BSSE Correction for Protonation Energy

[Kui Zhang <KZHANG%MIAMIU.BITNET ^at^ phem3.acs.ohio-state.edu>]

 Dear Netters,
      One of our projects is the calculation of protonation energy
 for some interesting compounds using high level ab initio calculations.
 Recently I have been trying to correct calculated protonation energies
 with BSSE.  I first used CH3NH2 as model compound to do testing
 calculations at HF, MP2 and MP4 levels with several basis sets.  The
 calculations were carried out using Gaussian 92 with MASSAGE method.
 The following is a set of calculated values of the BSSE correction for
 CH3NH2:
                                          BSSE Correction (kcal/mol)
 HF/6-31G*//HF/6-31G*                             0.60
 HF/6-31+G**//HF/6-31+G**                         0.84
 HF/6-31+G(2d,2p)//HF/6-31+G(2d,2p)               0.31
 MP2/6-31+G**//HF/6-31+G**                        6.87
 MP4/6-31+G**//HF/6-31+G**                        7.20
 MP2/6-31+G**//MP2/6-31+G**                      14.15
     It can be obviously seen that the MP and HF give a such huge
 difference that there is absolutely not comprimise between HF and MP
 values.  The same situation was also observed in another compound
 similar to CH3NH2.  Can any of you help me handle this problem?  My
 question is that which method I should trust in term of the BSSE
 computation, MP or HF?  I would greatly appreciate any direction,
 suggestion and comment.  Thank you very much in advance.
 Kui Zhang
 Department of Chemistry
 Miami University
 Oxford, OH 45056
 USA
 E-mail to: kzhang ^at^ miamiu.acs.muohio.edu