<div dir="ltr"><div><div><div><div><div><div><div><div><div>Dear QE Users,<br><br></div>I am trying to calculate the Hubbard U for bulk Fe2O3 using the linear response method of Cococcioni et al. <br></div><br>The problem I am having as of now is that some of the responses dn0/da and dn/da are not linear and do not intersect at 0.<br>
<br></div>I am following the approach suggested by Dr. Cococcioni in earlier posts several people.<br></div>1. SCF calculation<br></div>2. Grep the ethr from previous SCF and run the alpha perturbations for Fe and O.<br></div>
3. Calculate U by inverse of chi0 and chi matrices.<br><br></div>I am using QE version 5.0.2 and below are the options of the input criterion for the scf run:<br><br> &control<br> calculation = 'scf',<br> restart_mode='from_scratch',<br>
prefix = 'Fe2O3',<br> pseudo_dir='$ESPRESSO_PSEUDO',<br> outdir='$temp',<br> etot_conv_thr = 1.0D-6,<br> forc_conv_thr = 1.0D-5, <br>&system<br> ibrav = 0,<br> nat = 30,<br> ntyp = 4,<br>
ecutwfc = 35,<br> ecutrho = 280,<br> occupations = 'smearing',<br> smearing = 'mp',<br> degauss = 0.02,<br> nspin = 2,<br> starting_magnetization(1) = 0.5,<br> starting_magnetization(2) = -0.5,<br>
starting_magnetization(3) = 0.5,<br> lda_plus_u = .true.,<br> U_projection_type = 'atomic',<br> Hubbard_U(1) = 1.d-10,<br> Hubbard_U(2) = 1.d-10,<br> Hubbard_U(3) = 1.d-10,<br> Hubbard_U(4) = 1.d-10,<br>
/<br> &electrons<br> mixing_beta = 0.5,<br> conv_thr = 1.0d-8,<br>/<br>CELL_PARAMETERS {bohr}<br>9.531782 0.000000 0.000000<br>-4.765891 8.254766 0.000000<br>0.000000 0.000000 26.259456<br>ATOMIC_SPECIES<br> FE1 1.0 Fe.pbe-nd-rrkjus.UPF<br>
FE2 1.0 Fe.pbe-nd-rrkjus.UPF<br> FE3 1.0 Fe.pbe-nd-rrkjus.UPF<br> O 1.0 O.pbe-rrkjus.UPF<br>ATOMIC_POSITIONS {crystal}<br> FE3 0.000000000 0.000000000 0.000000000<br> FE2 -0.333333343 -0.666666687 0.119881678<br>
.....<br></div><div>.....other atoms' coordinates<br></div><div>.....<br>K_POINTS {automatic}<br> 6 6 6 1 1 1<br><br><br></div>This run is followed by the alpha perturbations. I am using alpha values ranging from +- 0.01 to +- 0.05 including 0.0 at intervals of 0.01.<br>
</div><div>Some responses are linear but most of them are not linear. I can attach the figures if needed!<br><br></div><div>Any suggestions are greatly appreciated!<br><br></div><div>Thank you,<br></div><div>Sai<br><br><br>
------<br><br></div><div>Sai Kumar Ramadugu,<br></div><div>Post-Doctoral Scholar,<br>University of Iowa.<br></div><br><div><div><br></div></div></div>