This difficulty was overcome in the present treatment by introducing the minus sign in (eq. 18). Physically speaking, the minus sign causes every point-group operation to be associated with a proper rotation of the Eulerian angles: if the point-group operation is itself a proper rotation, that same rotation is applied to the Eulerian angles; if the point-group operation is a sense-reversing operation, the proper rotation obtained by multiplying the sense-reversing operation by the molecule-fixed inversion operation is applied to the Eulerian angles. The latter prescription is followed whether or not the molecule-fixed inversion operation or the resultant proper rotation belongs to the point group. The minus sign in (eq. 18) leads ultimately to the minus sign in (eq. 19), and to the correlation of sense-reversing point-group operations with permutation-inversion operations. As pointed out earlier, however, these operations do not actually change the sense of the CH_{4} molecular framework or the sense of the molecule-fixed axis system.
Jahn's scheme suffers from two principal inconveniences. First, it is not actually possible mathematically to satisfy his postulated transformation properties for all spherical-top rotational wave functions simultaneously [13]. This is most easily seen by considering a trigonometric identity in the rotational wave functions, which for a choice of phase factors implied by setting u^{J}_{MK} = |kJm⟩ in (eq. 26), can be written
(eq. 37) |
It can be seen that an application of the molecule-fixed inversion operation i destroys this identity if the J = 1 functions are changed into their negatives while the J = 0 function remains unchanged.
As it happens, Jahn [6] was considering only vibration-rotation energies in the absence of nuclear spin interactions. Thus, he quite naturally chose to work with functions characterized by a given value of m, namely the simplest value m = 0. No inconsistencies need arise if consideration is restricted to m = 0 functions, since no identity analogous to (eq. 37) can be written involving only m = 0 functions. It is clear, however, that any treatment of the symmetry properties of hyperfine levels would be awkward in Jahn's scheme, since such a treatment involves "coupling" values for the quantum numbers m and m_{I} to give a total quantum number F and projection m_{F}.
The second inconvenience in Jahn's scheme concerns the impossibility of assigning a unique symmetry species to the laboratory-fixed components of the electric dipole-moment operator. Jahn's selection rules are such that the laboratory-fixed Z component of the dipole moment operator must be considered to be of species A_{2} when ΔJ = 0 transitions are to be calculated, but of species A_{1} when ΔJ = ±1 transitions are to be calculated. Again, for a molecule in free space and in the absence of nuclear spin interactions, no inconsistencies need ever arise, since J is rigorously a good quantum number. On the other hand, a discussion of allowed transitions including nuclear spin effects or involving perturbations by a magnetic field would be awkward in Jahn's scheme.
The two schemes actually lead to allowed transitions between the same sets of rovibrational levels for the υ_{3} and υ_{4} fundamental transitions, since Moret-Bailly's symmetry labels for the ground vibrational state of CH_{4} agree with those of the present treatment, while his labels for the υ_{3} = 1 or υ_{4} = 1 states differ from those of the present treatment by an exchange of subscripts 1 and 2 for all A_{1}, A_{2} and F_{1}, F_{2} species. While Moret-Bailly's scheme need lead to no inconsistencies, provided it is applied only to situations for which it was intended, it has two principal inconveniences.
The first inconvenience concerns selection rules for electric-dipole transitions. By exchanging all subscripts 1 and 2 in the υ_{3} = 1 and υ_{4} = 1 states, while leaving the υ = 0 state unchanged (with respect to the present treatment), allowed electric dipole transitions between υ = 0 rovibrational levels and υ_{3} = 1 or υ_{4} = 1 rovibrational levels occur only between states of the same symmetry. However, as can be seen from Figure 5, electric dipole transitions within the υ = 0 state [33-35, 48,49], or within the υ_{3} = 1 state [47, 58,59] must still occur between states of different symmetry whose direct product contains A_{2}. Thus, a generalization of Moret-Bailly's treatment to include a discussion of the recently observed pure rotational transitions in methane will be awkward, leading to different rovibrational selection rules for Δυ = 0 transitions than for Δυ = 1 transitions. Complications of this nature may also arise in discussions of an extended set of overtone, combination, and difference bands.
The second inconvenience concerns selection rules for perturbations caused by the Hamiltonian operator. An interchange of subscripts 1 and 2 for one vibrational state, while leaving those for another vibrational state unchanged, does not affect the selection rule that only states of the same symmetry can perturb each other, provided that only interactions within a given vibrational state are considered. When calculating vibration-rotation energies of CH_{4} using a transformed Hamiltonian approach [60], as Moret-Bailly did, such Δυ = 0 interactions are the only ones which need be considered. Brief reflection indicates, however, that interactions between rovibrational levels of different vibrational states may require different selection rules. For example, levels belonging to υ = 0 and levels belonging to υ_{3} = 1 can perturb each other in Moret-Bailly's scheme only if they have different species, i.e., only if their direct product contains A_{2}.
In the author's opinion, the treatment presented in this article, which is based conceptually on clearly specified algebraic transformations of variables and which has neither the two inconveniences associated with Jahn's treatment nor the two associated with Moret-Bailly's treatment is to be strongly preferred over the latter two alternatives, even though adoption of the present treatment will require some changes in the future publications of many authors on the spectrum of the methane molecule.