From owner-chemistry@ccl.net Tue Mar 13 08:37:00 2012
From: "Georg Lefkidis lefkidis!A!physik.uni-kl.de" <owner-chemistry~!~server.ccl.net>
To: CCL
Subject: CCL:G: AW: CASSCF does not produce spin densities
Message-Id: <-46471-120313083620-26296-aQUa0IikKpJmmXOjpmUtZw~!~server.ccl.net>
X-Original-From: "Georg Lefkidis" <lefkidis(a)physik.uni-kl.de>
Content-Language: de
Content-Transfer-Encoding: 8bit
Content-Type: text/plain;
	charset="iso-8859-1"
Date: Tue, 13 Mar 2012 13:36:10 +0100
MIME-Version: 1.0


Sent to CCL by: "Georg Lefkidis" [lefkidis^^physik.uni-kl.de]
Dear Josehp,

with all respect, but Frank Neese's statement is wrong. CASSCF is a post-HF,
multi-determinental, configuration-interaction method (it is a
multireference method [look e.g. in Chem. Phys. Lett., 93 (1982) 43-46,
Chem. Phys. Lett., 83 (1981) 362-68 ]). It provides wavefunctions which are
linear combinations of Slater determinants, and in fact is one of the best
methods around to calculate static correlations. The reason is that it not
only optimizes the coefficients of the determinental expansion but also the
MO themselves (look for STATEAVERAGE in Gaussian). In fact it provides
excellent results for transition metal elements (I myself have several
publications using CASSCF). Unfortunately Gaussian's output is somehow
cryptic, in that it produces the new optimized reference HF but it takes
some digging to find out the CI coefficients (but they are definitely there
- results, by the way, are calculated from the complete, internally stored
density matrix and not the simple HF output). It is true however that it
provides not (or very poor) spin-polarization results (SOC or hyperfine
splitting). As for the spin polarization, I personally calculate it on top
of Gaussian's results by extracting the complete wavefunction. It is indeed
a very tedious procedure, but surely possible.

Best regards
Georg 

-----Ursprüngliche Nachricht-----
Von: owner-chemistry+lefkidis==physik.uni-kl.de^ccl.net
[mailto:owner-chemistry+lefkidis==physik.uni-kl.de^ccl.net] Im Auftrag von
Frank Neese neese]|[mpi-muelheim.mpg.de
Gesendet: Dienstag, 13. März 2012 09:57
An: Lefkidis, Georg 
Betreff: CCL: CASSCF does not produce spin densities


Sent to CCL by: Frank Neese [neese_._mpi-muelheim.mpg.de] Dear Joseph,

there definitely is a spin density associated with a CASSCF wave function as
long as the total spin is larger than 0. It would be a major misconception
to think that spin-density only arises in a spin unrestricted framework.
I don't know about other programs, but at least the ORCA program will print
spin populations following a CASSCF calculation (please note that there is a
distinct difference between the spin density=a three dimensional function of
space and the spin-population= are more or less well justified distribution
of the spin density onto individual atoms) and also calculate expectation
values such as isotropic and dipolar hyperfine couplings on the basis of the
spin density. 
However, CASSCF is a Hartree-Fock level method - definitely not a high level
treatment. In my humble opinion, it is unlikely to produce accurate results
for iron-nitrosyl systems that are notoriously difficult. In particular, it
will not do well on isotropic hyperfine couplings as these require a proper
treatment of spin-polarization which you will not achieve at the CASSCF
level. It only properly comes into play once you include certain classes of
double excitations on top of the CASSCF reference wavefunction and also take
them into account in evaluating the density. 

I hope that helps a bit?! 
Best regards,
Frank
 

Am 12.03.2012 um 18:37 schrieb Joseph Katigbak jkatigba/./stevens.edu:

> 
> Sent to CCL by: "Joseph  Katigbak" [jkatigba{:}stevens.edu] Hello, I 
> am trying to determine the hyperfine coupling constants of some 
> nitrosyl compounds using a high level 
> Multiconfigurational/Multireference
> method like CASSCF. After running a CAS calculation using ROHF 
> molecular orbitals as guesses, I got energy convergence for my 
> compounds. However i noticed that the spin densities weren't being 
> printed out (I was looking specifically for the Isotropic Fermi 
> contact coupling). I have been told that CASSCF implements a 
> restricted SCF colution and therefore does not generate a spin 
> density. However I know from the literature that there are several 
> papers on iron nitrosyl systems that use CASSCF and have spin 
> densities. Is there any way for me to get the spin densities/Fermi contact
coupling term? or am i just doing it wrong?
> 
> 
> Thank you very much :D
> -J. Katigbakhttp://www.ccl.net/cgi-bin/ccl/send_ccl_messagehttp://www.ccl.net/chemistry/sub_unsub.shtmlhttp://www.ccl.net/spammers.txt