The anion¶
Another (difficult) case study example is the
The ground state¶
The ground state of
pam --noarch --get "DFCOEF=DFCOEF.UO6_6m_ground_1pass" --replace dampfactor=0.90 --inp=UO6_6minus.x2c.scf_damp.inp --mol=UO6.mol
(Files to download are UO6_6minus.x2c.scf_damp.inp
and UO6.mol
.)
Then, after we get the DFCOEF.UO6_6m_ground_1pass file, we decrease the damping level to 0.65. One could try also some lower value:
pam --noarch --put "DFCOEF.UO6_6m_ground_1pass=DFCOEF" --get "DFCOEF=DFCOEF.UO6_6m_ground_2pass" --replace dampfactor=0.50 --inp=UO6_6minus.x2c.scf_damp.inp --mol=UO6.mol
In this way we get converge molecular spinors saved in the DFCOEF.UO6_6m_ground_2pass file.
The anion¶
Now to the next state - the
First run cca 500 iterations with the damping factor of 0.65 and with the shift of virtual spinors of +0.65. This goes very slowly. The pass two continues from pass one. Then pass three from previous pass up to pass 4, which has another 500 iteration. Damping factors of these passes were decreasing from 0.65 to 0.25. Finally, in pass 5 the convergence was reached with the default DIIS method:
pam --noarch --put "DFCOEF.UO6_6m_ground_2pass=DFCOEF" --get "DFCOEF=DFCOEF.UO6_5m_2p3_1pass" --replace dampfactor=0.65 --inp=UO6_5m.x2c.scf_2p3_damp.inp --mol=UO6.mol
pam --noarch --put "DFCOEF=DFCOEF.UO6_5m_2p3_1pass=DFCOEF" --get "DFCOEF=DFCOEF.UO6_5m_2p3_2pass" --replace dampfactor=0.55 --inp=UO6_5m.x2c.scf_2p3_damp.inp --mol=UO6.mol
pam --noarch --put "DFCOEF=DFCOEF.UO6_5m_2p3_2pass=DFCOEF" --get "DFCOEF=DFCOEF.UO6_5m_2p3_3pass" --replace dampfactor=0.45 --inp=UO6_5m.x2c.scf_2p3_damp.inp --mol=UO6.mol
pam --noarch --put "DFCOEF=DFCOEF.UO6_5m_2p3_3pass=DFCOEF" --get "DFCOEF=DFCOEF.UO6_5m_2p3_4pass" --replace dampfactor=0.25 --inp=UO6_5m.x2c.scf_2p3_damp.inp --mol=UO6.mol
pam --noarch --put "DFCOEF=DFCOEF.UO6_5m_2p3_4pass=DFCOEF" --get "DFCOEF=DFCOEF.UO6_5m_2p3_diis_5pass" --inp=UO6_5m.x2c.scf_2p3.inp --mol=UO6.mol
(For downloading is the file UO6_5m.x2c.scf_2p3_damp.inp
). Note that user may encounter difficulties when reproducing SCF convergence of this tutorial case.