Faraday Discuss. Chem. SOC., 1986,82,99-110 Hydrogen-atom Photofragment Spectroscopy Photodissociation Dynamics of H20 in the 6-2 Absorption Band H. Joachim Krautwald, Ludger Schnieder and Karl H. Welge* Fakultaet fur Physik, Universitat Bielefeld, Bielefeld, Federal Republic of Germany Michael N. R. Ashfold School of Chemistry, University of Bristol, Bristol BS8 1 TS The photofragmentation dynamics of H20 molec$es following vacuum ultraviolet laser excitation at wavelengths within the B 'A1-X ' A , absorption band have been investigated using a novel form of photofragment transla- tional spectroscopy. Analysis of the nascent H atom time-of-flight spectra confirms the importance of the dissociation channel leading to ground-state H+OH products. In addition, it reveals that the OH(X) fragments are formed predominantly in their zero-point vibrational level with a highly excited, inverted rotational-state population distribution. A consideration of the topology of the various potential-energy surfaces sampled by the H20 molecules during this electronically non-adiabatic dissociation process sug- gests a likely explanation for this observed pattern of energy disposal.The past few years have witnessed continued improvement in the detail and sophistication being applied to studies of the dynamics of molecular photofragmentation processes. 172 The water molecule has long been popular as an (albeit complicated) model triatomic system, and aspects of its photodissociation behaviour have been revealed by a variety of e ~ p e r i m e n t a l ~ - ~ ~ and t h e ~ r e t i c a l ~ ~ - ~ ~ techniques.The recent classic studies of Andresen and coworkers, 13-16 supported by dynamical calculations from Schinke et aL,31-33 have done quch to clarify our picture of H20 photodissociation from its first excited singlet state ( A ' B 1 ) . Even so, experimental studies may still be hampered by the fact that excited-state predissociation precludes the use of laser-induced fluorescence (LIF) detection methods for quantitative population measurements of OH( X ) fragments carrying even moderate rotational and/ or vibrational e ~ c i t a t i o n . ~ ~ The complexities @crease manyfold at shorter excitation wavelengths. The second excited singlet state ( B ' A , ) has a very contorted potential-energy surface with a deep well in the HO-H dissociation coordinate resulting from a conical intersection with the ground-state Depending upon the Erecise excitation wavelength, photoabsorption may result in direct population of the B state, or in population of one of the more stable Rydeerg s_tates22v36-39 which can then predissociate via radiationless transition to either the A or B states.22 Four spin-allowed fragmentation channels have been identified for H 2 0 following excitation at wavelengths below 129 nm.These and their associated thermochemical thresholds are listed below:40 /H+OH(X *n) Do< 5.118 eV (1) H,+ o(' D ) H +OH( A 2Ec+) Do d 7.00 eV Do d 9.05 eV H20 + hv LH + H+ o ( ~ P ) Do d 9.61 eV. (4) Quantum-yield measurements have been reported for channels (2) and (4) following excitation of H20 at 121.67 and 123.6 nm,6 whilst several groups have provided estimates of the relative importance of channels ( 1 ) and (3).9-'2 All conclude that dissociation 99100 Hydrogen -a tom Photo fragment Spectroscopy to the ground-state products H + OH(X) 1s t,he dominant result following H20 photoab- sorption in the wavelength range of the B-X absorption band, despite the facts that (i) symmetry constraints dictate that-this asymptote may_ be Leached only as a result of non-adiabatic coupling from the B state to either the A or X state surface^^^-^' and (ii) there have been no reports of the successful observation of nascent OH(X) fragments following H20 photolysis at these wavelengths. In the present work we employ a new version of the technique of photofragment recoil spectros~opy,~~ originally pioneered by Wilson and coworkers,"2 to a study of H20 photodissGciat_ion at photon energies around 10 eV (to the high-frequency side of the peak of its B - X absorption).As with most other versions of the method, photofrag- ment time-of-flight (TOF) spectra are recorded and transformed to yield a spectrum of the photofragment translational energies from which, by energy conservation, the internal energy disposal may be deduced. The novel feature introduced in this work is the method of detecting the nascent photofragment. It (in this case the H atom) is selectively ionised by laser-induced excitation and its dynamics monitored via the ion. The general technique is blessed with a wide range of applicability and high sensitivity.A detailed description of its application to the detection of H atom photofragments appears e l ~ e w h e r e . ~ ~ The H atom TOF spectra so obtained show clearly resolved structure which, upon analysis, is found to be associated with the various different rotational states in which the OH(X) product is produced. This product is formed with a highly excited, inverted rotational state population distribution, thus accounting for the failure of all previous attempts to observe nascect QH( X ) fragments by LIF following photoexcitation of H,O within the region of its B-X transition. Previous studies of H20 photolysis at these excitation energies have revealed the nascent electronically excited OH( A) photofrag- ments to possess highly excited inverted rotational state population distributions also.17-23 In both cases a qualitative explanation for the high levels of product rotational excitation may be found by considering the topology of the one, or more, potential-energy surface( s) sampled as the molecule proceeds along the HO- H dissociation coordinate: a more quantitative rationale must await a more complete knowledge of the relevant potential-energy surfaces35'43-45 and of the possible non-adiabatic couplings between them. 26-30 Experimental Here we provide only a brief description of the experimental design and procedure. The experimental set-up, shown schematically in fig. 1, consists of a crossed molecular beam-laser beam arrangement with a time-of-flight spectrometer for analysis of H+ ions produced in the beam intersection region as described below.The parent molecule gas of interest, entrained in an inert carrier gas (usually a 1 : 2 mixture of H20 at 40 "C and Ar; stagnation pressure ca. 150 Torr) is expanded through a pulsed nozzle and then constrained to flow through a capillary of length 15 mm, inner diameter 2.5 mm before emerging into the interaction region within the reaction chamber. The gas beam is crossed at right angles to its axis by the dissociation (L,) and the H-atom detection (L,) laser beams, which counterpropagate through the chamber with a crossing angle of ca. 3". The ionisation laser pulse (pulse duration ca. 15 ns) is delayed 6 10 ns with respect to the photolysing laser pulse (duration ca. 10 ns). Hydrogen atoms produced uia dissociation channels ( l ) , (3) and (4) during the photolysis laser pulse are ionised selectively in the beam intersection region by resonant two-photon excitation: H(n=l)+vacuum ultraviolet (121.6nm) - H(n=2) with the first step at the Lyman-a line.46 The 121.6nm radiation is produced by ( 5 ) H( n = 2) + ultraviolet (364.8 nm) + H+ + eH.J. Krautwald, L. Schnieder, K . H. Welge and M. N. R. Ashfold 101 Fig. 1. Schematic diagram of the experimental arrangement for photoatom spectoscopy (PAS). frequency-tripling 364.8 nm laser light (pulse duration ca. 15 ns) in a cell containing krypton gas47-53 and focused into the reaction chamber (beam diameter ca. 0.2mm, photon flux ca. lo9 per pulse) so as to overlap spatially with the centre of the dissociating laser beam axis.The 364.8 nm radiation, traversing the reaction chamber as a slightly divergent beam, is back-reflected by a spherical mirror and has converged to a diameter of ca. 3 mm at the beam intersection region. The sum energy of a 121.6 and a 364.8 nm photon coincides very closely with the H-atom ionisation limit, resulting in practically zero recoil energy in the ionisation process. This is an essential requirement for this technique of photoatom spectroscopy (PAS), the utility of which depends on the validity of the assumption that the measured ion time-of-flight spectrum is an accurate monitor of the dynamics of the photolysis event by which the H atoms were formed. Experiments were conducted at three photolysis wavelengths A D = 122.0, 125.1 and 126.2 nm (10.16,9.91 and 9.82 eV, respectively), chosen because they could be generated with relatively high intensity (ca.10l2, 1013 and 10l2 photons per pulse, respectively) by four-wave mixing in mercury ~ a p o u r . ~ ~ Radiation emerging from the Hg mixing cell was focused into the interaction region within the reaction chamber, where it had a beam diameter of ca. 0.9 mm. The TOF spectrometer consists of two field-free sections (see fig. 1). The first comprises two parallel flat metal grids placed symmetrically about the plane containing the intersecting molecular and laser beams; this point of intersection defines the spec- trometer axis. The first part of the total flight path from the intersection is defined in length (do = 30 mm) and solid angle (0.52 Sr) by an aperture of 8 mm diameter in one plate, covered by a fine-mesh copper grid, Go.The second flight section is guided by a cylindrical tube and defined in length by grids GI and G, 401 mm apart. Go and G, are separated by a distance of 1.5 mm. This design allows for the possibility of manipulat- ing the drift velocity of the ions by applying a small potential to the second section. Ions passing through G2 are accelerated first by the grid electrode G3 to ca. 50V and then further to ca. 4 kV onto the cathode of the SEM detector (Johnston, type MM1-SG). Signals from the detector are recorded time-resolved by a transient digitiser (Biomation 6500,1024 channel capacity, 500 MHz maximum speed) are accumulated for a preselec- ted number of laser shots and transferred to a microprocessor for subsequent data102 Hydrogen-atom Photofragment Spectroscopy 10 20 flight time/ps 30 ps Fig.2. H+-ion TOF spectrum resulting from H20 photolysis at 125.1 nm. [ E ( R v ) = 9.9 eV.] handling. The maximum resolution obtained with this spectrometer is 15 meV (at a kinetic energy of 1.6eV), although in the present experiments resolution has been sacrificed (A& = 0.08 eV) in order to improve the signal strength. The measured TOF signal distribution S ( t) is converted into a translational energy distribution S( Etr) in the laboratory frame by the transformation S( t) d t = S( Etr) dE,, (6) where dEtr/dt = mZ2t-3, with time increments equal to the transient digitiser step size used. In the transformation from measured flight time to energy the electric potential distribution applied to the spectrometer sections was, of course, taken into account.The observed H atom TOF distribution is affected by the transmission of the spectrometer and the overall resolution in a complex fashion. Because these experiments are still in a preliminary state we refrain from detailed discussion of these effects, but some aspects of the problem are addressed in the following section of this paper. Results Only one dye laser is required to generate the necessary frequencies for the four-wave mixing process in Hg that gives rise to the 125.1 nm photolysis wavelength. Production of the other two photolysis wavelengths necessitated use of two dye lasers, the outputs of which were synchronised and spatially overlapped in the mixing cell.The resulting vacuum ultraviolet output intensities were an order of magnitude lower, and the photo- fragment translational energy spectra obtained at these wavelengths were of correspond- ingly poorer quality. Fig. 2 shows a complete H' ion TOF spectrum resulting from H20 photolysis at 125.1 nm. Signal was accumulated from 5000 laser shots at a 10 Hz repetition rate. To cover this time region with 20ns digitiser channel width the drift velocity of the H' ions in the second flight section of the TOF spectrometer was shifted by applying a 5 V potential between grids G, and G2. No signal was detectable at times d 11 ps, and by ca. 30ps the ion signal has fallen to a negligible level. Tests were made to ensure that no significant ( < 0.1 % ) ion signal derived from use of the vacuum ultraviolet probe or the vacuum ultraviolet photodissociation laser alone.Fig. 3 shows this TOF spectrum transformed into a spectrum of the total kinetic-energy release. The energy resolution is ca. 0.08 eV and, to a first approximation for the spectrometer set-up used, independent of the magnitude of the translational energy. Higher resolution (ca. 0.05 eV)H. J. Krautwald, L. Schnieder, K. H. Welge and M. N. R. Ashfold 103 N = kinetic energy/eV Fig. 3. Transformation of the TOF spectrum, shown in fig. 2, into a spectrum of total kinetic energy. Above the spectrum are the energies corresponding to different rotational levels, N, of the OH radical in the u”= 0 vibrational level of the ground (X) and the first excited ( A ) electronic states. 0 1 2 3 4 5 kinetic energy/eV Fig.4. Spectrum of total kinetic energy for H+ ions produced via H 2 0 photolysis at 122.0nm. [ E (hv) = 10.2 eV.] could be achieved (at the expense of signal-to-noise ratio) using lower laser powers. Fig. 4 and 5 display the corresponding translational energy release spectra for H atoms produced via H 2 0 photolysis at 122.0 and 126.2 nm, respectively. Although of poorer quality these spectra show a similar overall appearance to that obtained at the 125.1 nm photolysis wavelength. Energy conservation dictates that E ( hvD) + Ei,t( H20) - Do( HO - H) = E,, where104 Hydrogen-atom Photofragment Spectroscopy 0 1 2 3 4 5 kinetic energy/eV Fig. 5. Spectrum of total kinetic energy for H+ ions produced via H20 photolysis at 126.2nm.[ E ( h v ) =9.8 eV.] Thus, given Do( HO-H) = 5.1 l8 f 0.05 eV,55 it should, in principle, be possible to derive the population of OH radicals in a specific internal quantum state k by monitoring the number of H atoms ( Hk) in the corresponding translational-energy group. In practice we find that the position of the peaks in the translational-energy release spectrum allows unambiguous identification of the v and N quantum numbers of the OH states populated in the dissociation, but we recognise that interpretation of the individual peak heights requires caution because of possible experimental effects as indicated below. Shown above each translational-energy spectrum are the energies corresponding to the different rotational levels, N, of the OH radical in the u” = 0 vibrational levels of the ground ( X ) and first excited ( A ) electronic states.Energies for the ground-state levels were obtained by fitting the term values (averaged over the spin-orbit- and A-splittings) given by Dieke and C r o ~ s w h i t e ~ ~ to a third-order polynomial in N ( N + 1) and extrapolating to higher N. Term values for the A-state levels were taken directly from the 1iteratu1-e.~~ For best agreement between the observed and ‘calculated’ energy scales it was necessary to set the effective HO-H bond dissociation energy to 5.05 eV. About half of the offset (ca. 0.07 eV) from the literature Do value may be attributed to the contribution from E,,,(H,O) under the gas inlet conditions used in these studies. The rest may be due to experimental uncertainty.The three translational-energy spectra clearly demonstrate that the bulk of the detected H atoms are formed in association with rotationally excited but vibrationally cold OH(X) radicals. It is impossible to account for the structure resolved in these spectra in terms of vibrational excitation of the nascent OH fragments. However, we are not able to exclude the possibility that a small fraction of the OH(X) photoproducts are formed in vibrationally excited levels. For example, spectral simulations that assume the same overall rotational-state intensity distribution as that observed and attributed to formation of OH(X, u = 0) suggest that a contribution of s 20% from the dissociation channel leading to H + OH(X, v = 1 ) would be hard to discern with the present spectral resolution. We now turn to a consideration of the information that may be derived from the heights of the individual peaks resolved in the translational energy spectra. Necessarily this requires some (brief) discussion of factors that will cause differences from the idealised 1 : 1 correlation between the population of OH radicals formed in a specific quantum state and the number of H atoms detected with the appropriate correspondingH.J. Krautwald, L. Schnieder, K. H. Welge and M. N. R. Ashfold 105 I I I I I I I I I I I 1 - + c .- +-I + 3 ++ d P - a - m - + + 0 - .- 4 + + - * u - + + 2 - ++ ++ + - +++ 0 - . . . . . . . . . . . . . . . . . . . . . . . . I l l 1 I I I I L I I kinetic energy. Further experiments using this PAS technique will be necessary before it is possible to give a full appreciation of the various factors that might cause a translational-energy-dependent H-atom detection efficiency.Here we simply list some of the more obvious and comment on their likely effect on any observed TOF distribution. As in all TOF experiments there will be a tendency to miss the slowest particles; with the present set up this may come about as a result of the following factors. (i) The initial parent molecular velocity. The parent beam is perpendicular to the TOF axis; thus for the slowest particles the resultant H-atom velocity vector in the laboratory frame will fall outside the detection solid angle. Obviously the contribution from this loss mechanism increases with decreasing recoil energy; with the present experimental geometry we expect major loss of H+ ions with recoil energy < 0.3 eV.(ii) Collisional deflection, the probability of which will scale with the flight time. (iii) Space charge effects, the consequences of which are hard to assess quantitatively but, qualitatively, can be expected to lead to a reduction in the overall resolution and to have the greatest relative impact on slow particles. Proving experiments involving the much-studied 266 nm photodissociation of HI39 suggest that none of these effects are significant for 'slow' H+ ions with recoil energy 30.3 eV. Because of the relative sizes of the vacuum ultraviolet dissociation (L,) and probing (L,) beams the present experiment also discriminates against the fastest H atoms.Model calculations with the appropriate beam diameters and overlap conditions indicate that free flight out of the excitation-ionisation region will cause a 20% reduction in the detection efficiency for H atoms with 2 eV recoil energy (rising to 50% for 4 eV recoil energy) relative to that for H atoms with a translational energy of 1 eV. Finally, we recognise that the 121.6 nm bandwidth (Av == 50 GHz) used will introduce some degree of Doppler selection which could lead to a discrimination in favour of particular H-atom velocity groups. Further work is required before it will be possible to comment on this effect in detail. Most of the H atoms resulting from H20 photodissociation at 125.1 nm are associated with total translational energies < 3.5 eV.In view of the foregoing discussion we consider the overall shape of that part of the observed distribution assigned to the H + OH(X) dissociation channel to be genuine, including the maximum ( N " = 44) and the decline in population to higher N". Spectra obtained at the other excitation wavelengths peak and decline at very similar recoil energies. Fig. 6 shows a plot of the OH(X, = 0)106 Hydrogen-atom Photofragment Spectroscopy rotational-state population distribution resulting from photolysis at 125.1 nm. Popula- tions are derived by best-fit simulation of the experimental spectrum, assuming each H + OH( X ) , u = 0, N" product channel to have an associated energy peak of Gaussian lineshape (0.08 eV f.w.h.m.), and then correcting for the calculated effects of H-atom flight out of the detection region.The crucial observation that the OH( X, ZI = 0) fragments are formed with a highly excited, inverted rotational-state population distribution which, nevertheless, cuts off before the energetically allowed maximum is considered further in the next section. From this distribution it is possible to estimate a value of 0.68 for ( fR), the fraction of the total energy available to the H + OH(X) dissociation channel that is partitioned into product rotation. This, together with our estimate that vibrational excitation (fv) cannot exceed 0.02 leads to the conclusion that the fractior. appearing as translational energy ( fT) =: 0.3. Fig. 3-5 provide a base for quantitative estimation of the problems associated with the transmission loss of the slowest H atoms.Previous photofragment emission studies of the H+OH(A) product channel resulting from photolysis in this wavelength region have identified a very high level of rotational excitation in the OH(A) fragment, with a rotational state population distribution that peaks near the highest N' level permitted by energy c~nservation.~'~' Necessarily these OH( A) fragments are formed in conjunc- tion with near-zero kinetic-energy H atoms; these are not detected in the present experiment. Nevertheless it is possible for us to support previous of the branching between product channels ( 1 ) and (3) at these photolysis wavelengths. Knowing the form of the OH(A) rotational-state population distribution to be expected at this photolysis wavelength2' it is possible to get a reasonable match with the experi- mental spectrum shown in fig. 3 at energies down to ca.0.3 eV if we assume a branching ratio OH(A) : OH(X) = 0.10. Discussion A quantitative interpretation of the H atom time-of-flight distributions reported in this work is precluded by (i) uncertainties in the kinetic-energy dependence of the experi- mental H-atom detection efficiency (see previous seciion) and (ii) our incomplete knowledge of the full potential-energy surface for the B ' A , state of water, and of the various non-adiabatic couplings involving this state. Nevertheless, the MRD-CI calcula- tions of Theodorakopoulos et aL,35943,44 and especially those for the asymmetric stretching potentials of several of the singlet states of H20,35 are of sufficient quality that we can provide a qualitative rationale for tbe observed distributions. Results of these a6 initio calculations relating to the A and B states of water are shown in fig.7, which displays three-dimensional representations of the relevant potential-energy surfaces as a function of R(H0-H) and the angle LHOH. The topologies of these surfaces have been discussed previously25-30~35743-4s and will only be summarised here. For linear configurations there is a conical intersection between the 'II and E surfaces, the lower component of whigh gives the ground state in bent H20. Because of this conical intersection the upper ( B ) surface has a deep well in the asymmetric dissociation coordinate, with a minimum at R = 0.16 nm (see fig.7); its shape at shorter R (HO - H) distances is complicated ky Rydkerg-valence interaction^.^^ Further compli- cations arise from the fact that the A and B states both derive from an electronically degenerate 'I3 state in f,he linear configuration. As a result there is a 'seam of intersection' between the A and B surfaces at the linear geometry for all asymmetric stretching distances R < 0.16 nm (the geometry associated with the conical intersection). Bending, and the accGmpanying Renner-Teller effect;' lifts this degeneracy, with the result that, whilst the B state potectial-energy surface shows a minimum at linear configurations, the minimum in the A state surface at short R(H0-H) lies at a bent geometry comparable to that of the electronic ground state.1 +H. J. Krautwald, L. Schnieder, K . H. Welge and M. N. R. Ashfold 107 .OH(X) LHOHI" LHOH/" Fig. 1. Three-dimensional representations of parts of the potential-energy surfaces for the 2 'B1 and B ' A , states of H20 plotted as a function of LHOH bendisg angle and the HO-H dissociation coordinate [after ref. (33)]. A, small part of the ground ( X ) state surface, in ths region-of its conical intersection with the B-state surface, is included in the latter plot. The A- and B-state surfaces are degenerate for linear geometries, and are separated (horizontally) only for clarity of presentation. Shown also are three representative classical-trajectories for a configuration point moving away from the Franck-Condon region (F.C.) of the B-state surface, as discussed in the text.The present work lends strong support t,o previous suggestions 12,40 that the photofrag- mentation of H 2 0 molecules from the B electronic state occurs predominantly via dissociation channel (l), despite the fact that the adiabatic correlation for this state upon asymmetric distortion is with the products of channel (3). It also reveals the nascent OH(X) fragments to be formed with little vibrational excitation but with a highly inverted rotational-state population distribution. Such energy disposal is strik- ingly reminiscent of that o b ~ e r v e d ~ ~ " - ~ ~ for the dissociation pathway (3), leading to formgticn of electronically excited OH( A) fragments following excitation of H 2 0 within the B-X continuum. As in the latter, more widely s t ~ d i e d , ~ * ' ~ - ~ ~ case it is possible to interpret the observed product quantum-state distributions bx considering the more likely classical trajectories for dissociation originating on the B state surface. W,e identify three classes of trajectory starting from the Franck-Condon region of the B state surface, as displayed in fig.7, and consider the eventual fragmentation pattern associated with each. Necessarily therefore this discussion recognises no contri- bution from channel (4), to which our experiment is in any case insensitive. For each case the initial motion for which the potential has the steepest gradient is bending towards linearity. However, as the LHOH angle opens, the detailed trajectory of the configuration point will be sensitively dependent upon the balance between the asym- metric stretching (radial) and bending (angular) motions possessed by the molecule108 Hydrogen -a tom Photofragmen t Spectroscopy when prepared on the state surface.Those trajectories (type I) which experience the greatest radial acceleration may pass _directly outside [ i.e. large R( HO- H)] the conical intersection and thus remain on the B state surface where the curvature of the potential ensures strong rotational excitation of the OH(A) photoproduct as first suggested by F l o u q ~ e t ~ ~ and frequently demonstrated e~perimentally.~”~-*~ Mechanisms can be envisaged whereby all trajectories for which the radial acceler- ation is insufficient to carry the configuration point outside the conical intersection end up at the H + OH(X) dissociation limit.Those with sufficient associated stretching motion to be drawn into the deep w ~ l l arising as a result of the conical intersection (type 11) will emerge on the ground (X) state potential-energy surface in a region where it is ~ a l c u l a t e d ~ ~ to be gently repulsive along the dissociation coordinate but still to retain significant angular anisotropy, with a minimum at bent geometries. Indeed, for type -11 trajEctories passing through the conical intersection the angular properties of the B and X state potential are such that it is quite conceivable that all of the available energy could be partitioned into rotational excitation of the OH(X) fragment. On the other hand, trajectories which start out from the Franck-Condon region with little radial accelzrationitype 111) pass through linearity at least once at relatively short R ( HO-H).The B and A states are degenerate in this linear configuration, and Dixon3’ has shown how vibronic- inteEaction (the Renner-Teller effect) can provide a mechanism for the irreversible B - A radiationless transition. At short R(H0-H) the A state surface shows a marked angular anisotropy, with a_mini_mum near LHOH= 105” (see fig. 7). Thus type I11 trajectories which undergo B-A transfer at these shqrt R(H0-H) separations will experience an angular acceleration on both the B- and A-state surfaces before reaching the_ gently repulsive, isotropic longer range region of the dissociation coordinate of the A-state surface leading, again, to Ha+ OH(X).Here a significant difference beiween the A- and X-state surfaces becomes evident. The a6 initio calculations show the A-state surface to have lost all of its angular anisotropy by R b 0.16 nm, at which point its potential energy is still ca. 0.8 eV above the asymptotic limit for H+OH(X). Thus for trajectories that lead to eventual dissociation on the A-state surface this part of the total available energy cannot be available for product rotational excitation. Whilst recognising the inherent difficulties associated with detec- tion of the slowest fragments in any time-of-flight experiment, it is relevant to note that we observe no structure that is obviously attributable to rotationally excited OH(X) fragments within 0.8eV of the origin of the kinetic-energy spectra obtained at any of the three excitation wavelengths siudie_d.Such a result provides support for previous discussion^^^ of the importance of B - A radiationless transfer in promoting dissociation channel (1). Finally, we must consider the vibrational energy disposal in this fragmentation process. The measured kinetic-energy release spectra may be interpreted without assum- ing formation of any OH(X) fragments with 0”) 0, but this conclusion may require some revision if and when higher resolution TOF data becomes available. Such a low level of product_vibrational excitation is consistent with the very similar 0-H bond lengths in H20( X) and OH( X), and mirrors that reported previously for the correspond- ing dissociation channel leading to OH( A) fragrnent~.’.’~-~~ However, a6 initio MRD-CI c$culations for the symmetric stretching potentials of H2044 show the minimum in the B-state surface to lie at larger 0-H bond lengths.Franck-Condon considerations might thus suggest that dissociation proceeding via this surface would give rise to some product vibrational excitation. That such appears not to be the case for either H + OH dissociation channel is presumed to be because the slope of the potential for this symmetric motion is too shallow to compete with the combined bending/asymmetric dissociative motion. Further djscussion of this point must await the availability of a more complete picture of the B-state potential surface.H. J. Krautwald, L. Schnieder, K . H. Welge and M. N. R. Ashfold 109 Conclusion In this study we have demonstrated the utility and sensitivity of a new form of photofrag- ment translational spectroscopy by applying the technique to an investigation of the photodissociation dynaFic5 of the water molecule following excitation at various wavelengths within its 23-X absorption band.Measurements of the nascent H atom TOF spectra reveal a dominant role for the dissociative channel leading to ground-state H + OH products, but the OH(X) fragments are observed to carry very high levels of rotational excitation. Such energy disposal may be understood by considering the form of the various potential-energy surfaces sampled by the molecules as they evolve along this dissociation coordinate. The observed high level of fragment rotation explains the failure of previous attempts to monitor these OH(X) photoproducts by LIF (the corresponding levels in the first ekcited state of OH are heavily predissociated), and highlights one of the potential advantages of the new technique: its universality.Possible limitations, particularly the difficulties associated with efficient detection of those species that recoil with the fastest and the slowest velocities, are also considered. We are grateful to NATO for the award of a travel grant. M.N.R.A. thanks Prof. R. N. Dixon for numerous helpful discussions. References 1 M. Shapiro and R. Bersohn, Annu. Rev. Phys. Chem., 1982, 33, 409. 2 J. P. Simons, J. Phys. Chem., 1984, 88, 1287. 3 T. Carrington, J. Chem. Phys., 1964, 41, 2012. 4 K. H. Welge and F. Stuhl, J. Chem. Phys., 1967,46, 2432.5 F. Stuhl and K. H. Welge, J. Chem. Phys., 1967, 47, 332. 6 L. J. Stief, W. A. Payne and R. B. Klemm, J. Chem. Phys., 1975, 62, 4000. 7 T. G. Slanger and G. Black, J. Chem. Phys., 1982, 77, 2432. 8 M. T. Macpherson and J. P. Simons, Chem. Phys. Lett., 1977, 51, 261. 9 I. P. Vinogradov and F. I. Vilesov, Opt. Spectrosc., 1976, 40, 32; 1978, 44, 653. 10 L. C. Lee, L. Oren, E. Phillips and D. L. Judge, J. Phys. B, 1978, 11, 47. 11 L. C. Lee, J. Chem. Phys., 1980, 72, 4334. 12 0. Dutuit, A. Tabche-Fouhaile, I. Nenner, H. Frohlich and P. M. Guyon, J. Chem. Phys., 1985,83, 584. 13 P. Andresen and E. W. Rothe, Chem. Phys. Lett., 1982, 86, 270; J. Chem. Phys., 1983, 78, 989. 14 P. Andresen, G. S. Ondrey and B. Titze, Phys. Rev. Lett., 1983, 50, 486. 15 P. Andresen, G.S. Ondrey, B. Titze and E. W. Rothe, J. Chem. Phys., 1984, 80, 2548. 16 P. Andresen, V. Beushausen, D. Hausler, H. W. Lulf and E. W. Rothe, J. Chem. Phys., 1985,83, 1429. 17 H. Okabe, J. Chem. Phys., 1980, 72, 6642. 18 A. Gedanken, J. Mol. Spectrosc., 1980, 82, 246. 19 J. P. Simons and A. J. Smith, Chem. Phvs. Lett., 1983, 97, 1. 20 J. P. Simons, A. J. Smith and R. N. Dixon, J. Chem. SOC., Faraday Trans. 2, 1984, 80, 1489. 21 C. Fotakis, C. B. McKendrick and R. J. Donovan, Chem. Phys. Lett., 1981, 80, 598. 22 M. N. R. Ashfold, J. M. Bayley and R. N. Dixon, Chem. Phys., 1984,84,35; Can. J. Phys., 1984,62,1806. 23 A. Hodgson, J. P. Simons, M. N. R. Ashfold, J. M. Bayley and R. N. Dixon, Chem. Phys. Lett., 1984, 24 M. P. Docker, A. Hodgson and J. P. Simons, Mol.Phys., 1986, 57, 129. 25 S. Tsurubuchi, Chem. Phys., 1975, 10, 335. 26 F. Flouquet and J. A. Horsley, J. Chem. Phys., 1974, 60, 3767. 27 F. Flouquet, Faraday Discuss. Chem. SOC., 1977, 62, 143. 28 J. N. Murrell, S. Carter, I. M. Mills and M. F. Guest, Mol. Phys., 1981, 42, 605. 29 E. Segev and M. Shapiro, J. Chem. Phys., 1982, 77, 5604. 30 R. N. Dixon, Mol. Phys., 1985, 54, 333. 31 R. Schinke, V. Engel and V. Staemmler, Chem. Phys. Lett., 1985,116,165; J. Chem. Phys., 1985,83,4522. 32 R. Schinke, V. Engel, P. Andresen, D. Hausler and G. G. Baht-Kurti, Phys. Rev. Lett., 1985,85, 1180. 33 V. Engel, R. Schinke and V. Staemmler, J. Chem. Phys., 1986, in press. 34 K. P. Huber and G. Herzberg, Molecular Spectra and Molecular Structure, IV. Constants of Diatomic 107, 1; Mol. Phys., 1985, 54, 351. Molecules (van Nostrand Reinhold, New York, 1979).110 Hydrogen-atom Photofragment Spectroscopy 35 G. Theodorakopoulis, I. D. Petsalakis and R. J. Buenker, Chem. Phys., 1985,96, 217. 36 J. W. C. Johns, Can. J. Phys., 1963,41, 209. 37 H. T. Wang, W. S. Felps and S. P. McGlynn, J. Chem. Phys., 1977, 67, 2614. 38 E. Ishiguro, M. Sasanuma, H. Masuko, Y. Morioka and M. Nakamura, J. Phys. B, 1978, 11, 993. 39 P. Gurtler, V. Saile and E. E. Koch, Chem. Phys. Lett., 1977, 51, 386. 40 M. N. R. Ashfold, M. T. Macpherson and J. P. Simons, Top. Curr. Chem., 1979, 86, 1. 41 L. Schnieder, H. J. Krautwald, M. N. R. Ashfold and K. H. Welge, to be published. 42 K. R. Wilson, in Symposium on Excited State Chemistry, ed. J. N. Pitts Jr (Gordon and Breach, New 43 G. Theodorakopoulos, C. A. Nicolaides, R. J. Buenker and S. D. Peyerimhoff, Chem. Phys. Lett., 1986, 44 G. Theodorakopoulos, I. D. Petsalakis, R. J. Buenker and S. D. Peyerimhoff, Chem. Phys. Lerr., 1984, 45 V. Staemmler and A. Palma, Chem. Phys., 1985,93, 63. 46 H. Zacharias, H. Rottke, J. Danon and K. H. Welge, Opt. Commun., 1981, 37, 15. 47 D. C . Hanna, M. A. Yuratich and D. Cotter, Nonlinear Optics of Free Atoms and Molecules, Springer 48 R. Mahon, T. J. McIlrath, V. P. Myerscough and D. W. Koopman, IEEE J. Quantum Electron., 1979, 49 G. C. Bjorklund, IEEE J. Quantum Electron., 1975, 11, 287. 50 Y. M. Yiu, T. J. McIlrath and R. Mahon, Phys. Rev. A, 1984, 20, 2480. 51 D. Cotter, Opt. Commun., 1979, 31, 397. 52 R. Wallenstein, Opt. Commun., 1980, 33, 119. 53 H. Langer, H. h e l l and H. Rohr, Opt. Commun., 1980, 34, 137. 54 R. Hilbig and R. Wallenstein, IEEE J. Quantum Electron., 1983, 19, 1759. 55 D. R. Stull and H. Prophet, JANAF Thermochemical Tables (Natl Stand. Ref. Data Ser., US Natl Bur. 56 G. H. Dieke and H. M. Crosswhite, J. Quantum Spectrosc. Radiat. Transfer, 1962, 2, 97. York, 1970). 84, 164. 105, 253. Series in Optical Science, Vol. 17 (Springer, Berlin, 1979). 15, 444. Stand., Washington, D.C., 1979), vol. 9. Received 9th June, 1986