<p>Dear all,</p>
<p>I am studying
the chemisorption of molecules on iron. I would like the total spin of the
system to change, considering there are bonds forming/braking during the
dynamics, thus I switched on the SMEAR keyword. The SCF converges but the final
energy is higher than the very first step… and the wave function seems close
shell.. If I switch off the SMEARING, keeping the MAGNETIZATION fix to a value,
there are no problems. Can anyone help me? Hereafter the SCF part of
the input. </p>
<p>Thanks</p>
<p>Michele</p>
<p>&SCF</p>
<p> MAX_SCF
200</p>
<p> EPS_SCF
1.0E-6</p>
<p>
EPS_EIGVAL 1.0E-5</p>
<p> ADDED_MOS
200 200</p>
<p>
&DIAGONALIZATION T</p>
<p>
ALGORITHM STANDARD</p>
<p>
EPS_ADAPT 0.01</p>
<p> &END
DIAGONALIZATION</p>
<p>
&MIXING T</p>
<p>
METHOD KERKER_MIXING</p>
<p>
ALPHA 0.01 </p>
<p>
BETA 0.5</p>
<p>
NBUFFER 8 </p>
<p> &END
MIXING</p>
<p>
&SMEAR</p>
<p>
METHOD FERMI_DIRAC</p>
<p>
ELECTRONIC_TEMPERATURE 2300 </p>
<p>
FIXED_MAGNETIC_MOMENT -1</p>
<p> &END
SMEAR</p>
<p>
&OUTER_SCF</p>
<p>
EPS_SCF 1.0E-6</p>
<p>
MAX_SCF 20</p>
<p> &END
OUTER_SCF</p>
<p> &END SCF</p>
<p> </p>
<p> </p>