selfconsistent GW
Posted: Wed May 28, 2014 8:11 pm
Hi,
I was wondering that when we set ALGO = scGW in vasp, then is the quasiparticle energy and wavefunctions updated only, or the V_xc is also updated in order to match the quasiparticle energies from G to those from G_0. It seems to me that the V_xc is updated after every iteration. If so, is it similar to the approach by Schilfgaarde i.e. is the V_xc set to V_xc = sum_ij (psi_nk) Re {Sigma (E_i) + Sigma (E_j)} (psi_nk) in order to minimize the difference of the dressed QP and bare QP ?
In my calculations, V_xc is given by:
V_xc (KS) = KS-energies - (Expect(T + V_ion + V_H ) + Expect(T+V_H+V_ion)^1 + Expect(V_x)^1). I am not sure I understand this. I would really appreciate it if you could kindly help me.
Thanks and Regards.
I was wondering that when we set ALGO = scGW in vasp, then is the quasiparticle energy and wavefunctions updated only, or the V_xc is also updated in order to match the quasiparticle energies from G to those from G_0. It seems to me that the V_xc is updated after every iteration. If so, is it similar to the approach by Schilfgaarde i.e. is the V_xc set to V_xc = sum_ij (psi_nk) Re {Sigma (E_i) + Sigma (E_j)} (psi_nk) in order to minimize the difference of the dressed QP and bare QP ?
In my calculations, V_xc is given by:
V_xc (KS) = KS-energies - (Expect(T + V_ion + V_H ) + Expect(T+V_H+V_ion)^1 + Expect(V_x)^1). I am not sure I understand this. I would really appreciate it if you could kindly help me.
Thanks and Regards.