REACTION DYNAMICS OF ELECTRONICALLY EXCITED BARIUM ATOMS WITH FREE MOLECULES AND MOLECULAR CLUSTERS

In this review article we describe some recent results obtained in our laboratory. The successful combination of crossed molecular beam techniques and various laser excitation schemes has been used to study chemiluminescent reactions of ground and excited electronic states of barium with free molecules and molecular clusters. Studies include the identification of reaction products in cases where many chemiluminescent reaction channels are opened. The case of Ba(6s6plp1, 6s5d1D2, 6s5daDi) reacting with H20, methanol, ethanol, propanol-1, propanol-2, methyl-2, propanol-2, butanol-l, allyl alcohol, dimethyl ether, diethyl ether and diallyl ether is examined. A reaction mechanism is proposed which accounts for all these reactions. Studies reported in this review also include the unravelling of reaction dynamics where various forms of energy are mixed (electronic and kinetic energy). This is shown in studies of Ba( 1D2 and P) + 02 reactions. Finally the role of molecular clusters as reactant is examined. Evidence is provided that clusters of N20, H20 and CO2, in collision with Ba(1So, and 1P1), do not lead efficiently to both reactive and non reactive luminescent exit channels. The various chemiluminescent processes that are accessible energetically in Ba(6s6plp 1, 6s5d1D2, 6s5d3D)+H20 collisions at 0.25 eV are the following. Reactions from the 6s6p configuration of barium:


I. INTRODUCTION
Reactions of ground state alkaline earth atoms with oxygenated compounds have led to many studies during the seventies that are reviewed in Ref. 1. In particular, the detailed reaction dynamics of several reactant pairs has been investigated using crossed molecular beam or beam/gas equipments operating under the single collision regime. Since the late seventies, this field has moved towards more unusual situations. The first research direction is to investigate the reactivity of electronically excited alkaline earth atoms. [2][3][4][5][6][7] This very interesting field is rapidly growing since it is believed to bring new forms of reactivity for the alkaline earth atoms. As is well known indeed in photochemistry, the electronic excitation offers a unique possibility of changing the energy and symmetry of the valence electrons that participate to the reaction. This field profits also by the rapid advancement of laser technology and by the fruitful combination of crossed molecular beam and laser techniques. The second research direction is linked to the exploding amount of works devoted to cluster physics. It concerns the reactivity of ground state and excited state alkaline earth atoms with molecular clusters. [8][9][10][11][12] The interest of these studies is that molecular clusters may be considered as allowing for reaction dynamics studies in a solvated environment.
Over the past two years, this laboratory has been examining these situations using an apparatus which combines three techniques: crossed molecular beams, cw laser excitation and optical detection. The possibilities offered by the apparatus are reviewed in Section II together with a short description of the experimental setup. The paper then reviews some of the recent work performed in this laboratory. This concerns reactions of excited state barium with free molecules O2, H20 and various organic molecules (Section III), and reactions of ground state and excited state barium with molecular clusters N20, CO2 and n20 (section IV).

II. EXPERIMENT
The crossed beam machine and the techniques used for data acquisition and analysis have been described extensively in our former publications. 5'12'13'4 Various types of experiments can be performed with this machine. i) Analysis of chemiluminescence spectra resulting from reactive collisions as a function of the collision energy over a wide range extending between 0.1 and 1 eV. This is the main topic of the review.
ii) Laser induced detection of non luminescent reaction products. A preliminary result is shown here for the Ba(6s5dlD2) + H20 reaction.
iii) Total cross section measurements of reactive and non reactive luminescent processes as a function of the collision energy. Collisionally induced inelastic energy transfers to the 6s6p3p1 level of barium are reported here. iv) Differential cross section measurements as a function of the collision energy for non reactive processes. Laser induced fluorescence is combined with Doppler profile analysis for this purpose. (Not reported here, see Ref. 15). v) Studies of polarization effects in collision processes. (Not reported here, see Ref. 15).
The number of reactant pairs that can be studied with this machine is very large. The primary beam can operate with any alkaline or alkaline earth metal atoms. The secondary beam can support experiments with any reagent that is gaseous or liquid at room temperature. The only true limitation is that the reaction product must be either luminescent or suitable for laser induced chemiluminescence detection.
For the experiments reported below on excited barium atoms in reactions, the electronic excitation of Ba was performed by cw laser excitation with the laser tuned to the barium resonance line at 553 nm. Two excitation schemes were used. The first one is with the laser crossing the barium beam upstream of the reaction zone. Because of the optical pumping cycles schemed in Figure 1, most of the barium atoms are transferred radiatively to the metastable levels 6s5dXD2 and 6s5d3Di, and the reactivity of these levels is investigated. The second excitation scheme is with the 0000 10000 arlum Energy Diagram laser beam sent into the crossing zone of the particle beams. The reactivity of Ba(6s6plp1, 6s5dlD2, 6s5daD) is then explored.

III. REACTIONS OF ELECTRONICALLY EXCITED BARIUM ATOMS
The reaction dynamics of metal atoms is expected to be different whether the metal atom is electronically excited or not. The large excitation energy associated to electronic transitions and the change in the nature of the potential surface initiating the reaction are both enough reason to justify this expectation. The reactivity of excited Na atoms has been reviewed recently, and we recall here some of the more important aspects. 16 For instance in the case of Na reacting with HC1, a substantial change in the reaction mechanisms has been observed when the sodium atom is excited from the ground electronic state 3s to chosen excited states of higher and higher energy 3p, 5s and 4d. 17 Moreover, the excited Na + O2 reaction has shown a state selective behavior. The Na(4d) atoms react with O2 whereas Na(5s) atoms do not. TM A mechanism has been proposed for the Na(4d) / O2 reaction which involves a Na + O2-(AElIu) intermediate. 19 Some aspects of the scattering of electronically excited Cs atoms are also discussed in Ref. 20.
The chemiluminescent reactions of electronically excited alkaline earth atoms renew the subject a great deal, as compared to reactions of excited alkali atoms. Two valence electrons then participate to the reaction. In the entrance channel this fact offers possibilities of excitation to atomic levels with a rich variety of electronic configurations. In the exit channels several products with various electronic configurations are often accessible energetically. Colliding electronically excited alkaline earth atoms with molecules thus allows to examine collision dynamics where many reactive and non reactive channels are coupled and compete one with each other.
Studying reactions of alkaline earth atoms has also an experimental advantage as compared to studying reactions of alkali atoms. Chemiluminescent reaction channels are opened that allow the internal state of the reaction product (both vibrational and electronic) to be checked directly through the analysis of chemiluminescence spectra. This has been widely used for reactions of ground state atoms and is used again here for reactions involving electronically excited barium atoms. Among many interesting questions concerning the reactivity of excited barium atoms, the present article focusses on the following two" 1) What is the choice of the system when several chemiluminescence channels are opened?
2) Through which mechanism proceeds the reaction, when chemiluminescence is turned on by laser excitation.
III. 1 What is the Choice of the System when Several Chemiluminescence Channels are Opened? Interpretation in Terms of Reaction Mechanism This question was posed for the first time by Rettner and Zare in a situation where two different excited electronic states of the same reaction product were accessible (the states A and B of CaC1 in the reaction Ca(1P1) + HC1). 21 They showed that both states A and B are produced, but the branching ratio is affected by the alignment of the reacting 1p1 orbital.
The present article formulates the question differently. Situations are considered where several chemiluminescent reaction products are accessible energetically. Further questions then arise: are all the accessible reaction products formed, or is there selection amongst them? What is the electronic state of the products that are formed? In what follows, answers are given for excited Ba+H20, ROH and ROR reactions (R being an alkyl or allyl radical, and Ba excited to one of the levels 6s6p1p1, 6s5d1D2 and 6s5d3Dj). These answers allow us to propose a mechanism that accounts for all these reactions. 111.1.1 Ba(6s6plp1, 6s5d1D2, 6s5d3D2) / H20 The various chemiluminescent processes that are accessible energetically in Ba(6s6plp1, 6s5d1D2, 6s5d3D)+H20 collisions at 0.25 eV are the following. Reactions  Ba--OH 4.6 eV. 23 The luminescence spectrum shown in Figure 2 was obtained with the excitation scheme of barium where the laser beam is sent inside the reaction zone directly. Results of Figure 2 thus correspond to collisions of barium atoms excited to the 6s6p1p1, 6s5d1D2 and 6s5d3D levels. The collision energy is 0.25 eV. Assignments of the features observed in this spectrum has been discussed extensively in Ref. 6 The intense feature at 791 nm is the intercombination line Ba(6s6paPl --6s 2 1So). The 3P state of barium was populated by the following two collision processes: Ba(6s6plp1) + H20 Ba(6s6pap1) + H20 AH -0.67 eV (5) Ba(6sSdlD2) + H20 Ba(6s6pap1) + H20 AH +0.15 eV The other features at 755,832 and 875 nm in Figure 2 are assigned to chemiluminescence of BaOH through the transitions B2E + X2E +, A2I'I3/2 X2E + and A2I'I/2 X2E + respectively. No feature in Figure 2 can be assigned to BaO chemiluminescence. As a consequence, among the processes 1-4 given above, only 2a, 2b and 2c contribute to the chemiluminescence. These results are complemented by other information based on laser induced fluorescence detection. We have shown that the ground state BaOH is formed from Ba(6s5dlD2, 6s5d3Di)+ H20 collisions. A full account of this work will be published later. Related experiments have been performed at Berkeley using a crossed molecular beam machine and detecting the angular and velocity distribution of both ground state and excited state reaction products. Ba(6s5dD2, 6s5d3D) has been made react to H20 at various collision energies. Among the two energetically allowed products BaO and BaOH, only BaOH has been detected. 24 The observations made so far on excited Ba +H20 reactions can be summarized as follows: i) The excited levels 6s5d1D2 and 6s5d3D of Ba do not lead to observable chemiluminescence although chemiluminescence channels forming BaO are allowed energetically.
ii) The only reactive channel for the excited levels 6s5d1D2 and 6s5d3Dj of Ba is the formation of ground state BaOH. Formation of ground state BaO is not observed.
iii) The level 6s6p1p1 is the lowest excited state of Ba which leads to chemiluminescence. The chemiluminescence originates from BaOH(A2I-I1/2,A2II3/2, B2E 1/2) only, and not from the energetically favored channel forming excited BaO. 111.1.2 Ba(6s6plP1, 6s5d1D2, 6s5d3Dj) + alcohols Chemiluminescence of BaO, BaOH and BaOR is allowed energetically from reactions of excited barium with alcohols ROH (R being an alkyl or the allyl radical). Let us now see which product is formed. Figure 3 shows the luminescence spectra resulting from reactions of methanol and water with excited barium atoms. The barium atoms are excited with the pump laser beam sent into the reaction zone. Results of Figure 3 thus correspond to reactions of Ba(6s6plp1, 6s5d1D2, 6s5d3Dj). The collision energy is 0.25 eV. The dashed curve corresponds to the reactant CH3OH, and the full curve to H20. The curve for H20 is drawn for comparison purposes. It duplicates that shown in Figure 2.
The two spectra observed in Figure 3 are qualitatively similar. In particular the intense feature at 791 nm exists for both reactant indicating that both methanol and water are able to induce singlet-triplet transfers in Ba. Of more interest is the small but significant differences that exist between the two spectra: i) The feature at 755 nm exists for H20 only, ii) The feature at 832 nm for H20 is blue shifted to 826 nm for CH3OH, and its width is smaller for CH3OH than for H20. iii) The blue part of the feature at 875 nm shown in Figure 3 suggests also a difference in width and/or position between the two spectra. These differences are at the limit of the 12 nm spectral resolution of the experiment shown in Figure 3. They were confirmed by a 0.2 nm resolution experiment that was performed under a beam-jet configuration of the machine. 25 From the high resolution experiment, it can be concluded that the chemiluminescent emission with CH3OH peaks at 825.9 nm. It is assigned to the A-X transition of the methoxide BaOCH3. 26 Other experiments were run for a number of alcohols ROH (ethanol, propanol-1, propanol-2, methyl-2 propanol-2, n-butanol, and allyl-alcohol). In each case, the chemiluminescent product was assigned to the alkoxide BaOR rather than to BaOH or BaO. 111.1.3 Mechanism for excited Ba+H20, ROH reactions A reaction mechanism is discussed that accounts for the above observations. Formation of BaO has not been observed in reactions of excited barium with neither water nor alcohols. This is most likely due to the fact that formation of BaO would involve a substantial bond rearrangement. For instance, with water, formation of BaO supposes that the two OH bonds are broken and that two other bonds BamOH and HmH are formed. This is expected to induce barriers on the potential surface that hinder a reaction path leading to BaO formation. The reaction of excited barium either with water or alcohol thus follows another path.
A likely mechanism has been proposed in Ref. 6 for reaction with water. The reaction is initiated by an adduct: H Ba... O ( H The adduct converts by H migration into the insertion product HmBa--OH. Since H--BamOH has a large excess energy it fragments into BaOH+ H by rupture of the BaH bond. Let us mention that formation of BaH+OH by the other fragmentation path of HBaOH is not allowed energetically, and thus cannot be observed. This mechanism is easy to extend to alcohols. The collision of excited Ba with an alcohol ROH creates an adduct similar to (7):

R
The essential difference between the adducts (7) and (8) is that (8) is not symmetrical, and that two different migrations may be imagined. That of H, forming the insertion product HmBaOR, and that of R forming HO--Ba---R. The experimental observation shows that the alkoxide is formed preferentially. This implies that the insertion product HBa--OR predominates over HO--BamR. Our results thus show that migration of H predominates over that of R. In other words, excited barium atoms react with alcohols R--O--H through an insertion mechanism, and are more likely to insert into the O--H bond rather than into the R--O bond.
The difficulty of barium to insert into an O--R bond to form chemiluminescent products has been checked directly by reacting barium with dimethyl, diethyl and diallyl ethers (RO--R with R CH3, C2H5 and CH2 CHCH2 respectively).
In all cases, although formation of chemiluminescent products BaOR is allowed energetically, no chemiluminescence assignable to ether reactions were observed.
Interestingly, the present reaction mechanism that is proposed for chemiluminescent reaction channels of Ba(Os6plp1, 6s5dlD2 and 6s5daD) also applies for non chemiluminescent channels in Ba(6s5dlD2, 6s5daD) reactions. Exclusive or quasiexclusive formation of BaOCH3 is seen indeed in Ba(6s5dXD2 6s5daD) reactions with CHaOD, whereas no reaction product was seen in reactions with dimethyl ether. 24 These observations are in accordance with the expectations of the above model.

Through Which Mechanism Proceeds the Reaction, when Chemiluminescence is Turned on by Laser Excitation
Reaction of ground state barium with 02 is a typical example where the reaction goes through the formation of a long lived intermediate, although the Ba/O2 systems has few degrees of freedom. [27][28][29] The existence of the intermediate is interpreted by the occurrence of a deep well along the reaction path. 29 The present review aims to discuss whether this mechanism adequately describes chemiluminescence from Ba(6s6plp1, 6s5dlD2 and 6s5daD) + O2 reactions.
111.2.1 Ba(6s5dZD2, 6s5d3D) + 02 Table 1 shows that even for the low excited configuration 6s5d of Ba, many luminescent channels are opened above 0.13 eV collision energy. There has been some ambiguity whether all these channels participate to the chemiluminescence. It was thought indeed that chemiluminescence from BaO should originate from the excited singlet terms A12 + and A'H only. Our recent work 7 has brought the evidence that triplet states also participate to the chemiluminescence. Therefore, all the processes listed in Table 1 which are accessible energetically must be considered in order to account for the observed chemiluminescence spectra.
The observations can be summarized as follow: i) Above 0.15 eV collision energy, collisional population of Ba(6s6p3p) is accessible energetically, and the intercombination line 6s6p3p1 6s 2 xS0 is observed in the spectra.
ii) The remainder of the spectra is due to the chemiluminescent emission of BaO. A vibrational structure is clearly visible at collision energies between 0.11 and 0.32 eV. It is washed out at the three highest collision energies explored. experiments with those reported above it is possible to extract the contribution of 6s5d1D2 and 6s5d3D and get information on chemiluminescence from reactions of pure 6s6p1p1 barium atoms. Some results directly relevant to the reaction mechanism are reported here. A full account of the work will be reported later.
The experiments were performed between 0.1 and 1 eV. One chemiluminescence spectrum is shown in Figure 5 for the 0.11 eV collision energy. The spectrum is much less structured than the one shown in Figure 4 at the same collision energy with Ba(6s5dlD2) reacting. This results from the 0.83 eV increase of available energy when switching from 6s5d1D2 to 6s6pIP1.

Mechanism for excited Ba+02 chemiluminescent reactions
The spectra of Figure 4 and 5 together with those obtained at the other collision energies explored contain the desired information about the reaction mechanism and about its evolution with the collision energy. We have shown "in Ref The phase space model is relevant for discussing the excited Ba+O2 reactions. The details about phase space calculations are given in Ref. 7. However, it is important to recall that the model assumes the formation of a long lived collision complex, and that it assumes dissociation of the complex with equal probabilities among all the accessible quantum states of the system. The dynamical constraints of the model are minimum. They include conservation of energy and total angular momentum. No potential barrier, except centrifugal barrier, is assumed by the model to prevent the formation and the dissociation of the collision complex.
The first important result is that phase space calculations for Ba(6s5dlD2) + O2 collisions up to 0.58 eV reproduce the electronic and rovibrational state distributions that fit the observed chemiluminescence spectra. 7 A side result may suggest that predictions of the phase space model depart from the experimental observations above 0.6 eV collision energy.
The second important result is that phase space calculations never reproduce the experimental spectra for Ba(6s6pXPa) + 02 collisions. At this point, we are left with the conclusion that reaction of Ba(6s5dD2) with O2 leads to excited BaO molecules with a phase space distribution of internal and electronic states for collision energies up to 0.6 eV. This strongly suggests that the reaction goes through a long lived complex as assumed by the phase space model. In contrast, the mechanism is different for chemiluminescent reactions of Ba(6s6plp1) with 02. Two reasons may be invoked. Either the Ba(6s6plp)/02 system never really goes through a long lived complex, or the branching to chemiluminescence in the exit channel involved couplings that put a bias on the statistical energy distribution.
IV. REACTIONS OF GROUND STATE AND ELECTRONICALLY EXCITED BARIUM ATOMS WITH MOLECULAR CLUSTERS Large van der Waals clusters have attracted a growing interest over the past few years. A recent Faraday Symposium devoted to "Large Gas Phase Clusters" has shown the current interest in this field. 3 Most papers were concerned with (i) the fragmentation processes and intracluster reactions that follow ionization or electron attachment, (ii) structural properties and spectroscopy, and (iii) phase transitions.
All these topics concern "internal" properties of the van der Waals clusters.
Very little work has been performed on reactions of large van der Waals clusters with atoms and molecules. This is very surprising considering the large interest of such processes. They indeed yield information about chemical reactions in "solvated media". Pioneering work in the field was done on large CH3I clusters reacting with Rb. 8 The reaction product was identified as a heavy product, possibly a solvated RbI molecule or a long-lived Rb(CH3I)n complex. More recently, reaction between barium atoms and van der Waals clusters have been investigated, but that concerned van der Waals dimers of NO2, SO2 and CO2. 9-11A short reference is done to reaction of Ba with large CO2 clusters in which also suggests the existence of heavy solvated reaction products. Reactions of ground state and electronically excited barium atoms with large van der Waals clusters have been investigated in this laboratory. 12 The clusters were generated by the intense cooling due to the supersonic expansion of the molecular beam source. Low temperature, and large pressure in the stagnation chamber generating the beam stimulates cluster formation. It was found that the yield of cluster formation is tremendously larger when expanding a mixture of argon with the molecular gas that is to be clusterized. Argon is playing the role of an efficient cooling agent. In particular, by optimizing the expansion conditions of H20/Ar mixtures, it was possible to create beams containing argon and large H20 clusters, with no water monomer present in the beam. With the beam operating under such a "large cluster only" regime, one can estimate the size of the large clusters as 10 monomers or more.
Control of cluster abundance in the beam is an important question for these experiments. It is based on time-of-flight measurements of velocity distributions in the beam. The method is thoroughly explained in Ref. 31. Let us just recall that it allows us to determine the flux of free monomers remaining in the beam after the supersonic expansion.
We have shown that the reaction of ground state barium with large N20 clusters has a much weaker chemiluminescence yield than reaction with NzO monomers. A similar observation was done with Ba(6s6plp1) colliding large CO2 and H20 clusters, since no chemiluminescence was detected in these experiments. Interestingly, collisions with large CO2 and H20 clusters do not induce the transition 6s6p1p1 6s6p3p1 that is observed with CO2 and H20 monomers.
The results with N20 have been interpreted in Ref. 12 on the ground of electron attachment properties of N20 clusters. 32 Figure 6 pictures that (i) dissociative electron attachment to N20 clusters is possible with low energy electrons, and (ii) formation of solvated Oions is energetically the most favorable exit channel in dissociative electron attachment to N20 clusters. These two properties then allow us to propose that barium atoms colliding N20 clusters, enter into the cluster, and propose that reaction of barium with N20 clusters results in formation of a heavy product which probably is also BaO solvated by N20 molecules. The experiment shows that this product is not formed in a chemiluminescent state that emits inside the spectral range 400-900 nm where chemiluminescence from monomers is observed.
Interpretation of the collisions between barium atoms and clusters of CO2 and H20 goes along the same model. We are thus left with the conclusion that formation of heavy solvated products in these collisions is a fairly general situation. These observations and the interpretation model proposed here are entirely consistent with the results of Refs. 8,10 that are recalled above. Further work is in progress in this laboratory to analyse in more detail the formation mechanism of the heavy solvated compounds.

V. CONCLUSIONS
We provided a description of the various types of experiments that can be performed with the crossed molecular beam machine of this laboratory. The reactivity of excited barium atoms with a number of molecules and molecular clusters were reviewed, but no extensive presentation of the results was given since they are the subject of other publications.
The first subject presented here showed that reactions of excited barium atoms with water and alcohols are accounted by the same reaction mechanism where barium inserts into an OH bond. This mechanism accounts for the absence of (or very small) reactivity of excited Ba with ethers.
The second subject presented showed to which extent adding electronic energy to a reactive system changes the reaction mechanism. The Ba+ 02 reaction was chosen as an example. Ground state barium reacts through a long lived complex. Turning on chemiluminescence of the product BaO by exciting barium to the 6sSd1D2 level lets the reaction go through the same mechanism, at least if the collision energy is smaller than 0.6 eV. In contrast, the reaction mechanism changes when the chemiluminescence is induced by exciting barium to the 6s6paP1 level.
Finally, it was shown that large van der Waals clusters of N20, CO2 and H20 do not induce measurable reactive and non reactive luminescent processes when colliding Ba(6s 2 1S0, 6s6paP1) atoms. This was interpreted in terms of collisional formation of heavy solvated compounds.