Elastic constants of Cr

Hello everyone.

I followed the tutorial about elastic constants and I wanted to see if I could successfully apply that to another element e.g. Cr. I’ve made convergence studies (ecut,kpoints) and plotted the corresponding band diagram (which is in very good accordance with what can be found on materials project https://materialsproject.org/materials/mp-90/)

However, when I try to compute the elastic constants, using the converged values obtained during my previous runs (that you can find in Cerla.in attached), the results I obtain are completely off ! Indeed, as you can see in the output file attached, I have C_11= GPa, C_12= GPa, C_13= GPa and C_44= GPa while the correct values are C_11=499 GPa, C_12=139 GPa, C_13=139 and C_14=102 GPa (Cfr. materials project). For your information, I used a NC SR pseudopotential (PBE) found on pseudo dojo (http://www.pseudo-dojo.org/) to be consistent with my input file (ixc 11)

How can I solve this issue ?

Thanks for your help !
Crela.out (617 KB)
Crela.in (3.27 KB)

Have you used afterwards? https://docs.abinit.org/guide/mrgddb/ and https://docs.abinit.org/guide/anaddb/ ?

I don’t see numbers in your first list : where are you getting the c11 etc from? As mmazdz says, you can run anaddb to get the full elastic tensor in useful units. In your very specific case you don’t need the phonons as there is just 1 atom and no strain coupling. To be safe for any thing else, run with phonons (rfphon 1 natom) and all directions (rfdir 1 1 1) to get the relaxed ion C_ij

You should do things one step at a time:

  • first k point convergence (what you have with 10x10x10 is not enough for a metal)
  • your ecut of 100 is huge - check if you can not use something more reasonable (pseudo dojo typically converges at 35 Ha, at worst 45 or 50)
  • then relax the cell
  • occopt 3 is not cold smearing, it’s Fermi Dirac, so your calculation is actually super hot Cr electrons
  • why are you running the unoccupied wave functions calculation? For DFPT it should not be necessary.

Also check the magnetic state on the MP, it may be AFM or other (the ground state of Cr is complex)