Polarizability summary
In response to my recent post on determining polarizabilities,
I'll summarize with the following. Thanks to all the respondents.
The GAMESS program can calculate static linear polarizabilities and
hyperpolarizabilities with the runtype=FFIELD, or you can use runtype=TDHF
to get either static or time-dependent polarizabilities and
hyperpolarizabilities. Or, when doing an analytic hessian calculation,
set POLAR=.TRUE. in the $CPHF group and you will get the static linear
polarizabilities. All these methods are available at the ab initio
Hartree-Fock level, and I believe that FFIELD is also available with the
semiempirical methods (AM1, PM3, MNDO). GAMESS can be obtained free of
charge (including source code) from Mike Schmidt: mike -AatT- si.fi.ameslab.gov
Hope this helps. Paul
Dr Paul N. Day Commercial: (513) 255-6671, Ext 3106--
Visiting Scientist FAX: (513) 255-1128
Laser Hardened Materials Branch WL/MLPJ
E-Mail: daypaul -AatT- biotech.ml.wpafb.af.mil
Wright-Patterson AFB, OH 45433 (134.131.35.30)
-----
See Miller (JACS *112*, 8533 (1990)) for such a method. We've
implemented this in our ChemPlus QSAR Properties module, using additional
parameters from Bondi (_J. Phys. Chem._ *68*, 441 (1964)) and Shannon and
Prewitt (_Acta Cryst._ *B25*, 925 (1969)) to handle at least some
organometallics. For example, I optimized ferrocene with ZINDO/1 in
HyperChem and calculated its polarizability with ChemPlus as 14.50 A^3.
Joel Polowin, Ph.D. Manager, Scientific Support
Email to: polowin -AatT- hyper.com
Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2
(519)725-4040 Info requests to: info -AatT- hyper.com
Support questions to: support -AatT- hyper.com
http://www.hyper.com/
----
It sounds like you may be looking for a fast, approximate method. If,
however, you are willing to include ab initio approaches, the commercial
program PS-GVB put out by Schrodinger Inc. (http://www.psgvb.com) will
compute linear and first nonlinear polarizabilities at the HF level very
quickly. Not quickly compared to evaluating an empirical analytical
function but quickly in comparison to the same level of theory found in
other commercial codes. The program in general has been fine-tuned to
handle the complications which arise from transition metal calculations as
well (like in the initial guess of the wavefunction and range of effective
core potentials available). You can also hand-pick a series of uniform
zero-frequency fields to apply to the molecule of interest to do
calculations of higher order polarizabilities via a finite field
procedure.
Dr. Bryan Marten--
Ciba Pharmaceuticals
martebr2 -AatT- ussu.ciba.com
------
For ab-initio HONDO, by Michel Dupuis, IBM, Kingston is best. Also try
Gaussian or GAMESS. Semi-empirical try ZINDo, Mike Zerner, Quantum Theory
Project, Univ. of Florida. I am separately sending you my CNDO-EB-CHF +
data and output.
Good luck,
John
Dr. John Waite, e-mail: chem8 -AatT- york.ac.uk or
The National Hellenic Research Foundation,
rosen -AatT- cyclades.nrcps.ariadne-t.gr
Organic and Pharaceutical Institute, phone: ++30-1-7238958 (direct)
Vas. Konstantinou 48, phone: ++30-1-7247913(secrtry. Mary)
Athens 116-35, fax: ++30-1-7247913
Greece
or
NCRS "Democritos", phone: ++30-1-6513112-5 X219
c/o Dr. G.Kordas, e-mail john -AatT- john.nrcps.ariadne-t.gr
Material Science Institute,
Aghia Paraskevi,
Attikis,
Athens 153-10,
Greece
------
DGauss is a commericial density functional program that can compute
static optical polarizabilities. It is a part of the UniChem molecular
modeling package that is offered by the Oxford Molecular Group. It will be
able to predict polarizabilities for organometallics as well as organics,
and, since it computes these using analytic 2nd derivatives, it will
compute them relatively efficiently. If you'd interested, I can send
additional information regarding UniChem & DGauss.
Regards, Mark
Mark S. Stave, Ph.D.
Computational Scientist
e-mail: mstave -AatT- oxmol.com
Oxford Molecular Group, Inc. office: (612) 894-2510
-----
Gaussian 94 can compute polarizabilities and hyperpolarizabilities at
HF, DFT and traditional correlation levels. For the HF and DFT methods
these are analytic, for MP2 and beyond they are a mixture of analytic and
numerical.
Douglas J. Fox
Director of Technical Support
help -AatT- gaussian.com