Dear All,<br><br>i`m performing nscf calculations with quantum espresso.<br>these runs are given to Wannier90 to produce Wannier functions.<br>Here the input file (used after the scf run):<br><br> &control<br> calculation = 'nscf'<br>
! restart_mode = 'from_scratch'<br> prefix = 'fese'<br> tprnfor = .true.<br> pseudo_dir = './'<br> outdir = './'<br> iprint = 2<br>
wf_collect = .true.<br>/<br> &system<br> ibrav = 8<br> celldm(1) = 10.0733642480265<br> celldm(2) = 1.00000000000000<br> celldm(3) = 1.0291721306383<br> nat = 8<br>
ntyp = 5<br> ecutwfc = 25.0<br> ecutrho = 300.0<br> nbnd = 60<br> occupations = 'smearing'<br> smearing = 'methfessel-paxton'<br> degauss = 0.01<br>
nspin = 2<br> starting_magnetization(1) = 1.0<br> starting_magnetization(2) = 1.0<br> starting_magnetization(3) = -1.0<br> starting_magnetization(4) = -1.0<br> starting_magnetization(5) = 0.0<br> lda_plus_u =.true.<br>
Hubbard_U(1) = 1.d-8<br> Hubbard_U(2) = 1.d-8<br> Hubbard_U(3) = 1.d-8<br> Hubbard_U(4) = 1.d-8<br> Hubbard_alpha(1) = 1.d-8<br> Hubbard_alpha(2) = 1.d-8<br> Hubbard_alpha(3) = 1.d-8<br>
Hubbard_alpha(4) = 1.d-8<br> nosym=.true.<br>! noinv=.true.<br>/<br> &electrons<br> conv_thr = 1.0d-8<br> diagonalization = 'cg'<br>/<br>ATOMIC_SPECIES<br> [.....]<br>ATOMIC_POSITIONS { angstrom }<br>
[......]<br>K_POINTS crystal<br>1152<br> 0.00000000 0.00000000 0.00000000 8.680556e-04<br> 0.00000000 0.00000000 0.12500000 8.680556e-04<br> 0.00000000 0.00000000 0.25000000 8.680556e-04<br> 0.00000000 0.00000000 0.37500000 8.680556e-04<br>
0.00000000 0.00000000 0.50000000 8.680556e-04<br> 0.00000000 0.00000000 0.62500000 8.680556e-04<br> 0.00000000 0.00000000 0.75000000 8.680556e-04<br> 0.00000000 0.00000000 0.87500000 8.680556e-04<br>[....]<br>
<br><br>first problem.<br>in output file i get double number of points:<br><br>[....]<br> k( 2298) = ( 0.9166667 0.9166667 0.1214568), wk = 0.0008681<br> k( 2299) = ( 0.9166667 0.9166667 0.2429137), wk = 0.0008681<br>
k( 2300) = ( 0.9166667 0.9166667 0.3643705), wk = 0.0008681<br> k( 2301) = ( 0.9166667 0.9166667 0.4858274), wk = 0.0008681<br> k( 2302) = ( 0.9166667 0.9166667 0.6072842), wk = 0.0008681<br>
k( 2303) = ( 0.9166667 0.9166667 0.7287411), wk = 0.0008681<br> k( 2304) = ( 0.9166667 0.9166667 0.8501979), wk = 0.0008681<br><br>is this becasue spin polarized are treated with separated in k-points?<br>
<br>Once the QE nscf run is finished i start with Wannier90 but i get troubles as i should have 1152 k-points.<br><br>The file.pw2wan in Wannier90 looks like:<br><br>&inputpp <br> outdir = './'<br> prefix = 'fese'<br>
seedname = 'FeSe'<br> spin_component = 'up'<br> write_mmn = .true.<br> write_amn = .true.<br> write_unk = .true.<br>/<br><br>The file.win in Wannier90 looks like:<br><br>num_wann = 20<br>
num_iter = 500<br><br>dis_num_iter = 500<br><br>iprint = 3<br><br>!here we exclude all bands except the O2p bands<br>exclude_bands : 1-32,53-60<br><br>write_r2mn=.true.<br>hr_plot=.true.<br><br>kmesh_tol=0.00001<br>
<br>begin atoms_cart<br>ang<br>[.....]<br>[.....]<br>end atoms_cart<br><br>begin kpoint_path<br>[......]<br>end kpoint_path<br>bands_plot =T<br><br>!fermi_surface_plot = .true.<br><br>!To plot the WF<br>!wannier_plot = T<br>
!wannier_plot_supercell = 2<br>!wannier_plot_list = 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20<br>!restart = plot<br><br>guiding_centres = T<br><br>begin projections<br>Fe:dxy;dxz;dyz;dx2-y2;dz2<br>end projections<br>
<br>begin unit_cell_cart<br>bohr<br>10.0733642480265 0.0000 0.0000<br>0.0000 10.0733642480265 0.0000<br>0.0000 0.0000 10.3672257458371<br>end unit_cell_cart<br><br>mp_grid : 12 12 8<br><br>begin kpoints<br>0.00000000 0.00000000 0.00000000<br>
0.00000000 0.00000000 0.12500000<br>0.00000000 0.00000000 0.25000000<br>0.00000000 0.00000000 0.37500000<br>0.00000000 0.00000000 0.50000000<br>0.00000000 0.00000000 0.62500000<br>0.00000000 0.00000000 0.75000000<br>[.... here there are the same 1152 k-point of the nscf run]<br>
<br>Can you help to find out the problem?<br>If i do unpolarized spin calculations all is working fine.<br><br>thank you.<br><br>cheers<br><br>Gianluca<br><br><br><br><br>