Dear all,

Recently bad fit problems with ADF were reported by Xavi:
> THE CALCULATION:
> I have been computing the metal-oxide structure TeW6O24 (D3d symmetry),
> with the BP86 functional, a TZP basis set (basis IV in ADF) and the ZORA
> package for the core electrons. The core electrons contain all the
> non-valence shells, although a test with less core electrons for
> tungsten was carried out unsuccessfully.
>
> THE PROBLEM:
> Whereas the homologous compound with molybdenum converges properly,
> serious problems are detected with the tungstate. There is a WARNING at
> the end of the SCF procedure, in which a BAD FIT message shows that the
> number of expected electrons and the numerical integration of the
> density differ by more than 18 units. This problem could be related to
> the auxiliary basis, but we don't know how to overcome this. In addition
> of the unproper SCF-convergence, of course the geometry optimisation is
> unaffordable. Geometry steps behave rarely, with very strong energy
> changes in successive steps.
>
> We normally deal with very good fits in similar calculations, so we
> don't have an idea of the true problem here. I will thank any advice to
> solve the problem.

Possible solution:

I have come across similar problems in some of my calculations
if the SCF did not converge properly, namely that in one
or more of the cycles in the SCF convergence the
electron occupation numbers were in fact that of a highly excited state.
ADF is not always able to accurately fit such a highly excited
state, which then can give the BAD FIT message.
A solution might be in this case to fix the
electron occupation numbers with the occupations keyword,
which one might deduce from the molybdenum compound.
Or one could try to use some of the options
of the SCF block-key, in such a way that the SCF does converge.

Best regards,
Erik van Lenthe

----------------------------------------------------------------------
Dr. Erik van Lenthe SCIENTIFIC COMPUTING & MODELLING NV
Tel: +31 20 44 47615 Vrije Universiteit, Theoretische Chemie
secretary: 44 47519 De Boelelaan 1083
 fax: 44 47629 1081 HV Amsterdam, The Netherlands
 e-mail: vanlenthe@scm.com http://www.scm.com
----------------------------------------------------------------------
Received on 2002-12-17 12:45:59