LDIPOL with GW0
Posted: Mon Feb 08, 2016 6:04 pm
Dear all,
I converged a slab calculation using the dipole correction in z direction using LDIPOL and IDIPOL (algo=Exact at EDIFF=1.0E10-8) and then ran the ScGW0 calculations using the input as:
ALGO = SCGW0
ENCUTGW = 150
NOMEGA = 72
ISMEAR = -5
ISPIN = 2
SIGMA = 0.01
KPAR = 5
NPAR = 18
NBANDS = 108
LREAL = False
MAXMEM = 13500
NELM = 4
Next, I introduced in the the same INCAR for the ScGW0 calculations, the two variables for dipole correction, and ran the calculations again:
IDIPOL = 3
LDIPOL = True
And the results that I get are very different. The HOMO level shifts down by more than 1.2 eV which is a huge value, as you can see here http://snag.gy/2BvqC.jpg. [the x axis is the iterations over G and the orange and green denote the cases without and with dipole correction during the ScGW0 method]
Is it the right way for doing ScGW0 calculations for slabs to introduce dipole correction in the ScGW0 part of the calculation or should I not do it ? Any insights are appreciated.
I converged a slab calculation using the dipole correction in z direction using LDIPOL and IDIPOL (algo=Exact at EDIFF=1.0E10-8) and then ran the ScGW0 calculations using the input as:
ALGO = SCGW0
ENCUTGW = 150
NOMEGA = 72
ISMEAR = -5
ISPIN = 2
SIGMA = 0.01
KPAR = 5
NPAR = 18
NBANDS = 108
LREAL = False
MAXMEM = 13500
NELM = 4
Next, I introduced in the the same INCAR for the ScGW0 calculations, the two variables for dipole correction, and ran the calculations again:
IDIPOL = 3
LDIPOL = True
And the results that I get are very different. The HOMO level shifts down by more than 1.2 eV which is a huge value, as you can see here http://snag.gy/2BvqC.jpg. [the x axis is the iterations over G and the orange and green denote the cases without and with dipole correction during the ScGW0 method]
Is it the right way for doing ScGW0 calculations for slabs to introduce dipole correction in the ScGW0 part of the calculation or should I not do it ? Any insights are appreciated.