From CFAF@DS0RUS1I.bitnet Thu Mar  7 11:10:26 1991
Date: Thu, 07 Mar 91 17:09:17 MEZ
From: Andreas Savin <CFAF@DS0RUS1I.bitnet>
Subject: ECPs in basis set synopsis
To: "Dr. Jan K. Labanowski" <JKL@MPS.OHIO-STATE.EDU>
Status: R

 
Dear Dr. Labanowski,
 
your summary about basis functions is very helpful. We would like
to point out, however, that in the paragraph about ECPs there is
a minor error. You state that core potentials can only be specified
for filled shells. In fact, they can be also defined for open shell
cores (see, for example, the pseudopotentials for the lanthanides,
Dolg et al., Theor. Chim. Acta 75 (1989) 173 ).
 
We would like to use this opportunity to point out that pseudopotentials
and corresponding valence basis sets are available from us, not only
for the lanthanides, but for all the elements up to Uranium. A
comparison of our pseudopotentials with other ones shows their quality
(see, e.g., Andrae et al., Theor. Chim. Acta 78 (1991) 247).
 
With best regards,
 
Michael Dolg, Hermann Stoll, Andreas Savin, Heinzwerner Preuss

From zerner@qtp.ufl.edu Thu Mar  7 19:06:08 1991
Date: Thu, 7 Mar 91 19:05:43 EST
From: "Mike Zerner" <zerner@qtp.ufl.edu>
To: jkl@ccl.net
Subject: Re:  Correction to basis set intro Part 3/3
Status: R

Hi Jan.   Took your basis set write up on the plane with me yo New York,
and enjoyed reading it. May use it for the short course of which you are
an alumnus. 

One of the reasons that core orbitals make their appearance  in valence and 
other higher energy orbitals is due to the fact that all mo's must be orthogonal
If one examines these coefficients they are very much like those you would get
from simple Schmidt orthogonalisation to the lower lying core orbitals.

I don't understand the discusion above the 66-31 basis set for Si Table.
If a primitive occurs contracted  in a 3 combo , say, x+y+z,  and appears by
itself, say z, why is this linearly dependent? I.e think of three dimensional
space.  DId I read this wrong?

under 66-31G means that there is :"  should this not be S(6/6/3/1) rather than
S(6/6/3/2) ?

The difference between the 2p function on Li, and the 3d on S, is that the
former is needed for hybridization, and ignoring it ignores all of first year
chemistry! (The first solid state calculations on Li and alkali halides were
done by physists, and of course, did not contain the 2p!). The 3d function
on sulfur is usually polarization, but even this is somewhat controversial.

Hope all is well.  The only time I get to read in peace these days is on the
plane.

Regards,  Mike Z.