LASERS AND REACTIVE COLLISIONS: THE Cs* -H 2 REACTION

The detailed study of the collision between two reacting molecules is the necessary step for the understanding of elementary processes of reaction. Modification, breaking and creation of chemical bonds are determined by a number of parameters, internal energy, kinetic energy, relative orientation of reagents whose con-trol is essential before, during and after the collision. This is the aim of state-to-state chemistry to follow as closely as possible the evolution of these parameters and to evaluate the indi-vidual cross-sections. determined product activation angular analysis product dictribution information about geometry

es of reaction. Modification, breaking and creation of chemical bonds are determined by a number of parameters, internal energy, kinetic energy, relative orientation of reagents whose control is essential before, during and after the collision. This is the aim of state-to-state chemistry to follow as closely as possible the evolution of these parameters and to evaluate the individual cross-sections.
During the last two decades, the search for an "ideal" experiment where the whole set of parameters could be determined has made considerable progress with the advent of new sophisticated techniques. Among them, the technique of supersonic molecular beams offers unique possibilities, in particular with the good definition of internal and external energies of reagents which can be achieved. Use of two srossed beams of fixed geometry leads to product analysis after a unique collision has taken place, i.e., the analysic is free of complications due to successive encounters; variation of the kinetic energy allows for the determination of activation barriers; angular analysis of the product dictribution yields information about the geometry of the reactive collision.
Although extremely poweful, this technique is generally restricted to atoms and molecules in their ground state. Furthermore, chemical systems of interest for a useful comparison between theory and experiment are still relative to simple ones: this is due to the fact that the methods of Quantum Chemistry are able to provide accurate potential energy surfaces for systmes implying a small number of atoms only, typically up to three or four. Moreover, the dynamics of collision of these surfaces is still in its infancy since, even for simple systems, one is able to evaluate state-to-state cross-sections for a limited number of circumstances only, a collinear approach of non-rotating reagents for example.
In this context, use of tunable lasers is obviously of great interest for state-to-state chemistry. To schematize, state-selective excitation of reagents (electronic, vibrational, rotational) is easily accomplished and the subsequent effect on reactivity can be observed; second, large amounts of energy can be deposited into the system, hence the possible study of endoergic reactions; third, polarization of laser beams allows for the study of the orientation dependence of the cross-section; fourth, photodissociation products and unstable species (radicals) can be obtained in the zone of collisions by use of intense (pulsed) laser beams. Finally, the technique of laser-induced fluorescence offers two advantages in the detection of nascent products: a quasi-infinite energy resolution and a high sensitivity.
It comes therefore that the combined use of molecular beams and laser beams approaches the conditions required for an "ideal"  to the 7p state, the amount of energy which is available above threshold is drastically small:0.0016 eV and 0.024 eV for 7P I/2 and 7P3/2 respectively. Furthermore, the initial system (Cs* + H 2) is at least on the llth potential energy surface from the ground state, implying that many surface crossings must necessarily occur during the reaction. Under these conditions, is a unique collision able to promote the direct reaction: Cs* + H 2 CsH + H with simultaneously electronic deexcitation and bond breaking, or is it necessary to invoke more complicated processes whose energetics is more favorable? For instance, one could think to the foi-28 lowing two-collision processes  In the second chamber, which is maintained at low pressure by highvelocity pumping, a region of hydrodynamical regime is created, No signal could be detected for transitions involving the v"=l levels. The surprising result of these measurements however, is in the fact that for a 7PI/2 excitation of Cs atoms, the signal is larger than for a 7P3/2 one. Figure 4 shows an example of the data which has been obtained on the (v'=5, J'=7 v"=0, J"=6) transition. The signal is multiplied by a factor 6 by simply changing the wavelength of the first laser from A 4555 to=4593 .
According to the ratio of excited populations of cesium (n 7P3/2/ n 7PI/2 I.7) which can be deduced from the measurement of atomic fluorescence, this means that the reactive cross-section is roughly 10 times larger for a 7PI, 2/ excitation of Cs atoms. and (6s). In this manner, the suggested mechanism of reaction should be a type of harpooning leading to a molecular "zwitterion" and one atom. The interesting feature is that Cs(7Pl/2) with H 2 2+ generates states which have the same symmetry as that of the Cs(7P3/2) generates 2 states which are not ionic states, whereas relevant for the jump onto the ionic surface, in a collinear approach of the reagents. Similar considerations were already sug-33,34 gested to explain this "fine structure" effect Other data can be extracted from these measurements. First, observation of a signal for J" values up to 13 in the v"=0 level (see Figure 3) indicates that the relative kinetic energy is necessarily transformed into rotational energy during the course of the collision since the only potential energy is not sufficient. present experimental uncertainties, this value must be considered as an order of magnitude only, but the same experimental data indicate that a two-collision process would require cross-sections -13 2 of the order of I0 cm for each collision to lead to the same signal. Consequently a two-collision process is rather unlikely to occur in this experiment.

Possible Developments
As explained above, determination of the rotational distribution of products is of major importance in such sytem at threshold. Second, comparison berween the efficiency of photon excitation versus kinetic energy is made possible by use of the technique of "seeded" beams; addition of a light carrier gas in the beam of cesium should allow, at least partially, for compensation of the endoergiticity of reaction. Also, vibrational excitation of hydrogen could be possible by electronic bombardment; unfortunately, this technique is not yet selective in energy.
Third, the technique of Doppler tuning developed for atomatom scatering could yield the angular distribution of products.
In this technique, the laser beam which induces the fluorescence of CsH molecules is sent into the collision chamber, in the direction of relative velocities at the center of mass of the system.
Then, there is a unique relation between the angle of deflection of CsH molecules and the detunign of the laser with respect to line-center: only respond those molecules whose velocity projection on the laser beam obeys the Doppler relation. However, this technique suffers from an important loss in sensitivity due to the angular resolution it-self.
To conclude, the crossed-beam experiment with laser excitation of reagents and laser detection of products which has benn descibed here, appears to have versatile applications in many aspects of reactive collisions. It should be emphasize once more that the necessary need for interpretation implies Quantum Chemistry calculations which are still limited tD simple situations. The author is pleased to acknowledge with thanks his colleagues of Laboratoire Aim4 Cotton, Orsay, and Laboratoire de Physique Quantique, Toulouse, for participation to the experiment and interpretation of the data.