STATE-SELECTED PHOTODISSOCIATION DYNAMICS OF FORMALDEHYDE

The study of photofragmentation dynamics allows unimolecular chemical reactions to be examined in a state-resolved manner. Such state resolution can reveal a great deal about the potential energy surfaces which control these elementary events, and can serve as a crucial test for theories of reaction dynamics on those surfaces. The internal energy (v,J) distributions of photofragments have often been determined by laser-induced fluorescence, in cases here the fragments absorb in the visible or near-ultraviolet re1,2 gions accessible with commercial tunable dye lasers If the fragments do not have such convenient electronic spectra, the techniques of nonlinear optics must be used to provide the necessary sensitivity and spectral resolution. In the research presented here, such techniques have, for the first time, completely determined the distribution of energy among all the possible degrees of freedom in the photodissociation of a tetratomic molecule. In addition, rotational state resolution has been obtained in both entry and exit channels for the photofragmentation.


i. INTRODUCTION
The study of photofragmentation dynamics allows unimolecular chem- ical reactions to be examined in a state-resolved manner.Such state resolution can reveal a great deal about the potential ener- gy surfaces which control these elementary events, and can serve as a crucial test for theories of reaction dynamics on those sur- faces.The internal energy (v,J) distributions of photofragments have often been determined by laser-induced fluorescence, in cases here the fragments absorb in the visible or near-ultraviolet re- 1,2 gions accessible with commercial tunable dye lasers If the fragments do not have such convenient electronic spectra, the tech- niques of nonlinear optics must be used to provide the necessary sensitivity and spectral resolution.In the research presented here, such techniques have, for the first time, completely deter- mined the distribution of energy among all the possible degrees of freedom in the photodissociation of a tetratomic molecule.In ad- dition, rotational state resolution has been obtained in both entry and exit channels for the photofragmentation.

FIGURE
Schematic diagram of potential surfaces and photophysical processes for formaldehyde 4 Formaldehyde is an ideal system in which to study photofragment (v,J) distributions for several reasons.To carry out state-selected photodissociation of formaldehyde fol- lowed by state-specific product detection of CO, two separate pulsed ii tunable laser systems were needed The experimental concept was simple.The photolysis laser was tuned onto a formaldehyde absorp- tion peak.A short time after the photolysis pulse, the probe la- ser was fired.The probe laser wavelength was then scanned to ob- tain a fluorescence excitation spectrum of the photochemical CO.
Figure 2 shows the overall schematic.
Tunable ultraviolet radiation for state-selective formaldehyde photolysis was provided by a commercial system consisting of an excimer laser (Lumonics TE-861, I0 Hz, 15 ns) operating on XeCl at 308 nm which pumped a dye laser (Lambda-Physik 2002E).The dye -I laser was etalon-narrowed at 0.09 cm bandwidth and pressuretuned to produce tunable radiation in the 339 nm region.Formalde- hyde absorption lines were identified by fluorescence excitation 15 16 spectra using line lists provided by Ramsay The vacuum ultraviolet probe laser system was patterned after FIGURE 2 Schematic of photofragmentation experiment for CO(v,J) measurements.the original Toronto system 17.The third harmonic of a Nd:YAG laser (Quanta-Ray DCR-IA) pumped two dye lasers.One of these lasers (Lambda-Physik 2002) was fixed at a wavelength of 430.9 nm, cor- responding to a two-photon resonance in magnesium ( ).The other (o2 2 dye laser (Quanta-Ray PDL-I) was tuned between 470 -510 nm or between 565 600 nm, depending on the band of CO being studied.
Energy from each of the dye lasers was typically 1.0 1.5 mJ per 5 ns pulse.The dye beams were combined in a Glan Prism and focus- ed into the center of magnesium heat-pipe oven.Tunable VUV radia- tion (to 8 2o + 2 was generated in the focal region by re- 18 sonantly enhanced four-wave mixing Carbon monoxide was excited + 19 via the A X Z transition Fluorescence at right angles to the plane of intersection of photolysis and probe lasers was detected by a solar-blind photomultiplier tube (EMR 542G-09-18).
From spectra of room temperature CO, a detection limit of 108 3 molecules per cm per quantum state was inferred.Absolute pulse energies were not measured, but the observed CO detection effi- 11 12 ciency corresponded to between I0 and 10 photons per pulse.To obtain a spectrum of the photochemical CO, the photolysis laser was tuned onto the desired formaldehyde absorption.Then, the la- sers were synchronized to give a delay of 150 500 ns between photolysis and probe pulses.At the formaldehyde pressures of 0.05 0.i0 Torr used, this was a sufficently short delay to allow unre- laxed rotational destributions of CO to be obtained.The probe laser was then tuned to obtain a laser-induced fluorescence spectrum of the photochemical CO.A typical spectrum is shown in Fig. 3.

B. Results
Carbon monoxide rotational distribution following H2CO photolysis are shown in Figure 4.The CO has a remarkable amount of rotation- al excitation, with no detectable population in J states below 20.
Room temperature CO, by contrast, has a rotational distribution  Top trace is for CO(v=0) produced by photolysis on the rQI(3)E+rQI(4)0 transitions of H2CO.Pressure was 0.05 Torr and delay 150 ns.Middle trace is for CO(v=0) produced by photolysis on the PRI(15)0 transition of H2CO at 29490.5 cm -I Conditions identical to those in the top trace.Bottom trace is for CO(v=l) following pRI(15)0 photolysis of 0.i0 Torr of H2CO with 150 ns delay.The three traces have been displaced by 2 verti- cal units.The curves have all been normalized to the same area, with one vertical unit corresponding to a probability of 0.0178 for formation of a given J state of CO (from Ref. ii, with permission).
peaking at J 7. CO (v=l) has nearly the same rotational distri- bution as CO(v=0).The CO has very little vibrational excitation, in agreement with earlier infrared experiments9; no signals were seen for CO(v > I). Figure 5 shows that increased rotational excita- tion of H2CO leads to a slightly broader CO rotational distribu- tion, without changing the most probable J value.As Figure 6 de- monstrates, the CO rotational distribution cannot be characterized by a temperature.
CO Rotational Quantum Number FIGURE 5 The effect of initial H2CO angular momentum on the CO(v=0,J) distribution.The hand-drawn lines through the data points in the top two traces of Figure 4 show a broader distribution for the higher initial angular momentum ( Ero (cm -I 99 FIGURE 6 Boltzman plot for the data in the top trace of Fig. 4. The distribution cannot be characterized by a temperature (from Ref. ii, with permission).

HYDROGEN (v,J) DISTRIBUTION BY CARS SPECTROSCOPY
The vibrational and rotational distribution of the H 2 was deter- mined using Coherent Anti-Stokes Raman Scattering(CARS), in a 12,13 series of experiments carried out in France The experimental concept was the same as that of the CO experiments described above, 20 with the tunable VUV system replaced by the CARS spectrometer The spectrometer has a detection limit for H 2 of approximately ii 3 I0 molecules per cm per quantum state, making photolysis at the low pressures of formaldehyde used in the CO experiments impos- sible.Fortunately, it was observed that the (v,J) states of H 2 formed in formaldehyde photolysis have slow relaxation rates, so that distributions obtained from about 3 Torr of formaldehyde with 7 Torr of helium, with a 150 ns delay between photolysis and probe lasers were almost unrelaxed.The photolysis laser was a Quantel YAG-pumped, frequency-doubled dye laser.
The complete (v,J) distribution following H2CO photolysis at -i 29496 cm is shown in Figure 7.The hydrogen has substantial vib- rational and modest rotational excitation.Because the photolysis laser only excited formaldehyde molecules in odd K states (ortho-

DISCUSSION
The available data can be combined to give a complete account of the energy disposal in this photodissociation.For each degree of freedom the experimental energy distribution is summarized in Table I.
Because the angular momentum of the CO is much larger than both the initial angular momentum of the H2CO and the angular Impact Parameter (,) FIGURE 8 Calculated impact parameter distributions assuming uncorrelated (v,J) states of the two products.Because the distance from the carbon atom to the center of mass is 0.6 in both the transition state and the free CO molecule, most of the impact parameters cor- respond to a point of impact outside the carbon atom.momentum of the H 2 product, the fragments must fly apart with a large amount of orbital angular momentum.The data presented above can be used, along with simple angular momentum conservation, to calculate a distribution function for the product impact parameter, b.The result is shown in Figure 8. Inspite of the uncertainty about the relative directions of CO and H 2 rotation, the most probable impact parameter is clearly quite large, indicative of a 22 bent transition state with the hydrogen having a point of impact a fraction of Angstrom outside the nucleus of the carbon atom.

CONCLUSION
The experimental characterization of the dynamics of formaldehyde photodissociation is now nearly complite.The use of well-established techniques of nonlinear optics has made this the most thor- oughly studied and well-understood polyatomic photofragmentation.
Application of these techniques to other problems in reaction dynamics will provide a sensitive test for both theoretical potential energy surfaces and dymanical calculations using those surfaces.
The relevant energy levels and dissociation limits are shown in Figure I.Its well- 3 resolved and well -understood UV spectrum allows for state-specific excitation with a narrowband UV dye laser.The mechanism of 4 the dissociation, after decades of work, is now well established The molecule internally converts to high vibrational levels of the ground electronic state and then dissociates in the absence of 5-7 8 collisions.The fragments, H 2 and CO are both detectable in low concentrations using available laser spectroscopic techniques.Early work on this photodissociation measured a vibrati.onaldistri- 9 bution of the CO Later the translational energy distribution I0 wasmeasured by time-of-flight mass spectroscopy the most recent experiments are reviewed here and complete the picture by giving 11-14 (v,J) distributions for both fragments 2. CO (v,J) DISTRIBUTIONS BY VACUUM ULTRAVIOLET LASER-INDUCED FLUORESCENCE A. Experimental

FIGURE 3
FIGURE 3  Portion of laser-induced fluorescence spectrum of car- bon monoxide in the (2,0) band produced in the photo- lysis of 0.05 Torr of H2CO with the photolysis laser fixed on the rQI(3)E +rQl(4)0 transitions at 29515.2 cm and 150 ns delay(from Ref.  ii, with permission).

4 FIGURE 7
FIGURE  7 Quantum-state distribution for the H 2 (v,J) fragment from formaldehyde dissociation as determined by CARS spectroscopy (Ref.13)formaldehyde), only odd J states of H (ortho-hydrogen) were pro- 2 duced, as earlier Raman experiments 12"21 had demonstrated.Nuclear spin is conserved during the dissociation.
from Ref. Ii, with permission).