Dear ADF users,

For those of you who perform ADF calculations with point charges:
a serious bug has been detected which leads to an incorrect
reordering of these charges, leading to incorrect results.
(Thanks to Lasse Jensen and Serguei Patchkovskii for their bug reports).

Below I will explain what goes wrong, how one can check if this error
has occurred in your previous calculations and what work-around to
this bug is available. If you are not using point charges, the following
is probably not of interest to you.

The ADF output prints which point charges are used in two
places, both in the beginning of the output. The first print is a
copy of the point charges specified in the input file.
The second print gives the reordered point charges with which the
actual calculation will be performed. You can therefore easily check
if your calculation has used the correct point charges or not.

ADF organizes the point charges into groups with the same
charge, and then reorders the point charges such that all
point charges with the same charge occur together. This is
needed for the analysis of the symmetry of the molecule,
where the point charges need to be taken into account.

However, the value of the charge itself is not modified
accordingly. This bug is fixed in the current development
version which will become the next release, but not in
ADF2002.02 and earlier versions.

An example of what goes wrong is given here:

If the point charges in the input file are as follows (with charges of +1
and -1)

EFIELD
1 1 1 -1.0
2 2 2 1.0
3 3 3 -1.0
4 4 4 1.0
end

The position and value of the charges will initially be copied in the
output file in unmodified order:

    Electric Field:
 *** SPECIAL *
       Point Charges : X Y Z (Angstrom) Q
                     ------------------------- ----------
               1) 1.0000 1.0000 1.0000 -1.00000000
               2) 2.0000 2.0000 2.0000 1.00000000
               3) 3.0000 3.0000 3.0000 -1.00000000
               4) 4.0000 4.0000 4.0000 1.00000000
                     ------------------------------------------
       Total Charge 0.00000000

However, not much later in the program the reordering takes place
and the final charges which will be used in the SCF are printed:

 Point Charges
 =============

                           X Y Z (angstr) Q
                   ------------------------------ ----------
                       1.0000 1.0000 1.0000 -1.0000
                       3.0000 3.0000 3.0000 1.0000
                       2.0000 2.0000 2.0000 -1.0000
                       4.0000 4.0000 4.0000 1.0000

Here it can be seen that the second and third point charges have
switched places, because ADF wants the atoms with charges +1
together. However, the charges have not been modified accordingly.
This leads of course to unintended results.

WORK-AROUND
===========

A simple work-around exists for this problem.

If the charges were given in the input file in the following order:

EFIELD
1 1 1 -1.0
3 3 3 -1.0
2 2 2 1.0
4 4 4 1.0
end

then they would already be in the order that ADF desires and they will
not be rearranged.

We apologize for inconveniences caused by this bug. We have added
suitable example calculations in our internal set of test calculations
to avoid further related problems in future versions of ADF.

Best regards,
Stan van Gisbergen

------------------------------------------------------------------------
Dr. S.J.A. van Gisbergen SCIENTIFIC COMPUTING & MODELLING NV
Phone: +31-20-4447626 Vrije Universiteit, Theoretical Chemistry
Fax: +31-20-4447629 De Boelelaan 1083
E-mail: vangisbergen_at_scm.com 1081 HV Amsterdam, The Netherlands
                               http://www.scm.com
------------------------------------------------------------------------
Received on 2002-08-27 13:21:10