slab LOCPOT figure sees slope in both bulk and vacuum region
Posted: Thu Oct 02, 2014 5:30 am
Hello.
I am doing some calculations on a 5 layer (more than 30 Å) CH3NH3PbI3 (cubic unit cell) 5 layer (the same distance as the bulk region) vacuum slab system. KPOINTS is at 7 x 7 x 1. LVHAR = T. The bulk calculation has been relaxed, and the three I atoms are not on the exact edge centers. For the slab there is no added passivating atom, nor extra relaxation on the surface. I did not switch on dipole related tags.
It turns out the plane averaged curve has slopes in both regions. The slope in vacuum seems very smooth. I added two other I atoms on one surface to make the two surfaces more "complete" and resembling each other, and the figure still shows two regions with slopes, but slopes of different rates; the slopes are steeper. I also played around with LREAL, recommended lowering of AMIX, and elongating the c parameter, but nothing changes this behavior.
I had this scenario with BN before, when I misplaced one N to a B location around the surface area. So my tentative conclusion is that the potential is heavily dependent on the surface condition, but this statement is not physical. In the bulk region, or in vacuum, should I expect the (macroscopic average) electrostatic potential to converge? Or, what kind of surface effect could have lead to such a potential pattern? Charges?
Thank you.
I am doing some calculations on a 5 layer (more than 30 Å) CH3NH3PbI3 (cubic unit cell) 5 layer (the same distance as the bulk region) vacuum slab system. KPOINTS is at 7 x 7 x 1. LVHAR = T. The bulk calculation has been relaxed, and the three I atoms are not on the exact edge centers. For the slab there is no added passivating atom, nor extra relaxation on the surface. I did not switch on dipole related tags.
It turns out the plane averaged curve has slopes in both regions. The slope in vacuum seems very smooth. I added two other I atoms on one surface to make the two surfaces more "complete" and resembling each other, and the figure still shows two regions with slopes, but slopes of different rates; the slopes are steeper. I also played around with LREAL, recommended lowering of AMIX, and elongating the c parameter, but nothing changes this behavior.
I had this scenario with BN before, when I misplaced one N to a B location around the surface area. So my tentative conclusion is that the potential is heavily dependent on the surface condition, but this statement is not physical. In the bulk region, or in vacuum, should I expect the (macroscopic average) electrostatic potential to converge? Or, what kind of surface effect could have lead to such a potential pattern? Charges?
Thank you.