Gaussian98 "alter" keyword

[Elena Molteni <molteni # - at - # unisi.it>]

 Dear all,
I need to perform a single point energy calculation on a system
 containing
 Fe(II) in high spin state (S=2) coordinated to two oxygens of
 one  carboxylate and 4 nitrogens belonging to 4 histidine residues.
 I had really hard times in obtaining convergence both using DFT, UHF
 methods. I use Gaussian 98.
 I tried the followings:
reducing the system to the miminum number of atoms (cutting the
 ligands and
 saturating valences with hydrogens).
 My aim was to use this output as a starting checkpoint file for the
 extensive calculation.
 Unfortunately the results I obtain even with the small system are
 hardly
 believable. The frontier orbitals do not resemble d orbitals very much
 and
 the spin density is not concentrated on iron (only 1/4 of it is on Fe).
 Moreover cutting the histidines is not very beautiful because i loose
 the
 aromaticity and this would change electronic environment a lot.
 I also tried the QC method to allow the big system to converge but the
 energy continues to jump a lot.
 I was told that i could take the output of a calculation converged with
 the
 wrong unpaired electron population and artificially put the population
 of
 the level with the unpaired electron into one level which is basically
 formed by d orbitals.
 That is artificially put unpaired electrons onto d orbitals.
 If I understood well this option is called alter.
Does anybody know how to solve this problem and if this is the way i
 can
 perform my calculation?
 thanks a lot
 Elena Molteni
 e-mail:molteni # - at - # unisi.it