About "Config" keyword

[H.-J. Werner]

In the case of determinants, + means alpha spin and - beta spin.
In the case of CSFs, + means up in the branching diagram and
- down. For instance, ++ - + neans that the first 2 electrons
are coupled to a triplet, this triplet plus the third electron
forms a doublet, and finally the last electron + the doublet 
are coupled to a triplet. You can find more about branching
diagrams in many textbooks. I cannot tell you why the coefficient
of the last CSF is zero without knowing the orbitals. Probably
this is due to symmetry.
H.-J. Werner
> 
> Dear Prof. Werner (and listmembers),
> 
> Thank you for your reply. This is not what I am asking though. I would
> like to know how to correctly READ the CI vector when using CSFs. For
> instance, in the example I have mentioned in my initial message (see
> below), when using determinants I read:
> 
> ++ - +       0.1704248
> 
> as "1a1(alpha)2a1(alpha)1b1(beta)1b2(alpha)".
> 
> When using CSFs, the same line is:
> 
> ++ - +       0.2783026
> 
> This corresponds to one spin-adapted linear combination, isn't it? 
> Which one is it? And which one is, for instance: 
> "+- + +       0.0000000"?
> 
> Ioannis Kerkines
> 
> 
> On Thu, 6 Sep 2001, H.-J. Werner wrote:
> 
> > Hi,
> > one can only specify "orbital configurations", i.e., the occupancy
> > for each orbital. The program will then generate all possible
> > spin-adapted CSFs. It is not possible to use subsets of CSFs
> > for a given orbital configuration. The input for your example
> > would read
> > 
> > select
> > con,2,0,1,1
> > con,0,2,1,1
> > con,1,1,1,1
> > 
> > but in fact this would generate all possible CSFs for 3Sig-.
> > 
> > Note that determinants can only be used for CASSCF but not
> > for selected MCSCF (the select card triggers the use of CSFs).
> > Also note that the closed-shell (inactive) orbitals are not
> > included on the con cards as shown above.
> > 
> > Instead of the orbital occupation numbers, one can also
> > specify the orbitals itself, see e.g., 
> > 
> > testjobs/h2o_explicit.test
> > testjobs/o3_explicit.test
> > 
> > and the manual.
> > 
> > H.-J. Werner
> > 
> > > 
> > > Hello,
> > > 
> > > I am really puzzled by the use of the "Config" keyword in CASSCF
> > > calculations. It is supposed to work with CSFs instead of Slater
> > > determinants. I cannot understand how to read the CI vector in the case of
> > > CSFs. For example, a single point CASSCF (4,4) triplet sigma minus
> > > calculation 1st with determinants and 2nd with CSFs produces the
> > > following
> > > output:
> > > 
> > >  CI vector
> > >  ---------
> > > 
> > >  20 + +       0.8726238
> > >  02 + +      -0.3497846
> > >  -+ + +      -0.1704248
> > >  ++ - +       0.1704248
> > >  ++ + -       0.1704248
> > >  +- + +      -0.1704248
> > > 
> > >  TOTAL ENERGIES                     -1815.56889064
> > > 
> > > 
> > >  CI vector
> > >  ---------
> > > 
> > >  20 + +       0.8726238
> > >  02 + +      -0.3497846
> > >  ++ - +       0.2783026
> > >  ++ + -       0.1967896
> > >  +- + +       0.0000000
> > > 
> > >  TOTAL ENERGIES                     -1815.56889064
> > > 
> > > 
> > > I can easily understand why there is a difference in coefficients between
> > > CSFs and determinants, but how does one read for instance "++ - +" in the
> > > CSF case? How does one know to which spin-adapted linear combination does
> > > it belong to? What is the convention, since "++ - +" is written in both
> > > cases?
> > > 
> > > Best Regards
> > > Ioannis Kerkines
> > > 
> > > 
> > 
> > 
> > --
> > Prof. Hans-Joachim Werner
> > Institute for Theoretical Chemistry
> > University of Stuttgart
> > Pfaffenwaldring 55
> > D-70569 Stuttgart
> > Tel: (0049) 711 / 685 4400
> > Fax: (0049) 711 / 685 4442
> > email: werner at theochem.uni-stuttgart.de
> > 
> 
> 


--
Prof. Hans-Joachim Werner
Institute for Theoretical Chemistry
University of Stuttgart
Pfaffenwaldring 55
D-70569 Stuttgart
Tel: (0049) 711 / 685 4400
Fax: (0049) 711 / 685 4442
email: werner at theochem.uni-stuttgart.de