CCL: Spin contamination increased by annihilation of the first-order contaminant



Dear all,

I was quite surprised when, during stability analysis of my wavefunction and subsequent reoptimization of it in unrestricted space, the spin contamination was not descreased by removal of the first-order contaminant:


  SCF Done:  E(RHF) =  -2277.17663511     A.U. after    1 cycles
            NFock=  1  Conv=0.39D-08     -V/T= 2.0012
 [..]
 The wavefunction has an RHF -> UHF instability.
 The two search directions are symmetric.
 [..]
 SCF Done:  E(UHF) =  -2277.19705082     a.u. after   42 cycles
            Convg  =    0.4967D-05                   164 Fock formations.
              S**2 =  2.1080                  -V/T =  2.0009
 <Sx>= 0.0000 <Sy>= 0.0000 <Sz>= 0.0000 <S**2>= 2.1080 S= 1.0356
 <L.S>= 0.000000000000E+00
 Annihilation of the first spin contaminant:
 S**2 before annihilation     2.1080,   after     8.9661
 

Is there some way to interpret this result?

Looking forward to Your help,
Igors Mihailovs (engineer)
Institute of Solid State Physics
University of Latvia