[Wien] about a warning at case.SCF

[Peter Blaha]

> while I was calculating the total energy of single
> nitrogen atom, what I did completely was based on the
> instruction of Pro.Blaha: FCC, a=b=c=20bohr, and 1k
> point. I used the RK = 8; lmax = 12. but it is still
> not easy to converge. and I found a warning at the
> case.scf which red as:
> :WARN : RESULT OF INTEGRATION:    1.00000; SHOULD BE:
>   2.50000.
> who can tell me what does it mean? and how to improve.

In atoms one has very often degenerate eigenvalues, which are however,
only paritally occupied. Thus it happens often that the occupation
of these formally degenerate states cannot be determined properly by the
tetrahedron method.

Used a broadening method (GAUSS or TEMP, 0.002), which will at the end
occupy all degenerate states partially.

In addition: N should be done spin polarized!! This will occupy the 3
spin-up p orbitals fully, the spin-dn states are empty and I would not
expect any convergence problems.

                                      P.Blaha
--------------------------------------------------------------------------
Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
Phone: +43-1-58801-15671             FAX: +43-1-58801-15698
Email: blaha at theochem.tuwien.ac.at    WWW: http://info.tuwien.ac.at/theochem/
--------------------------------------------------------------------------