Occupation numbers input parameters
This file contains the occupation numbers for each of the Kohn Sham orbitals. It can be used to enforce certain fractional occupancies, for instance to treat an atom of spherical symmetry.
input.occ for closed shell systems
For spin averaged calculations (nspin=1, as defined in input.dft), the syntax for the file input.occ is the following:
norb (an integer for the number of orbitals) 1 occup(1) 2 occup(2) ... norb occup(norb)
All occupations occup must be between zero and two. Non-integer occupancies are supported and may be written as fractions or floating point numbers.
Example: A spherical carbon atom
4 orbitals 1 2 2s 2 2/3 2px 3 2/3 2py 4 2/3 2pz
Note: The last word on each line is just a comment that will not be parsed by BigDFT.
input.occ for spin polarized systems
For calculations with collinear spin polarization (nspin=2, as defined in input.dft), the syntax for the file input.occ is the following:
nup ndown (two integers for the orbitals per spin channel) 1 occup(1) (these are the spin up orbitals) 2 occup(2) ... nup occup(nup) nup+1 occup(nup+1) (these are the spin down orbitals) nup+2 occup(nup+2) ... nup+ndown occup(nup+ndown)
All occupations occup must be between zero and one. The orbitals are enumerated first according to their spin and then according to the energetic ordering. Non-integer occupancies are supported and may be written as fractions or floating point numbers.
Example: A spherical polarized carbon atom
4 1 up and down orbitals 1 1 2s up 2 2/3 2px up 3 2/3 2py up 4 2/3 2pz up 5 1 2s down
Note: The 2px, 2py and 2pz down orbitals are unoccupied here and can thus be omitted. A calculation for this occupancy will require nspin=2, mpol=2 in input.dft and a polarization of 2 on the atom in posinp.xyz. This should qualitatively allow to reproduce the LSDA atomic reference data of the NIST , with deviations due to the choice of the functional and pseudopotential (ixc=-001007 versus ixc=1, the LDA functionals used for the NIST data and for the HGH pseudos, respectively).