Experimental and Theoretical Study of the Kinetics of Dimerization of Ammonia at Low Temperatures

The kinetics of the dimerization of NH<inf>3</inf> in helium and nitrogen bath gas within the supersonic flow of a Laval nozzle were investigated at very low temperatures. Experimentally, the fraction of the NH<inf>3</inf> monomer, f<inf>monomer</inf>, remaining in the flow at 91 K in N<inf>2</inf> and at 35 K in He for a total bath gas density [M]∼5 × 10¹⁶ molecules cm^-3 was monitored using fluorescence from the electronically excited NH<inf>2</inf> photofragment formed following NH<inf>3</inf> photolysis at 213 nm. No dimerization was observed up to [NH<inf>3</inf>] = 1 × 10¹⁵ molecules cm^-3 for 160 mm downstream of the 91 K N<inf>2</inf> nozzle, nor up to [NH<inf>3</inf>] = 5 × 10¹⁴ molecules cm^-3 for 150 mm downstream of the 35 K He nozzle. Dimerization was observed at higher [NH<inf>3</inf>], being more pronounced at lower temperatures. For the C<inf>s</inf> and C<inf>2h</inf> conformers of the NH<inf>3</inf> dimer, calculations at the CCSD(T)/aug-cc-pVTZ level gave a zero-point vibrational-energy corrected binding energy of −7.52 and −7.33 kJ mol^-1, respectively. Energy-grained master equation calculations based on statistical rate theory using the open-source MESMER package were used to calculate rate coefficients for dimerization, k<inf>dimer</inf>, over the temperature range T = 25-300 K and [M] = 10¹³-10²² molecules cm^-3 for He and N<inf>2</inf>. k<inf>dimer</inf> displayed a negative T dependence and a positive [M] dependence and was found to be sensitive to changes in the low-lying vibrational frequencies of the NH<inf>3</inf> dimer, for example, the inclusion of a hindered rotor potential for the internal twisting mode, which alters the density of states. Using the axial profiles of T, [M], and velocity for the Laval nozzles, the calculated values of k<inf>dimer</inf> were used to calculate f<inf>monomer</inf> in the flow for comparison with the experiment. At higher [NH<inf>3</inf>], when dimerization was observed, the calculations significantly underestimated the degree of dimerization taking place in the flow, with a significant increase in the calculated value of k<inf>dimer</inf> required to match the experiment. The reasons for the discrepancy are discussed, for example, errors in the calculation of the density of states for the NH<inf>3</inf> dimer and the average energy removed per collision by the bath gas at very low temperatures.