about dipole moment and electronic polarization

[zj]

Dear all,

It seems that vasp can get the polarization by two ways: one is berry phase, the other is calculating the dipole moment of a cell.

I always can not understand the Dipole tag well. It correct the energy by considering the monopole/dipole and quadrupole. In the OUTCAR file, it gives the dipole moment. However, its value does not seem correct.

For the berry phase, an error always occurs in the OUTCAR:

Error in subroutine BERRY: did not find all determinants

Matrix CMK is not an nxn matrix for
ISTR = 1 j = 0

Sometimes, modifying the k points can avoid this error.

Does anyone encounter this error?

Thank you in advance.

Best wishes

[admin]

the main reason for that error is that the system is not insulating (ie the number of occupied bands is not the same for all k-points), which is a prerequisite requirement for the calculation of the Berry phases (as pointed out in the online manual)
the reasons for that mainly may be:
-- the electronic structure of cell you calculate REALLY has no gap at the
Fermi level
-- the system is not at its equilibrium geometry (high pressure)

[zj]

Thank you.

[xianghjun]

I also encounter the same error when I calculate the
polarization of the well-known semiconducting GaN:
Error in subroutine BERRY: did not find all determinants

Matrix CMK is not an nxn matrix for
ISTR = 1 j = 0

What is the problem and how to solve it?
Thanks a lot.