Absolute Infrared Intensities of the Fundamental Absorption Bands in Solid CS2

Absolute infrared intensities have been measured for ν2 and ν3 of CS2 in the solid phase. Path lengths were measured by counting interference fringes as the solid is condensed onto a cold window. The index of refraction of the solid film is estimated from the intensity of the fringe pattern on different windows of known index of refraction. The problem of choosing a suitable background is discussed. Although two different kinds of films were formed, the intensities were identical within experimental error. The intensities found for ν2 and ν3 were 850±100 and 80 000±14 000 darks, respectively. These values are quite comparable to the intensities found in liquid CS2, and are both higher than found in the gas phase. Hence, the intensity sum rule does not hold for this crystal. The observed intensification is compared with that predicted from the simple field effect, using an elliptical cavity, and found to agree quite well. Hence, we conclude that CS2 molecules do not interact strongly in the crystal in ways which affect the intensity. The ``ideal'' behavior is tested further in a comparison of the observed frequency shift with that calculated from the Kirkwood—Bauer—Magat theory; again the agreement is quite good for ν3 where the predicted shift is large. For ν2, with a smaller predicted shift, the agreement is poor, indicating that other forces are at least as large. Finally, the infrared spectra and x-ray powder pattern are used in combination to estimate the crystal structure. Both are consistent with an orthorhombic space group D2h12, with two molecules per unit cell (a=5.45, b=8.16, and c=3.74 Å). Although this space group is not determined uniquely, the crystal splitting was calculated from the dipole—dipole model, assuming that the structure was approximately correct in locating the molecules. The calculated splitting for ν3, where it is large, agrees very well with the observed splitting; for ν2, where the calculated dipole—dipole splitting is small, the agreement with experiment is poor.