Faraday Discuss. Chem. SOC., 1986, 82, 13-24 Product Correlations in Photofragment Dynamics Gregory E. Hall, Natarajan Sivakumar, Rachel Ogorzalek, Gunjit Chawla, Hans-Peter Haerri? and Paul L. Houston* Department of Chemistry, Cornell University, Ithaca, New York 14853, U.S.A. Itamar Burak School of Chemistry, Sackler Faculty of Science, Tel-Aviv University, Tel-Aviv, Israel John W. Hepburn Centre for Molecular Beams and Laser Chemistry, University of Waterloo, Waterloo, Ontario N2L 3G1, Canada Correlations between either scalar or vector quantities measured in the study of photodissociation dynamics can serve to provide a very detailed picture of the dissociative event. This article discusses the use of Doppler profile and time-of-flight spectroscopy to learn about the correlation between the separate internal energies of two recoiling fragments, to study the way in which the internal energy distribution of a fragment varies with its recoil direction and to determine the angle between a photofragment's recoil velocity direction and its rotation vector.Two new techniques are introduced. High-voltage switching of the potential applied to a time-of-flight mass spectrometer is used to map the velocity distribution of photofragments onto their arrival time distribution. Probing of photofragments by polarized light with sub-Doppler resolution is used to determine the degree of angular correlation between their rotation vector and their recoil velocity vector. The sophistication with which the dynamics of photodissociation events can be probed has increased significantly in the past few years.Measurement of the internal distribution of photofragments (electronic, vibrational, rotational, lambda doublet, fine structure etc.) has become almost determination of the recoil speed distribution by time-of-flight 13925,26 or Doppler techniques22 has blossomed and even measurement of vector properties, such as the direction of a fragment's recoil 13,25,26 or the orientation of its rotation ~ e c t o r ~ ~ ~ ~ ' ~ ~ " with respect to the electric vector of the dissociation source, has become more widely employed. While the collection of measurements concerning these parts of the dissociation event is certainly impressive, one may nonetheless ask whether the whole picture might not be more than the sum of its parts. Could the dissociation event be brought into even better focus by considering the correlations between measurable parameters? As we shall see, there is often as much information in these correlations as in all the parts by themselves. What type of correlations might we expect? Correlations between two scalar proper- ties include those, for example, between the vibrational and rotational energy or between the electronic and vibrational-rotational energy of a particular fragment.Such correla- tions are routinely measured in the course of determining the internal energy for the selected fragment. A more important scalar-scalar correlation is that between, say, the internal energy of a fragment and its recoil speed. Knowledge of the speed of a state-selected fragment allows one to determine the internal energy of the unobserved fragment (assuming there are only two) and often to infer its state (e.g.when the unobserved fragment is an atom in one of two electronic states). Vector-scalar correla- tions are also important. Knowledge of the direction of recoil for a state-selected t Present address: Ciba-Geigy, R. and A. Research Facility, WFM 185.003, CH-1701 Fribourg, Switzerland. 1314 Correlations in Photodissociation fragment with respect to the electric vector of the photolysis source allows one to decompose the overall spatial anisotropy into components for individual internal levels. A bimodal rotational distribution, for example, might be found to display different anisotropy parameters for its high- and low-J components, perhaps indicating that two surfaces are involved.Finally, we can consider vector-vector correlations. An example is the correlation between the recoil velocity vector of a diatomic fragment and its rotation vector. If these vectors were found to be perpendicular, one might infer that dissociation took place from a planar intermediate, whereas if they were found to be parallel one might infer that dissociation involved a torsional mode. Thus, knowledge of the correlations between observables in a photodissociation can lead to a much more detailed picture of the process than might be gained from the observables by themselves. This paper is devoted to examination of such correlations. The relationship between the internal energy and the recoil speed will be used to decide whether a given CO( u, J) product of OCS dissociation is produced in coincidence with S('D) or S(3P).The correlation between the internal energy and the recoil direction will be used to show that OCS and CD31 dissociate on a time scale short compared to rotation and to show that dissociation of the former molecule probably involves more than one excited surface. Finally, the angular correlation between the recoil velocity vector and the angular momentum vector of the CO product in the photolysis of OCS and glyoxal will be used to show that these molecules dissociate from a planar configuration. Experimental Two different experimental arrangements were used in these studies. In each case the sample was prepared by nozzle expansion and photolysed with one of several available laser sources.Products were detected in one case by laser-induced fluorescence and in the other by multiphoton ionization. For studies of OCS and glyoxal, a pulsed nozzle source without collimation was used to prepare the sample and a tunable vacuum- ultraviolet laser was employed to probe the CO photo fragments by laser-induced fluores- cence. For studies of methyl iodide, a well collimated molecular beam was used to prepare the sample and multiphoton ionization using a time-of-flight mass spectrometer was employed for detection of the I and CD3 products. Each of the systems is described in greater detail below. OCS and Glyoxal Studies Photolysis of OCS and glyoxal was performed in a pulsed supersonic jet apparatus similar to that described el~ewhere.~~ Mixtures were prepared by flowing helium over OCS or glyoxal, which was held in a trap at a specified temperature.The seeded gas was then expanded from a total pressure of 1300Torrt through a 0.5 mm diameter pinhole using a pulsed nozzle assembly (Newport). A KrCl excimer laser (Lumonics TE861-4) or an Nd:YAG pumped dye laser/wavelength extension package (Quanta Ray, DCR-2A, PDL-1, WEX) was used to generate the 222 nm light for dissociation of OCS. An excimer pumped dye laser (Lambda Physik EMG-101, FL2001E) was used to excite glyoxal on the 8; band of its first excited singlet state. In both cases, the laser intersected the molecular jet ca. 1.25 cm from the nozzle source. The CO(X 'Z, v, J) product of either dissociation was probed by laser-induced fluorescence using a tunable vacuum-ultraviolet source based on four-wave mixing in magnesium v a p o ~ r .~ ~ The V.U.V. source has been described in detail in recent studies of the photodissociation of g l y ~ x a l ~ ~ and OCS35 and is very similar to that used previously for detection of Br and C0.36-41 t 1 Tom = 101 325/760 Pa.G. E. Hall et al. 15 The molecular jet, the dissociation laser, and the V.U.V. source propagated in mutually orthogonal directions for the studies of OCS, while for studies of glyoxal the dissociation and V.U.V. sources were propagated in antiparallel directions perpendicular to the molecular jet. Fluorescence from CO was detected at 45" from the probe laser and 90" from the jet by an EMR solar-blind photomultiplier tube (5416-09-17).A second solar-blind photomultiplier (EM1 G-26E3 14LF) detected a reflection of the tunable V.U.V. light; its signal was used to normalize the spectra of CO for variations in the probe laser intensity. CDJ Studies Photolysis of CD31 was performed in an apparatus consisting of a pulsed supersonic nozzle source, two skimmers and a standard time-of-flight mass ~pectrometer.~~ Its design is similar to systems described previously43 and will be reported in detail else- where.'"' Methyl iodide was dissociated using the 266nm output of a quadrupled ND:YAG laser (Quanta-Ray DCR-l), whose linear polarization could be rotated with a waveplate. Two-photon resonant/ three-photon ionization of the I and CD, fragments was induced with a doubled Nd:YAG pumped dye laser (Quanta-Ray, DCNA, PDL-1).The relevant experimental conditions are as follows: source pressure, 8% CD31 in 1900 Torr helium; source temperatures, 298 K; molecular beam diameter at laser intersection, 1.6 mm; beam velocity 11 80 m s-'; flight tube length, 105 cm; extraction field, 500 V/2.54 cm; acceleration field, 1500 V/ 1.27 cm; dye laser wavelength, 334-310 nm (DCM); dye laser energy/pulsewidth, 1 mJ/ 10 ns; photolysis energy/pulsewidth, 5 mJ/ 10 ns. The photolysis and probe lasers were made to intersect each other and the molecular beam at a location 8 mm upstream from the centre line of the mass spectrometer and were timed so that ionization was induced <50 ns after photolysis. The extraction and acceleration voltages were held at ground potential during the photolysis and for an adjustable time delay, r d , thereafter.During this delay the ionized fragments recoil with the velocity imparted by the dissociation. Following the delay Td, voltages were applied to the grids of the time-of-flight mass spectrometer during a switching time of ca. 0.05 ps. This voltage caused the ionized photofragments to accelerate toward the detector, where their arrival time was a sensitive function of their position and velocity at the time Td. Thus, the shape of the appropriate mass-spectral peak in the time-of-flight distribution provided information concerning the recoil velocity of photofragments in the particular state to which the ionization laser was tuned?' Results and Discussion Correlation between Internal Energies of Two Photofragments Dissociation of OCS at 222 nm can result in two possible sets of products: ocs + CO(u, J ) + S ( , P ) + CO(u,J)+S('D). Ca.19 880 cm-' of energy is available to recoil and CO internal excitation in the former case, while only 10640cm-' is available in the latter case. Direct monitoring of the S( ' D ) and S ( , P ) atoms by V.U.V. laser-induced fl~orescence~'~~ has shown that the S( ' D ) branching fraction is 0.85. The CO product is formed almost exclusively in the o = 0 level, with a rotational distribution shown in fig. 1. It is tempting to speculate that the smaller peak at higher J in the rotational distribution corresponds to those CO fragments produced in coincidence with the less probable S(,P) channel and that the larger peak at lower J corresponds to the dominant S('D) channel.Indeed, such a hypothesis was16 Correlations in Photodissociation I 2 5 - h 10 1 0- i3 0 0 A Q a 0 b 0 8 0 0 0 35 45 55 65 75 rotational level ( J ) Fig. 1. Rotational distribution of CO( u = 0) produced in the 222 nm photodissociation of OCS. A, 0,O indicate measurements on P-, Q-, and R-branch transitions, respectively. Note that there are two peaks in the distribution, one near J = 56 and one near J = 65. Since the peak areas are roughly in the 85/15 ratio, corresponding to S('D)/S(3P), one might expect that the peak near J = 56 corresponds to CO( = 0) produced in coincidence with S('D), while that near J = 65 corresponds to CO( u = 0) produced in coincidence with S(3P). Analysis of the Doppler profile of the individual rotational lines can confirm or disprove this hypothesis. proposed by us in an earlier p~blication.~~ It is possible to test this hypothesis by measuring the Doppler profiles of individual CO rotational lines, e.g. J = 56 and J = 66, which are near the maxima of the two peaks.If all the CO were produced in coincidence with S( ' D ) , for example, then the Doppler linewidth of the J = 66 line would be narrower than that of the J = 5 6 line, since this product would have a slightly higher internal energy (by 2375 cm-') with correspondingly less energy for recoil. On the other hand, if CO(J = 66) were produced in coincidence with S(3P), while CO(J = 56) were produced in coincidence with S('D), there would be 10 640-2375 = 8265 cm-' more energy avail- able for recoil of CO(J=66) than for CO(J=56); i.e.the Doppler width of the CO(J=66) would be broader than that for CO(J=56). The data, shown in fig. 2, demonstrate clearly that both CO molecules are produced in coincidence with S('D). Thus, it is possible through measurements of the magnitude of the recoil velocity to determine the correlation between the internal energies of the two photofragments. Correlation between the Internal Energy of a Photofragment and its Recoil Direction Similar correlations are possible between scalar and vector quantities, for example between the internal energy of a photofragment and its recoil direction. Consider the photodissociation by linearly polarized light of a molecule into two fragments. It is well known that the probability of finding the recoil velocity vector for the fragments at an angle of 0 with respect to the polarization vector of the dissociating light can be described by the function 1 +/3P2(cos @).47i48 A value of p is derived from the Doppler profile of a single internal energy state of the fragments, measured with a tunable probleG.E. Hall et al. 17 I 1 1 1 -0.50 -0.25 0.00 0.25 0.50 detuning/ cm-’ Fig. 2. Doppler profiles of the CO( v = 0, J = 56) ( a ) and CO( v = 0, J = 66) ( b ) produced in the 222 nm photodissociation of OCS. Q-branch transitions were used and the circularly polarized probe laser propagated in a direction perpendicular to both the electric vector and the propagation direction of the linearly polarized photolysis laser.The Doppler width of the J = 6 6 line is narrower than that for the J = 56 line, so it cannot be the case that J = 66 is produced in coincidence with S(3P), while J = 56 is produced in coincidence with S(’D). In fact, the widths and shapes for both lines are in very good agreement with those predicted (smooth curves) if both fragments are produced in coincidence with S( * D ) . laser propagating perpendic2larly to the direction of the dissociation laser and at an angle 8’ with respect to the E vector of the dissociation laser. For fragments of sharply defined kinetic energy rnvi/2, the probability of finding a recoil velocity vector at an angle x with respect to the probe beam propagation direction is given by49 mx7 8’)K 5 d 4 W X , 8‘, 4 ) (1) where wx7 87 4 ) = [1 +PP,(COS 811 cos 8 = cos 8’ cos x +sin 8’ sin x cos 4 ( 2 ) (3) with and 4 is the azimuthal angle of o about the probe direction as measured from the plane containing the probe direction and the polarization vector of the dissociation laser.The result of integration over + is4’ D(X, el) oc 2 T [ 1 + pp2(c0S x) P,(CO~ 8’11 (4) where the component of the velocity responsible for the Doppler detuning is vocosx. If the fragment has no angular momentum or if J is uncorrelated with 0 Doppler lineshapes can be analysed simply with eqn (4) to determine p. In this way, the angular18 Correlations in Photodissociation I 1 1 1 d -0.50 -0.25 0.00 0.25 0.50 detuning/cm-' Fig. 3. Q-branch Doppler profiles for the CO(v = 0, J = 66) ( a ) , CO(v = 0, J = 57) (b), and CO( u = 0, J = 46) ( c ) products of 222 nm OCS photodissociation.The circularly polarized probe laser propagated in a direction perpendicular to that of the photolysis laser, but parallel to its polarization vector. The recoil velocity distribution is characterized by p = 1.8 for J = 66, p = 0.7 for J = 57, and p = -0.5 for J = 46. The change in p with J suggests that more than one dissociative potential surface is involved. distribution of product recoils may be measured for a particular internal energy level of the fragment. By combining such correlated measurements for many internal energy levels one can separate the internal energy distribution into component parts having similar recoil velocity distributions. The discovery of more than one component may imply that the dissociation involves more than one excited surface.The OCS photodissociation offers an example of the correlation between internal energy and fragment recoil direction. Fig. 3 displays the Q-branch Doppler profiles for three CO(u = 0, J) lines (J = 66, 57, 46) following dissociation of OCS with linearly polarized light aligned parallel to the propagation direction of the circularly polarized probe laser. The profile at the highest J shows a pronounced dip in the centre, indicative of excitation via a near-parallel OCS transition followed by prompt fragmentation. However, as lower rotational values of J are probed, the Doppler profile shifts to oneG. E. Hall et al. 19 22 .o 22.5 23 .O arrival time/ps Fig. 4. Arrival time distribution of the I+ ion from multiphoton ionization of I* following 266 nm dissociation of CD31 with light polarized parallel ( a ) or perpendicular ( b ) to the acceleration axis of the time-of-flight mass spectrometer.which is more peaked at the centre frequency, indicative of a more perpendicular transition in the OCS. One way to summarize these results qualitatively is to note that the recoil velocity distribution changes with the rotational level of the photofragment from a nearly cos2 8 distribution at high J to a more sin2 8 distribution at lower J. An alternative description is that the rotational distribution of the CO product varies strongly with the recoil angle used for detection: the rotational distribution observed along the direction of the electric vector of the photolysis light is peaked at high J, while that observed perpendicular to the direction of the electric vector is peaked at lower J.Since in addition to this correlation between internal energy level and recoil direction, there is a strong correlation between recoil direction and angular momentum vector (see below), we will postpone for the moment a more quantitative description. offers a second example of the correlation between internal energy and fragment recoil velocity. In this experiment the CD3, I and I* fragments were probed by multiphoton ionization using the pulsed extraction technique described earlier. Fig. 4 displays the arrival time spectrum of the I' ion produced by multiphoton ionization of I* at 31 1.4 nm following photolysis of CD31 with 266nm light polarized either parallel or perpendicular to the centre line of the time-of-flight mass spectrometer.Because the CD31 dipole moment is parallel to the symmetry axis, and because the dissociation is rapid compared to the CD31 rotational period, the CD3 and I or I* fragments recoil predominantly along the polarization axis of the dissociation laser. Ionization of the state-selected fragment occurs immediately following the photolysis pulse (within 50 ns), and the concentration of ions is low enough that space-charge effects can be ignored. Thus, while the field is held at zero, the ions recoil with the velocity distribution imparted by the dissociation. When the field is switched on after the delay Td, the ions are accelerated toward the detector, where their arrival time is a measure of their position and velocity at Td.For The 266 nm photodissociation of CD31 to yield CD3 + I, I* (I* =20 Correlations in Photodissocia tion the ‘parallel’ geometry, the products which have recoiled forward and backward along the polarization vector have different velocities and positions along the field direction and thus arrive at the detector at different times. For the ‘perpendicular’ geometry, on the other hand, both sets of products have the same velocity and position along the field direction and thus reach the detector at the same time. Detailed analysis of the arrival time distribution can be used to obtain the velocity distribution of the state-selected fragment. Preliminary analysis suggests that the I* fragment recoils with an anisotropic distribution characterized by p = 2.0.Although the analysis of the corresponding data for the I fragment is complicated by the relative strength of the I+ signal caused by the probe laser alone, the results also indicate a high degree of anisotropy for the I fragment. For the particular time-of-flight tube employed in these experiments (105 cm) it was not possible to image all of the CD3 fragments. In principle, however, with a shorter flight tube it should be possible with this technique to measure the velocity distribution of CD3 fragments in a selected vibrational and rotational level. Correlation between a Photofragment’s Velocity and Angular Momentum Vectors Correlations between vector quantities can also yield important information about the photodissociation process, as shown by consideration of an example when the two vectors u and J , describing the recoil velocity and the angular momentum of a product, respectively, have an angular correlation.Suppose that the fragment with angular momentum J is a symmetric-top molecule and let the projection of J onto u be described by a distribution P(M,). If there is an angular correlation between u and J, then the projection of J onto u will produce a non-statistical population of Mu levels. E.g. if J is constrained to be perpendicular to u, then P ( Mu = 0) = 1 and all other probabilities are zero. Now suppose that the molecule is probed by laser-induced fluorescence with light propagating along an axis z which makes an angle x with u. The projection of P(Mu) onto z to give P’(MJ) will produce a distribution of M, values that depends on the angle x, i.e.one that changes with the velocity probed by the Doppler technique. To pursue the above example, if J is perpendicular to u, then molecules moving with u parallel to z will have only M, = 0, while those moving with u perpendicular to z will have a broad distribution of MJ values. Since polarized light will interact differently with different Mj distributions, the actual LIF Doppler profile will be different from that predicted by eqn (4). The change in the Doppler profile caused by the u-J correlation will depend on several factors: the polarization characteristics of the probing light, the angle and polarization acceptance of the detector and the nature of the absorption and fluorescence transitions.Measurement of the Doppler profile as these factors are varied can be used to uncover the degree of u-J angular correlation. The detailed mathematical relationship between the Doppler profile and the degree of u-J correlation has been treated elsewhere50i51 so we will concentrate here on some applications. Consider again, for example, the dissociation of OCS at 222 nm. Conserva- tiop of angular momentum for this dissociation gives the following relation: where L is the orbital angular momentum of the half collision. Since the OCS is dissociated from very low rotational levels (Jocs=O) and since Jco is so large that JhY = 1 and Js = 2 can be ignored, we find that L = -J Because L must be perpendicular to u, it follows that J is perpendicular to u.The lower row of panels in fig. 5 displays the experimental data (dots) and the convolution of the laser linewidth (0.14 cm-’ f.w.h.m.) with the Doppler profiles predic- ted by with an anisotropy parameter of p = 0.6 (solid lines). The upper row of panels shows the convoluted profiles expected in the absence of any u-J correlation-0.5 0.0 0.5 0.0 0.5 0.0 -0.5 0.0 0.5 detuning/cm-' Fig. 5. Experimental and calculated data for the Q( 59) and P( 59) lines of CO produced in the photodissociation of OCS and probed by laser-induced fluorescence on the CO A 'II 4- X 'X transition using circularly polarized light. The dissociation and LIF probe lasers were orthogonal. Upper row: Doppler profiles expected in the absence of v-J correlation (solid lines); the Gaussian laser linewidth is shown in the dashed curve.Bottom row: Experimental profiles (dots) and profiles calculated for v l J (%olid curves). From left to right the panels show the Q( 59) line with the electric vector of the dissoci%ting light E aligned perpendicular to the propagation vzctor of the probe light z ( e l = 90°), the Q(59) line with El(< ( O ' = O'), the third panel is for the P(59) line with E L 2 (O'=90°) and the last panel is for the P(59) line with Ell2 ( 6 ' = 0'). Calculations were made with values of p=O.6 for the recoil anisotropy, 0.14cm-' for the f.w.h.m. laser linewidth, and 1232ms-' for the recoil velocity, as determined by energy and momentum conservation.22 Correlations in Photodissociation -1.0 -0.5 0.0 0.5 1.0 -1.0 -0.5 0.0 0.5 1.0 detuning/cm-' Fig. 6.Calculated and experimental profiles for CO produced in the photodissociation of glyoxal. Top row: calculation assuming a Boltzmann distribution of CO speeds and assuming u l J . Middle row: calculation assuming a Boltzmann distribution of CO speeds and assuming ullJ. In each of the top two rows the left panel is for an R- or P-branch transition. Bottom row: experimental data for Q(41) and P(41) transitions. for /3 = 0.6. The laser linewidth is given as the dotted line in the first panel. It is clear from the figure that lineshapes in the presence of u-J angular correlation (solid lines, bottom row) are qualitatively different from those when there is no correlation (solid lines, top row), and that there are differences in the lineshapes depending on the transition probed and the angle 6'.An attempt to fit these data without including the effect of u - J correlation would lead to different apparent values of /3 for different 6', as well as incorrect total linewidths. Inclusion o'f the u - J correlation is necessary to obtain reasonable fits to the data. As a second example, consider the dissociation of the planar glyoxal molecule (trans-CHOCHO) to give CO and a variety of possible other products: CHOCHO + CO+H,CO + 2CO+H, -+ CO+CHOH. In the case of OCS it was possible to predict beforehand that J is perpendicular to u, but in the dissociation of a molecule with more than three atoms such a prediction is not feasible. However, there are two interesting limits that can be considered.If the glyoxal retains its planar configuration throughout the dissociation, we would expect that the CO would rotate in the plane and that J would be perpendicular to u. On the other hand, if the dissociation is accompanied by substantial out-of-plane bending and if it is primarly this torsional motion that leads to the rotational excitation of the CO, then one would expect J to be aligned more nearly parallel to u.G. E. Hall et al. 23 Analysis of the Doppler profiles allows one to distinguish between these two cases. In general, as we have seen above, the Doppler profile will contain information concern- ing both the distribution of recoil velocities and the correlation of J and u. Fragmentation in glyoxal is caused by predissociation of the first-excited singlet level on a time scale (ca.1 ps) which is extremely long compared to the parent molecule’s rotational period. It is expected, therefore, that the distribution of recoil velocities will be essentially isotropic. Thus, the Doppler profile of the CO is determined simply by the u-J correlation and by the distribution of recoil speeds. In the limiting case of a single recoil speed for a given internal state and for the counterpropagating photolysis/ probe configuration employed, the predicted51 Doppler profiles of the CO A + X lines are as follows. (a) For Q lines with the u l J constraint as well as for P or R lines with the u IJJ constraint, a relative minimum occurs at the centre frequency. (b) For P or R lines with the u llJ constraint as well as for Q lines with the u l J constraint, a relative maximum occurs at the centre frequency.Of course, because there are several possible unprobed products ( H2C0, H2, CHOH) and because these may each have a distribution of internal energies, the recoil speed of the CO will not be sharp. The composite Doppler profile of CO molecules for a particular rotational transition will be an integral over the speed distribution of a speed-dependent profile function, whose shape for the particular transition is described above. E g . , a Boltzmann distribution of speeds would give rise to the composite profiles illustrated in the top two rows of fig. 6, depending on whether v is aligned perpendicular or parallel to J. The actual data, shown in the bottom row of fig.6, clearly indicate that the predominant orientation of J is perpendicular to u. Thus, by careful analysis of the Doppler profile, it is possible to determine whether the transition state in a photodissociation is planar or non-planar. Using similar techniques, Dubs et aZ.52 have recently inferred a planar transition state in the photodissociation of (CH3)*NN0. Conclusion The application of advanced laser-based techniques to the study of molecular photodis- sociation can yield correlations between scalar and/ or vector properties. Often, these correlations provide a much more detailed picture of the photodissociation than the scalar or vector properties taken alone. Specific correlations between the internal energy levels in the departing fragments, between the internal energy of a fragment and its recoil direction and between the recoil velocity vector and the rotational angular momentum vector have been demonstrated.It seems likely that such correlations will play an increasingly important role in studies of molecular photodissociation. This work was supported by the U.S. National Science Foundation (CHE-8314146) and the U.S. Air Force Office of Scientific Research (F49620-83-K-0012). The research employed lasers funded through the U.S. Department of Defense Instrumentation programme (DAAG29-84G-0076) and the Dow Chemical Foundation. J.W.H. gratefully acknowledges support from NSERC (Canada) and NATO (grant no. 637/83). The partial financial support of H-P.H. by the Swiss National Science Foundation is also gratefully acknowledged.References 1 P. Andersen, G. S. Ondrey, B. Titze and E. W. Rothe, J. Chem. Phys., 1984,80, 2548. 2 P. Andresen, G. S. Ondrky and B. Titze, Phys. Rev. Lett., 1983, 50, 486. 3 P. Andresen and E. W. Rothe, Chem. Phys. Lett., 1982, 86, 270. 4 P. Andresen and E. W. Rothe, J. Chem. Phys., 1983 78,989. 5 M. P. Docker, A. Hodgson and J. P. Simons, Mol. Phys., to be published. 6 A. Hodgson, J. P. Simons, M. N. R. Ashfold, J. M. Bayley and R. N. Dixon, Chem. Phys. Lett., 1984, 107, 1.24 Correlations in Photodissociation 7 A. Hodgson, J. P. Simons, M. N. R. Ashfold, J. M. Bayley and R. N. Dixon, Mol. Phys., 1985,54, 351. 8 J. P. Simons and A. J. Smith, Chem. Phys. Lett., 1983, 97, 1. 9 J. P. Simons, A. J. Smith and R. N. Dixon, J. Chem. Soc., Faraday Trans.2, 1984, 80, 1489. 10 R. Vasudev, R. N. Zare and R. N. Dixon, Chem. Phys. Lett., 1983,96,399; J. Chem. Phys., 1984,80,4863. 11 G. E. Hall, N. Sivakumar and P. L. Houston, J. Chem. Phys., 1986, 84, 2120. 12 R. J. Donovan and J. Konstantatos, J. Photochem., 1972, 1, 7 5 . 13 J. H. Ling and K. R. Wilson, J. Chem. Phys., 1975, 63, 101. 14 M. J. Sabety-Dzvonik and R. J. Cody, J. Chem. Phys., 1977, 66, 125. 15 A. P. Baronavski and J. R. McDonald, Chem. Phys. Lett., 1977,45, 172. 16 S. T. Amimoto, J. R. Wiesenfeld and R. H. Young, Chem. Phys. Lett., 1979, 65, 402. 17 W. M. Pitts and A. P. Baronavski, Chem. Phys. Lett., 1980, 71, 395. 18 A. P. Baronavski, Chem. Phys., 1982, 66, 217. 19 W. Krieger, J. Hager and J. Pfab, Chem. Phys. Lett., 1982, 85, 69. 20 W. H. Fisher, T.Carrington, S. V. Filseth, C. M. Sadowski and C. H. Dugan, Chem. Phys., 1983,82,443. 21 I. Nadler, H. Reisler and C. Wittig, Chem. Phys. Let?., 1984, 103, 451. 22 I. Nadler, D. Mahgerefteh, H. Reisler and C. Wittig, J. Chem. Phys., 1985, 82, 3885. 23 W. J. Marinelli, N. Sivakumar and P. L. Houston, J. Phys. Chem., 1984,88, 6685. 24 J. V. V. Kasper and G. C. Pimentel, Appf. Phys. Lett., 1964, 5, 231. 25 S. J. Riley and K. R. Wilson, Faraday Discuss. Chem. Soc., 1972, 53, 132. 26 M. J. Dzvonik and S. C. Yang, Rev. Sci. Instrum., 1974, 45, 750. 27 S. L. Baughcum and S. R. Leone, J. Chem. Phys., 1980, 72, 6531. 28 R. K. Sparks, K. Shobatake, L. R. Carlson and Y. T. Lee, J. Chem. Phys., 1981, 75, 3838. 29 H. W. Hermann and S. R. Leone, J. Chem. Phys., 1982,76,4759.30 G. N. A. van Veen, T. Bailer, A. E. de Vries and N. J. A. van Veen, Chem. Phys., 1984,87, 405. 31 M. D. Barry and P. A. Gorry, Mof. Phys., 1984, 52, 461. 32 R. D. Bower, R. W. Jones and P. L. Houston, J. Chem. Phys., 1983, 79, 2799. 33 S. C. Wallace and G. Zdasiuk, Appl. Phys. Lett., 1976, 28, 449. 34 J. W. Hepburn, N. Sivakumar and P. L. Houston, in Laser Techniques in the Extreme Ultraviolet, ed. 35 N. Sivakumar, I. Burak, W-Y. Cheung, P. L. Houston and J. W. Hepburn, J. Phys. Chem., 1985,89,3609. 36 J. W. Hepburn, D. Klimek, K. Liu, J. C. Polanyi and S. C. Wallace, J. Chem. Phys., 1978, 69, 431 1. 37 J. W. Hepburn, D. Klimek, K. Liu, R. G. Macdonald, F. J. Northrup and J. C. Polanyi, J. Chem. Phys., 38 J. W. Hepburn, K. Liu, R. G. Macdonald, F. J. Northrup and J. C. Polanyi, J. Chem. Phys., 1981, 75, 39 J. W. Hepburn, F. J. Northrup, G. L. Ogram, J. M. Williamson and J. C. Polanyi, Chem. Phys. Lett., 40 P. Ho and A. V. Smith, Chem. Phys. Lett., 1982,90, 407. 41 D. J. Bamford, S. V. Filseth, M. F. Foltz, J. W. Hepburn and C. B. Moore, J. Chem. Phys., 1985,82,3032. 42 W. C. Wiley and I. H. McLaren, Rev. Sci. Instrum., 1955, 26, 1150. 43 T. G. Dietz, M. A, Duncan, M. G. Liverman and R. E. Smalley, J. Chem. Phys., 1980,73,4816. 44 H-P. Haem, R. Ogorzalek and P. L. Houston, in preparation. 45 G. E. Hall, H-P. Haerri, R. Ogorzalek and P. L. Houston, in preparation. 46 N. Sivakumar, G. E. Hall, P. L. Houston, J. W. Hepburn and I. Burak, in preparation. 47 R. N. Zare and D. R. Herschbach, IEEE Proc. Nutf Aerosp. Electron. Con$, 1963, 51, 173. 48 S. Yang and R. Bersohn, J. Chem. Phys., 1974, 61, 4400. 49 R. Schmiedl, H. Dugan, W. Meier and K. H. Welge, 2. Phys., 1982, A403, 137. 50 G. E. Hall, N. Sivakumar, P. L. Houston and I. Burak, Phys. Rev. Lett., 1986, 56, 1671. 51 G. E. Hall, N. Sivakumar, P. L. Houston and I. Burak, in preparation. 52 M. Dubs, U. Bruhlman and J. R. Huber, J. Chem. Phys., 1986, 84, 3106. S. E. Harris and T. B. Lucatorto (AIP Conf. Proc. No. 119, 1984), pp. 126-134. 1981, 74, 6226. 3353. 1982, 85, 227. Received 19th May, 1986