Does the McConnell-I model really work? An ab initio study of the magneticcharacter of some intermolecular contacts

The performance of the first model proposed by McConnell to rationalize the nature of the magnetic interactions in purely organic molecular crystals ( the McConnelL-I model) has been analyzed using ab initio methods. As a firs t step, we analyzed the form of deriving the basic formula of the McConnell -I model from a general Heisenberg spin Hamiltonian. We reached the conclus ion that it is not possible in general, although the two show some resembla nce. When the same procedure is applied to the well known example of the [2 .2]paracyclophanes, we found that the agreement with the experiment comes f rom the fortuitous cancellation of terms which is not likely to hold in gen eral. We then studied the high-low spin energy difference for the H2NO dime r and the methyl-allyl radicals for different relative orientations of the two monomers using CASSCF calculations and the 6-31G(d) basis set. It is fo und that there are regions in which McConnell-I model does not reproduce th e ab initio results. So, we can conclude that the McConnell-I does not work in general.