Hi all,
here a summary of the answers for my question regarding protonation
energies of adenine.
Thanks to everyone. The problem wasn't really a problem, merely a typo in my
orginal input, I assume. I have calculated the H+ now, and gaind a energy of
+288.82 kcal/mole, which is in rather good agreement with the other
asummptions. All I have to do now is to account for the zero point energy
(stated by Lou Noodleman) and, of course the BSSE.

----
 my original question: Dear
all,
We have tried to investigate the protonation energies of Adenine with ADF and
encountered a small, but yet annoying problem. We have obtained values, that
are about 280 kcal/mole higher than results in G98 and gas phase
measurements. These two are in line. 
My question is, how do I calculate the protonation energy properly?    
F.Glahe
-----
From: Reinaldo Pis Diez <pis_diez_at_quimica.unlp.edu.ar>
         Note that "bonding energy" is referred to as the energy difference
between the system under study and the corresponding fragments the program
uses to build up the guess density. It shouldn't be taken as a true bonding
energy.  
         Well, I ran a job with a hydrogen atom (using basis set II) having a +1
charge and 0 for the difference in the alpha and beta populations and it
was ok. I finally got a proton and the bond energy wrt the (spherical) H
fragment is about 12.12 eV which is incredibly high by the way. I'm not
sure if you're looking for something like this but, if yes, I can send you
the inputs asap. 
         If you need other type of information, please send me a clarifying mail
to see if I can help you.    
         Regards,
-----
From:Serguei Patchkovskii <patchkov_at_ucalgary.ca>
280 kcal/mole sounds suspiciously close to the bond energy of H+ 
relative to the fictious spherically-and-spin-symmetric hydrogen
atom reference used by ADF. Per chance, you didn't forget to 
subtract bonding energy of H+, did you? Unlike more conventional 
quantum chemistry programs (like G98), it is NOT zero in ADF ...
Cheers,
/Serge.P
-----
From: Lou Noodleman <lou_at_scripps.edu>
Dear Frank Glahe,
Your problem is almost certainly the reference state in ADF, which is
the sum of the spin restricted atoms. Thus, protonated adenine has an
additional spin restricted H-atom in the reference state.
 The ionization potential of the "restricted"
H atom is 12.633eV=291.3 Kcal/mol., calculated for a VBP potential-for example,
see C.L. Fisher et al., J.Phys.Chem. 100,13498, 1996.
 You must correct for this process:
H(res)-> H+ +1e-(at infinity). (Note that the corresponding spin-polarized
H-atom has an IP=13.58eV, close to exp. 13.6058eV). You can calculate these
IP's yourself, using the ADF programs. It may be helpful to have a small
fraction of 1e-, say one-millionth, in your calculation of the "bare proton"
(H+) state, since the program is expecting a non-zero electron density.
All of this is well known to those doing protonation (or deprotonation)
calculations with ADF.
Lou Noodleman
Molecular Biology
The Scripps Research Institute
La Jolla, CA
----
From: Lou Noodleman <lou_at_scripps.edu>
Dear Frank Glahe,
Another term you may have missed is the Zero-Point-Energy (ZPE) term.
The ZPE term for breaking an N-H or O-H bond (by deprotonating) is
about -7 or -8 kcal/mol. It can be obtained by calculating the vibrational
frequencies for both species, and then evaluating the ZPE term. The remaining
vibrational energy difference can be obtained similarly (as discussed in
the book: Ab initio molecular orbital theory by Hehre, Radom,
 von Schleyer, Pople). Our group has done a fair number of gas phase
deprotonation as well as solvation calculations both for organic and
metal complexes, evaluating also pKa's. You can look at: Richardson et al.,
Int.J.Quantum Chem. 61,207, 1997, Chen et al., J.Phys.Chem. 98,11059, 1994,
Konecny et al., Inorg. chem. 38,940, 1999, Li et al., Inorg. Chem. 38,929,1999.
We are continuing to work on these difficult, but important problems as
 are other groups. There is a recent review in Chem.Rev. by Cramer and Truhlar,
and work by groups like Truong and Stefanovich and C. Lim. A considerable
literature is now developing on these topics.
Lou Noodleman
Scripps
----
From: "Louwen, J. (Jaap)" <Jaap.Louwen_at_akzonobel.com>
With energies differences of that kind something is obviously wrong. Without
knowing further details, it is hard tell what went wrong, but on a general
note I would stick to the principle: rely as much on compensation of error
as possible.
That is: do not try to compute the energy of the reaction Ad + H(+) ->
AdH(+) but refer to a reaction like:
Ad + H3O(+) -> AdH(+) + H2O
This has the definite advantage that the number of bonds doesn't change (and
the character doesn't change much), so there will substantial cancellation
of model errors. From the energy of the reaction last given and an
experimental value for the protonation of H2O, you should be able to derive
proper values.
Incidentally, one of the soures of errors might be the basis set
superposition error (BSSE) - only if you don't use ghost atoms, of course.
However, those errors tend to be less significant in DFT than in
Hartree-Fock.
Regards
Jaap N. Louwen
It just occured to me: could it be that you take the energy of the proton as
zero? Note that if you compute the energy of AdH(+) in ADF, you get the
energy difference between AdH(+) and its fragments (at least in my version
of ADF). The proper fragment for H is the hydrogen atom, NOT the proton, so
you need to account for the difference in energy between proton and H
fragment.
Regards
Jaap
-----
From: Matthias Bickelhaupt <bickel_at_chem.vu.nl>
Lieber Frank,
Ich habe eben die Email-Korrespondenz ueber H(+) verfolgt.
Du kannst einfach eine ADF-Rechnung fuer H(+) als "Molekuel"
machen und bekommst dann die korrekte Energie fuer die Berechnung
E(AH+) - E(A) - E(H+).
Gruss,
Matthias
-----
From: Cory Pye <cpye_at_crux.stmarys.ca>
In principle, you could get the analytic solution difference (0.5 au) between 
the energies of the H+ and H moieties and use Hess's law to figure it out. 
Usually, the energies are very close for the H atom (with H+ as the reference 
state, as in Gaussian). For ADF, with the H atom itself as the reference state,
the H+ should be -ve this value. I don't know if DFT behaves similarly to HF 
when it comes to H atom. 
----
  --
****
WWW   http://www.chemie.uni-dortmund.de/groups/lippert/
MAIL  uch480_at_pop.uni-dortmund.de
TEL   +49 231 755 3747
FAX   +49 231 755 3797
****