[CPMD-list] More on HCTH

[Erik Santiso]

Hi again. I've fixed the problem I mentioned before in the HCTH routine, and now it is converging to a lower energy in the LSD case (as it should). However, the energy still seems incorrect. I have another question regarding the LSD_HCTH routine in lsd_func.f. When finding the opposite- and parallel- spin contributions to the correlation energy density to be used in the HCTH equations, the program is finding ecab from:

ecab = ec(rho,0)-ec(rhoa,1)-ec(rhob,1)

This comes from equation (7) of JCP 107, 8554. Now, this is not the exact equation in the paper. For a non-LSD calculation it would be right because the spin polarization of the total system is zero. However, for the LSD case I think the first term should be obtained using equation (8) of the PW paper (PRB 45, 13244) instead of being ec(rho,0). Is this right?

Thanks again,

Erik.

-----------------------------------------
Instant psychoanalysis - Just add coffee.