My Community

You may try adding "LORBIT = 10" in the INCAR file

I presume that you used either POT_PBE Cr or Cr_pv PP's that have ENMAX 227.082 and 265.683 respectively. In both case, your main target should be to find the equilibrium volume against ENCUT, minimizing the Pulay stress tensor by zeoring (or almost) the "external pressure", which can be found in OUTCAR. The straightforward way is to select different lattice constants and do ion relaxation plus cell shape variation by setting ISIF = 4 (fixed vol) through EOS fits at different ENCUT's, after you have obtained a reasonably convergence in K-mesh at about 30% higher cutoff, which is ENCUT=1.3*max (Enmax in POTCAR). Having obtained equilibrium volume, relax again with ISIF =4 with ENCUT = 1.3*max(ENMAX in POTCAR), and again re-relax moving CONTCAR to POSCAR. You should use ISMEAR= 0 (Gaussian integration) with SIGMA=0.01 for any relaxation. Having obtained an equilibrium geometry after "re-relaxation", do a single-point calculation (NSW=0) with ISMEAR = -5 (Tetrahedron integration with Blochl's correction) and without SIGMA to get the total energy. This strategy should work.

However, due to having background in pure first principle theories, I do not myself love EOS fit as it is "emperical". So, what I rather do after I found a convereged K-mesh with 30% higher ENCUT is, I bravely do several full relaxations (ions+cell vol and shape) with congugate gradient search (IBRION = 2) followed by Newton's technique (IBRION = 1) and with different step sizes (POTIM) with 30% higher ENCUT. This has so far given me an excellent cell geometry to start with subsequent volume relaxation (within 1% deviation from eq. vol). Now I choose several ENCUT and relax the volume with carefully chosen IBRION and POTIM, and I analyze convergence in Vol. Having obtained an equilibrium Vol, I relax twice with ISIF = 4, convereged K-mesh and 1.3*max(ENMAX in POTCAR).

If you have time, you could play around a bit following these mad strategies and let me know what you have ended up with if you dared!

Cheers,
Sankh

1) yes, as there are periodic boundary conditions, all direct coordinates have to be in the range between 0 and 1 (in units of a).
2) please note that you have to set ISPIN = 2 if your system is AFM. (and use a supercell that is large enough to allow for AF ordering) If you do not set this parameter, all states are double-occupied.

Dear Coleague
I am new too in comutational material science.
Please, how can I get the structure file " POSCAR" for Cr2O3?????!!!!!.
Thanks a lot.