Collision-induced optical double resonance

In a previous paper, we demonstrated that a set of new resonances can accompany the conventional infrared Lamb-dip or double-resonance transitions in C${\mathrm{H}}_3$F gas. These satellite lines have a different origin from the primary resonances even though they are similar in intensity and linewidth. The traditional double resonance, for example, requires that a molecule interact simultaneously with two radiation fields, causing a transition from an initial to a final state through an intermediate level. Here, the double-resonance concept is extended to the situation of two coherently driven optical transitions that do not share a common level but are coupled by molecular collisions that tip the angular momentum vector while preserving the molecular velocity and rotational energy. Thus, velocity-selective population changes are communicated from one transition to another through collisions. Collision-induced double resonance is observed with Stark tuning as a series of sharp lines, free of Doppler broadening and can be explained in the same order of perturbation theory as the ordinary double-resonance experiment. Each satellite corresponds to a specific level structure involving one or more collision-induced transitions among the space-quantized $M$ states. Virtually all characteristics of these satellite resonances, either in Lamb-dip or double-resonance experiments, are in agreement with the theory presented. While purely optical coupling mechanisms can yield double-resonance behavior also, such effects are estimated to be too weak. For the symmetric top molecule C${\mathrm{H}}_3$F, the dipole-dipole interaction of collision pairs dominates and induces the observed reorienting transitions. The corresponding cross section for low-angular-momentum states ($J,K=4,3\mathrm{} or 5,3$) is found to be ∼ 100 ${\mathrm{A}}^2$ whereas for high-angular-momentum states ($J,K=12,2\mathrm{}$), the cross section is calculated to be about 100 times smaller, and satellite resonances are not even detected. A treatment is presented also for the case where the collisionally coupled $M$ states require a sequential transfer of population over intermediate levels. The problem of velocity smearing is discussed in this context where the energy exchange in a collision is small and also for more energetic collisions that may involve rotational quantum jumps.