I've got an interesting issue when checking how does the k-point grid affect the DOS of Lu2O3 (primitive cell).
Namely, I've performed several almost identical calculations which differed in k-point grid only.
The grids were:
2x2x2, 3x3x3, 4x4x4, 5x5x5, 6x6x6, 7x7x7 and 8x8x8.
Different Fermi energies were obtained:
0.224, 0.219, 0.223, 0.288, 0.291, 0.288, 0.223, respectively.
The values in bold are non-systematically larger, which resulted in significant shift of DOS plot (obviously) in respect to the other calculations. The shape of DOS plot was rather the same and bandgaps were almost identical.
My question is, is that normal? Or is that a bug?
Hi,
I've got an interesting issue when checking how does the k-point grid affect the DOS of Lu2O3 (primitive cell).
Namely, I've performed several almost identical calculations which differed in k-point grid only.
The grids were:
2x2x2, 3x3x3, 4x4x4, 5x5x5, 6x6x6, 7x7x7 and 8x8x8.
Different Fermi energies were obtained:
0.224, 0.219, 0.223, 0.288, 0.291, 0.288, 0.223, respectively.
The values in bold are non-systematically larger, which resulted in significant shift of DOS plot (obviously) in respect to the other calculations. The shape of DOS plot was rather the same and bandgaps were almost identical.
My question is, is that normal? Or is that a bug?
Last edit: Andrew Shyichuk 2016-11-16
Dear Andrew,
nothing to worry, I sometimes find the same for I guess the reason is the numerical integration in combination with fine numerical details.
best
Michael
Thanks.