CCL: MOPAC and metals

[David Gallagher <gallagher.da_-_gmail.com>]

 Sent to CCL by: David Gallagher [gallagher.da],[gmail.com]
 Hi Jose,
What you are asking is certainly possible when
 optimizing with molecular mechanics methods,
 however, with quantum methods such as MOPAC, the
 only things you can really control are the
 starting positions of the atoms, the number of
 electrons (net charge), and the multiplicity
 (although, it is possible to lock atom positions,
 bond lengths, angles and dihedrals).  Even the
 bonds you see on the screen are ignored by a
 quantum mechanics program.  Once the optimization
 is complete, MOPAC decides where the electrons
 should go and so effectively controls the
 bonding, partial charges and oxidation
 states.  Do not be misled by some graphical
 representations of the bonding, but look at the
 final atom distances, calculated bond orders,
 partial changes, and molecular orbitals, etc.
Regarding adding counter ions, this is usually
 fine.  Also, unless you are simulating gas phase,
 you should use the appropriate solvent field
 (e.g. COSMO) or explicit solvent molecules,
 particularly when there are charged or polar species present.
 David Gallagher
 CAChe Research.com
At 04:34 AM 2/2/2011, Jose Goncalo Deira Duarte
 de Campos Justino jgcj|fct.unl.pt wrote:
> Sent to CCL by: "Jose Goncalo Deira Duarte de
>  Campos Justino" [jgcj:_:fct.unl.pt]
>  Dear all,
>  I've been thinking about using mopac2009 to model some iron-porphyrin
>  systems. The aim so far is to check how the software behaves. I'm quite
>  familiar with mopac by now, but the metal atom has been triggering some
>  questions.
>  Formally, I have an iron (IV) atom complexed equatorially by the 4 N atoms
>  > from porphyrin (which has a radical cation delocalized over the aromatic
>  part and two anionic carboxylate groups) and, apically, by a HS-
>  (thiolate) group and an O atom (formal charge -2). This should be neutral.
>  (for those of you who might recognize, it's derived from the P450 compound
>  I active site).
>  Playing with mopac and mozyme and related keywords from the manual, I
>  never get a Fe(IV) state but a Fe(II) state. Is there anyway to set  this
>  in the input file?
>  My best try so far got me a total charge of -3 with very good geometries,
>  so i threw in 3 K+ atoms, and the geometry remained excellent. What do you
>  feel about the validity of this approach ?
>  Paralell, is there any particular way to specify dative bonds in mopac
>  such as those of Fe to the porphyrin N atoms ? Is SETPI the best option
>  for these, or totally non-sense ?
>  Any input (any at all, really) is welcome!
>  All the best,
>  Gonçalo