Faraday Discuss. Chem. SOC., 1987, 84, 145-157 Reactive Scattering of Electronically Excited Alkali-metal Atoms with Molecules J. M. Mestdagh,"? B. A. Balko, M. H. Covinsky, P. S. Weiss,S M. F. Vernon,§ H. Schmidt7 and Y. T. Lee Materials and Molecular Research Division, Lawrence Berkeley Laboratory, and Department of Chemistry, University of California, Berkeley, California 94720, U.S.A. Representative families of excited alkali-metal reactions have been studied using a crossed-beam apparatus. For those alkali-metal-molecule systems in which reactions are also known for ground-state alkali metal and involve an early electron-transfer step, no large differences are observed in the reactivity as Na is excited. More interesting are the reactions with hydrogen halides (HCl); it was found that adding electronic energy into Na changes the reaction mechanism. Early electron transfer is responsible for Na(5S, 4 0 ) reactions, but not for Na(3P) reactions.Moreover, the NaCl product scattering is dominated by the HC1- repulsion in Na(5S, 4 0 ) reac- tions, and by the NaCl-H repulsion in the case of Na(3P). The reaction of Na with O2 is of particular interest since it was found to be state-specific. Only Na(4D) reacts, and the reaction requires restrictive constraints on the impact parameter and the reactants' relative orientation. The reaction with NO2 is even more complex, since Na(4D) leads to the formation of NaO by two different pathways. However, the identification of NaO as product in these reactions has yet to be confirmed. '' For a given elementary chemical reaction, specific forms of energy (i.e.translational, vibrational, rotational, and electronic) deposited in reactants are known to affect reactive scattering processes in different ways. Adding electronic energy to reactants is of special interest since in this way a relatively large amount of energy can be deposited in a single excitation step and, more importantly, the potential-energy surface on which the reaction is initiated can be selected by exciting reactants to states of different symmetries. The reactivity and reaction dynamics of such excited reactants often differ substantially from those of the corresponding ground-state molecules.' Reactive scattering of alkali-metal atoms with molecules has been studied extensively over the past two decades.* A wide variety of experimental techniques has been used to determine the dynamics of these reactions, as well as the effect of reactant translational, vibrational and rotational e~citation.~ Until recently the effect of electronic excitation remained The reactions of excited Na atoms with simple molecules have been systematically investigated in our laboratory using the crossed-molecular-beam m e t h ~ d .~ - ~ In this paper some examples of excited Na atom reactions will be presented to illustrate the effect of electronic excitation on sodium reactivity. t Permanent address: Service de Physique des Atomes et des Surfaces, CEN Saclay, 91 191 Gif sur Yvette $ Present address: AT & T, Bell Laboratories, f 1C-402, 600 Mountain Avenue, Murray Hill, NJ 07974, fi Present address: Department of Chemistry, Columbia University, New York, NY 10027, U.S.A.ll Permanent address: Braun A.G. Forschung, Frankfurter Str. 145, D6242, Kronberg, Federal Republic Cedex, France. U.S.A. of Germany. 145146 Reactive Scattering of Excited Alkali-metal Atoms molecular beam source Na beam secondary rotating molecular polarization detector I mass spectrometer b e a y rotator Fig. 1. Schematic diagram of the crossed-molecular-beam apparatus. First, systems are considered in which the reaction is exothermic and is initiated by an electron transfer from a sodium atom to a molecule; a representative example of this is the reaction In reactions ( l ) , n,l represents the electronic state of Na: ground-state 3 s or excited states [e.g.Na(3P)I. The AH: is for the reaction of Na(3S). The second type of reaction is where ground-state Na is known to react, but a simple electron-transfer model cannot be used to account for observed features of the collision dynamics. The chosen example is where the excited state n,l is 3P, 5s or 4 0 . The points which will be considered about processes ( 1 ) and (2) concern changes in the Na reactivity as well as changes in the reaction dynamics that are associated with depositing electronic energy into the Na. The remainder of the examples concern reactions which are substantially endothermic and energetically unfavourable for ground-state Na. The goal is to understand if and by what mechanism electronic excitation is able to turn on the reaction.One example that will be discussed is Na( n,l) + C1, --* NaCl+ C1, AH: = -40.4 kcal mol-'. ( 1 ) Na( n,l) + HCl -+ NaCl+ H AH: = 4.7 kcal mol-' (2) Na(n,l)+02 --* NaO+O, AH: = +58 kcal mol-'. (3) Na( n,l) + NO2 --* NaO + NO, (4) The somewhat more complicated process AH: = +9 kcal mol-' will be described in greater detail. Experiment and Analysis The experimental set-up has been described e l ~ e w h e r e . ~ - ~ . ~ Briefly, two supersonic beams and one or two lasers are crossed orthogonally under single-collision conditions. A mass-spectrometric detector rotates about the collision region in the plane defined by the two reactant beams as shown schematically in fig. 1. Several types of experimentsJ. M. Mestagh et al. 147 1 h m Y .- c $ 0.5 v - c M .* m 0 0 30 60 00 laboratory angle/" Fig.2. NaCl product angular distribution for the reaction Na(3S,3 P ) + C12 at 19 kcal mol-' collision energy. The solid lines are the best fit to the data generated from centre-of-mass angular and recoil energy distributions. 0, 3 P ; 0, 3s. are performed: (i) product angular distributions are measured by rotating the detector, (ii) product velocity distributions are measured by a time-of-flight technique, (iii) the dependence of the chemical reactivity on the nature of the excited state of sodium is investigated for Na(3P), Na(4D) and Na(SS), and (iv) the effect of collision energy is studied by changing the beam velocities. The experimental angular and velocity distributions are fitted using a forward convolution method which varies the assumed centre-of-mass angular and translational energy distributions and averages over the variation in experimental conditions.For cases in which the reaction cross-section is found to depend strongly on collision energy, the energy dependence of the reaction cross-section is specifically taken into account in the convolution process. Results Na(n,l) + C12 + NaCl + C1 Angular distributions of the NaCl products were measured for Na(3S,3P), at three nominal collision energies: 3, 6 , and 19 kcal mol-'. Typical results are shown in fig. 2 for 19 kcal mol-' collision energy. The results for reactions of Na(3S) agree with the general understanding of ground-state alkali-metal-halogen reactions.*l They show predominant forward scattering and narrow recoil energy distributions.The new feature brought out by fig. 2 is that no large difference is observed between the product angular distributions associated with the reactions of Na(3S) and Na(3P), even though the electronic excitation lowers the ionization potential of Na by 2.1 eV.148 Reactive Scattering of Excited Alkali-metal Atoms 0 Table 1. Features of the centre-of-mass distributions fitted to the laboratory angular distributions for Na(3S,3 P) + CI, reactions 1 i i ' + 1 I i l collision % of forward peak recoil Na level energy" scattering energy 4 3 W 4 3 S ) 3 s 6.0 3P 6.0 3s 19.0 3 P 19.0 83 76 85 82 0.6 1.2 0.8 1 .o 1.58 1.16 " Energies are in kcal mol '. Fig. 3. NaCl product angular distributions for the reaction Na(3S,3 P,5S,40) + HCI at 5.6 kcal mol-' collision energy.The solid lines are just for guiding the eyes. 0, 3 s ; 0, 3P; A, 5s; m, 4 0 . The predominance of forward scattering and narrow recoil energy distributions allows one to derive centre-of-mass distributions directly by fitting the laboratory angular distributions on the basis of an assumed centre-of-mass distribution." Best fits are shown in fig. 2 for a 19 kcalmol-' collision energy. A summary of features of the centre-of-mass distributions is found in table 1. Na( n,l) + HCl + NaCl + H This system was studied at collision energies of 3.4, 5.6 and 16.3 kcal mol-' for reactions with Na( 3S,3P,5S,4D). The laboratory angular distributions of the product NaCl are shown in fig. 3 for a 5.6 kcal mol-' collision energy. As can be seen, there are important changes in the Na reactivity when switching from Na(3S) to Na(4D).J.M. Mestagh et al. 149 n 0 30 60 90 laboratory angle/" Fig. 4. NaO product angular distribution for the reaction Na(4D) + O2 at 18 kcal mol-' collision energy. The solid lines are the best fit to the data generated from centre-of-mass angular and recoil energy distributions. These angular distributions, along with the measured laboratory velocity distributions of the product NaCl, allow one to extract a centre-of-mass NaCl flux distribution. As discussed in ref. (5), (7), (9), all the reactions of excited Na with HCl exhibit predominant backward scattering with low recoil energy; these characteristics are most apparent in the reactions of the two highest excited states, 5 s and 4 0 .Na(n,l)+O, -+ NaO+O Process (3) for Na (3S,3P,5S,4D) was studied over a wide range of collision energies: 8-22 kcal mol-'. The most interesting feature of the results is the state specificity; no reaction was found for Na( 3S,3P,5S). Backwards-scattered product was observed in the Na(4D) reactions when the average collision energy was >11 kcal m ~ l - ' . ~ The non-reactivity of Na(5S) is striking because process (3) has about the same exothermicity with either Na(5S) as the reactant or Na(4D) (36.9 us. 41.5 kcal mol-I). The high activation energy for the reaction of Na(4D) is also surprising in view of the large exothermicity of the reaction. Representative results are shown in fig. 4 and 5 for the reaction of Na(4D) at a collision energy of 18 kcal mol-'.Na(n,l) + NOz -+ NaO + NO Angular distributions for the Na+ NO, reaction were measured for Na(3S,3P,5S and 4 0 ) . The detected product in these experiments was the Naf that had fragmented from the NaO during electron bombardment; this means that all the recorded distributions have a Na elastic scattering contribution. No Na(3S) or Na(3P) reaction was observed above this background. The Na(5S) and Na(4D) states, however, do react with NOz,150 1 - n 2 0.5 c1 0 - Reactive Scattering of Excited Alkali-metal Atoms - recoil energy/ kcal mol-' 0 0.5 1 1.5 2 I ' I I I I 1 0 90 O / O 180 1 h w a, 0.5 - 0 Fig. 5. Best-fit centre-of-mass angular distribution (full line, left and lower scale) and recoil energy (dashed line, right and upper scale) of NaO in Na(4D) +O, reactions at 18 kcal mol-' collision energy./ I / / 0 30 60 90 laboratory scattering angle/" Fig. 6. NaO product angular distributions for the reaction Na(3S,3P,5S,4D) + NOz at 19 kcal mol-' collision energy. 0, 3P; ., 4 0 ; 0, 5s.J. M. Mestagh et al. 151 Na b+++H+l 1 x lo5 cm s-' Fig. 7. NaO product centre-of-mass contour plot for the reaction of Na(4D) + NOz at 19 kcal mol-I collision energy. as can be seen in fig. 6. The product angular distributions for both these states are not the same. The most obvious distinction is in the reaction cross-sections which are in a ratio of ca. 12 to 1 in favour of Na(4D). The products also scatter differently. The Na(4D) product is more backwardly scattered and has a narrower energy distribution, as evidenced by the sharper peak.The nominal collision energy was varied between 5.5 and 19 kcal mol-' so as to determine the Na(4D) + NOz reaction threshold. The measured speed ratios of the reactant beams, which ranged from 3.2 to 5.2, allowed the spread in these energies to be calculated while the signal levels in the measured angular distributions gave an estimate of the relative reaction cross-sections at each nominal energy. A cross-section vs. collision energy plot could then be constructed to match these observations. An adequate fit was obtained when a step function with a threshold of 20 kcal mol-' was selected. It must be mentioned, however, that this method is not very sensitive to the exact form of this function. Product velocity distributions were measured for the Na(4D) + NOz reaction at a collision energy of 19 kcal mol-'.The laboratory angular and velocity distributions were calculated and fitted using an assumed product recoil energy distribution [ P( E ) ] and centre-of-mass angular distribution [ T ( O)]. The best T ( 0) and P( E ) generated the contour map shown in fig. 7. Two different NaO products are apparent in fig. 7. One (product 1) is very similar to the NaO observed in process (3). It is backward-scattered, has very little energy in translation and has a very narrow distribution in velocity. The other type of NaO (product 2) is very different. It is forward-scattered, has much more energy in translation, and the velocity distribution is much broader. The relative cross-sections for formation152 Reactive Scattering of Excited Alkali-metal Atoms of each NaO have been estimated to be almost equal.However, product 1 is kinematically enhanced for detection in the laboratory reference frame, and is therefore detected preferentially in the present experiment. Discussion In the reaction of alkali-metal atoms with halogen-containing molecules, some essential features can be modelled by the electron transfer 'harpoon' mechanism. In this type of reaction, an ion-pair complex is formed early by transfer of the Na valence electron into the lowest unoccupied molecular orbital (LUMO) of the molecule. This electron transfer is normally assumed to be rapid compared to the nuclear motions. The negative ion is formed on a dissociative repulsive region of the potential. When the ion dissociates, the negative fragment combines with Na+ via coulombic forces to form the reaction product.* Reactions of Excited Na Atoms with Halogen Molecules The prominent feature seen in fig. 2 and table 1 is that no large difference appears in the collision dynamics and, consequently, in the reaction mechanism when Na is excited.This observation is an indication that an electron transfer of the Na electron to the LUMO of the molecule is the initial step.2 The shape of the LUMO is what determines the collision dynamics; it has been observed that the dynamics are very different when switching from halogen to organic halide molecules, but are only weakly sensitive to the identity of the alkali-metal reactant.2 This accounts for the similarity in collision dynamics for reactions with different Na states; exciting Na affects the valence electron orbital, but not, of course, the LUMO of the molecule.Experiments indicate that the total reactive cross-section of process ( 1 ) does not increase very much when Na is excited. This result is somewhat surprising. The higher the Na excitation, the lower the ionization potential; greater Na reactivity would then be expected, since the electron transfer initiating the reaction should be possible at larger Na-molecule distances. However, the shift to large interatomic distances of the crossing between covalent and ion-pair surfaces reduces the interaction between them. This makes the excited states of Na less reactive than they would otherwise be. Such a reduction of the coupling element between covalent and ion-pair surfaces has been demonstrated in a recent calculation using the DIM method on alkali-metal-Br, sur- faces." Apparently, these two factors, the earlier electron transfer and the lower coupling, compensate each other to some extent in these reactions. No large change is observed in the centre-of-mass recoil energies of NaCl produced in reaction ( 1 ) as Na is excited.The additional 48.4 kcal mol-' corresponding to excitation of Na to the 3 P level is thus reIeased into internal excitation of the product NaCl. This would suggest ion-pair formation occurring at a larger Na-C1, distance, which leads to the formation of more highly vibrationally excited NaCl. Reactions of Excited Na Atoms with Hydrogen Halides The dominant features of the excited Na+ HC1 reactions are the decreasing product recoil energy and the increasing reactive cross-section with increasing electronic energy.This can be observed directly in fig. 3 and has been checked quantitatively by considering the centre-of-mass product angular and velocity distribution. For the reaction of Na(5S, 4 0 ) most excess energy goes into the vibrational excitation of NaCl, in contrast to the reaction of Na(3P), where a large fraction of this energy goes into the translational energy of the products. Process (2) thus provides a situation where the reaction dynamics change as the electronic excitation of the reactant is increased.J. M. Mestagh et al. 153 Table 2. Covalent ion-pair non-adiabatic curve-crossing radii for various electronic states of Na for the Na+ HCl systema ionization crossink radius Na level potential/ eV /A 3s 5.1 3P 3.0 5s 1 .o 4 0 0.8 2.4 3.7 7.8 8.6 ~~ a The electron affinity of HCl is taken to be -0.82 eV.14 Reactions of ground-state alkali-metal atoms with hydrogen halides have never been adequately explained in terms of electron transfer ‘harpoon’ models.The electron transfer in these cases does not take place at a long distance because hydrogen halides have negative electron affinities.” However, excited Na atoms have much lower ioniz- ation potentials. This makes early electron transfer possible (see table 2), so such ‘harpoon’ models should provide a useful framework to discuss reactions of excited Na atoms. HCl molecules are known to be dissociated into C1- + H by low-energy elec- t r o n ~ , ~ ~ thus a dissociative electronic attachment to HCl is expected to initiate the reactions of Na( 3 P, 5 S, 4 0 ) . Let us consider the reaction of Na(5S,40) first.Based on the covalent and ion-pair curve crossing, the electron transfer initiating the reaction occurs at large Na-HCl separations. The H atom departs very quickly after the electron transfer. The Na’ and C1- ions are then left at a separation approximately equal to the electron jump radius. NaCl is thus formed with a large vibrational excitation, and the hydrogen atom is no longer present to release the excess energy. The early departure of H is thus able to explain the low product recoil energy found experimentally in reaction (2) for Na(5S, 4 0 ) . Turning to Na(3P), the electron transfer occurs at a smaller distance, which is less than the sum of van der Waals radii of HCl and Na, and the stages of electron transfer, H departure and NaCl formation, are probably not well separated.The dynamics of H atom departure in this reaction are no longer dominated by the dissociation of HCl-. When the electron transfer occurs at a small distance between’HCI and Na, the electronic configuration is closer to HCl--Na+. The repulsive departure between H and NaCl and the stronger coupling among all three atoms are thus responsible for a large fraction of the excess energy in reaction (2) for Na(3P) going into product recoil energy. When the electron is to transfer at a long distance, a model such as the direct interaction model with products distributed as in photodissociation (DIPR-DIP) should be able to describe the reaction dynamics well.14 This model characterizes the reaction by sequential two-body steps: electron transfer from Na to HCl, departure of H from C1-, and, finally, association of Na+ and C1- to form NaCl.The above discussion for process (2) suggests that the DIPR-DIP model should explain Na(5S,4D) reactions well, but not the reaction of Na(3P) with HCl. This has been confirmed more quantita- tively in ref. ( 5 ) , (7) and (9). Reactions of Excited Na Atoms with O2 Molecules The most prominent feature in reaction (3) is that Na(40) reacts with 02, but not Na(5S). Since the radiative decay of the 5s state populates the 4P,4S. . . levels, it is clear that these levels also do not react to produce NaO.154 Reactive Scattering of Excited Alkali-metal Atoms The strong back scattering of the NaO products with respect to the incoming Na atoms is evidence of a direct reaction (i.e.no long-lived collision complex is formed). The especially narrow NaO centre-of-mass angular range seen in fig. 5 suggests that there are restrictive constraints on the impact parameter and relative orientation required for the reaction to proceed. The narrow product recoil energy distribution in fig. 5 shows that very little of the excess energy of this reaction goes into translation. It has been proposed in ref. (6) that electronic excitation of NaO, 0 or both carries away most of the excess energy. Alexander has derived analytic forms for the lowest non-adiabatic potential surfaces of the Na+-Oy system.” Two reactive pathways have been proposed in this work as possibly involved in Na + O2 reactions: and Na+02 -+ Na+...O,(X’II,) -+ NaO+O ( 5 ) Na+O, -+ Na+-..0,(A21Z,) + NaO+O It was suggested in ref.(6) that pathway (5) is probably not involved in the reaction since this ‘harpoon’-like pathway would be unlikely to be associated with a large activation barrier as was observed experimentally for process (3). Recently, calculations have been performed where the reaction pathways (5) and (6) are included in a multiple ionic-covalent curve-crossing model. l 6 These calculations point out that pathway ( 5 ) is actually associated with much smaller reactive cross-sections than pathway (6). The competition of the inelastic channel Na(4D+ n,l) with the reactive channel along pathway (5) is one of the main reasons why (5) makes such a small contribution to the reaction.Consistent with experimental results, the cross-section associated with pathway (6) has a large threshold energy, and involves impact parameters < 1 A. This makes pathway (6) a likely candidate to explain process (3). It is worthwhile to recall that dissociative electronic attachment on 0, molecules giving 0.- + 0 fragments is a resonant process which proceeds through the same A *nu level of 0, as that involved in reactive pathway (6).” Reactions of Excited Na Atoms with NO2 Molecules Process (4) is different from all the others that have been discussed so far in that two products are formed from the reaction of Na(4D).One of these products, product 1, is very similar to the NaO formed in reaction (3). Both exhibit backward scattering and have a small, narrow recoil energy distribution. This suggests that, as with the oxygen reaction, product 1 formation is a direct reaction with certain strict impact-parameter and orientation requirements. The other product observed in the Na(4D)+N02 reaction, seen in fig. 7, is very different from product 1. This is most apparent in the angular distributions. Product 2 is forward scattered suggesting a direct reaction with a large impact parameter. The recoil energy distributions also differ. Product 2 is formed with a much broader distribution of recoil energies and the mean translational energy of this distribution is 5.5 kcal mol-’ larger than that of product 1.The differences between the two products formed in process (4) suggest that two different mechanisms are involved in their production. This is supported by the observa- tion that Na(5S) seems to react with NO, to produce product 2 but not 1, which can be seen from the Na(4D) and Na(5S) product angular distributions in fig. 6. What is most noticeable is the large difference in the reaction cross-sections. Since the Na(4D) and Na(5S) states are very close in energy (98.8 vus. 94.9 kcal mol- ’ ) and both put the reactions well above the reaction endothermicity and threshold barrier of 20 kcal mol-I, this difference in reactivity could be the result of different reaction pathways. This point is furthered by the different peak positions and widths in the two angular distributions.J. M.Mestagh et al. 155 Sholeen and Herm have studied the reaction Li(2S) + NO, which is similar to process (4) for ground-state Na.18 The reaction products they found share many of the charac- teristics of product 2. Their experiments showed the formation of LiO(X ’II) and LiO( A &), which were forward-scattered and exhibited a large, broadly distributed recoil energy (the peak energy was 45% of the reaction exothermicity or 3.8 kcal mol-I). They saw no low-translational-energy, backward-scattered product, although they did not try the experiment using electronically excited Li. Sholeen and Herm proposed a Li+...NO,(’B, or ‘Al) intermediate to explain their data. The Li+. - .NO;(’B,) ion-pair seems more plausible.It does not provide as deep a well in the reaction potential surface as the Li+. - .NO,( ‘ A , ) intermediate. Also, this choice could explain the large product recoil energies observed; NO;(3B1) results from the transfer of the Li electron to an antibonding orbital (26,7r) on NO2. Sholeen and Herm, however, did not choose one pathway over the other.I8 The mechanism proposed in both the Li + NO, and Na + O2 experiments is reaction through an excited ion-pair intermediate. Since one of the Na+ NO2 products is similar to that of the Na+02 and the other to that of the LiO, it seems possible to propose that process (4) proceeds through the following intermediates: Na+. - .NO,( ‘ A , ) (7) Na+...NO,(’B,) (8) Nat...NOy(lBl). (9) Based on the comparison of the two products found to those found by others, the Na+ NO, reaction might proceed via intermediate (9) to give product 1 and through intermediate (8) to form product 2.Since NO2 is a bent molecule, a near-collinear collision between the Na and the 0-N bond in the NO, would result in a forward-scattered product with a fairly large impact parameter. This approach geometry, then, is consistent with the observations for product 2 formation. Problems, however, arise when one attempts to determine the Na(4D)-N02 collision geometry necessary to form product 1. A small impact parameter, which was what was seen experimentally, would position the Na away from the oxygen with which it is supposed to be colliding. Also, one would expect forward scattering for this type of collision.Physically, how is the NaO of product 1 formed? One answer which must be considered is that NaO is not being formed. As mentioned in the experiment, Na’, not NaO’, is detected. The possibility thus exists that Rydberg sodium atoms are being produced. If this were true, the similarity between product 1 and the Na + 0, product would cast some doubt on the Na + O2 mechanistic assignment. Another explanation is that NaO, + N is forming. The Na(4D) + NO2 --* NaO, + N reaction channel is energetically possible (AH is ca. -10 kcal mol-I). If the Na approached the O2 along the C2” symmetry axis, one could expect the product to be backward-scattered as is observed. Experiments carried out thus far have not given a definitive answer to the question of what this backward-scattered product is.Future experiments are being planned to resolve this. Other Interesting Features The experiments reported here have been extended in order to study how rotating the laser polarization, and thus the excited Na(P or D ) orbital, affects the Na reactivity. Owing to space limitations, only a few of these results are summarized below [see ref. (7) and (9) for more details]. The Na(4D)+HCl reaction is enhanced when the Na(4D) orbital is aligned along the relative velocity vector. This behaviour is expected if the reaction proceeds through a long-range electron-transfer mechanism in collinear Na. - 421- H geometry. For the156 Reactive Scattering of Excited Alkali-metal Atoms Na(4D) + 0, system, the favourable alignment for the reaction changes with the scatter- ing angle, and corresponds to the Na(4D) orbital being perpendicular to the molecular axis Na-0-0.Finally, the polarization effects encountered in the Na(4D) + NOz reaction suggest that product 2 but not product 1 formation is affected by the Na(4D) orbital alignment. This reinforces the conclusion reached in the previous section that products 1 and 2 are formed by different reaction mechanisms. Conclusion Representative families of excited alkali-metal atom reactions have been studied using a crossed-beam apparatus, and the reactivity of various excited states of Na has been investigated. For those systems in which reactions are also known for the ground-state alkali metal and involve an early electron-transfer step (e.g. reactions of Na with C12), no large differences are observed in the reactivity as Na is excited.This can be understood in terms of two competing effects; the lower ionization potential of excited Na increases the Na reactivity, which is compensated by a lower coupling between the covalent and ion-pair non-adiabatic curves of the Na-halogen system. The dynamics of the reaction remain unaffected by the Na electronic excitation, since the reaction behaviour is entirely determined by the shape of the LUMO of the molecular reactant. Similar observations, which are not reported here, have been made for reaction of excited Na atoms with organic halide molecule^.^ More interesting are the reactions of Na atoms with hydrogen halides. For ground- state Na and Na(3P), the reaction does not proceed via early electron transfer, and the NaCl-H repulsion dominates the product scattering. The excitation of Na to the 5s and 4 0 levels changes the reaction mechanism. The reaction then proceeds via an early dissociative electron attachment of the Na valence electron on HCl, and the HCI- repulsion dominates the NaCl product scattering.The reactions of excited Na with oxygenated compounds are of particular interest. The reaction with O2 was found to be state-specific [only Na(4D) reacts with 02], and seems to involve electron transfer to O2 so as to form the excited state A *nu of O,, i.e. the state responsible for dissociative electronic attachment on 02. The reaction with NO2 is even more complex, since Na( 4 0 ) leads to the formation of NaO by two different pathways.It must be mentioned, however, that the identification of NaO as product in these reactions has yet to be confirmed. This work was supported by the Director, Office of Energy Research, Office of Basic Energy Sciences, Chemical Sciences Division of the U.S. Department of Energy under contract no. DE-AC03-76SF00098. B.A.B. acknowledges a National Science Foundation Graduate Fellowship, J.M.M. thanks the Centre National de la Recherche Scientifique (France) for financial support. Some of the lasers used were on loan from the San Francisco Laser Center supported by the NSF under grant no. CHE79-16250 awarded to the University of California at Berkeley in collaboration with Stanford University. References 1 R. J. Buss, P. Casavecchia, T. Hirooka, S. J. Siebener and Y. T. Lee, Chem. Phys. Lett., 1981,82, 386. 2 R. R. Herm, in Alkali Halide Vapors, ed. P. Davidovits and D. L. McFadden (Academic Press, New York, 1979), p. 189. 3 M. W. Geis, H. Dispert, T. L. Budzynski and P. R. Brooks, in State to State Chemistry, ed. P. R. Brooks and E. F. Hayes, ACS Symp. Ser. No 56 (American Chemical Society, Washington, DC, 1977), p. 103. 4 P. S. Weiss, J. M. Mestdagh, H. Schmidt, M. F. Vernon, M. H. Covinsky, B. A. Balko and Y. T. Lee, in Recent Advances in Molecular Reaction Dynamics, ed. R. Vetter and J. Vigue (1985). 5 M. F. Vernon, H. Schmidt, P. S. Weiss, M. H. Covinsky and Y. T. Lee, J. Chem. Phys., 1986 84, 5580.J. M. Mestagh et al. 157 6 H. Schmidt, P. S. Weiss, J. M. Mestdagh, M. H. Covinsky and Y. T. Lee, Chem. Phys. Lett., 1985, 118, 7 P. S. Weiss, J. M. Mestdagh, M. H. Covinsky, B. A. Balko and Y. T. Lee, to be published. 8 G. Rahmat, F. Spiegelmann, J. Verges and R. Vetter, Chem. Phys. Lett., 1987, 135, 459. 9 P. S. Weiss, Ph.D. Thesis (Lawrence Berkeley Laboratory, University of California, 1986). 539. 10 Y. T. Lee in Atomic and Molecular Beam Methods, ed. G. Scoles and U. Buck, Oxford University Press, New York, 1986. 11 E. M. Goldfield, E. A. Gislason and N. H. Sabelli, J. Chem. Phys., 1985,82, 3179; E. M. Goldfield, A. H. Kosmas, and E. A. Gislason, J. Chem. Phys., 1985, 82, 3191. 12 B. A. Blackwell, J. C. Polanyi and J. J. Sloan, Chem. Phys., 1978, 30, 299. 13 J. N. Bardsley and J. M. Wadehra, J. Chem. Phys., 1983, 78, 7227. 14 P. J. Kuntz, M. H. Mok and J. C. Polanyi, 1969, J. Chem. Phys., 1969, 50, 4623. 15 M. H. Alexander, J. Chem. Phys., 1978, 69, 3502. 16 J. M. Mestagh, D. Paillard and J. Berlande, J. Chem. Phys., submitted for publication. 17 D. S. Belic and R. I. Hall, J. Phys. B, 1981, 14, 365. 18 C. M. Sholeen and R. R. Herm, J. Chem. Phys., 1976, 64, 5261. Received 19th May, 1987