IDIPOL=3 and slab calculations
Posted: Fri Jul 23, 2010 2:51 pm
I am running some calculations with a charged adsorbate above a slab, and I am wondering about how to use the DIPOL flags correctly (...or if I need to). The slab is in the a-b direction, so I think that IDIPOL=3 is the flag that I need. The part that I am not sure about is how to set the DIPOL=X X X flag. The center of mass of the system is positioned at 0.5 0.5 0.5.
I have tried both setting DIPOL to be the COM of the adsorbate, and to be the COM of the whole system of slab+adsorbate. In both cases there is a very big difference between corrected and uncorrected energies ( >20 eV with a vacuum slab of 20 \AA ); looking at the OUTCAR file the trace of the quadrupole is very large: ~900, and this appears to be the reason for the big correction. Since my system is charged, I assume the DIPOL corrections are necessary. However, the un-corrected energies in fact decrease linearly with the inverse of the vacuum slab (over a range of 10-45 \AA), so is it unreasonable to simply extrapolate these values to 1/(vacuum slab) --> 0? Presumably the corrected and uncorrected values ultimately converge to the same number (if DIPOL is set correctly), so if the uncorrected value is converged it is OK to use this number for the total energy?
I have tried both setting DIPOL to be the COM of the adsorbate, and to be the COM of the whole system of slab+adsorbate. In both cases there is a very big difference between corrected and uncorrected energies ( >20 eV with a vacuum slab of 20 \AA ); looking at the OUTCAR file the trace of the quadrupole is very large: ~900, and this appears to be the reason for the big correction. Since my system is charged, I assume the DIPOL corrections are necessary. However, the un-corrected energies in fact decrease linearly with the inverse of the vacuum slab (over a range of 10-45 \AA), so is it unreasonable to simply extrapolate these values to 1/(vacuum slab) --> 0? Presumably the corrected and uncorrected values ultimately converge to the same number (if DIPOL is set correctly), so if the uncorrected value is converged it is OK to use this number for the total energy?