I calculate the interaction energies between some molecules with ADF in
the department of Chemistry at the University of Joensuu . I have
previously used
Gaussian to compute different kind interaction energies for different
molecules. In Gaussian this process is done with Total Energies and
process to compute interaction energies is like
following (representing the method roughly ): You first do geometry
optimization to one molecule and do the same to
another molecule. After that you do geometry optimization to whole
system ( first molecule + second molecule). Then
you get interaction energy from the following equation: DE (interaction
energy)= E (energy of the whole
system)-E1(energy of the first molecule)-E2 (energy of the second
molecule).
    The question is this: How you calculate this interaction energy with
ADF. Is the method same as in Gaussian meaning
that you now use only 'Total Bonding Energies' (as Total Energy ) and
single-atom fragments all the time OR is the
method quite much different than in Gaussian? I mean that you first
optimize both molecules separately to get
something starting point. Then you optimize the whole system. After that
you cut the optimized system so that you take
the first molecule and calculate single-point calculation to that
molecule to get one fragment file. Then you do the same to
the another molecule to get the second fragment file. Finally you have
two different fragment files, which you use for
calculation the interaction energy of those two molecules via single
point calculation to whole previously optimized
system.
    I have used ADF very little time as you notice about these
questions, but anyway the second question is like this.
Which is the best coordinate method (Z-matrix or Cartesian) in those
interaction energy calculations if the interaction
energy is calculated via fragment files (if molecules are big enough).
This question sounds little curious but meaning is this.
Now I optimize this 2-moleculesystem in Z-matrix coordinate system,
because I must use the ADF subkey optim
selected and I have understood from ADF 2.3 manual (Am I wrong?), that
you can't use this key if you operate with
cartesian system. However, if the fragment-file-method is the correct
way for the interaction energy calculations, I must
use cartesian coordinate system in the final single-point calculation
(specially in my case), because, if I have understood
right, the ADF is very exact, that when you start fragment-file
calculation (I mean, that fragment file contains now more
than 3 atoms), the atomic coordinate places of the fragment files must
be exactly the same as in input file or at least the
symmetry must be same. Could I use the same coordinate system for whole
calculation time??? I ask this therefore, that
I have noticed, that ADF gives little different energies depending on
coordinate system, which you have selected.
    Anyway, I hope that somebody has had strenght enough to read to this
point of this email. But finally I have only one
little question left at this time. I tried calculate today a
single-point calculation with one molecule with ADF. There was an
error warning in the logfile after five minutes counting .
It sounds like this: TOO MANY TRIES. HULL3
                            WARNING: not all scratch files were closed
                            archec: can't find IB2TIDS
                            END
In the output file this warning message is after COMPUTATION-section or
precisely after where you can read General
Accuracy Parameter : .... Anyway error message became before
SCF-section.
    There is very much queries in this email now, but I hope that
somebody would be so kind and would answer these
questions. Somebody could say, that these questions are obvious, but ...
for me they are very important.

Please send the answer to next email address: porento_at_cc.joensuu.fi
Received on 1999-08-16 13:38:47