Abstract
In a study of internal rotation in methanol, the millimeter wave spectra of CH3OH, CD3OH, and CH3OD have been investigated between 90 and 200 Gc/sec. In the analysis of the spectra, torsion–vibration–rotation interactions were treated as adjustable parameters in semiempirical formulas. Kivelson's formula for a‐type ΔK = 0 transitions was tested over a wide range of quantum numbers. It reproduced the CH3OH and CD3OH spectra quite well, but the approximations used in the calculations appear to start breaking down for the larger asymmetry of CH3OD. For assignment of b‐type ΔK = ± 1 transitions, a method was developed based on the wide spectral range of the millimeter wave spectrometer. Sufficient b‐type data were obtained for CH3OH to permit a test of Kirtman's formula for origins of Q branches. Convergence difficulties in the Q‐branch least‐squares fit prompted a re‐examination of the theory, which revealed an interesting linear relation coupling six of the parameters. This relation shows that for any molecule the torsional barrier terms V3 and V6 , the two moments of inertia about the near‐symmetry axis, and two of the adjustable interaction parameters cannot be independently determined from the spectrum of a single isotopic species. This casts some doubt on values previously reported for V6 in other molecules and adds further uncertainty to V3 determinations. The effective V3 obtained by ignoring torsion–vibration–rotation interactions was found to decrease slightly on deuteration, values of 375.6, 371.8, and 370.3 cm− 1 being obtained for CH3OH, CD3OH, and CH3OD, respectively. Although it is impossible to determine V6 unambiguously from the spectrum of CH3OH alone, a qualitative argument suggests that V6 lies between 0 and −0.8 cm−1. The geometry of the molecule has been completely determined from the experimental data.