Combined Infrared and Microwave Determination of Torsional Parameters

Considerable discrepancies are frequently observed between the values of torsional barriers determined by microwave splitting and those obtained from far‐infrared transitions, if the usual rigid‐top–rigid‐frame model is employed. If a nonrigid model which includes first‐order coupling with the internal vibrations is used, the requirement of consistency between the microwave and infrared data permits barrier Fourier coefficients $V_3$ and $V_6$ to be evaluated without any prior choice of a value of $I_{\mathrm{red}}$ , the reduced moment of the top. This has hitherto been the principal source of uncertainty in the determination of barrier parameters. The torsional parameters ${\mathrm{F,\; I}}_{\mathrm{red}}\mathrm{,\; S,\; \alpha ,}\mathrm{and}\beta$ are also independently evaluated; from these $I_{\alpha }$ is determined. Comparison of the “self‐consistent” and “a priori” values of parameters such as $I_{\alpha }$ indicates the extent of interaction with other normal vibrations. The limitation of the treatment involved by the neglect of second‐order vibrational interaction terms, such as centrifugal distortion, can be assessed in cases where microwave data from torsionally excited states is available. The method is applied to CH₃CHO, CD₃CHO, and CH₃CH = CH₂, for which accurate far‐infrared spectra have been newly determined.