Measurement of Ammonia Hyperfine Structure with a Two-Cavity Maser

Hyperfine structure of the $J-K=1\mathrm{}-1,2-2,3-3, \mathrm{and} 3-2$ inversion transitions in ${\mathrm{N}}^{14}$ ${\mathrm{H}}_3$ and the 1-1, 2-2, and 3-3 transitions in ${\mathrm{N}}^{15}$ ${\mathrm{H}}_3$ has been measured with a two-cavity maser spectrometer. This device employs Ramsey's method of separated oscillating fields to obtain a molecular resonance linewidth of 350 cps. The theory of Gunther-Mohr et al. and of Gordon has been extended to include all terms off-diagonal in ${\mathbf{F}}_1\equiv {\mathbf{I}}_N+\mathbf{J}$ in an attempt to explain some discrepancies between the previous theory and our measurements. The interactions included in this treatment are the nitrogen quadrupole interaction, the nitrogen ${\mathbf{I}}_N·\mathbf{J}$ interaction, the hydrogen I · J interaction, the hydrogen-nitrogen spin-spin interaction, and the hydrogen-hydrogen spin-spin interaction. The strengths of these interactions are treated as adjustable parameters in least-squares fit programs which determine the parameters by fitting the experimental data. There are still significant deviations between theory and experiment for the 1-1, 3-3, and 3-2 transitions of ${\mathrm{N}}^{14}$ ${\mathrm{H}}_3$ and for the 1-1, 2-2, and 3-3 transitions of ${\mathrm{N}}^{15}$ ${\mathrm{H}}_3$. The largest discrepancies occur for the ${\mathrm{N}}^{14}$ ${\mathrm{H}}_3$ 3-2 transition and the 1-1, 2-2, and 3-3 transitions in ${\mathrm{N}}^{15}$ ${\mathrm{H}}_3$ where the quadrupole interaction vanishes. The discrepancies are greater than 1 kHz in these cases. According to the theory, pairs of satellites of the ${\mathrm{N}}^{15}$ ${\mathrm{H}}_3$ 2-2 transition should occur at the same frequency, and these are all split by 4 kHz. In addition, the values of the ${\mathrm{N}}^{15}$ ${\mathrm{H}}_3$ coupling parameters do not agree with theory.