THE AMMONIA DIMER REVISITED

Author Institution: Missouri University of Science and Technology; Rolla, MO 65409-0010; Radboud University, 6525 AJ Nijmegen, The NetherlandsThe conclusion from microwave spectra by Nelson, Fraser, and Klemperer nderline{\textbf{83}} 6201 (1985)} that the ammonia dimer has a nearly cyclic structure led to much debate about the issue of whether (NH$_3$)$_2$ is hydrogen bonded. This structure was surprising because most \textit{ab initio} calculations led to a classical, nearly linear, hydrogen-bonded structure. An obvious explanation of the discrepancy between the outcome of these calculations and the microwave data which led Nelson \textit{et al.} to their "surprising structure'' might be the effect of vibrational averaging: the electronic structure calculations focus on finding the minimum of the intermolecular potential, the experiment gives a vibrationally averaged structure. Isotope substitution studies seemed to indicate, however, that the complex is nearly rigid. Additional data became available from high-resolution molecular beam far-infrared spectroscopy in the Saykally group nderline{\textbf{97}} 4727 (1992)}. These spectra, displaying large tunneling splittings, indicate that the complex is very floppy. The seemingly contradictory experimental data were explained when it became possible nderline{\textbf{101}} 8430 (1994); E.~H.~T.~Olthof, A.~van der Avoird, P.~E.~S.~Wormer, J.~G.~Loeser, and R.~J.~Saykally \textit{J.~Chem.~Phys.} nderline{\textbf{101}} 8443 (1994)} to calculate the vibration-rotation-tunneling (VRT) states of the complex on a six-dimensional intermolecular potential surface. The potential used was a simple model potential, with parameters fitted to the far-infrared data. Now, for the first time, a six-dimensional potential was computed by high level \textit{ab initio} methods and this potential will be used in calculations of the VRT states of (NH$_3$)$_2$ and (ND$_3$)$_2$. So, we will finally be able to answer the question whether the conclusions from the model calculations are indeed a valid explanation of the experimental data