can LOCPOT be used to derive electric field?

[tholme]

Can the information in LOCPOT be used to reverse engineer the electric field, or does it somehow contain contributions from exchange, correlation, and kinetic energy? Can I simply view the LOCPOT as the scalar potential and use the equation
E=- grad (phi)

Thanks.

[tholme]

As a follow-on question, can we calculate the energy stored in electric fields using either CHG, LOCPOT, or both? Can we use the equation for energy density:
w=1/2 rho(x) * phi(x)
where rho is the charge density from the file CHG and phi is the potential from LOCPOT and integrate over all space (e.g. Jackson, Classical Electrodynamics, p.41, eq. 1.53)?
Alternately, can we compute the gradient of the potential and use the equation
w=epsilon_0/2 |grad phi|^2
and integrate over the supercell (e.g. Jackson p.41 eq. 1.54)?

Thanks for your input.

[tlchan]

You can check the VASP manual about the LOCPOT here. By default, it should contain the Hartree potential and the local part of the pseudopotential. You can take derivative if you like, if such quantity correlates to what you're studying.