Abstract
The infrared and Raman spectra of the toluenes C6H5CHD2 and C6D5CHD2in the aliphatic CH stretching mode range have been recorded in a large temperature range (17 to 165 K) for both crystalline phases α and β. At very low temperature, the β form spectra show three bands; each of them is assigned to the vibration of a CH oscillator localized in a different site. Three groups of bands are also observed in the α phase spectra: a single band at higher frequency and two doublets at lower frequency. This splitting is assigned to the existence of two types of molecules in the unit cell, involving six different CH vibrators. A quantum theory of these spectra is carried out, assuming an anharmonic coupling of the CH stretching mode with the CHD2 torsion. As a consequence of this coupling, in the adiabatic approximation, the vibrational energy depends on the conformation and can be considered as an additional torsional potential. This latter has no ternary symmetry so that the total torsional potential has three principal unequal wells that correspond to three different locations of the CH oscillator. Therefore, no tunneling effect appears, which is in agreement with the classical interpretation. Furthermore, this theory ascribes the temperature dependence of the relative intensities of the νCH bands to the population density of the first torsional levels in the vibrational ground state and suggests that, at very low temperature, the isotopic system gets ordered. At higher temperature, a strong relaxation of the νCH vibration bands is observed. This relaxation is much stronger than that of the aromatic ring modes. Thus the relaxation process is essentially due to the influence of the anharmonic coupling between the CH stretching mode and the τCHD2 mode. Two mechanisms are considered: the first one involves Markovian jumps of the system from an equilibrium position to another one, the second one involves fluctuations of the CH vibration around each of these equilibrium positions. NMR and neutron scattering data have already been analyzed on the basis of the first process. Starting from the residence times so determined, the computations show that this mechanism is an efficient relaxation process, but indicate that it is not sufficient to fit the experimental profiles. This fit is obtained rather by using the second model with parameters of reasonable physical values; thus, the second process is also efficient. A better treatment of the relaxation process would be to elaborate; it would have to include both mechanisms and to take into account motions of the methyl group with different amplitudes.