[Wien] GHOST BANDS error
fabien.tran at vasp.at
fabien.tran at vasp.at
Thu Jul 13 11:17:02 CEST 2023
Hi,
As mentioned by Peter Blaha, this is not possible to do a calculation
with a hybrid functional using XC_HYB_LDA_*, XC_HYB_GGA_* or
XC_HYB_MGGA_* from Libxc. Actually, without entering into detail, this
is not possible to use Libxc at all for a hybrid calculation (the
results would be wrong).
The only hybrid functionals that can be used are listed in the user's
guide under "The available functionals" in section "Unscreened and
screened hybrid functionals".
On 08.07.2023 13:39, Peter Blaha wrote:
> As was mentioned before, you can't use the XC_HYB_GGA_XC_* switches.
>
> In addition: The manual says for NBAND: use "at least" NE+1. Then
> it should run through.
>
> But this does not mean that such a calculation is good. You should use
> more bands for the second variation HF step, depending on your case.
>
> You have above EF 5 empty Ti-d bands and a Ba-s band. So the minimum
> for NBAND is NE+6, larger values are better.
> Am 08.07.2023 um 09:06 schrieb Natalia Andreeva:
>
>> Dear Professor Blaha,
>>
>> I use wien21.
>> I calculate the BaTiO3 structure, atom 1 is Ba, RMT(Ba) = 2.44;
>> RMT(Ti) = 1.72; RMT(O) = 1.55.
>> First I run the SCF calculation on WC, it converges successfully.
>> Results of scf calculation for WC potential:
>> Insulator, EF-inconsistency corrected
>> :GAP (global) : 0.144257 Ry = 1.963 eV (accurate value if
>> proper k-mesh)
>> Bandranges (emin - emax) and occupancy:
>> :BAN00010: 10 -0.230168 -0.162160 2.00000000
>> :BAN00011: 11 -0.225681 -0.146894 2.00000000
>> :BAN00012: 12 0.215762 0.365039 2.00000000
>> :BAN00013: 13 0.248128 0.370283 2.00000000
>> :BAN00014: 14 0.266208 0.370564 2.00000000
>> :BAN00015: 15 0.318315 0.438250 2.00000000
>> :BAN00016: 16 0.347963 0.440022 2.00000000
>> :BAN00017: 17 0.363037 0.457065 2.00000000
>> :BAN00018: 18 0.427422 0.529925 2.00000000
>> :BAN00019: 19 0.461775 0.531282 2.00000000
>> :BAN00020: 20 0.472709 0.539353 2.00000000
>> :BAN00021: 21 0.683610 0.849512 0.00000000
>> :BAN00022: 22 0.701665 0.850244 0.00000000
>> :BAN00023: 23 0.702723 0.850771 0.00000000
>> :BAN00024: 24 0.852848 1.043061 0.00000000
>> :BAN00025: 25 0.862116 1.102879 0.00000000
>> Energy to separate low and high energystates: 0.16576
>> :NOE : NUMBER OF ELECTRONS = 40.000
>> :FER : F E R M I - ENERGY(TETRAH.M.)= 0.5393529588
>> :GMA : POTENTIAL AND CHARGE CUT-OFF 12.00 Ry**.5
>> :POS001: ATOM -1 X,Y,Z = 0.00000 0.00000 0.00000 MULT= 1 ZZ=
>> 56.000 Ba
>> lmax 10
>>
>> This is a hybrid calculation, I use XC_HYB_GGA_XC_B1WC in case.in0,
>> after which I run init_hf_lapw, select nband = 21 (since after
>> calculating on WC nb_occ = 20). When I delete a line for l = 1 from
>> in1, an error still appears.
>>
>> Thank you in advance,
>> With Best Regard,
>> Natalia Andreeva
>>
>> On Fri, Jul 7, 2023 at 6:15 PM Peter Blaha
>> <peter.blaha at tuwien.ac.at> wrote:
>>
>> I need some more information:
>>
>> Are you using wien23 or an earlier version ?
>>
>> What is your system (struct file), in particular what is atom 1; its
>> RMT, ...
>>
>> Such a large EF is rather unusual (except for 5f systems) and for
>> sure the resulting E-parameters as you showed below must lead to
>> ghostbands. Usually it is good to remove the LO for this atom and
>> l-value, at least temporarily, but you must check at what energy
>> would you expect the semicore state and you must not miss it, if it
>> is a real low energy state. ....
>>
>> Is this actually a hybrid calculation ? Does it happen in the first
>> iteration ? Did you start a hybrid calc. from a converged PBE
>> calculation ? ....
>>
>> Am 07.07.2023 um 13:48 schrieb Natalia Andreeva:
>>
>> Dear Professor Blaha,
>>
>> I tried to run calculations for hybrid functionality from libxc, but
>> I get a GHOST BANDS error. The case.scf2 file states that for atom 1
>> with l=1, the energy parameters need to be changed.
>>
>> :WARN : QTL-B value eq. 32.08 in Band of energy 1.68736 ATOM= 1 L= 1
>> :WARN : You should change the E-parameter for this atom and L-value
>> in case.in1 (or try the -in1new switch)
>> The Fermi level in scf2 is
>> :FER : F E R M I - ENERGY(TETRAH.M.)= 1.8321578843
>> I tried setting it to 1.632 instead of the default 0.3 in case.in1,
>> but that didn't fix it.
>> Then I tried to remove the values that have similar energy values in
>> case.scf1 from the case.in1 file (reducing the number of n choices),
>> but this did not lead to a positive result.
>> :E1_0001: E( 1)= 0.7000
>> APW+lo
>> :E1_0001: E( 1)= 0.4094 E(BOTTOM)= 0.409 E(TOP)= -200.000 3 -1 222
>> LOCAL ORBITAL
>> When trying to change the energy parameters and set the in1new flag,
>> an nband error appeared.
>>
>> Thank you in advance,
>>
>> Best Regards,
>> Natalia Andreeva.
>>
>> _______________________________________________
>> Wien mailing list
>> Wien at zeus.theochem.tuwien.ac.at
>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>> SEARCH the MAILING-LIST at:
>>
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>>
>> --
>>
> -----------------------------------------------------------------------
>> Peter Blaha, Inst. f. Materials Chemistry, TU Vienna, A-1060 Vienna
>> Phone: +43-158801165300
>> Email: peter.blaha at tuwien.ac.at
>> WWW: http://www.imc.tuwien.ac.at WIEN2k: http://www.wien2k.at
>>
> -------------------------------------------------------------------------
>> _______________________________________________
>> Wien mailing list
>> Wien at zeus.theochem.tuwien.ac.at
>> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
>> SEARCH the MAILING-LIST at:
>>
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
> --
> Best Regards,
> Natalia Andreeva.
>
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
>
> --
> -----------------------------------------------------------------------
> Peter Blaha, Inst. f. Materials Chemistry, TU Vienna, A-1060 Vienna
> Phone: +43-158801165300
> Email: peter.blaha at tuwien.ac.at
> WWW: http://www.imc.tuwien.ac.at WIEN2k: http://www.wien2k.at
> -------------------------------------------------------------------------
> _______________________________________________
> Wien mailing list
> Wien at zeus.theochem.tuwien.ac.at
> http://zeus.theochem.tuwien.ac.at/mailman/listinfo/wien
> SEARCH the MAILING-LIST at:
> http://www.mail-archive.com/wien@zeus.theochem.tuwien.ac.at/index.html
More information about the Wien
mailing list