The potential energy hypersurfaces for the triple inversion, from chair to boat and alpha to beta conformations, are explored theoretically, by using the quantum semiempirical method AM 1, in 3-azabicyclo[3.3.1]nonan-9-one and its N-methyl derivative. Both compounds are precursors of rigid analogues of the potential of gamma-butyric acid receptor antagonists (GABA(B)). It is found that the chair-chair a conformer is by far the most stable structure for the non-methylated compound. In contrast, N-methylation changes the conformational preference to the beta-structure, the chair-chair beta conformer being the most stable. The inversion barriers are calculated to be relatively low, so that the conformers must be found in thermodynamic equilibrium. The theoretical results compare well with the available experimental results.
