422372 0 0 0
>> N 2.481849 5.745456 10.422372 0 0 0
>> N 6.645429 5.745456 10.422372 0 0 0
>> N 10.809008 5.745456 10.422372 0 0 0
>> N 2.481849 10.238900 10.422372 0 0 0
>> N 6.645429 10.238900 10.422372 0 0 0
>> N 10.809008 10.238900 10.422372 0 0 0
>> W 2.481849 3.461867 10.333236 0 0 0
>> W 6.645429 3.461867 10.333236 0 0 0
>> W 10.809008 3.461867 10.333236 0 0 0
>> W 2.481849 7.955311 10.333236 0 0 0
>> W 6.645429 7.955311 10.333236 0 0 0
>> W 10.809008 7.955311 10.333236 0 0 0
>> W 2.481849 -1.031577 10.333236 0 0 0
>> W 6.645429 -1.031577 10.333236 0 0 0
>> W 10.809008 -1.031577 10.333236 0 0 0
>> N 0.400059 3.944664 10.052766 0 0 0
>> N 4.563639 3.944664 10.052766 0 0 0
>> N 8.727218 3.944664 10.052766 0 0 0
>> N 0.400059 8.438108 10.052766 0 0 0
>> N 4.563639 8.438108 10.052766 0 0 0
>> N 8.727218 8.438108 10.052766 0 0 0
>> N 0.400059 -0.548780 10.052766 0 0 0
>> N 4.563639 -0.548780 10.052766 0 0 0
>> N 8.727218 -0.548780 10.052766 0 0 0
>> W 0.400059 1.617195 10.000000 0 0 0
>> W 4.563639 1.617195 10.000000 0 0 0
>> W 8.727218 1.617195 10.000000 0 0 0
>> W 0.400059 6.110639 10.000000 0 0 0
>> W 4.563639 6.110639 10.000000 0 0 0
>> W 8.727218 6.110639 10.000000 0 0 0
>> W 0.400059 10.604083 10.000000 0 0 0
>> W 4.563639 10.604083 10.000000 0 0 0
>> W 8.727218 10.604083 10.000000 0 0 0
>> C 10.809008 2.628097 19.750000
>> H 10.177000 1.997000 20.383000
>> H 11.443000 1.997000 19.117000
>> H 10.177000 3.263000 19.117000
>> H 11.443000 3.263000 20.383000
>>
>> _______________________________________________
>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
>> users mailing list users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>
>>
>> _______________________________________________
>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
>> users mailing list users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>>
>>
>>
>
> _______________________________________________
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users
>
--
----------------------------------------------------------------------
Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL
Director, National Centre for Competence in Research NCCR MARVEL, EPFL
http://theossrv1.epfl.ch/Main/Contact http://nccr-marvel.ch/en/project
More information about the users
mailing list