Re: CCL:RE2:G98: ONIOM and Amber on TM systems
Valentine Ananikov,
Indeed, at the moment it is not very easy to study TM using ONIOM(xx:MM)
and ONIOM(xx:AM1/PM3). These two cases are a bit different, though. For
the semiempirical methods, the high level region of the system must be
well-defined with the low level method, and unavailable parameters for the
TM pose a real problem. For ONIOM(xx:MM) combinations, there is a
work-around. Of course, ideally one would be able to specify additional
parameters in the input deck, but this is currently not possible. As
mentioned in an earlier posting, MM bonded terms in the high level region
that are further than 3 bonds away from the low level (MM) region cancel.
This is simply the result of these terms being exactly the same in the MM
model calculation and in the MM real system calculation, but of opposite
signs in the ONIOM energy expression. The solution is to disconnect all
the atoms in the model region that are further away then three bond from
the MM region (whatever the parameters for those bonded terms are doesn't
matter, so you might as well not include those bonded terms anyway). No
bonds will be defined, and G98 will not protest that there are no
parameters for those bonds (and associated angles). In order to do this,
use the "Geom=Connectivity" keywords. Of course, atoms closer than
three
bonds to the MM region must not be disconnected. This solves the problem
with the bonded terms correctly, but the Vanderwaals parameters for the TM
are still missing. If the TM is quite far away from the MM region (which
it will be, since it has to be at least three bonds anyway), the exact
values of the VDW parameters hardly affect the results. Again, the MM VDW
interactions within the high level region will cancel, and the VDW
parameters of the TM are only used to describe the interaction of the high
level layer with the low level layer. Just use something like
"Fe-H-charge" as amber atomtype in G98, and effectively the VDW
parameters
of H will be used for Fe.
Thom
BTW. the 1996 ONIOM paper used a different implementation, where there was
more flexibility to adjust MM parameters.
______________________________________________
Dr. Thom Vreven
Chemistry Department
1515 Pierce Drive, Atlanta, Georgia 30322, USA
Phone (404) 727-2381 Fax (404) 727-7412
E-mail thom %! at !% euch4e.chem.emory.edu
______________________________________________
On Sun, 23 Jul 2000, Valentine Ananikov wrote:
>
>
>
> Hi Stefan and CCl members!
>
> Your description seems to be reasonable, but if one assumes so,
> this means that current G98 ONIOM implementation is not suitable
> at all for running oniom(ab-initio:Amaber) or oniom(ab-initio:AM1/PM3)
> calculations for transition metal containing systems.
> Moreover, very unlikely in the nearest future due to known
> difficulties of MM and semiempirical treatment of TM systems.
>
> Sounds a bit pessimistic. ( I know about UFF, but it does
> not seem to work well, especially for charged molecules)
>
> >From the other hand, in JPC 1996,100,19357 oniom(b3lyp:HF:MM3)
> calculations for platinum complexes are described. However,
> I am not aware of any MM3 parametrization for platinum.
>
> is there a way to repeat such calculations in G98?
>
>
> regards, Valentin.
>
>
>
> > Hi Valentine,
> >
> > The ONIOM method requires 3 calculations for
> > a molecule: a low level calc of the whole molecule
> > (ll/all), a low level calc of the high level part (ll/part)
> > and a high level calc of the high level part (hl/part).
> > The ONIOM energy is defined as
> > E(ONIOM) = E(ll/all) - E(ll/part) + E(hl/part)
> >
> > Although the low level energy contributions of the
> > high level part seem to cancel out, that is not
> > completely true. In order to have well behaved
> > high level parts, broken bonds are saturated with
> > hydrogen. This causes a difference in the MM
> > energy terms between ll/all and ll/part that reaches
> > at least three atoms inside the high level part
> > (dihedral angle with the saturation H). For that
> > reason (and programming reasons, I assume),
> > you need MM parameters for all atoms in the
> > molecule.
> >
> > In the last few days there was a thread concerning
> > ONIOM energies. One of the mails pointed to an
> > article that explains the ONIOM method in more
> > detail.
> >
> > Stefan
> > __________________________________________________________________
> > Dr. Stefan Fau | fau %! at !% qtp.ufl.edu
> >
>
> ====================================================================
> , , , ,
> Valentine P. Ananikov |\\\\ ////| /////|
> NMR Group | \\\|/// | ///// |
> ND Zelinsky Institute of Organic Chemistry | |~~~| | |~~~| |
> Leninsky Prospect 47 | |===| | |===| |
> Moscow 117913 | | | | | | |
> Russia | | A | | | Z | |
> \ | | / | | /
> e-mail: val %! at !% cacr.ioc.ac.ru \|===|/ |===|/
> http://nmr.ioc.ac.ru/Staff/AnanikovVP/ '---'
'---'
> Fax +7 (095)1355328 Phone +7 (095)9383536
> ====================================================================
>
>
>
>
> -= This is automatically added to each message by mailing script =-
> CHEMISTRY %! at !% ccl.net -- To Everybody | CHEMISTRY-REQUEST %! at
!% ccl.net -- To Admins
> MAILSERV %! at !% ccl.net -- HELP CHEMISTRY or HELP SEARCH
> CHEMISTRY-SEARCH %! at !% ccl.net -- archive search | Gopher:
gopher.ccl.net 70
> Ftp: ftp.ccl.net | WWW: http://www.ccl.net/chemistry/ | Jan: jkl %! at !% ccl.net
>
>
>
>
>
>
>