My Community

I usually calculated the cohesive enrgy by dividing the VASP total energy by number of atoms in bulk supercell.
Is that a right method?
In my bcc-Fe calculation, the lattice constant was 2.74A, approximately same as the experimental value. But the cohesive energy, 7.8 eV/atom is much higher than exp. value, 4.1 eV/atom.
For USPP-LDA and PAW-GGA, these overestimating tendency are commonly appeared.

no, you have to subtract the energies of the free atoms (please have a look in the manual on how to calculate the ground state configuration of free atoms).

E.g.,
E_coh = E_tot^bulk (Fe) - E_tot^free atom(Fe)

E_tot^fee atom is obtained by performing a spin-polarized total energy calculation for a free atom in a large box, e.g., orthorhombic box with 10x11x12 A and one k-point (1/4,1/4,1/4).

Juarez L. F. Da Silva