Molecular Beam Studies of Stratospheric Photochemistry Citation Moore, Teresa Anne (1998) Molecular Beam Studies of Stratospheric Photochemistry. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/nkqt-qj82. https://resolver.caltech.edu/CaltechTHESIS:07302025-155308477 Abstract Photochemistry of chlorine oxide containing species plays a major role in stratospheric ozone depletion. This thesis discusses two photodissociation studies of the key molecules ClONO 2 and ClOOCl which were previously thought to only produce Cl-atom ( ozone depleting) products at wavelengths relevant to the stratosphere. The development of a molecular beam source of ClOOCl and the photodissociation dynamics of the model system Cl 2 O are also discussed. In the first chapter, the photochemistry of ClONO 2 is examined at 308 run using the technique of photofragment translational spectroscopy. Two primary decomposition pathways, leading to Cl+ NO 3 and ClO + NO 2 , were observed, with a lower limit of 0.33 for the relative yield of ClO. The angular distributions for both channels were anisotropic, indicating that the dissociation occurs within a rotational period. Chapter two revisits the photodissociation dynamics of Cl 2 O at 248 and 308 nm, on which we had previously reported preliminary findings. At 248 nm, three distinct dissociation pathways leading to Cl + ClO products were resolved. At 308 nm, the angular distribution was slightly more isotropic that previously reported, leaving open the possibility that Cl 2 O excited at 308 run lives longer than a rotational period. Chapter three describes the development and optimization of a molecular beam source of ClOOCl. We utilized pulsed laser photolysis of Cl 2 O to generate ClO radicals, and cooled the cell to promote three body recombination to form ClOOCl. The principal components in the beam were Cl 2 , Cl 2 O, and ClOOCl. In the fourth chapter, the photodissociation dynamics of ClOOCl are investigated at 248 and 308 nm. We observed multiple dissociation pathways which produced ClO + ClO and 2Cl + O 2 products. The relative Cl:ClO product yields are 1.0:0.13 and 1.0:0.20 for ClOOCl photolysis at 248 and 308 run, respectively. The data at 308 nm was more difficult to interpret because of extensive interference from Cl 2 and Cl 2 O byproducts. The upper limit for the relative yield of the ClO + ClO channel was 0.19 at 248 nm and 0.31 at 308 nm. These results substantially confirm the current assumption but decrease somewhat the efficiency of the ClOOCl ozone-depleting catalytic cycle. At 248 run, ClOOCl photolysis exhibited novel dissociation dynamics which appeared to depend on the symmetry of the excited state. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Okumura, Mitchio Thesis Committee: Beauchamp, Jesse L. (chair) Kuppermann, Aron Lewis, Nathan Saul Okumura, Mitchio Defense Date: 6 October 1997 Record Number: CaltechTHESIS:07302025-155308477 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:07302025-155308477 DOI: 10.7907/nkqt-qj82 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 17573 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 31 Jul 2025 19:18 Last Modified: 31 Jul 2025 19:36 Thesis Files PDF - Final Version See Usage Policy. 43MB Repository Staff Only: item control page

MOLECULAR BEAM STUDIES OF STRATOSPHERIC PHOTOCHEMISTRY Thesis by Teresa Anne Moore In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1998 (Defended October 6, 1997) 11 ACKNOWLEDGMENTS Many people contributed to this work and to my graduate career at Caltech, and rm sure I have left some of them off of this short list. To those persons and to the people listed below, I extend my thanks for their help, support, and encouragement. Without them, this work would not have been possible. First I must acknowledge my husband Dustin Laurence, who stood by me during this - whole endeavor, especially when I left him for nine months (March - December 1996) to finish the experiment after the beam machine moved to Montana. He gave up the most so that the ClOOCI experiment could be completed. Also thanks to my advisor Mitchio Okumura, who showed me how to be a scientist, and to my "unofficial" advisor Tim Minton, on whose Crossed Molecular Beams Apparatus all of these experiments were done. They were the major influences on my scientific career. Thanks to Janet Minton, who welcomed me into their household in Montana for several months until I got a place of my own, as well as on several of the "just one more experiment" trips near the end of the ClOOCl experiment. Thanks to Chrissy Minton, who gave up her bedroom during these later trips, and to Michelle Minton who shared her room with Chrissy so I could have someplace to stay. Also thanks to James Seale, who came in to take data in the mornings so I could catch a few hour's nap after a long night of data taking; to Donna Garton, who helped set up and take data for the ClOOCl elastic scattering experiments; and to postdoc Thomas Schindler, who came up to Montana for several weeks to help take ClOOCl photolysis and scattering data and to let me visit home for a few weeks. These people were a great help in the last few 111 months of the ClOOCl experiment, when I was about ready to give up and go home. Alex Ide also helped in the preliminary setup for the ClOOCl experiment in Montana. The entire Okumura group deserves recognition for their support; Yi-Bin Cao, JongHo Cho~ Matt Johnson, and Chrissy Nelson were senior group members who helped me get started, also Keith Kuwata, Jim Spotts, Alex Wong, and Julio Lobo listened to countless practice talks and helped out in many ways. Postdoc Bermi (Bernd-Michael) Haas and undergrad PC (Pratap Chakrabarty) helped in the very early ClOOCl source design phase. Chrissy Nelson was the senior graduate student on the early ClON02 experiments and helped me learn the CMBA. Undergrad Jessie Haldeman helped make Cl20 and ClON02 for the early photolysis experiments at 193 and 248 nm. Our group secretary, Lena Lenore, was also helpful and encouraging. Other people at Caltech and JPL also played a part: Yuk Yung, who let me use his model to look at the effect of changing the ClON02 branching ratio; Stan Sander and his postdocs (Troy, Chip Miller, Scott Nickolaisen, possibly others) for helpful advice; Randy Friedl for discussions about Cl20 and ClON02 synthesis and ClOOCl cracking patterns, among other things; Greg McDermott at SRI who showed us a much better way to make N 20 5; and probably many others who I can't think of at the moment. During my undergraduate years, I was greatly influenced by and owe a lot to several people: Lee Spangler, my undergrad research advisor, who gave me my start in research; and Brad Mundy and Dallas Johnson, who made freshman chemistry lecture and labs sufficiently interesting that I switched my major from chemical engineering to chemistry. Mr. Frank (math) and Mr. Schladweiler (chemistry and biology) at Roundup High IV School were major influences on my decision to go to college and major in a technical field. I also owe a great deal to Mr. Helgeson, my fifth grade teacher, who took me out of the "dumb" group and put me in the "smart" group. Who knows where I would be today without that statement of faith in my ability? And of course, to my parents and grandparents, who always believed in me and supported me in all of my college and career choices. V ABSTRACT Photochemistry of chlorine oxide containing species plays a major role in stratospheric ozone depletion. This thesis discusses two photodissociation studies of the key molecules ClONO2 and ClOOCl which were previously thought to only produce Cl-atom ( ozone depleting) products at wavelengths relevant to the stratosphere. The development of a molecular beam source of ClOOCl and the photodissociation dynamics of the model system Cl2O are also discussed. In the first chapter, the photochemistry ofClONO2 is examined at 308 run using the technique of photofragment translational spectroscopy. Two primary decomposition pathways, leading to Cl+ NO3 and ClO + NO 2, were observed, with a lower limit of0.33 for the relative yield of ClO. The angular distributions for both channels were anisotropic, indicating that the dissociation occurs within a rotational period. Chapter two revisits the photodissociation dynamics of Cl2O at 248 and 308 nm, on which we had previously reported preliminary findings. At 248 nm, three distinct dissociation pathways leading to Cl + ClO products were resolved. At 308 nm, the angular distribution was slightly more isotropic that previously reported, leaving open the possibility that Cl2O excited at 308 run lives longer than a rotational period. Chapter three descnbes the development and optimization of a molecular beam source ofClOOCI. We utilized pulsed laser photolysis of Cl2O to generate ClO radicals, and cooled the cell to promote three body recombination to form ClOOCI. The principal components in the beam were Cl2, Cl2O, and ClOOCI. In the fourth chapter, the photodissociation dynamics of ClOOCl are investigated at VI 248 and 308 nm. We observed multiple dissociation pathways which produced CIO + CIO and 2Cl + 0 2 products. The relative Cl:ClO product yields are 1.0:0.13 and 1.0:0.20 for ClOOCl photolysis at 248 and 308 run, respectively. The data at 308 nm was more difficult to interpret because of extensive interference from Cl2 and Cl20 byproducts. The upper limit for the relative yield of the ClO + ClO channel was 0.19 at 248 nm and 0.31 at 308 nm. These results substantially confirm the current assumption but decrease somewhat the efficiency of the ClOOCI ozone-depleting catalytic cycle. At 248 run, ClOOCl photolysis exhibited novel dissociation dynamics which appeared to depend on the symmetry of the excited state. Vll TABLE OF CONTENTS Acknowledgements 11 Abstract V Table of Contents Vil Introduction 1 References 6 Chapter 1. Dissociation Dynamics of CIONO2 and Relative Cl and CIO Product Yields following Photoexcitation at 308 nm Abstract 8 A. Introduction 9 B. Experiment 11 C. Results and Analysis 13 D. Discussion Excited State Lifetime 19 Product Yields 22 E. References 25 F. Figures and Captions 27 Chapter 2. Photodissociation of Cl2O at 248 and 308 nm Abstract 35 A. Introduction 36 Vlll B. 248 run 37 C. 308 run 38 D. References 40 E. Figures and Captions 41 Chapter 3. CI00CI Molecular Beam Source Abstract 44 A. Introduction 45 B. Source Design Parameters 49 C. Time-of-Flight Mass Spectrometric Analysis 52 D. Optimization of Source Conditions 57 E. Elastic Scattering 60 F. References 66 G. Figures and Captions 68 Chapter 4. Photodissociation Dynamics of CI00CI and the Relative Yield of the 00 + CI0 Product Channel at 248 and 308 nm Abstract 85 A. Introduction 86 B. Experiment 88 C. Results and Analysis Analysis Overview 93 lX Relative Scaling of Primary and Secondary Distributions 95 Relative Yields 96 Overview of Results 97 ClO Contribution 97 Cl20 Contribution 98 ClO + ClO Channels 99 Concerted Cl + 0 2 + Cl Channel 102 Cl + ClOO Channels 104 Slow Signal at mlz = 35 106 Overview of Results 107 Contributions from Other Components in the Beam 107 ClOOCl Photolysis 108 248 run 308 run Relative Yields D. Discussion 110 112 ClO Photolysis Signal at 248 run 114 Transition Dipole of the Concerted Channel and Effects of Excited State Symmetry (248 run) Comparison with Previous Experiments 115 116 Stratospheric Implications 118 E. References 120 F. Figures and Captions 123 1 Introduction Significant ozone depletion has been occurring over Antarctica every spring since the late 1970's, as reported by Fannan and coworkers in 1985 .1 Since then, considerable progress has been made in trying to understand the mechanisms ofthis ozone depletion. The current models link ozone depletion to catalytic cycles of active chlorine species.2 Chlorine enters the stratosphere primarily as a component of chlorofluorocarbons (CFCs), which are stable in the troposphere but photodissociate in UV light once they get above the ozone layer. 3 These reactive chlorine species can catalytically destroy ozone through the mechanism Cl + 0 3 - ClO + 0 2 ClO + 0 - Cl + 0 2 . Most of the active chlorine species react to form more stable reservoir species such as hydrogen chloride and chlorine nitrate: 4 Cl + CH4 - HCl + CH3 ClO + NO2 - ClONO2 . These reservoir species can photodissociate or undergo other gas phase reactions in the unperturbed stratosphere. The Antarctic "ozone hole" develops in the perturbed stratosphere in the early polar spring. After the autumnal equinox, the polar regions fall into darkness and the solar UV heating ceases. Thermal radiation quickly cools the polar stratosphere to temperatures much lower than those of the mid latitudes. This cooling causes a pressure gradient which creates a polar vortex, effectively isolating the air over the pole from the rest of the stratosphere. 5 Temperatures fall low enough to form polar stratospheric clouds (PSCs) of nitric acid 2 hydrates, possibly nitric acid trihydrate, 6 and liquid aerosols consisting of ternary solutions of sulfuric acid, nitric acid, and water. The formation of these clouds dehumidifies and denitrifies the stratosphere, trapping most of the N0x as nitric acid. Chlorine nitrate can react with water or adsorbed species such as HCI on the surface of the PSCs: 7 CI0N02 (g) + HCI (s) - Cl2 (g) + HN03 (s) CI0N02 (g) + H 20 (s) - H0CI (g) + HN03 (s). In this reaction scheme, the reservoir chlorine species are ·converted to active forms of chlorine which escape back into the stratosphere while the N0x products are trapped in the PSCs as nitric acid. Therefore, the reactive chlorine species can not react with N0x to reform the reservoir species chlorine nitrate, and the chlorine is essentially all converted from inactive to active forms. The low temperature and high CI0 concentrations during the polar night allow for three body recombination to form the dimer, CI00CI. When the polar night is over, photochemistry begins to occur and the active chlorine species can engage in photochemical catalytic ozone depleting cycles. In particular, CI0OCI photolysis is a key step in the catalytic ozone-depleting cycle8 CI0 + CI0+M- CI00CI +M CI00CI +hv CI00 + Cl CI00+M Cl+ 0 2 + M 2[ Cl+ 0 3 CI0+02 ] net: which is thought to be responsible for a majority of Antarctic springtime ozone depletion. Current models assume that only CI-atom products are formed from CI00CI photolysis. 3 Photodissociation of ClOOCl to form CIO products would result in a null cycle, which would not deplete ozone. The branching ratio of CIOOCI photolysis into Cl and ClO products is critical in determining the effectiveness of this mechanism for destroying ozone. CIOOCI has several possible photolysis channels: 9 CIOOCI + hv - CIO + CIO .6.flo = 17.1 kcal/mo! - CIOO + Cl = 21.2 kcal/mo! - 2CI + O2<3~-g) = 25-_9 kcal/mo! - 2Cl + Oi( 1.6.g) = 48.4 kcal/mo! - Cl+ 0 + CIO = 80.5 kcal/mo! - Cl2 + 0 2 = - 31.2 kcal/mo!. 0 Dissociation to produce solely Cl atom products requires that the stronger Cl-O bond break preferentially over the peroxy bond. The bond-selective nature of this dissociation has been rationalized by the argument that an n - o * transition localized on the ClO chromophore is excited.8 Ab initio calculations 10 have confirmed the conjecture that the excitation involves the promotion of an electron to a ClO antibonding sigma orbital. However, a localized transition does not necessarily imply that preferential Cl-O bond fission should occur. Two experiments have found that the quantum yield for Cl atom production from ClOOCl photolysis is unity or greater, but large uncertainties remain in both measurements. Cox and Hayman 11 examined steady-state photolysis of Cl2O at 254 nm and modeled the disappearance of Cl2O and the appearance ofClOOCl in a static photolysis cell held at 203233K. They reported a yield of3.8 ± 1.6, which is unphysically high. In a more direct study, Molina et al. 12 reported a Cl atom quantum yield of 1.03 ± 0.12 from ClOOCl photolysis at 4 308 nm in a flow cell. They used resonance fluorescence to detect Cl atoms and calibrated the Cl signals relative to Cl2 photolysis. No ClO was inferred from titration with NO. Their value for the quantum yield depends on the ratio of the ClOOCl absorption cross section at 245 and 308 run, for which they used a value of 22, which is larger than more recent literature values. If the cross section ratio of 13 from the 1997 JPL evaluation of photochemical data13 were used, the corresponding Cl atom yield would be reduced to 0.6. The absorption spectrum has also been recently measured by Huder and DeMore, 14 who observed a smaller cross section in the long-wavelength tail. Using their 245/308 nm ratio of 18.7, Molina's quantum yield for Cl would become 0.88. The mechanism in which the stronger bond is preferentially broken has been rationalized by the argument that the excitation of the n-o* transition localized on the ClO bond in ClONO2 leads only to Cl + NO3 products. Early gas cell experiments of the photodissociation of ClONO2 only detected products from the Cl + NO3 dissociation channel, despite the fact that the ClO + NO 2 channel lies 14 kcals/mol lower in energy_1s,16,11,1s,19 We examined ClONO2 photolysis at 193,248, and 308 nm in a molecular beam study, with direct detection of both Cl and ClO photoproducts, and found a substantial yield of CIO products. 20,21,22 Our results for 308 nm photolysis are detailed in Chapter 1. The observation ofClO products from ClONO2 photolysis has since been corroborated in kinetics experiments by other groups. 23' 24' 25 Our ClONO2 results demonstrate how molecular beam techniques can complement the kinetics experiments by allowing direct detection of primary products and by giving 5 insights into the dissociation dynamics. The photodissociation dynamics of ClONO2 cannot be described by a simple bond-selective ClO picture even when a localized transition is excited, and the photolysis of ClOOCl is therefore likely to be equally complex. Given the uncertainties in the previous ClOOCl photolysis experiments and the importance of the ClOOCl photolysis mechanism for polar stratospheric ozone depletion, we endeavored to design a molecular beam source for ClOOCl (described in Chapter 3). Our ClOOCl source was based on photolysis ofC12O to provide ClO, so a significant Cl2O contamination was also present in the beam, requiring study of the Cl2O photolysis as well. Also, because Cl2O photolysis produces Cl + ClO in a I : I ratio, the results can be utilized to determine the relative detection efficiencies for Cl and ClO, which was necessary to detennine the relative yields for Cl and ClO producing channels from ClOOCl photolysis. Chapter 2 examines the results from the Cl2O photolysis experiments. The results of our ClOOCl photolysis experiment are presented in Chapter 4. We did indeed observe both Cl and ClO products from ClOOCl photolysis, although Cl products did dominate. We obtained an upper limit for the relative yield ofClO products of31% at 308 nm and 19% at 248nm. We also observed slow signal that could correspond to excited ClOOCl* which lives longer than a rotational period, which might be quenched under stratospheric conditions. Because of the extreme importance of the ClOOCl ozone depleting cycle, these new results should be included in models of the stratosphere. 6 References 1. J.C. Farman, B. G. Gardiner, and J. D. Shanklin, Nature, 315, 207 (1985). 2. J. G. Anderson, D. W. Toohey, W. H. Brune, Science, 251, 39 (1991). 3. M. B. McElroy and R. J. Salawitch, Science, 243, 763 (1989). 4. M. R. Schoeberl and D. L Hartmann, Science, 251, 46 (1991). 5. M.A. Tolbert and A. M. Middlebrook, J. Geophys. Res., 95, 22423 (1990). 6. M.A. Tolbert, M. J. Rossi, R. Malhotra, and D. M. Golden, Science, 238, 1258 (1987). 7. M. J. Molina, T.-L. Tso, L. T. Molina, and F. C.-Y. Wang, Science, 238, 1253 (1987). 8. L. T. Molina and M. J. Molina, J. Phys. Chem. 91, 433 (1987). 9. Heat of formation for ClOOCl from S. L. Nickolaisen, R.R. Friedl, and S. P. Sander, J. Phys. Chem. 98, 155 (1994). 10. J. F. Stanton and R. J. Bartlett, J. Chem. Phys., 98, 9335 (1993). 11 . R. A. Cox and G.D. Hayman, Nature 332, 796 (1988). 12. M. J. Molina, A. J. Colussi, L. T. Molina, R. N. Schindler, and T.-L. Tso, Chem. Phys. Lett. 173, 3 IO (I 990). 13. W. B. DeMore, S. P. Sander, D. M. Golden, RF. Hampson, M. J. Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb, and M. J. Molina, Chemical Kinetics and Photochemical Data/or Use in Stratospheric Modeling, Evaluation Number 12, JPL Publication 97-4, JPL (1997). 14. K. J. Hurler and W. B. DeMore, J. Phys. Chem. 99, 3905 (1995). 15. J. S. Chang, J. R Barker, J.E. Davenport, and D. M. Golden, Chem. Phys. Lett., 60, 385 (1979). 16. S. M. Adler-Golden and J. R. Wiesenfeld, Chem. Phys. Lett., 82, 281 (1981). 17. W. J. Marinelli and H. S. Johnston, Chem. Phys. Lett., 93, 127 (1982). 18. J. J. Margitan, J. Phys. Chem., 87, 674 (1983). 19. J.P. Burrows, G. S. Tyndall, and G. K. Moortgat, J. Phys. Chem., 92, 4340 (1988). 7 20. T. K. Minton, C. M. Nelson, T. A Moore, and M. Okumura, Science, 258, 1342 (1992). 21. C. M. Nelson, T. A. Moore, M. Okumura, and T. K. Minton, Chem. Phys., 207, 287 (1996). 22. T. A Moore, M. Okumura, M. Tagawa, and T. K. Minton, Faraday Discuss. Chem. Soc., 100,295 (1995). 23. AR Ravishankara, Faraday Disc. Chem. Soc., 100, 333 (1995); J.B. Burkholder, Abs. Pap. Chem. Soc., 210, 51 (1995). 24. G. S. Tyndall, C. S. Kegley-Owen, J. J. Orlando, and J. G. Calvert, J. Chem. Soc., Faraday Trans. 93, 2675 (1997). 25. L. Goldfarb, A-M. Schmoltner, M. K. Gilles, J.B. Burkholder, and A. R Ravishank:ara, J. Phys. Chem. A, 101, 6658 (1997); R J. Yokelson, J.B. Burkholder, R W. Fox, and AR. Ravishankara, J. Phys. Chem. A, 101, 6667 (1997). 8 Chapter 1.* Dissociation Dynamics of CIONO2 and Relative Cl and CIO Product Yields foil owing Photoexcitation at 308 nm Chlorine nitrate photolysis at 308 run was investigated with a molecular beam technique. Two primary decomposition pathways, leading to Cl + NO3 and ClO + NO 2, were observed. The branching ratio between these two respective channels was determined to be 0.67±0.06:0.33±0.06. This ratio is an upper limit, because some of the ClO photoproducts may have undergone secondary photodissociation. The angular distributions of the photoproducts with respect to the direction of polarization of the exciting light were anisotropic. The anisotropy parameters were ~ = 0.5±0.2 for the Cl+ NO3 channel and ~ = 1.1±0.2 for the ClO + NO 2 channel, indicating that dissociation of ClONO2 by either pathway occurs within a rotational period. Weak signal at mass-to-charge ratios of 35 and 51, arising from products with laboratory velocities close to the beam velocity, was observed. While this signal could result from statistical dissociation channels with a total relative yield of 0.07 or less, it is more likely attributable to products from ClO secondary photodissociation or from dissociation of clusters. * This chapter has been published. The reference is: Teresa A. Moore, Mitchio Okumura, Masahito Tagawa, and Timothy K. Minton, Faraday Discuss. Chem. Soc. 100,295 (1995). 9 Introduction Chlorine nitrate (ClONO 2) plays a key role in stratospheric chemistry as a reservoir for reactive chlorine and nitrogen oxides. 1•2 Photolysis is a major decomposition process for ClONO2, 3 and has been the subject of many studies. 4•5•6•7•8•9•10•11 •12 At wavelengths less than 250 nm, photodecomposition follows two primary channels:9 ClONO2 + hv ... ClO + NO 2 6H = 26 kcal/mol (1) ... Cl + NO 3 6H = 40 kcal/mol (2) Channel (2) can lead directly to loss of two ozone molecules via the following reactions:13 NO3 + hv ... NO + 0 2 (3a) NO + 0 3 ... NO 2 + 0 2 (3b) Cl + 0 3 ... ClO + 0 2. (3c) Toumi et al. have recently shown in model studies that this catalytic cycle13 including steps (2) and (3a-c) becomes the dominant ozone loss mechanism in the latter stages of springtime Arctic ozone depletion. 14 In this model, the extent of ozone loss depends upon the quantum yield of channel (2), especially at A>300 nm. Toumi et al. used the value of 0.9 for this quantum yield, which was recommended by the 1992 JPL evaluation of kinetic data. 15 However, results from our laboratory at shorter wavelengths (193 and 248 nm) showed the Cl + NO3 product yield to be significantly lower, 9 a conclusion later corroborated by other groups.10. 11 ,12 A lower yield at longer wavelengths would diminish the predicted contribution of the ClONO2 catalytic cycle to Arctic ozone depletion and reduce the predicted magnitude and duration of the ozone loss episode. Sander and co-workers have recently obtained evidence from flow tube studies that 10 excited ClONO2 formed by wavelengths longer than 300 run lives long enough to be quenched under stratospheric conditions.12 Quenching of ClONO2 photolysis would profoundly affect the partitioning of the Cl and NOx budgets and would largely eliminate the ClONO2 catalytic cycle as a candidate loss mechanism for stratospheric ozone. In their experiment, Sander and co-workers observed products from channels (I) and (2) when ClONO2 was excited by a broadband Xe lamp; however, at wavelengths longer than 300 run they found that the total quantum yield was significantly quenched at N2 buffer gas pressures above 50 Torr. Under their conditions (approximately 50 Torr and 200 K), the mean time between gas-kinetic collisions is 3 ns, so an excited molecule must have a lifetime on the order of nanoseconds or longer to be quenched rather than dissociated. Previous flow tube studies revealed no direct evidence of a relatively long-lived state following UV excitation ofCIONO2 . 4•8•10•11 However, recent ab initio calculations by Grana et al. do predict the existence of a metastable state. 16 These authors performed high level configuration interaction (Cl) and coupled-cluster calculations to predict the vertical excitation energies and oscillator strengths for transitions to eight singlet and eight triplet states of ClONO2 between 400 run and 160 run. They also found a bound, metastable triplet state, with a nonplanar geometry, lying 62 kcal/mol above the ground state equilibrium configuration. If CIONO2 were formed in this state either by direct excitation or by relaxation from an initially prepared state, the excited ClONO2 could conceivably persist for nanoseconds or longer. The fate of chlorine nitrate excited at wavelengths greater than 300 nm has important implications for stratospheric ozone depletion. To address questions concerning the 11 dissociation pathways and the possibility of a long-lived excited state, we extended our previous photodissociation studies of ClONO2 to a longer wavelength. With the use of molecular beam methods, we can directly detect photoproducts, measure their relative yields, and determine angular and translational energy distributions in the absence of collisions. Furthermore, we can infer the timescale for dissociation from the anisotropy of the product angular distribution. In our earlier experiments, we observed comparable yields of Cl and ClO products from ClONO2 photolysis at both 193 and 248 nm.9 We calibrated the relative detector sensitivity for Cl and ClO by performing an experiment on Cl2O photolysis at 308 nm and found Cl:ClO branching ratios of0.57:0.43 and 0.67:0.33 for ClONO2 photolysis at 248 and 193 nm, respectively. Furthermore, we observed anisotropic angular distributions for the two channels and inferred that the dissociation is prompt, occurring within a rotational period (1:ro1~ 9 ps at the rotational temperature of the beam). Here we present results from a similar investigation of ClONO2 photolysis at 308 nm. Experiment The molecular beams apparatus and methods have been described in detail elsewhere. 9' 17 Briefly, a molecular beam was crossed with a laser beam in a low pressure region at the center ofrotation of a rotatable mass spectrometer detector18 (see fig. 1). Collisionless conditions allowed for detection of primary dissociation products recoiling away from the beam direction. Center-of-mass (c.m.) product translational energy and angular distributions were derived from the arrival time distributions of the neutral photoproducts over a calibrated flight length as a function of detector angle away from the beam direction. 12 Continuous molecular beams were formed by passing H 2 over ClONO2 at -50°C or He over Cl2O at - 74 ° C and expanding the mixture through a O.1 mm diameter quartz nozzle maintained at +5° C. ClONO2 was seeded in hydrogen rather than helium in order to obtain a faster beam, which enhanced the signal through kinematic focusing. Stagnation pressures of the gas mixtures were 300 Torr for the ClONO2 experiment and 200 Torr for the Cl2 O experiment. The ClONO2 data was obtained in two sets, with corresponding beam velocities 1350 mis and 1580 mis, as determined by beam time-of-flight (TOF) distributions taken with the aid of a chopper wheel. The Cl2O beam velocity ( 1480 mis) was determined by examining the arrival time of the laser-induced depletion of signal at a mass-to-charge ratio (mle) of 37 with the detector directly viewing the molecular beam. 19•20 Cl2O was synthesized by oxidation of Cl2 on HgO and was purified by pumping on the sample at -l l8°C (ethanol slush bath). 21 ClONO2 was synthesized from the reaction ofC12 O with N 2 O 5 and purified first by cannulation and then by pumping on the sample at -95° C (methanol slush bath). 22 Purity was checked by a survey of beam TOF distributions at all masses where signal was observed. The only impurities in either molecular beam could be attributed to dimers and/or clusters. A small peak in the beam TOF at mle = 143 (ClONO2 • NO/) indicated that clusters were present in the ClONO2 beam; however, the product signals observed were insensitive to the relative magnitudes of the cluster signals. The molecular beam was crossed in the interaction region with a laser beam at 308 nm from a Lambda Physik EMG 103 excimer laser operating at 200 Hz. The laser beam was randomly polarized in a plane perpendicular to the scattering plane and parallel to the molecular beam axis. The laser beam was focused to a spot size of 2. 5 mm x 2. 5 mm, where 13 it intersected the molecular beam. The laser pulse energy used for the ClON02 experiment was 170 mJ. The consequent laser fluence was high enough that secondary photodissociation of the primary ClO products was a concern; however, the low absorption cross section15 of ClON02 at 308 nm ( a = 1.6x 10-20 cm2) required the use ofrelatively high laser fluences to make the primary photoproduct signals observable. For the Cl20 experiment, a pulse energy of 126 mJ was used for data collection at mle = 51, and the laser was attenuated with a stack of quartz plates to 36 mJ/pulse for data collection at mle = 35 in order to minimize secondary photodissociation of the ClO product. TOF distributions were collected at mle = 35(CJ+), 51(Cl0+), 30(NO+), 46(NO/), 16(0+), and 81(ClONO+), and angular distributions of product signals detected at mle = 35, 51 , and 46 were measured. Results and Analysis Analysis of TOF and angular distributions was carried out by a forward convolution technique.20•23•24•25 For each channel, trial c.m. translational energy P(ET) and angular w(0) distributions were transformed to the laboratory frame of reference to obtain a predicted number density TOF distribution N(t) that could be compared with the experimental data. The P(ET) and w(0) distributions were then adjusted iteratively until the predicted TOF distributions agreed with the observed distributions at all angles. The data were similar to our earlier results at 193 and 248 nm. 9 A single peak was prominent at mle = 51 (see figs. 2 and 3), which we assigned to ClO from the ClO + N0 2 channel by analogy with our previous results. The mle = 3 5 TOF distribution contained two main components, the slower of which matched the peak at mle = 51, indicating that it must 14 arise from fragmentation of ClO to Ci+ in the ionizer. 26 We assigned the fast component at mle = 35 to Cl from the Cl+ N03 channel. We used the TOF spectra taken at mle = 35 and 51 to derive the c.m. translational energy distributions for the two dissociation channels (fig. 4). We verified our assignments of the two dissociation channels by using the P(Er) and w(6) distributions derived from the Clo+ and Cl+ data to predict the TOF distributions for the respective counterfragments, N02 and N03 (fig. 5). No signal was -observed at mle = 62, the N03 parent mass, but the TOF spectrum collected at mle = 46 (N02+) was fitted well with the distribution predicted for the N03 fragment by the P(Er) and w(0) distributions for the Cl + N03 channel. Apparently, the NO/ signal comes not from N02, but from N03 which fragments to N02+in the ionizer. The bimodal TOF distribution at mle = 30 (NO+) could not be fit from the P(Er) and w(0) distributions for either the Cl + N03 or ClO + N02 channels alone, but it could be fit with a sum of contributions from the predicted TOF distributions for both N03 and N02 fragments. Hence, both N03 and N0 2 photofragments lead to NO+ in the ionizer (similar ionizer fragmentation patterns were observed in our previous study.)9 The ability to fit both counterfragments of each dissociation channel with the use of the same P(ET) and w(0) distributions thus confirms our original assignments of the two primary photodissociation channels. Laboratory angular distributions of Cl 0, Cl, and N03 were obtained from the integrated TOF data detected at mle = 51, 35, and 46, respectively. At mle = 51 and 46, where the TOF distributions corresponded only to a single photoproduct, the signal for each detector angle was simply integrated. The Cl angular distribution was extracted from the mle = 35 TOF 15 distributions by integrating the calculated TOF distributions for the Cl+signal arising from the Cl+ N03 channel (obtained from the optimized P(Er) and w(0) distributions). The c.m. angular distribution must have the form w(0) ex 1 + PPi(cos0), where P 2 denotes the second-order Legendre polynomial and e the angle between the electric field vector of the excitation photon and the product recoil direction in the c.m. reference frame.23•27 Pis the anisotropy parameter and must lie in the range -1 ;5; P;5;;2_ We obtained p = 1.1±0.2 for the ClO + N02 channel from the ClO angular distribution and p = 0.5±0.2 for the Cl + N03 channel from both the Cl and N03 angular distributions. The fits used to determine the anisotropy parameters are shown in fig. 6. The broad, slow signal at mle = 35 may be explained by a number of possible mechanisms: spontaneous secondary dissociation, dissociation of clusters formed in the beam expansion, secondary photodissociation of primary ClO photoproducts, or a third primary channel leading to products with low c.m. translational energies. We can rule out spontaneous secondary dissociation because the P(Er) distribution (fig. 4B) implies that no ClO photoproducts have enough internal energy to dissociate. Clusters were present in the beam, but we expect the contribution from dissociation of clusters to the mle = 35 TOF to be small because changes in the stagnation pressure did not significantly alter the magnitude of the slow signal relative to the fast peaks. The most likely explanation for the tail at mle = 35 is secondary photodissociation of ClO. The ClO absorption cross section is larger than that of ClON02 at 308 nm, and at the high laser fluences used, some secondary photodissociation of ClO is expected. However, we could not model the secondary photodissociation, because the signal-to-noise ratio was too low. We therefore could not 16 eliminate the possibility of a third primary dissociation channel forming products with low translational energies. Direct dissociation on a triplet surface to ClONO + 0(3P) could conceivably contribute to a slow signal at mle = 51 and 35 through fragmentation to Clo+ or er in the ionizer. ClONO might also fragment to NO 2+ or NO+ and give rise to a slow signal at mle = 46 and 30. The energy available for translation after breaking the O-N bond is only about 26 kcal/mol, and the ClONO fragment will recoil slowly from the relatively light O-atom counterfragment. Therefore, the minimum flight time possible at a detector angle of 30° is 133 µs (not including ion flight time), and any signal present should manifest itself as a peak in the slow tail of the product TOF distributions. The only TOF distributions which showed an obvious hint of slow signal were the er distributions. If all slow signal at mle = 35 not assigned to the dominant channels did come from ClONO, then the ClONO + 0 yield would be - 0.07. Even this small yield is suspect because 1) no photofragment signal was detected at the CIONO parent mass (mle = 81 ), 2) no obvious slow signal was seen at other masses, and 3) ClONO can only be formed by a spin-forbidden process at an excitation wavelength of308 nm. Statistical dissociation of ClONO2 could also lead to products with low translational energies. The c.m. translational energy distributions for products of a statistical dissociation process have a maximum near zero and decay rapidly with Er~ therefore, the observed signal would have low c.m. velocities. At the smallest detector angle used here, 15°, we can detect Cl products from the Cl+ NO3 channel with total c.m. translational energies Er~ 0.7 kcal/mol and ClO products from the ClO + NO 2 channel with Er c!: 1.4 kcal/mol. The experiment 17 should therefore be sensitive to statistical dissociation channels with available energies of a few kcal/mol above threshold. We can estimate the maximum possible yield from statistical dissociation by fitting the slow component to a statistical model, but the modeling will depend upon the final dissociation channel. Several dissociation channels are energetically available to ClONO2 with 92.8 kcal/mol of internal energy (i.e., the photon energy of308 nm light). While statistical dissociation of ClONO2 on the ground electronic state potential energy surface will lead predominantly to ClO and NO 2 in their ground electronic states, an excited state of ClONO2 may correlate to a higher energy asymptote and dissociate to form excited state products. The statistical decay rate will be slower and the translational energy release smaller for channels with lower available energy. We therefore considered dissociation via two excited state channels which consume most of the 308 nm photon energy: ClONO2 - Cl (2P) + NOl4 2E) t::,,.H = 83 kcal/mol (4) - ClO (a 4 :E-) + NOi(X 2A 1) t::,,.H = 77 kcal/mol (5) We proposed the existence of a bound ClO quartet state in a previous study of photoproducts from Cl2O dissociation, 17 and this hypothesis was confirmed theoretically by Langhoff, who computed the energy and geometry of this state.28 In order to estimate the maximum possible yield of these channels, we first performed RRKM calculations to obtain a c.m. translational energy distribution for each channel. 29 We then compared the slow component of the TOF distribution at the relevant mle to the TOF distribution predicted by the RRKM P(ET) distribution. We obtained upper limits on the yields by assuming that the signal in the slow components arose solely from statistical dissociation (see fig. 2 and 3). We estimated a yield 18 of0.05 for dissociation to ground state ClO and N0 2 products, based on the signal at mle = 51. We then determined the ClO contribution to the slow component at mle = 3 5 by using the same ClO ionizer fragmentation pattern observed for the main peaks. We assumed that the remainder of the slow signal at mle = 35 corresponded to Cl products from statistical decay to ground state Cl and N03 products and obtained a maximum yield of 0.02 for this channel. Thus, the maximum total relative yield for statistical dissociation to ground state products via either channel is 0.07. For the excited state products, we found a maximum possible total yield of0.03 based on the data collected at 15° (fig. 3). However, the RRKM P(Er) distributions for the excited state product channels predict negligible contributions to the slow signal at larger detector angles, because the P(Er) distributions fall off quickly with ET. Our inability to achieve a consistent fit of all the data under the assumption of excited state products makes it highly unlikely that channels (4) or (5) play any role in ClON02 photodissociation at 308 nm. In any case, these yields are upper limits because it is improbable that all of the slow signal arises from statistical dissociation. Secondary photodissociation, observed unambiguously in our previous experiments at shorter wavelengths, may even account for all of this broad underlying signal at mle = 35. Regardless of the actual product channel assumed, we conclude that any statistical dissociation process is at most a minor channel. We performed a photodissociation experiment on Cl20 at 308 nm in order to calibrate the relative detection efficiency for Cl and ClO and thereby determine the branching ratio between the two channels in the photodissociation of ClON02. The Cl20 data (not shown) were analogous to previous observations and were fit with the same P(E-r) and w(0) 19 distributions as before. 17 The relative yields for the two photodissociation channels of ClONO2 following excitation at 308 run were detennined to be 0.67±0.06 for the Cl+ NO3 channel and 0.33±0.06 for the ClO + NO 2 channel. Our results may underestimate the ClO yield because the primary ClO product has a significant probability of undergoing secondary photodissociation during the laser pulse~ therefore, the Cl:ClO branching ratio of0.67:0.33 is an upper limit. Discussion Excited-State Lifetime The anisotropy of the angular distributions allows us to determine whether the main products associated with the dominant peaks were formed by prompt dissociation, i.e., occurring in less than a rotational period, ,: rot· If the excited molecule has time to rotate before dissociating, then the molecule will change orientation, causing the angular distribution to become more isotropic. Rotational averaging thus reduces the absolute magnitude of P, though not necessarily exactly to zero. We can estimate upper limits on the minimum values of p for the two channels with the use of expressions that Yang and Bersohn have derived for calculating Pwhen the dissociating molecule is long-lived. 30 We assume 1) the transition dipole moments are parallel to the axes of the respective dissociating bonds, 2) the geometry of ClONO2 does not change on going to the dissociative state, and 3) ClONO2 can be approximated as a symmetric top (K = -.89). We find that the upper limits on the anisotropy parameters for a long-lived state ofClONO2 are P= 0.07 for dissociation to CIO + NO2 and P= 0.13 for dissociation to Cl + NO3. Effects such as vibrational averaging and breakdown 20 of the symmetric top assumption will reduce P further. The calculated values lie below the uncertainties in the anisotropy parameters for the two main channels, P = 1.1±0.2 for ClO + NO2 and P= 0.5±0.2 for Cl+ NO3. The observed anisotropies thus indicate that dissociation via either channel occurs within 'trot· Assuming a typical rotational temperature of ~50 Kin the molecular beam, 'tro, would be approximately 9 ps. We conclude that the main products observed in this experiment are formed in less than ~9 ps, which is shorter than the mean collision time (~5 ns) under stratospheric conditions, and thus cannot arise from a long-lived excited state. We could not rule out assignment of the slow signal at mle = 3 5 and 51 to a third dissociation channel, such as statistical decomposition, with low c.m. translational energies; however, the contribution from such a process will be small. Dissociation via a statistical process should be dominated by fission of the weakest bond, making ground state ClO and NO 2 the most likely major products if ClONO2 dissociates on its ground electronic state surface. We derived an upper limit of 0.05 for the yield ofthis channel from the mle = 51 data (figs. 1 and 2). However, the RRKM lifetime for ground state ClONO2 with 92.8 kcal/mol of internal energy will be on the subpicosecond timescale. To account for a nanosecond lifetime, we would have to invoke a rate limiting step, such as internal conversion or intersystem crossing from an initially prepared state. Statistical dissociation of metastable ClONO2 could occur on a nanosecond timescale only if the available energy is small (on the order of 10 kcal/moI or less for CIONOi), e.g. dissociation to a channel near threshold leading to excited state products. Dissociation to channels (4) and (5) may have lifetimes as long as nanoseconds, but we did not have enough information to calculate accurate decay rates for 21 processes so close to threshold. The bound triplet state predicted by Grana et al. is a candidate for this metastable state, but there is no information on its possible decomposition pathways. Although the translational energy release would be even lower than for dissociation to ground state products, our experiment still places an upper limit of 0.03 on the total relative yield for the plausible product channels (4) and (5) near 93 kcal/mol (see fig. 2). Thus, the slow product signal could account for only a small fraction of the total observed products; furthermore, all of these estimates are upper bounds, because secondary photodissociation ofClO will likely contribute substantially to the observed slow signal at mle =35. Given our beam velocity, laser spot size, and the solid angle viewed by the detector, we can only detect fragments from excited ClONO2 molecules that dissociate within 2.5 µs. Our sensitivity allows us to detect molecules with dissociation lifetimes 1: ~ 10 µs; metastable states which live longer would go unobserved in our experiment. Such lifetimes are significantly longer than typical mean collision times of 1-10 ns in flow tube experiments, where unit quantum yields are observed for ClONO2 photolysis at buffer gas pressures of 510 Torr.7•10•11 If dissociation occurs through a state with a lifetime of 10 µs or longer, then the metastable ClONO2 must undergo > I 03 gas kinetic collisions without being quenched at these pressures. Thus, the molecular beam and flow tube studies are together consistent with a long-lived metastable only if that state has a lifetime 1: > IO µsand is very inefficiently quenched. Our results would also be consistent with the experiment of Sander and co-workers if a metastable state of ClONO2 is prepared by excitation at wavelengths longer than 308 nm. 22 Sander and co-workers photolyzed chlorine nitrate with a broad band, filtered Xe lamp whose output extended from ca. 300 run to visible wavelengths. 12 Such long wavelength radiation would be absorbed in the 370-380 run region, where Molina and Molina3 observed a weak absorption band that was distinct from absorption bands deeper in the ultraviolet. This band lies close to 362 nm, the wavelength that Grana et al. predict for the lowest vertical triplet excitation. 16 This state is predicted to be directly dissociative, but may lead by curve crossing to the bound metastable state. Additional experiments at longer Wavelengths are necessary to examine the possibility that this band involves a transition leading to formation of a longlived state; however, the small absorption cross-section, nine times lower at 362 nm than at 308 nm, will make these measurements difficult. Product Yields We found channels (1) and (2) to dominate ClONO2 photochemistry at 308 run; the total yield of all other possible channels must be ~0.07. The Cl/ClO product branching ratio was determined to be 0.67/0.33. The relative yield of 0.33 for ClO is a lower limit, because some of the ClO products may have undergone secondary photodissociation. This branching ratio does not agree with a number of earlier studies done at A > 300 run by Smith et al., 4 Chang et al., 13 Knauth and Schindler,6 and Margitan. 7 The last three experiments all found near unit quantum yields of Cl atoms and failed to detect ClO products, even for photolysis at wavelengths less than 300 run. The JPL evaluation, 15 which recommends quantum yields of 0.9 ± 0.1 for the Cl+ NO 3 channel and 0.0 for the ClO + NO 2 channel, was based primarily on the results of the most direct study, that of Margitan. However, our results are more 23 consistent with recent unpublished flow tube studies10•11 •12 which detect both channels with similar relative yields. The decreased yield for dissociation of the weakest bond, Cl0-NO2 , in the present study relative to our earlier results suggests that photons at 308 nm excite transitions favoring fission of the Cl-0 bond. Grana et al. calculate that the lowest singlet state 11A" possesses ClO n-o* character and has a vertical excitation wavelength of 284 nm with an oscillator strength approximately 1% of the band strength at 190 nm. The next highest state, the 21A" state, is predicted to be a 1t-1t * transition localized on the NO 2 group at 252 nm but with zero oscillator strength. While excitation to the 11A state could explain the higher Cl yield II at 308 nm relative to 248 run, it is not consistent with the observed angular distributions II which are indicative of a parallel transition. To excite a state of A symmetry, the transition moment must be out-of-plane and perpendicular to either dissociation coordinate in the ground state geometry. A perpendicular transition would result in a negative value of P unless the dissociating bond rotates significantly out of plane prior to dissociation. There are many states of both symmetries at slightly higher energies, and excitation at 308 nm could instead readily excite the Franck-Condon wing of a dissociative A' state. It is possible that there are simultaneous parallel and perpendicular transitions to multiple excited states or a transition to a mixed state with the dipole transition moment aligned midway between the two bond axes, making the net P small and positive. The complexity of the ClONO2 excited state manifold suggests that photodissociation of ClONO2 cannot be described by a simple bondselective picture but rather that curve crossings and the dynamics of nuclear motion following excitation may largely determine the final product branching ratio. 24 The dominance of direct dissociation processes at excitation wavelengths as long as 308 run implies that it is important to consider a catalytic cycle involving ClONO2 photolysis in models of stratospheric ozone depletion. However, our finding of a significant yield for the ClO + NO 2 channel will lead to a decrease in the ozone-depleting efficiency of the ClONO2 cycle from what has been predicted. At mid-latitudes, where this cycle contributes only 3-5% of the ozone loss, 2 the new product yield ratio will not affect the predicted ozone depletion. The impact will be greater on model predictions of ozone loss in the Arctic polar spring as well as in the Antarctic collar region, where Toumi et al. have shown that this cycle plays a major role.14 Acknowledgments The experiments were performed at the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration (NASA), and were supported by the NASA Upper Atmosphere Research Program under contract NAGW-3893 and by E. I. duPont de Nemours and Co., Inc. We acknowledge support for T.A.M. from a NASA Graduate Research Fellowship and for M.T. from the Murata Overseas Scholarship Foundation. 25 References 1. F. S. Rowland, J. E. Spencer, and M. J. Molina, J. Phys. Chem., 1976, 80, 2711; ibid., 2713. 2. P. 0. Wennberg, R. C. Cohen, R. M. Stimpfle, J. P. Koplow, J. G. Anderson, R. J. Salawitch, D. W. Fahey, E. L. Woodbridge, E. R. Keim, R. S. Gao, C.R. Webster, R. D. May, D. W. Toohey, L. M. Avallone, M. H. Proffitt, M. Loewenstein, J. R. Podolske, K. R. Chan, and S. C. Wofsy, Science, 1994, 266, 398. 3. M. Molina and L. T. Molina, J. Photochem., 1979, 11, 139. 4. W. S. Smith, C. C. Chou, and F. S. Rowland, Geophys. Res.-Lett., 1977, 4,517. 5. S. M. Adler-Golden and J. R. Wiesenfeld, Chem. Phys. Lett., 1981, 82, 281. 6. H. D. Knauth and R. N. Schindler, Z. Naturfor., 1983, 38A, 893. 7. J. J. Margitan, J. Phys. Chem., 1983, 87, 674. 8. J. P. Burrows, G. S. Tyndall, and G. K. Moortgat, J. Phys. Chem., 1988, 92, 4340. These authors proposed a hypothetical metastable state of ClONO2 excited at 254 run, because the Cl + NO3 products could not account for all the photolysis yield. It is likely that the missing product is in the CIO + NO 2 channel, which they did not detect. 9. T. K. Minton, C. M. Nelson, T. A Moore, and M. Okumura, Science, 1992, 258, 1342. 10. A R. Ravishankara, private communication. 11. J. J. Orlando, C. S. Kegley-Owen, G. S. Tyndall, and J. G. Calvert, XX/st Informal Conference on Photochemistry, Toronto, Ontario, Canada, May 1994. 12. S. L. Nickolaisen, S. P. Sander, R. R. Friedl, L. Jaegle, M. A. Allen, Y. L. Yung, C. R. Webster, R D. May, G. C. Toon, B. Sen, and J.-F. Blavier, XX/st Informal Conference on Photochemistry, Toronto, Ontario, Canada, May 1994. 13. J. S. Chang, J. R Barker, J.E. Davenport, and D. M. Golden, Chem. Phys. Lett., 1979, 60, 385. 14. R. Toumi, R. L. Jones, and J. A Pyle, Nature, 1993, 365, 37. 15. W. B. DeMore, S. P. Sander, D. M. Golden, R. F. Hampson, M. J. Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb, and M. J. Molina, Chemical Kinetics and Photochemical Data/or Use in Stratospheric Modeling, JPL Publication 92-20, 1992. 26 16. A. Grana, T. J. Lee, and M. Head-Gordon, J. Phys. Chem., 1995, 99, 3493. 17. C. M. Nelson, T. A. Moore, M. Okumura, and T. K. Minton,J. Chem. Phys., 1994, 100, 8055. 18. Y. T. Lee, J. D. McDonald, P.R. LeBreton, and D.R. Herschbach, Rev. Sci. Jnstrum., 1969, 40, 1402. 19. T. K. Minton, P. Felder, R. J. Brudzynski, and Y. T. Lee, J. Chem. Phys., 1984, 81, 1759. 20. P. Felder, Habi/itationsschrift, University of Zurich, Switzerland, 1993. 21. G. H. Cady, Inorg. Synth., 1957, 5, 156. 22. M. Schmeisser, Jnorg. Synth., 1967, 9, 127. 23. G. E. Busch and K. R. Wilson, J. Chem. Phys., 1972, 56, 3626; ibid, 3638. 24. T. K. Minton, G. M. Nathanson, and Y. T. Lee, J. Chem. Phys., 1987, 86, 1991. 25. A. M. Wodtke and Y. T. Lee, J. Phys. Chem., 1985, 89, 4744. 26. The 4 µs difference in peak positions of the ClO signals detected at mle = 51 and mle = 3 5 reflects the mass dependence on the ion flight time from the ionizer of the mass spectrometer to the Daly-type ion counter. The ion flight time was determined to be t = 3.5 (mle) 112 µs. 27. R. N. Zare, Mo/. Photochem., 1972, 4, 1. 28. S. R. Langhoff, unpublished results. 29. P. J. Robinson and J. A Holbrook, Unimolecu/ar Reactions, Wiley-Interscience, 1972. 30. S.-C. Yang and R. Bersohn, J. Chem. Phys., 1974, 61, 4400. 27 Figure Captions Fig. 1. Schematic diagram of the molecular beam photodissociation experiment. The numbers correspond to pressures in Torr for the various regions. 9 is the detector angle. Fig. 2. TOF distnbutionsNa,J collected at a detector angle of30° for (A) Ci+ and (B) CIO+ fragments. The molecular beam velocity was 1580 mis. Fits to the dominant product - channels, Cl + N03 and ClO + N02, are calculated with optimized center-of-mass translational energy P(Er) and angular w(6) distributions (see figs. 4 and 6). The long-dashed line at slow times in (B) is a hypothetical fit based on the assumption that ground electronic state ClO and N02 products are formed by a statistical (RRKM) process. The long-dashed line at slow times in (A) is also a hypothetical fit based on statistical (RRKM) decay that leads to ground electronic state products. This "fit" includes contributions to a Ci+ signal from both ClO + N02 and Cl + N03 channels. The upper limit on the total yield from both channels is determined from the fits in (A) to be 0.07. Fig. 3. TOF distributions collected at a detector angle of 15° for (A) Ci+ and (B) Clo+ fragments. The molecular beam velocity was 1350 mis. Fits to the main peaks are calculated with the same P<Er) and w(6) distributions used to achieve the analogous fits in fig. 2. Two kinds of hypothetical fits are shown at longer flight times. One is based on the assumption of statistical decay to ground state products, and the other is based on the assumption of statistical decay to excited state products. As in fig. 2, the hypothetical fits at mle = 3 5 include dissociation along both ClO + N02 and Cl + N03 product channels. Statistical decay 28 to excited state products via channels (4) and (5) leads to low product translational energies and thus a smaller possible contribution to the signal than dissociation to ground state products. The upper limit on the total yield of ground state products is 0.07, while the upper limit on the total yield of excited state products at this detector angle is O.03 . The total yield of all slow products cannot exceed 0.07. At detector angles of30° and larger, the excited state products cannot be detected because their center-of-mass velocities are too low. Fig. 4. Optimized center-of-mass translational energy P(Er) distributions for the two dominant photodissociation channels. For the Cl+ N03 channei I.ET) = 32.9 kcal/mol, which is 62% of EMI. For the ClO + N02channel, (ET)= 26.0 kcal/mol, which is 39% of EMI. Fig. 5. TOF distributions collected at 30° for (A) N0 2+and (B) NO+ fragments. The peak in (A) is fit with the Pffrr) and w(0) distributions for the Cl + N03 channel under the assumption that all the signal at mle = 46 arises from N03 which fragments to N0 2+ in the ionizer. The bimodal distribution in (B) apparently contains signal from both N03 and N02 products which fragment to NO+ in the ionizer; this distribution is fit with two components corresponding to both Cl + N03 and ClO + N0 2 channels. Fig. 6. Semi-logarithm plot of the laboratory angular distributions for the two photodissociation channels observed. The circles are the experimental data, and the lines are the calculated distributions using values of the anisotropy parameter ~ shown. The solid lines are the best fit to the data; the dashed lines are calculated using upper and lower limits for the parameter ~. 29 TO ION COUNTING SYSTEM ~ / QUADRUPOLE MASS FILTER MOLECULAR BEAM Fig. l 30 1 A 30° ------·····- CIO - - · · - · - Cl + NO 2 + NO 3 RRKM (gnd. st. prods.) ·-.!J§ o~~-~~~~llt■e~~~~~~~ .. • .Q ca ~ .., .._.. 1 B mle = 51 (CIO+) z - CIO+NO 2 - RRKM (gnd. st. prods.) 0 0 cP (0 o 0 o 0 ~~~~:::....:..;;;._~~~~'--'e:IJ~~~-"-ff\ 000 100 Figure 2. 0 0 0 200 Flight time, µs 0 300 31 15° mle = 35 {Cl+) 1 - - · Cl+ NO3 - - - - - - CIO + NO2 - - RRKM (gnd. st. prods.) - -- RRKM (ex. st. prods.) 0 B 0 m/e = 51 (CIO+) CIO + NO2 - - RRKM (gnd. st prods.) - - - RRKM (ex. st. prods.) 0 0 0 ~~~"-~~~~~~~.,.,._._,..,,.....,.,... 0 00 100 0 200 Flight time, µs Fig. 3 0 300 32 A ·-~c::, .c ... ca • B CIO + N02 0 10 20 30 40 Energy, kcal/mo I Fig. 4 50 33 1 A 0 30° Oo 0 l! 0 ·= .a ...ca C • .... 1 "' ..-... .._.. mle = 30 (NO+) B z ... .... . CIO + NO2 - - - Cl+ NO3 I •.. I t I •.. I •.. 0 0 I I 0 100 Figure 5. 200 Flight time, µs 300 34 0.8 A 0.4 0.2 p = 0.3 p = 0.5 p = 0.7 ..-.. a:> ..._., .. 0.8 CIO + NO2 3 0.6 ·-l!C::::s • .c ca ~ 0.4 __ pp== 0.2 ,_,,__ 20 Figure 6. 0.9 1.1 p = 1.3 40 Detector Angle 0 60 35 Chapter 2.* Photodissociation of Cl2O at 248 and 308 nm Molecular beam studies of Cl2O photolysis at 248 and 308 nm have been repeated and the analysis refined. At 248 nm, three distinct dissociation pathways that led to Cl + ClO products were resolved. At 308 nm, the angular distribution was slightly more isotropic than previously reported, leaving open the possibility that Cl2O excited at 308 nm lives longer than a rotational period. * This chapter has been previously published. The reference is: Teresa A. Moore, Mitchio Okumura, and Timothy K. Minton, J. Chem. Phys. 107, 3337 (1997). 36 Introduction Studies of the photodissociation dynamics of Cl2O ( dichlorine monoxide) can aid in understanding the photochemistry of the more complex chlorine oxides which play a major role in stratospheric ozone depletion. As we reported in a previous publication, 1 formation of Cl+ ClO products (!:Jl0° = 32.9 kcal/mol) is the only channel observed for Cl2O photolysis at 308 nm and is the major channel at 248 nm. We have repeated the Cl2O photolysis experiment at 248 and 308 nm and present here a refinement of the analysis for the Cl+ ClO channel. At 248 nm, three distinct dissociation pathways that led to Cl + ClO products were resolved. At 308 nm, the angular distribution was slightly more isotropic than previously reported. The experimental method and apparatus are described elsewhere. 2 Briefly, a molecular beam was crossed with an excimer laser beam at the center of rotation of a rotatable mass spectrometer detector. Collisionless conditions allowed for detection of primaiy dissociation products recoiling away from the beam direction. Photoproduct time-offlight (TOF) distributions were measured as a function of detector angle. The new data were recorded on the same apparatus under similar conditions and differed primarily in source conditions and laser polarization. For the new experiment, a continuous molecular beam of Cl2O was formed by passing a gas mixture of 2% SF6 in helium over the swface of liquid Cl2O at - 73 ° C and expanding the mixture through a 0.1 mm quartz nozzle held at or below -60 ° C with a stagnation pressure of 190 Torr. In all experiments, the laser beam was nearly randomly polarized in a plane perpendicular to the scattering plane and parallel to the molecular beam axis. In the current experiment, there was a slight net 37 polarization(~ 58%) perpendicular to the scattering plane, whereas previously there was a slight net polarization parallel to that plane. The differences in the beam conditions and the laser polarization were taken into account in the analysis. Analyses of product TOF and angular distributions were carried out by the forward convolution technique used previously. For each channel, center of mass (c.m.) translational energy P(Er) and angular w(0) distributions were transfonned to the LAB frame and adjusted iteratively until the predicted LAB TOF distributions agreed with the observed distributions at all LAB angles. The c.m. angular distributions were fit to w(8) cc I + PPi(cos8), where the anisotropy parameter p varies from - 1 to 2 for dissociation with the recoil vector perpendicular and parallel to the transition dipole, respectively. 248nm Representative TOF distributions of products from 248 nm photolysis are shown in Figure 1. The mlz = 51 (ClO+) data could not be fit at all angles with a single P(ET) distribution as was done in the previous work. Instead, the three observed peaks at mlz = 51 were fit with three separate components in the P(ET) distribution (Fig. 2), each with different anisotropy parameters. The fastest component had a large translational energy release with a maximum at 49 kcal/mol (Eirn = 33 kcal/mol). The angular distribution was fit by P = 0. 7 ± 0. 2, indicating a prompt, parallel dissociation. The middle component had an average translational energy release of 1..£-r) = 36 kcal/mol ((F,in1) = 46 kcal/mol) and anisotropy of P = 1.5 ± 0.2. A fit to the slowest component gave p = 1.2 ± 0.2 and a P(ET) distribution with a sharp cutoff with a minimum Er of 15 ± I kcal/moI. The cutoff was due to opening of the Cl20 - Cl + Cl + 0 channel, because the internal energy of the ClO product exceeded the 38 bond dissociation energy (D 0 = 63.4 kcal/mol) for products with Er< 19.9 kcal/mot. The 5 kcal/mot discrepancy could be attributed to a) spin-orbit excitation (2.5 kcal/mot) of one or both Cl atom products, 3 b) a centrifugal barrier to dissociation of rotationally excited CIO products (in the rigid radical approximation, 4 the CIO centrifugal barrier could be as high as 7.3 kcal/mot), or c) a combination of these causes. The mlz = 35 signal (Fig. 1A, C) comprised direct Cl photoproducts, Cl atoms from secondary dissociation ofCIO, CIO fragments which cracked in the ionizer, and background from cluster photolysis. In order to fit the Cl counterfragment, we extended the slowest P(ET) distribution to lower translational energies. We concluded that = 4% of the CIO underwent spontaneous dissociation. The secondary Cl signal from this dissociation was modeled and included in the fit. Contributions from primary CIO products were fit assuming 32%, 3'1°/o, and 49% of the ClO fragmented to Ci+ in the ionizer, increasing linearly with ClO internal energy. Photodissociation of ClO was also observed at higher laser power, but was minimized by keeping laser fluences low. A small underlying signal remained, which we assigned to products from photolysis of Cl2O clusters. In summary, we have now resolved three components in the P(ET) distribution with differing anisotropies, each leading to Cl + ClO products. These results support our conjecture that formation of Cl + ClO products involves several excited state potential energy surfaces. 308nm At 308 nm, the mlz = 51 (CIO+) TOF spectra were fit by a P(Er) distribution very similar to the one published previously, with a peak at 33 kcal/mol and a slow shoulder 39 extending down to 6.5 kcaJ/mol. The angular distribution, fit with p = 0.4 ± 0.3, was slightly less anisotropic than previously reported. A long-lived excited state of Cl20 was recently proposed by Nickolaisen et al. s to explain quenching at high pressures of the ClO yield from Cl20 photolysis at A> 300 run in flow tube studies. They postulated a metastable excited state which must live longer than the mean collision time at 5 Torr of~ 50 ns. Such a lifetime would lead to complete rotational averaging of the anisotropic angular distribution, resulting in P ~ 0 .19 if Cl20 dissociates from its ground state geometry. 6 This value of Pis half of the observed value, but within our uncertainty. Our results thus indicate that 't < 9 ps at 308 nm, but our uncertainties preclude us from ruling out a longer-lived state. We acknowledge support of NASA grant NAGW-3893, and assistance from C. M . Nelson and J. W. Seale. 40 References 1. C. M. Nelson, T. A. Moore, M. Okumura, and T. K. Minton, J. Chem. Phys. 100, 8055 (1994). 2. Y. T. Lee, J. D. McDonald, P.R. LeBreton, and D.R. Herschbach, Rev. Sci. Instrurn. 40, 1402 (1969); M. J. O'Laughlin, B. P. Reid, and R. K. Sparks, J. Chem. Phys. 83, 5647 (1985). 3. Cl* products were observed with 18% yield from Cl20 photolysis at 248 nm by A. J. Chichinin, Chem. Phys. Lett. 209,459 (1993). 4. G. E. Busch and K. R. Wilson, J. Chem. Phys. 56, 3626 (1972). 5. S. L. Nickolaisen, C. E. Miller, S. P. Sander, M. R. Hand, I. H. Williams, and J. S. Francisco, J. Chem. Phys. 104, 2857 (1996). 6. S.-C. Yang and R. Bersohn, J. Chem. Phys. 61, 4400 (1974). 41 Figure Captions 1. TOF number density distributions N(t) from Cl2O photolysis at 248 nm for A) Ci+ products collected at a detector angle of 30°, B) CIO+ products at 30°, C) Ci+ products at 50°, and D) Clo+ products at 50°. The solid line is the overall fit to the data. The dashed, dotted, and dot-dashed lines correspond to the three components in the c.m. P(Er) distribution (Fig. 2). The long-dashed line in A) and C) is a hypothetical fit based on photolysis of Cl2O clusters . in the beam. The dot-dot-dashed line in A) and C) corresponds to the spontaneous secondary dissociation of ClO. Stimulated secondary dissociation of ClO is not included. 2. Optimized c.m. P(Er) distribution for the Cl+ ClO channel from Cl2O photolysis at 248 nm. The light-dashed line in the slowest component shows the cutoff from the mlz = 51 data, corresponding to internal energies beyond the ClO dissociation limit. The anisotropy parameters and average translational energies for each component are: dashed line, P= 1.2 ± 0.2, 1£-r) = 17.8 kcal/mol; dotted line, p = 1.5 ± 0.2, {,£,T) = 35.9 kcaVmol; and dot-dashed line, p = 0. 7 ± 0.2, (ET) = 51.1 kcaVmol. independent of angle. The scaling factor for each component is 42 A 100 150 200 er, 30° C 250 100 150 Flight time, µs Figure 1 200 250 300 43 . . u, ...., ·C: T\,' :::) • I I II \ .•• . .c ~ <( I I \ I I I •• :\ •\ II •: \ 20 30 I : ..-..- urQ I 0 I 10 I . I ·..'\ '·.j I : , \ \ ,, I .... I 40 I • 50 Energy, kcal/mol Figure 2 - I 60 44 Chapter 3. ClOOCl Molecular Beam Source A molecular beam source of ClOOCl was designed in order to study ClOOCl photolysis using photofragment translational spectroscopy. We created a molecular beam of ClOOCl by photolyzing Cl2O at 308 nm in the stagnation region and cooling the reaction cell to promote three body recombination to form ClOOCl. The main species in the beam were Cl2, ClOOCl, and Cl2O, with minor species 0 2 and undimerized ClO also present. 45 A. Introduction Knowledge of the branching ratio between Cl and ClO products from ClOOCl photolysis is crucial to our understanding of polar stratospheric ozone depletion. Because of the uncertainties in the previous measurements, we endeavored to study ClOOCl photolysis via photofragment translational spectroscopy, a powerful technique capable of direct detection of both Cl and ClO photolysis products that can provide an absorption cross-section independent method of determining the Cl:ClO branching ratio. Thus, we had to design a molecular beam source of ClOOCl. Previous researchers have utilized different methods to produce ClOOCl, forming ClO either chemically or photolytically under conditions where three-body recombination is favored over the ClO self-reaction. These conditions usually included cooling to temperatures below 250 K, increasing the pressure, and selecting the buffer gas to increase the rate of dimer formation.1 ClO has been formed by photolysis of Cl2O, 2, 3 by flash photolysis of Cl2Cl2O, Cl2-O3, or Cl2-O2 mixtures, 4•5 by flash radiolysis of Cl2O-Cl2 mixtures, 6 by the chemical reaction of Cl with 0 3, 7•8•9•10•11 •12 Cl2O, 7•10•11 •12 or OClO, 7•10•11 or F + Cl2O, 11 or by photolysis ofOClO. 6 Static or flow cells were used in all previous experiments involving ClOOci in which typical concentrations of ClOOCl were on the order of 1-10 mTorr, with a maximum ofa few hundred mTorr.2, 3 In order to design a molecular beam source of ClOOCl that could be used for a photolysis experiment, several problems had to be overcome. considerations are as follows: The major design 46 • ClO source ClO can be produced by any of the methods described above. The ClO source chosen determines what other species besides ClOOCl will be present. Table I lists other possible species that could be present in a ClOOCl source which could give rise to Cl or ClO photolysis signal, along with their absorption cross sections at 248 and 308 nm. Table I. Absorption cross sections for various species that could be formed in a ClOOCl source. All cross sections are taken from the 1997 JPL evaluation, 13 except for the ClOOCl cross sections which are from ref 3, with the values from ref 13 in parentheses. Species 0(248 nm), cm2 0(308 nm), cm2 Cl2 1.1 X 10- 2! 1.7 X 10-19 ClO 3.2 X 10-! 3 2.0 X 10-!9 Cl20 1.7 X 10-! 3 4.6 X 10-!9 OClO ~ 10-19 ~ 1.5 X 10-18 ClOO 2.95 X 10-! 7 ClOOCl 6.2 (6.3) X 10-! 3 3.4 (4.9) X 10-l 9 Cl20 3 1.2 X 10-l 7 2.4 X 10-18 • ClO dimerization The possible ClO + ClO reaction channels are ClO + ClO + M -+ ClOOCl + M ClO + ClO -+ ClOO + Cl -+ Cl + Cl + 0 2 -+ Cl +OCIO. At low temperature, below 250 K, ClOOCl formation is the dominant ClO reaction pathway. The identity of the third body is significant, 1 with reaction rates following the expected trend 47 for third body quenching efficiency except for 0 2 and Cl2 . Oxygen results in a lower rate of ClOOCl fonnation, and chlorine in a higher rate. Thus, cooling the reaction cell and careful choice of the bath gas can encourage dimer fonnation. • ClOOCl concentration The kinetics experiments produced ClOOCl concentrations on the order of milliTorr, whereas for a beam experiment to be feasible, concentrations on the order of Torr are required. Major loss mechanisms for ClOOCl include 1) reaction with Cl or O in sources where these atoms are continuously present, 2) photolysis of ClOOCl in sources which employ continuous photolysis to produce ClO, and 3) thennal decomposition and the ClO self reaction in sources at temperatures above 250 K. The rate constant 13 for the three body association of ClO to fonn ClOOCl is [2.2 X 10-32 (T/3oor 3•1]cm6/molec2/s. If a ClO concentration of 3 x 1016 molec/cm3 (1 Torr) could be fanned, for a background pressure of6 x 1018 molec/cm3 (200 Torr) and a temperature of220 K, then 3 x 1020 molec Cl0OCl/cm3/s could be formed. In this case, to produce 1 Torr of ClOOCl, the residence time in the reaction chamber must be at least 100 µs (not accounting for loss processes). Thus, high production ofClO, long residence times, and minimization ofloss processes are necessary to obtain the high ClOOCl concentration required. We chose Cl2O photolysis at 308 nm as our ClO source, because the other species likely to be present besides ClOOCl are residual Cl2O, ClO, and Cl2, with little or no OClO or Cl2O3. Cl2 and ClO are diatomics, and will produce sharp well-defined peaks at mlz = 3 5 in a photolysis experiment. We have extensively studied Cl2 O photolysis, 14 which produces 48 signal at Ci+ and CIO+, and can model that in our photodissociation experiment. With the exception of a small concentration of clusters, no species likely to be present in the beam besides CIOOCI and Cl2O can produce CIO photolysis signal. We wanted to avoid CIO sources that would have significant concentrations of OCIO and Cl2O3 to minimize the number of species that could contribute to the Clo+ photolysis signal so that we could obtain a more accurate CIO/CI branching ratio. Also, the absorption cross sections of OCIO and Cl2O3 at 308 run and ofCl2O3 at 248 run are significantly higher than that for ClOOCI, so that the photolysis signal from these components could mask the ClOOCI photolysis signal. A molecular beam of CIOOCl was formed by photolyzing Cl2O in a quartz cell and expanding the reaction mixture through a 0.1 mm nozzle. At 308 run, the only Cl2O photolysis products formed are Cl and ClO. The Cl atoms produced also reacted with Cl2 O, producing ClO and Cl2 : Cl2O + hv - Cl + ClO Cl + Cl2O - Cl2 + CIO ClO + ClO + M - ClOOCl + M net: 2Cl2O - ClOOCl + Cl2 A pulse of ClOOCl/Cl2O/Cl 2 and other reaction products exited the nozzle, effectively creating a pulsed ClOOCl source with a continuous background of Cl2O. Given our choice of ClO source, our solutions to the problems posed in the previous section are as follows: • Continuous gas flow allowed for long residence times, on the order of milliseconds. • Pulsed laser photolysis in the stagnation region allowed for high pressures, but avoided 49 steady state effects like ClOOCl photolysis and attack by Cl. • Cooling the source to T < 240K minimized the ClO self reaction and ClOOCl thermal decomposition. • High stagnation pressure promoted rapid three body recombination, but could be kept low enough to avoid significant formation of clusters in the beam. • Minimization of a) back diffusion ofphotolyzed gas and b) mixing of cooled gas with incoming warm gas keeps the ClOOCl concentration from being diluted by the Cl20 and buffer gas. The chemistry of this source was modeled using the reactions and photochemistry shown in Table II, including the species Cl20, Cl, ClO, 0, ClOOCL Cl2, 0 2, OClO, ClOO, and Cl20 3. All rate constants, branching ratios, and photodissociation cross sections were taken from the 1994 JPL report, 13 except the branching ratio for ClOOCl photolysis, for which we used an adjustable parameter for the ClO + ClO yield. Model calculations were carried out using values of0.0 to 0.5 for this parameter. The model predicted that for pressures on the order of IO0T, temperatures around 220K, and initial Cl20 concentrations of~ 10 T, the output from this source would be primarily Cl2, Cl20, and ClOOCl, with some ClO and 0 2, and negligible amounts of OClO and Cl20 3. Figure I shows the results of a model calculation for the initial conditions Ps1ag = 200T, P020 = IOT, T = 200K, residence time trea = 50 µs, laser spot size= 0.09 cm2, and laser energy= 0.5 mJ/pulse, with the ClO + ClO yield set to 0.2. B. Source Design Parameters Preliminary source development was carried out in a test chamber at Caltech, and optimization of the operating parameters was completed on the 3 5" rotating detector Crossed 50 Molecular Beams Apparatus at Montana State University. A schematic diagram of the source Table II. Chemistry and photochemistry modeled for the ClOOCl source. Chemistry Photochemistry ClO + ClO + M ... ClOOCl + M Cl2O + hv ... Cl + ClO ClO + OClO + M ... Cl2O3 + M Cl2 + hv ... Cl + Cl Cl + ClOOCl ... Cl2 + ClOO OClO + hv ... 0 + ClO Cl + Cl2O ... Cl2 + ClO ClOO + hv ... 0 + ClO - Cl + OClO ... ClO + ClO Cl2O3 + hv ... ClO + OClO Cl + ClOO ... Cl2 + 0 2 ... ClO + ClO 0 + ClOOCl ... ClO + OClO ClOOCl + hv ... Cl+ ClOO ... ClO + ClO ClO + hv ... Cl + 0 0 + Cl2O ... ClO + ClO 0 + ClO ... Cl + 0 2 0 + OClO ... ClO + 0 2 ClO + ClO ... ClOO + Cl ... Cl2 + 0 2 ... Cl+ OClO ClOO + M ... Cl + 0 2 ClOOCl + M ... ClO + ClO is shown in Figure 2. The shape and dimensions of the reaction chamber were very important to the beam composition. The chamber diameter helped define the reaction volume, and the shape helped control the mixing of the laser processed gas with the unprocessed gas. Several 51 reaction chamber configurations were tested. The best configuration was a tapered cell starting from an ID of 3 mm and tapering down to about 0.1 mm over a length of about 1 cm, with a 0. 1 mm nozzle at the tip. The cell protruded 1 cm from the cooling jacket, with a constriction of the inner diameter to~ Imm at the point where the cell left the cooling jacket. This constriction was quite important because it reduced back diffusion of the ClOOCl into the unprocessed Cl 20, which would result in a dilution of the ClOOCl signal and thus a reduction of the peak ClOOCl concentration. Another constriction was placed at the entrance to the cooling jacket, to reduce mixing of the cold Cl20 with the warm incoming Cl20 . The 3 inches of the 1/4 inch OD quartz tubing immediately prior to the cell tip was encased in a cooling jacket held at 205-220 K to facilitate ClO dimer formation and reduce side reactions. At temperatures below 205 K, condensation of Cl20 in the cell and the formation of clusters in the molecular beam occurred. At higher temperatures, the absolute concentrations of ClOOCl were reduced. Cold pentane was circulated through the cooling jacket to maintain the temperature. Pentane was used as the cooling fluid because its low viscosity allows it to be easily pumped at these low temperatures. After every experiment, the pentane bath was allowed to warm up, and the water separated out. The pentane was stored in metal containers until the next experiment. The cell was mounted using two sets of three point Vespel contacts set in a Nema G-10 base in order to minimize movement when cooled and thus keep the nozzle aligned with the plane of the detector. Great care had to be taken to tighten the contacts enough to hold the cell in place but not so much that the cell cracked. The temperature was measured by thermocouples epoxied to the stainless steel Swagelock nuts of the input and output ports on the vacuum jacketed liquid feedthrough 52 through which the cold pentane flowed. Helium was kept flowing through the nozzle after the beam was turned off until the nozzle reached room temperature, in order to prevent condensation of pump oil on the cold nozzle which could lead to clogging. C. Time-of-flight Mass Spectrometric Analysis The beam passed through two regions of differential pumping before entering the main chamber. The source chamber pressure was ~ 2 x 10-4 Torr with the beam on. The beam exited the source chamber through a 0.5 mm skimmer which was located 6 mm downstream from the nozzle. The beam passed through a second differentially pumped chamber, with a pressure of< 10- 5 Torr, and exited through a slit aperture (1.66 mm horizontal by 2.2 mm vertical) which defined a molecular beam cross section of 1.92 mm by 2.55 mm at the interaction region (the intersection point of the ClOOCl molecular beam and the photolysis laser or scattering molecular beam, and the point around which the detector rotates). This aperture also reduced the effusive background entering the detector at small viewing angles. The pressure in the main chamber was ~ 10- 7 Torr. The rotatable mass spectrometer detector had three regions of differential pumping, with pressures in the third region, which housed the ionizer, reaching~ 10- 10 Torr. Time-offlight (TOF) mass spectra of the source output were obtained by rotating the mass spectrometer to look directly at the beam. From the interaction region, the beam traveled 34.5 cm to a Brink electron bombardment ionizer. The nominal electron energy used for examining the beam was lowered to 30 eV in order to reduce fragmentation of the ClOOCl in the ionizer. Ions were mass analyzed with an Extranuclear quadrupole mass filter, and detected with a Daly-type ion detector. Table III lists the ionizer conditions used to obtain 53 all of the source TOF spectra shown. The firing of the source laser was used to trigger the multi-channel scaler, so that the measured flight time of the ClOOCl signal was the sum of the residence time of the gas in the nozzle after the laser fired, the flight time from the nozzle to the ionizer, and the time for the resulting ion to travel through the mass filter to the ion detector (ion flight time). The ion flight time for an ion with mass-to-charge ratio mlz was empirically determined to be 3.14 (mlz)½ µs for these ionizer conditions. The source TOF signal consisted of the background signal from the pure Cl2O beam-before and after the laser processed pulse and the processed region (Cl2O depletion and ClOOCl peak). Table III. Low electron energy ionizer conditions for examining the source beam. Ionizer Component Voltage Aluminum coated cathode - 30 kV Photomultiplier tube -2000 to-2200 V First and second exit lenses -900V Electron energy 30eV Filament emission current 2.1 mA Mass spectrometer resolution !lm ~ 2.2 amu (at mlz = 51) Extractor -lOV Lens -400V Ion energy 35 eV TOF spectra of the molecular beam were obtained at mlz = 35 (Ci+), 5 l(ClO+), 67 (CIOO+), 70 (Cl/), 86(Cl2O+), 102 (ClOOCi+), 118(Cl2O/), 121(Cl2O•Ci+), and 137 (Cl2O•CIO+, CIOOCI•Ci+). Typical source data are shown in Figure 3. No signal was observed for cluster masses above mlz = 137, up to the mass spectrometer limit of mlz = 200. 54 Although ClOOCl is the lowest energy isomer of Cl2O2, 15•16•17•18 two other isomers, C1ClO2 and ClOClO, exist. These isomers do not form stable parent cations when ionized, 19 so that the signal at mlz = 102 was not from ionization of these isomers, which would appear at mlz = 67. ClOOCl is the dominant reaction product in flow tube experiments, 19 although there is some evidence that C1ClO2 can be formed heterogeneously, especially if the walls of the flow tube are coated with ice. 20 In our experiment, the walls of the reaction cell were dry; furthermore, the shape and arrival time (after accounting for the ion flight time) of the mlz = 67 peak was very similar to that of the mlz = 102 peak, indicating that the two peaks arose from the same species. Thus, we assumed that the only Cl2O2 isomer formed in our experiment was ClOOCI. It is possible that some ClOClO was formed by clustering of CIO in the expansion, but the low concentration of ClO (< 30 mTorr) remaining at the nozzle suggests that if ClOClO is present in our experiment, the concentration must be small compared to the ClOOCl concentration. Even with the low electron energy used, the ClOOCl still fragmented in the ionizer, with major cracks at mlz = 67 (CIOO+) and mlz = 51 (Clo+). Cl2O contributed some signal at ClO+, but the cracking pattern of Cl 2O was measured separately with the source laser turned off: and Cl2O was accounted for at mlz = 51 based on the mlz = 86 signal. While it is possible that OClO could contribute signal to the mlz = 67 peak, this is unlikely because 1) the peaks at mlz = 67 and 102 have the same shape and arrival time after accounting for ion flight time, 2) the model of the source chemistry indicates that the OClO concentration in the source is very low, 3) an elastic scattering experiment (described later in this chapter) showed no evidence ofOClO present in the beam, and 4) no photolysis signal corresponding to OClO 55 was observed for photolysis at 308 nm.21 The mass spectrometer resolution was set such that the large Cl2 signal observed at mlz = 70 contributed< 5% to the signal at mlz = 67. Thus, we assumed that all of the signal at mlz = 67 arises from ClOOCl that has fragmented in the 10ruzer. Table IV shows the fragmentation pattern for Cl 2, Cl2O, and ClOOCl for the ionizer conditions used. The background signal at mlz = 5 l and 86 was assumed to be entirely from the Cl2O background and was used to determine the relative amou-nts of the Clo+ and c12O+ ion fragments of Cl2O. The remaining signal at the bottom of the depletion at mlz = 86 was assumed to be undissociated Cl2O, and the remaining signal at mlz = 5 l was assumed to be ClO, Cl2O, and ClOOCl, with the Cl2O contribution scaled by the Cl2 O fragmentation pattern. The ClO accounted for ~ 6% of the mlz = 5 l signal, based on subsequent analysis of the ratio of the ClO : Cl2O photolysis signal at 248 nm. The mlz = 70 background was attributed to residual Cl2 impurities in the Cl2O, and the peak signal to Cl2 formed in the beam. The SF+ and SF/ cracks from SF6 (also present in the buffer gas) contributed negligible signal to the mlz = 5 l and mlz = 70 TOF spectra for this electron energy and were neglected. Almost every species present in the beam contributed signal to mlz = 35, so the contributions are difficult to sort out, but at this electron energy the mlz = 3 5 signal is very small compared to the other fragment masses. Table IV. Fragmentation pattern for Cl,, Cl,O, and ClOOCl. mlz = 51 Species mlz= 35 Cl2 1 Cl2O ~.05 1 ClOOCl ~.l 3.4 mlz= 67 mlz = 70 mlz = 86 mlz= 102 14 1.1 4.5 1 56 A peak at mlz = 118, which could correspond to Cl2O3+ or to cluster peaks including Cl2O•O/ or ClOOCl•O+, was observed, rising and reaching a maximum at slower times than the ClOOCl peak. This peak was at least an order of magnitude smaller than the ClOOCl parent peak. The observation that this peak was slower and shaped differently than the ClOOCl peak indicated that the signal at mlz = 102 did not arise from Cl2O3 fragmenting in the ionizer. A depletion in the signal at mlz = 121 (Cl2O•Ci+) was observed, corresponding to the depletion in the Cl2O. At mlz = 137 (Cl2O•ClO+, ClOOCl•Ci+), a small peak was observed, indicating the formation of some ClOOCl-containing clusters. This peak did not start rising until after the maximum ClOOCl signal. The cluster signals were small compared to the Cl2O and ClOOCl signals, and we attempted to minimize them in spite of the need for a cold nozzle which is conducive to cluster formation. Figure 4 shows mlz = 102 and the higher mass peaks 118, 121, and 13 7. The photolysis laser delay was set to sample only the maximum ClOOCl signal, where the cluster peaks (including Cl2O3+) were small or negligible. An estimate of the relative amounts of each species in the molecular beam could be made by examining the source TOF spectra. The background signal arose only from Cl2O, any residual Cl2 impurities, and Cl2O clusters (which were a small fraction of the total signal and can be ignored in this estimate). At the flight time corresponding to the CIOOCl peak, the signal arose from all the species formed from the photolysis and ensuing reactions: Cl2, Cl 2O, ClOOCl, and ClO. From the background signal and the fragmentation patterns, the relative contributions at each mass could be sorted out. Then we added up the signal at all of the ion fragments for each species to get the total signal arising from that species. We estimated the relative detection efficiencies for each species by estimating the ionization cross 57 sections from the polarizability.22 Based on the total signal for each species and the relative detection efficiencies, we obtained the ratio ClO : Cl2 : Cl2 O : ClOOCl ~ 0.14 : 14 : 1.0 : 2.4 for the relative amounts of the major species formed in the source. Based on these amounts and the absorption cross sections (using Huder and DeMore3 for ClOOCl), we could predict the relative contributions to the photolysis signals expected at 248 and 308 nm: 248 nm: Cl2 : Cl 2 O : ClOOCl ~ 0.008 : 1.0 : 8.8 308 nm: ClO : Cl2 : Cl2O: ClOOCl ~ 0.3: 5.2: 1.0: 1.8. These predictions were based on average values for the relative amounts of each species in the beam over all of the experiments performed; however, each experiment was carried out under slightly different conditions, so the photolysis signal contributions can vary depending on the actual ratios of the species for that given beam. No estimate for the ClO photolysis contribution at 248 nm is provided, because the observed photolysis contribution was used to obtain the relative ClO concentration in the beam. D. Optimization of Source Conditions A carrier gas swept over liquid Cl2O to produce a continuous beam of Cl2O which was then photolyzed at 308 nm. The Cl2 O synthesis, purification, and Cl2 :Cl2 O ratio are discussed in Appendix 1. The photolytically produced CIO had an average _internal energy of 28 kcal/mo I, and the dimer bond dissociation energy is 17 .1 kcal/mo1, so the excess internal energy needed to be quenched in order for dimerization to occur. Different buffer gases were tested, including pure helium and nitrogen buffer gases, as well as CH4 , CH3 F, CH2 F2 , CHF3 , and SF6 mixtures in helium. Figure 5 shows TOF spectra for a pure helium beam, He + 2% 58 CF~ and He+ 2% SF6. A mixture of2% SF6 in helium was chosen as the best mixture, as it allowed for the highest absolute amounts ofClOOCi+ signal for the desired ClOOCl+: Cl2 O+ ratio. Increasing the amount ofSF6 led to a slower beam velocity and lower absolute ClOOCl concentrations for the same ClOOCi+:Cl2O+ratio. We tested to be sure that the SF6 did not cause any spurious signal by carrying out a photolysis experiment using pure helium as the buffer gas. The absolute stagnation pressure used was ~190 Torr, ranging from 180 to 200 T. Higher stagnation pressures led to formation of clusters, and lower stagnation pressures led to a weaker expansion and poorer beam quality (lower speed ratio). The Cl2O was held at -73° to -74° (vapor pressure 10-12 Torr). Higher Cl2O concentrations caused an increase in cluster formation, and lower concentrations led to reduced ClOOCl concentrations. The 308 nm output from a pulsed Lambda Physik EMG 103 MSC excimer laser was focussed with a 1 m focal length lens and sent into the source chamber through a fused silica window and then directed onto the tip of the reaction cell by a 90° prism. The laser spot size was varied by changing the position of the lens, with the best results obtained for a spot size ~ 8mm wide by 2 mm high, so that a section of the cell 2mm long by the cell diameter at the laser spot position was illuminated. The laser spot could be moved along the cell closer to or further from the tip to vary the reaction time between when the laser fired and when the reaction mixture exited the cell. The spot position was changed by raising or lowering the rear legs of the laser, which could be done in a reproducible manner. The laser was aligned to the cell by removing the top flange of the source chamber, inserting a white business card under the cell, and firing the laser for 10-20 shots until the card showed a bum mark. The 59 outside diameter of the cell was 5 mm, so 3 mm of the laser spot overlapped the edges of the cell. The laser position was adjusted until the "shadow" made by the cell was centered on the bum mark on the business card. The laser position from the tip of the cell was determined by moving the laser spot along the cell until it passed the tip and the cell no longer caused a shadow to form in the bum mark. A typical experiment was run with the center of the laser spot about 5 mm from the tip of the cell. During the experiment, care was taken to not move the spot within 1 mm of the tip to avoid blowing the tip off the cell. The laser power was varied by attenuating the laser beam with quartz plates to change the extent of Cl2O dissociation, with typical laser energies of 130-170 mJ/pulse. Ninety to ninety-five percent depletion of the Cl2O signal gave the highest ClOOCi+:Cl2O+ ratio. At laser powers causing > 95% Cl2O depletion, the ClOOCl parent signal was also reduced, because of secondary photodissociation of ClO and attack of ClOOCl by Cl atoms. At lower laser powers, the amount of Cl2O remaining undissociated (> 10%) was much greater than the amount of ClOOCl formed. Figure 6 shows the mlz = 86 and 102 spectra for laser energies of 130 mJ/pulse (A, B) and 160 rnJ/pulse (C, D). The source laser was operated at 40 Hz, as that was the maximum repetition rate of the photolysis laser used to do the photolysis experiments. Some data was also obtained at a repetition rate of 20 Hz. The absolute ClOOCl signal did seem to increase at the lower repetition rate, possibly because the reaction cell could maintain a lower temperature. The increase in signal at 20 Hz was not enough to compensate for the reduction in counting time obtained by going to the higher repetition rate. The laser power and position were optimized before each set of experiments while the 60 other operating conditions remained consistent from experiment to experiment. The laser parameters were adjusted until the ClOOCl peak observed at mlz = 102 was O.1 to 1. 0 times the Cl 20 signal remaining at mlz = 86. For the conditions described, the pulse of ClOOCl, Cl20, Cl2, and undimerized ClO that exited the nozzle was about 1-3 ms wide, with the ClOOCl pulse maximum spanning ~200 µs. Typical residence times in the cell after the laser pulse were 100 - 500µs. E. Elastic Scattering In order to further characterize the source beam, we performed an elastic scattering experiment in which we scattered the ClOOCl beam off a neon beam and measured the angular distribution of the deflected products. Angular distributions were measured for a given mlz to determine what masses contributed to the signal. This technique provides a measure of mass that is independent of mlz and fragmentation patterns. A neon beam produced by expanding neon at a stagnation pressure of 1250 Torr through a 0.1 mm nozzle at 22°C was crossed by a ClOOCl beam at an angle of90° in the interaction region. The beam velocity of the ClOOCl beam could not be measured directly or inferred from diatomic photolysis signal. However, the beam velocity for all of the photolysis experiments as well as that measured separately with a chopper wheel remained fairly consistent, ranging from 9.0 to 9.5 x 104 cm/s. The source conditions for the scattering experiment were closest to those for the preceding few experiments, for which the beam velocity was 9.05x 104 cm/s; thus, we used this value as the ClOOCl beam velocity. Because we did not measure the velocity directly, this value may have some error, which would slightly change the angles corresponding to the edges of the elastic Newton circle. The 61 deflected molecules were detected by rotating the detector off the ClOOCl beam axis. Distributions were obtained both with the beams crossing and with the neon beam blocked by a beam flag. The background distributions (beam flagged) were subtracted from the total signal (beam unflagged) to obtain the scattered signal, which consisted of signal from the Cl2 O background as well as the Cl2O depleted, ClOOCl rich portion of the beam. The distributions were obtained with a resolution of 100 µs/point, so different ranges could be taken to characterize the ClOOCl portion of the beam. Ranges of 100, 200, 400, and 600 µs were used. Data for the ClOOCl beam scattering off neon were obtained for mlz = 51 (ClO+), 67 (OClO+), 86 (Cl2O+), and 102 (ClOOCi+), with angular ranges of 19° - 49°, 20° - 36°, 25° 31 °, and 13° - 25°, respectively. Experiments were also done for a pure Cl2O beam from a cooled nozzle (ClOOCl source laser turned off) crossed with neon and for argon crossed with helium. For the Ar+ He case, there are no vibrations or rotations for the collision energy to be partitioned into. The Ar beam velocity was 5.54 x 104 emfs and the He beam velocity was 1. 75 x 10 5 emfs. From the Newton diagram (Fig. 7), the Ar signal can not be scattered beyond the edge of the Newton circle, or a lab angle of 37°. The angular distribution shown in Figure 8 shows this sharp cutoff in signal as expected, centered about the laboratory velocity corresponding to the edge of the elastic Newton circle. In the pure Cl2O + Ne case, the Cl2O beam velocity was 7.7xl04 emfs and the Ne beam velocity was 7.8x 104 emfs. From the Newton diagram (Fig. 9), the maximum lab angle at which scattered Cl2O can be observed is 32.4° . However, unlike the pure elastic Ar+ He scattering data, the angular distribution (Fig. 10) shows that the scattered signal has 62 essentially disappeared before this angle, and the drop in signal occurs over a longer angular range, 10° instead of~ 6° . These differences indicate that the scattering is inelastic, with some of the energy partitioned into Cl20 internal modes. If we took the midpoint of the portion where the signal is rapidly falling (i.e. the region which would correspond to the edge of the elastic Newton circle if the scattering was elastic), we obtained an angle of26.5°. Assuming that this angle corresponded to the edge of the inelastic Newton circle, we obtained a new c.m. velocity (and -thus translational energy) of the scattered Cl20 and, using conservation of energy, determined the amount of the collision energy partitioned into Cl20 internal energy. The initial total c.m. translational energy Er was 9.8 kJ/mol. Using 26.5° as the edge of the inelastic Newton circle, the final Er becomes 5.0 kJ/moL indicating that 4.8 kJ/mol was partitioned into internal energy (EmJ. Thus, the change in translational energy was 49% of the initial translational energy, or llEiET = 0.49. With the source laser turned on, we examined the Cl20 background at both mlz = 51 and 86 well after the ClOOCl portion of the distribution. The beam velocity in this region was 8.4x 104 cm/s, as measured previously with a chopper wheel, and the edge of the elastic Newton circle (see Newton diagram, Fig. 11) corresponded to a laboratory angle of30.4°. Again, the scattered signal is already gone by this angle. At mlz = 86, we did not collect data at small enough angles to see where the drop in signal begins, but at mlz = 51 we were able to obtain this information (Fig. 12 A,B). The midpoint of the drop, as explained before, came at 24°, which we then took as the edge of the inelastic Newton circle. Using this value, the amount of energy partitioned into Cl20 internal energy was tEm,) = 6.2 kJ/mol. The initial 63 ET was 10.75 kJ/mol, so ~EJET = 0.58. This value is higher than that for the pure Cl2O beam. However, the collision energy was higher because the Cl2O beam velocity was higher, so the fraction partitioned into internal energy would not be expected to remain the same. For the ClOOCl beam scattering experiment, all ClOOCl beam components will contribute to the scattered signal. The ClOOCI beam velocity was 9.05xl04 emfs, and the Newton diagram including elastic Newton circles for CIO, OCIO, Cl2O, CIOOCI, and Cl2O 3 is shown in Fig. 13. The mlz = 86 signal can have contributions from Cl2 O3, CIOOCI, or Cl2O. Unfortunately, we did not take data at small enough angles to observe the Cl2O3 or ClOOCl contributions, or to observe the portion of the angular distribution before the signal started to drop. The angular distribution is shown in Fig 14 A, where it is apparent that 1) there is signal after the CIOOCl elastic Newton circle cutoff, and 2) there is a small shoulder (~ 10% of the peak signal before the drop) at 29°, just before the edge of the Cl2O elastic Newton circle. We can obtain information on the Cl2O internal energy partitioning from the mlz = 51 angular distribution, shown in Fig. 14 B. The Cl2O+ : ClOOCi+ ratio was high for this experiment, so the mlz = 51 signal arose mainly from Cl2O. Also, the signal drop observed occurs near the ClOOCl elastic Newton circle; however, the ClOOCI scattering would be expected to also be inelastic and thus be observed at smaller angles, where no data was taken. Ifwe assumed the scattered signal observed at mlz = 51 arises solely from Cl2O, and that the same fraction of the collision energy was partitioned into internal energy as in the background Cl 2 O distribution (~EJET), then we predicted that the edge of the inelastic Newton circle would be at 23°. From the mlz = 51 angular distribution, the edge of the Newton circle 64 should be at~ 23.5°, which would lead to a llEiET of0.63. The beam velocity was higher in the ClOOCl-rich region than in the Cl20 background, which could account for the higher fraction of collision energy partitioned into internal energy. The mlz = 51 angular distribution also shows a small fraction of scattered signal remaining out to 43°, well beyond the elastic Newton circle of every species in the beam except for ClO. This signal was small compared to the Cl20 signal(~ 5%) and indicated the presence of a small concentration of ClO in the beam. The mlz = l 02 signal can consist of contributions from ClOOCl, Cl20 3, and any higher cluster masses. The edge of the elastic Newton circle is at 24.8° for ClOOCl and 21.3° for Cl20 3. Like the Cl20 signal, the mlz = l 02 angular distribution (Fig 14 C) showed a gradual drop in signal which was gone by 25° (but not 21 ° ). There was a slight shoulder peaking at 22°, just before the edge of the elastic ClOOCl Newton circle, similar in width and magnitude to the shoulder near the Cl20 Newton circle in the mlz = 86 angular distribution. The midpoint of the sharp drop in signal occurred around the laboratory angle 19° - 19.5°. Ifwe assumed that the mlz = l 02 signal arose from ClOOCI, the resulting collision energy was 11. 8 kJ/mol, and the amount partitioned into internal energy was 6. 7 k:J/mol, indicating that llE/Er = 0.56. If we instead assumed that the signal arose from Cl20 3, the collision energy was 13.1 k:J/mol, and the amount partitioned into internal energy was 4.5 kJ/mol, indicating that llE/Er = 0.34. Because Cl20 3 has more internal degrees of freedom than Cl20, we would expect that the fraction of the collision energy partitioned into internal energy would be greater, not less, than that for the Cl20 case. The Cl20 3 parent signal was too weak to obtain scattering data. The amount of energy partitioned into the ClOOCl internal energy was 65 comparable to the Cl 2O case if we assumed that the mlz = 102 signal was from ClOOCl. From the energy partitioning results, we inferred that the mlz = 102 signal arises from ClOOCI. Themlz= 67 angular distribution, shown in Fig 14 D, only includes angles from the last 4 ° of the portion where the signal was dropping rapidly out to the edge of the OClO elastic Newton circle, but not the portion showing the beginning of the drop in signal. The drop in the angular distribution occurs before the edge of the ClOOCl elastic Newton circle. The average angular range over which the signal drops is~ 8°, so we assumed that the edge of the inelastic Newton circle was~ 20°. If the signal arose from OClO, then the ll.ErfET would be 0.94, with nearly all of the collision energy being partitioned into OClO internal energy. The edge of the inelastic Newton circle for the mlz = 102 scattered signal was ~ 19.5°, which strongly indicates that the mlz = 67 and mlz = 102 signal arise from the same species. A very small shoulder (~ 6% of the highest signal observed, or :::: 3% of the estimated peak) was observed right at the edge of the ClOOCl elastic Newton circle, and no signal was observed beyond this shoulder. If shoulder arose from OClO scattering, most of the collision energy would again be partitioned into internal energy, with ll.ErfEr = 0. 72. These considerations indicate that the mlz = 67 signal does not arise from OClO, which provides indirect evidence that its reaction product Cl2O3 is also unimportant in our source. The other two isomers of Cl2O2 could contribute to the mlz = 67 scattered signal; however, we did not obtain data at small enough angles to observe much of the angular range where these species could contribute signal. It is possible that the different Cl2O2 isomers might result in different fractions of the collision energy partitioned into internal energy. In 66 this case, a careful study of mlz = 67 angular distributions and comparison to mlz = 102 data (to which only ClOOCl contributes) could give some information on the relative concentrations of each isomer in the beam. In conclusion, the scattering experiment showed that the components of the ClOOCl beam scattered inelastically off of the neon beam, with over half of the collision energy being partitioned into internal energy. Evidence was observed for ClOOCl, Cl2 0, and ClO in the laser processed portion of the beam, with OClO and Cl20 3 unlikely to be major components in the beam. No scattering data for Cl2 was obtained, although it is a major component in the beam. More data was needed to be able to make more firm conclusions on the species present in the beam and their relative concentrations. F. References 1. S. L. Nickolaisen, R.R. Friedl, and S. P. Sander, J. Phys. Chem. 98, 155 (1994). 2. R. A. Cox and G.D. Hayman, Nature 332, 796 (1988). 3. K. J. Huder and W. B. DeMore, J. Phys. Chem. 99, 3905 (1995). 4. S. P. Sander, R.R. Friedl, and Y. L. Yung, Science 245, 1095 (1989). 5. N. Basco and J.E. Hunt, Int. J. of Chem. Kinetics Vol XI, 649 (1979). 6. T. Ellermann, K. Johnsson, A. Lund, and P. Pagsberg, Acta Chemica Scandinavica 49, 28 (1995). 7. L. T. Molina and M. J. Molina, J. Phys. Chem. 91, 433 (1987). 8. M. J. Molina, A. J. Colussi, L. T. Molina, R. N. Schindler, and T.-L. Tso, Chem. Phys. Lett. 173, 310 (1990). 9. M. Schwell, H.-W. Jochims, B. Wassermann, U. Rockland, R. Flesch, and E. Ruhl, J. Phys. Chem. 100, 10070 (1996). 67 10. J. B. Burkholder, J. J. Orlando, and C. J. Howard, J. Phys. Chem 94, 687 (1990). 11. M. Birk, R. R. Friedl, E . A. Cohen, H. M. Pickett, and S. P. Sander, J. Chem. Phys. 91, 6588 (1989). 12. M. Trolier, R. L. Mauldin III, and A. R. Ravishankara, J. Phys. Chem . 94, 4896 (1990). 13 . W. B. DeMore, S. P. Sander, D. M. Golden, R. F. Hampson, M. J. Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb, and M. J. Molina, Chemical Kinetics and Photochemical Datafor Use in Stratospheric Modeling, Evaluation Number 12, 1997, also Evaluation Number 11, 1994. 14. C. M. Nelson, T. A. Moore, M. Okumura, and T. K. Minton, J. Chem. Phys. 100, 8055 (1994), and T. A. Moore, M. Okumura, and T. K. Minton, J. Chem. Phys. 107, 3337 (1997). 15. T. J. Lee, C. M. Rohlfing, and J.E. Rice, J. Chem. Phys. 97, 6593 (1992). 16. J. F. Stanton, C. M. L. Rittby, R. J. Bartlett, and D. W. Toohey, J. Phys. Chem.95, 2107 (1991). 17. M. P. McGrath, K. C. Clemitshaw, F. S. Rowland, and W. H. Hehre, Geophys. Res. Lett. 15, 883 (1988). 18. F. Jensen and J. Oddershede, J. Phys. Chem. 94, 2235 (1990). 19. M. Schwell, H.-W. Jochims, B. Wassermann, U. Rockland, R. Flesch, and E. Ruhl, J. Phys. Chem. 100, 10700 (1996). 20. R. N. Schindler, Proceedings of the 4. Statusseminar des Ozonforschungsprogramms, Bonn, 1994, as reported in Ref 19. 21. The OCIO photochemistry at 308 nm was modeled using translational energy distribution and angular distributions obtained from Floyd Davis, unpublished data. 22. R. E. Center and A. Mandi, J. Chem. Phys. 57, 4104 (1972). 68 G. Figure Captions Figure 1. Model prediction of the concentrations of the species in the ClOOCl beam for the initial conditions Pw.g= 200T, Pa20 = l0T, T= 200K, trca = 50 µs, laser spot size 0.09 cm2 , laser energy 0.5 ml/pulse, and ClO + ClO yield 0.2. Figure 2. Schematic of the ClOOCl source cell. Figure 3. Typical TOF spectra of the molecular beam at mlz = 35 (Ci+), 51(ClO+), 67 (CIOO+), 70 (Cl/), 86(Cl2O+), 102 (ClOOCi+), 118(Cl2O/), 121(Cl2O•CJ+), and 137 (Cl2O•ClO+, ClOOCl•Ci+). Figure 4. Source TOF spectra of the ClOOCl parent mass and the higher cluster masses. The solid vertical line indicates the time where the photolysis laser was fired for the photodissociation experiments. Figure 5. Comparison of the ClOOCl parent TOF spectra for different buffer gas compositions. A) Pure helium; B) He+ CF3H; C) He+ SF6. Figure 6. TOF spectra for the ClOOCl and Cl2O parent masses showing the effect of laser power on the ClOOCl concentration and the Cl2O:ClOOCl ratio. A) mlz = 102, 130 ml/pulse; B) mlz = 86, 130 ml/pulse; C) mlz = 102, 160 ml/pulse; D) mlz = 86, 160 ml/pulse. The lines show the region of the pulse where the Cl2O+: ClOOCi+ ratio is the greatest. 69 Figure 7. Newton diagram for argon scattering off of helium. The edge of the Newton circle for Ar scattering corresponds to a laboratory angle of 37.0° . Figure 8. Angular distribution for argon scattered off of helium. The line indicates the laboratory angle (3 7. 0°) corresponding to the edge of the elastic Newton circle. Figure 9. Newton diagram for a pure Cl20 beam scattering off of neon. The edge of the Newton circle for Cl20 corresponds to a laboratory angle of 32.4°. Figure 10. Angular distribution for Cl20 scattered off of neon. The solid line (32.4°) corresponds to the edge of the elastic Newton circle. The dotted line indicates the angle (26.5°) which we took to correspond to the edge of the inelastic Newton circle. Figure 11 . Newton diagram for Cl20 in the background region of the ClOOCl beam scattering off of neon. The edge of the Newton circle for Cl20 scattering corresponds to a laboratory angle of 30.4°. Figure 12. Angular distribution for mlz = 86 (A) and mlz = 51 (B) for Cl20 in the background region of the ClOOCl beam scattered off of neon. The solid line (3 0. 4 °) indicates the edge of the elastic Newton circle; the dotted line indicates the angle (24°) which we took to correspond to the edge of the inelastic Newton circle. 70 Figure 13. Newton diagram for the ClOOCl beam scattering off of neon. Newton circles for CIO (48.8°), OCIO (37.5°), Cl2O (29.6°), CIOOCI (24.8°), and Cl2O3 (21.3°) are shown. Figure 14. Angular distributions for mlz = 86 (A), mlz = 51 (B), mlz = 102 (C), and mlz = 67 (D) from the ClOOCl beam scattering off of neon. The solid lines indicate the edges of the elastic Newton circles as marked, and the dotted lines indicate the angles we took as the edges for the inelastic Newton circles. 71 4.0e+17 2.5e+17 \ \ \ \ \ \ 2.0e+17 Cl 2 U) .!! a \ oE 1.5e+17 \ Cl) C: \ ic ~ \\ CIOOCI \ 8 1.oe+11 I ____\_ ___ __ , - - ~~-~~r--T=:_-~---~-- \.,,,,--/ // // 5.0e+16 I II ""' ' '-.... . . . . ___ / -... ..__ -- --- -- · I,------------ -- - ·· - ·· ---- -- -- ----- -·-··- --------~ I .. - CIO I 1 -- Cl 20 3 OCIO 0.0e+0 0 5 10 15 20 25 30 Residence time, µs Figure 1. 35 40 45 72 308 nm (20 ns pulse) cooling in ) G cooling out ) 180 T seed gas 10 T Cl 2 0 Figure 2. 73 8000 rnlz = 35 8 rnlz = 86 1e+5 8e+4 7000 0 6e+4 6000 5000 0 4e+4 <o 2e+4 4000 4000 3500 3000 2500 2000 1500 1000 80000 60000 g 40000 0 .! 20000 C :, 0 CJ 600 500 400 300 200 100 ~ 16000 'iii C Cl) - C 14000 12000 10000 8000 6000 4000 2000 f oao fD o<c? z- 1.0e+5 8.0e+4 6.0e+4 o~~ o ~o cP8?J'(;ID 0 0 (D~ 4.0e+4 2.0e+4 0.51.01.5 2.0 2.5 3.0 3.5 4.0 4.5 rn/z = 137 0 flight time, ms 0o c8 $00 & 't?# ~8:no~o o 0 0~ Figure 3. 00 f!)~ 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 flight time, ms 140 120 100 80 60 40 20 40 30 20 10 74 6000 5000 mlz =102 mlz =118 mlz =121 4000 3000 0 2000 1000 .! C ::::, 0 0 .. ~ ·-C U) ...,Cl) C 300 r:9.0 250 &°oo 200 cfD 0 0 150 0 100 50 0 160 140 120 100 80 60 40 0 ooo (§)~o CQ9 CcPoCf:r6 oo c9 {P 0 0 0 8 0 40 c9 c9 30 20 0 0 mlz 0 =137 c9coo o c9 Cooc9 00 0 oO ~ O <Y o OO~ ooc9 c9 o 2.5 3.0 0 0.5 Figure 4. 0 0 50 10 O 1.0 1.5 2.0 flight time, ms 0 0 75 7000 pure He buffer gas Cl 20+/CIOOCI+ = 1.1 8:> 6000 0 5000 O 0 4000 0 cf 0 3000 <9 ~ 2000 1000 7000 .!! C ::::, 0 u.. ~ ·Cl) C .! C 6000 5000 4000 3000 2000 1000 7000 He + SF6 buffer gas Cl 20+/CIOOCI+ = 1.1 6000 5000 4000 3000 2000 1000 ~u,.., 0.5 1.0 1.5 2.0 flight time, ms Figure 5. 2.5 3.0 76 6000 laser 130 mJ/pulse laser 160 mJ/pulse cI 20+1c1ooc1+ = 2.6 cI 20+1c1ooc1+ = 1.2 A mlz = 102 5000 4000 .! C ::::, 0 0 .. 0 0 0 5000 ~ 4000 \ 3000 2000 2000 1000 1000 ~ 1.2e+5 ·"'....C 1.0e+5 C mlz = 86 mlz = 86 Cl) - mlz = 102 of\ \ 3000 6000 C 0 1.2e+5 1.0e+5 8.0e+4 8.0e+4 6.0e+4 6.0e+4 0 4.0e+4 0 2.0e+4 2.0e+4 0.0e+0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 O.Oe+O 0.5 1.0 1.5 2.0 2.5 3.0 3.5 flight time, ms Figure 6. 4.0e+4 77 "' ....... E 0 ~ 0 '\"""' >< V . LC) LC) II vHe = 1•75 X 105 cm/s Figure 7. 78 Ar + He Elastic Scattering 2.2 0 2.0 g 1.8 □ 1.6 1.4 -"' 1.2 "E ::::, -eca 8 Q □ Q □ □ CJ ii- 1.0 ·;;; C Jg C □ 0.8 0.6 0.4 0.2 0 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 Lab angle, degrees Figure 8. 79 32.4° X . I'I'- ll 4 vNe = 7.8 X 10 cm/s Figure 9. 80 Cl2O Elastically Scattering off Neon 1.4 Cl2O in He+ SF6 nozzle cold laser off mlz = 86 0 1.2 1.0 en ...., 0 C: ::, 0 0 0.8 >~ en C: 0 0.6 0 ...., Q) 0 C: 0 0 0 0 0 0.4 0: 0 0.2 0 0.0 5 10 15 20 25 Lab angle, degrees Figure 10. 30 35 81 ~ E (.) v 0 ~ r. &/ ~ 7 •7. '5'+ ~ X 7oc5' V. ~~ CX) II 0 ~ -:::,.u 4 VNe = 7.8 X 10 cm/s Figure 11. 82 Cl 20 Elastically Scattering off Neon Background after CIOOCI Pulse 1.2 ,-------- - -- - - - - - - - . - - -- ----. mlz = 86 B 1.0 08 • o □ 6.05 -10.05 ms: 8.55 -9.05 ms l 0 0.6 0.4 8 u, ~ C: ::, 0.2 g . ... 8 .c co 0.0 e ~ ~ ~ mlz = 51 u, C: sC: 1.0 0.8 g 0.6 0.4 g 0.2 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 Lab angle, degrees Figure 12. 83 CIOOCI + Neon Elastic Scattering CIOOCI 37.5° CIO >< LO 0 . 0) II Figure 13. 4 vNe = 7.8 X 10 cm/s 84 1.2 ,_ A 1.0 0.8 0.6 0.4 0.2 ._ 0.0 1.0 ,- ) mlz = 86 Cl 2O CIOOCI 0 p □ b ~ ~ 6 'v o □ ~ range 1: 100 µS range 2: 200 µs range 3: 400 µs range 4: 600 µs 0 ~~ ~ ~x B ~Cl~Ofl mlz = 51 Cl 2O - 0.8 - OCIO 0.6 tn :t:: 0.4 - ::::::, 0.2 ._ C: ..c .... C'O ~ e 0.0 ~ ·u; 1.0 - c6 D C: C: 0.4 ,- n ~ i"'1 .. jol ... - l", mlz = 102 o· !~ 0.2 ._ 1.0 - f, CIOOCI D €' e rJ ~: ,- 0 0.0 1.2 Ct2p3 f} Jg 0.8 ,-06 0.6 CIO §! D ~~~ oaf;J - D mlz = 67 CIOOCI ~ij OCIO 0.8 ,0.6 0.4 ._ e 0.2 - ~ 0.0 ~8 i!3 ~ ~ 15 Figure 14. 20 25 30 A l:d 35 Lab angle, degrees 40 45 50 85 Chapter 4. CIOOCI Photolysis at 248 and 308 nm We investigated the photodissociation of CIOOCI at 248 and 308 nm using photofragment translational spectroscopy, with mass spectrometric detection of products. Time-of-flight and angular distributions of all observed products were collected in order to determine the primary and secondary photolysis channels and their relative yields. At 248 nm, Cl, ClO, and 0 2 products were observed from the photodissociation of ClOOCI (after subtracting out signal arising from Cl2O photolysis). Both ClO and Cl products were produced from multiple competing channels. At 308 nm, the data was more difficult to interpret because of extensive interference from Cl2 and Cl2O byproducts, but there was evidence for both Cl and ClO products from ClOOCl photolysis. The upper limit for the relative yield for ClO products is 31 % at 308 nm and 19% at 248 nm. 86 Introduction ClOOCl photolysis is a key step in the catalytic ozone-depleting cycle1 ClO+ClO+M- ClOOCl+M (1) ClOOCl + hv CIOO+Cl (2) ClOO+M Cl+O2 +M (3) 2[ Cl+ 0 3 ClO + 0 2 ] (4) net: This cycle is thought to be responsible for a majority of Antarctic springtime ozone depletion. Current models assume that only Cl-atom products are formed from ClOOCl photolysis. Photodissociation of ClOOCl to form ClO products would result in a null cycle, which would not deplete ozone. The branching ratio of ClOOCl photolysis into Cl and CIO products is critical in determining the effectiveness of this mechanism for destroying ozone. Recent theoretical work2-3-4.5 has shown that there are three thermodynamically stable isomers of Cl2O2, with dichlorine peroxide (ClOOCl) being the most stable, and chloryl chloride (C1ClO2) and chlorine chlorite (ClOClO) ~ 1 and~ 10 kcal/mol less stable.2 The ClOOCl isomer is thought to be the main dimerization product in flow tube experiments, although there is some evidence for formation of the C1ClO2 isomer if the walls of the flow tube are coated with water ice. 6 The ground state structural parameters7 of the ClOOCl molecule are r 00 = 1.42 A, ra0 = 1.70 A, Laoo = 110.1 °, and dihedral angle= 81.0°, as shown in Figure 1. 87 Possible photolysis channels for ClOOCl are8 ClOOCl + hv ... ClO + ClO Afio0 = 17.1 kcal/mol (5) ... ClOO + Cl = 21.2 kcal/mol (6) ... 2Cl + 0 2(3~ -g) = 25. 9 kcal/mol (7) ... 2Cl + Oi( 1 Ag) = 48.4 kcal/mol (8) ... Cl +O+ ClO = 80.5 kcal/mol (9) ... Cl2 + 0 2 = -31.2 kcal7mol. (10) Two experiments have found that the quantum yield for Cl atom production from ClOOCl photolysis is unity or greater, but large uncertainties remain in both measurements. Cox and Hayman9 examined steady-state photolysis of Cl20 at 254 nm and modeled the disappearance of Cl 2 0 and the appearance of ClOOCI in a static photolysis cell held at 203-233K. They reported a yield of 3.8 ± 1.6, which is unphysically high. In a more direct study, Molina et al. 10 reported a Cl atom quantum yield of 1.03 ± 0.12 from ClOOCl photolysis at 308 nm in a flow cell. They used resonance fluorescence to detect Cl atoms and calibrated the Cl signals relative to Cl2 photolysis. No ClO was inferred from titration with NO. Their value for the quantum yield depends on the ratio of the ClOOCl absorption cross section at 245 and 308 nm, for which they used a value of 22, which is larger than more recent literature values. If the cross section ratio of 13 from the 1997 JPL evaluation of photochemical data11 were used, the recalculated Cl atom yield would be reduced to 0.6. The absorption spectrum has also been recently measured by Huder and DeMore, 12 who observed a smaller cross section in the long-wavelength tail. Using their 245/308 nm ratio of 18.7, Molina's quantum yield for Cl would become 0.88. 88 A recent experirnent13 by Schmidt, Benter, and Schindler examined ClOOCl photolysis in the wavelength range 240-260 run as well as at 308 run. They photolyzed a ClOOCl beam produced by OClO photolysis and recombination of ClO in the ion source of a reflectron equipped with a time-of-flight mass analyzer. Products were ionized by Resonant Enhanced Multi-Photon Ionization (REMPI). Cl products were detected, but no ClO(v = 0-5) products were observed. ClO (v>5) states can not be detected via REMPI. They reported a quantum yield for the Cl + ClOO channel of <l> = 0.65 ± 0.15 at 25ff nm, and an average Cl translational energy of 0.42 ± 0.05 eV. We have examined the photodissociation ofClOOCl at 248 and 308 run in a molecular beam experiment with mass spectrometric detection. Direct detection of both Cl and CIO photoproducts provides a cross-section independent means of determining the branching ratio between Cl and ClO products. The relative detection efficiencies for Cl and ClO were calibrated by performing Cl20 photolysis experiments at 308 run, where the only photolysis products are Cl and ClO in a 1: 1 ratio. Experiment The experiments were performed on a crossed molecular beams apparatus 14 modified for laser photodissociation.15' 16' 17 The ClOOCl beam, described in the previous chapter, was crossed at right angles with the focused UV radiation from a pulsed excimer laser in the main vacuum chamber. The photodissociation products were detected off the beam axis by a mass spectrometer that is rotatable in the plane of the laser and molecular beams. Only photofragments which scatter out of the molecular beam are detected; undissociated 89 molecules do not contribute to the scattered signal. Product time-of-flight (TOF) distributions were collected at various detector angles by monitoring the signal from the mass spectrometer at a single mass as a function of time and accumulating the signal over many laser pulses at each angle. Photofragments traveled 33.3 or 33.5 cm from the intersection of the two beams to the Brink electron bombardment ionizer. The shape and position of the electron cloud for this type of ionizer depend on the ionizer conditions used. Two ionizer settings were used to obtain data. The initial setting, with which almost all of the 308 nm data and the preliminary 248 run data were taken, was optimized on axis with an argon beam using the small detector aperture. The other setting was obtained by optimizing the mlz = 51 signal from Cl20 photolysis using the large aperture for off-axis detection. The electron energies were nominally 80 and 160 eV for the first and second settings, respectively. The mass resolution of the mass spectrometer was set so that ~m ~ 2 amu, in order to increase the signal. Pulses from the Daly-type ion counter were sent to a LeCroy 8818 discriminator and then recorded by an EG&G Ortec ACE-MCS multichannel scaler with a dwell time of 2 µs per channel. We determined the velocity distributions of the parent ClOOCl molecules by several methods. The beam was chopped with a slotted disk which was synchronized with the source laser pulse and then delayed to sample the ClOOCl pulse. TOF distributions were collected at various ion fragment masses and then converted to a velocity distribution N(v) which was fit with an assumed functional form, N(v) cc v2 exp[-(v/a. - S)2], where N(v) is the number density of molecules with speeds in the range v to v + dv, a is the width of the speed distribution, and Sis the ratio of the stream velocity V8 to the width a .18 The TOF wheel was 90 only used for beam velocity measurements are was removed for source optimization and the photodissociation experiments. Since the beam velocity depended on the section of the ClOOCl pulse sampled and the ClOOCl beam conditions, the actual beam velocity was determined independently for each experiment by examining the diatomic Cl2 photolysis signal at mlz = 35 for the 308 nm experiments and the diatomic ClO photolysis signal at mlz = 35 for the 248 nm experiments. The beam velocities obtained in this manner agreed very well with each other, with an average beam velocity of9.1 x 104 emfs (a= 0.83-0.86, S = 10.610.9). The measured flight time was the sum of the time for the photofragment to travel to the ionizer and the ion flight time from the ionizer to the detector. The ion flight times were determined to be 4.68(m/z) 112 µsand 3.2(m/z) 112 µs for the first and second ionizer settings, respectively. Corrections for ion flight time were made in the fitting routine, so the abscissa of all TOF spectra shown are total flight times. The ultraviolet radiation source for the ClOOCl photolysis was a Lambda Physik EMG 150 excimer laser operating in oscillator only mode, which was used to provide 248 nm (KrF) or 308 nm (XeCl) radiation. Typical pulse energies used were ~ 70 mJ/pulse at 248 nm and 20-100 ml/pulse at 308 nm at a repetition rate of 40 Hz. The laser light was partially preferentially polarized (~56%) in the plane perpendicular to the scattering plane and parallel to the molecular beam. The rectangular-shaped laser beam was focused with a combination of a CaF 2 cylindrical lens (focal length/= +600 mm) and a UV grade fused silica planoconvex lens (f= +300 mm) to a~ 1.5 mm wide x 3mm high spot at the interaction region. At 308 run, the delay between the firing of the source laser and the photolysis laser 91 was set by maximizing the photodissociation signal at mlz = 35 from Cl2 . Since the maximum Cl2O depletion and the Cl2 and ClOOCl peak positions in the source TOF spectra coincide, maximizing the Cl signal from Cl2 photolysis should coincide with the maximum Cl2O depletion and thus the highest ClOOCl:Cl2O ratio. At 248 nm, the photolysis laser delay was based on the beam shape and timing, and checked for consistency by comparing with delays set for similar beam profiles at 308 run. Since the maximum ClOOCl signal spanned several hundred microseconds, the delay only had to be within ± 100 µs of the actual peak center. TOF distributions of the photoproducts from photolysis at both wavelengths were collected at mlz = 16, 32, 35, 51, 67, and 70, corresponding to O+, 0/, Cl+, ClO+, ClOO+, and Cl/, respectively. These distributions were recorded at detector angles from 15° to 50° and were averaged for 100,000 to 2,000,000 laser pulses, depending on signal level. No significant signal was observed at mlz = 16 (O+), 67 (ClOO+), or 70 (Cl/) for either wavelength, or at mlz = 32 (02+) at 308 run. The mlz = 32 signal-to-noise ratio was low because of the large background 0 2 signal, and the lower ClOOCl absorption cross section at 308 run made it impossible to observe any 0 2+ signal. Laboratory angular distributions for mlz = 35 and 51 were recorded at detector angles 15°, 30°, 40°, and 50° in both the 248 and 308 run experiments, and for mlz = 32 over the same angular range in the 248 run experiment. In order to eliminate systematic error resulting from possible long-term drifts, the detector was rotated back and forth over this angular range, and TOF distributions were collected for 50,000 or 100,000 laser pulses at each increment. The resulting TOF distributions at each angle were averaged for a total of 330,000 pulses for mlz = 32 and 250,000 pulses for mlz = 35 and 51 at 248 nm, and 500,000 92 pulses at mlz = 51 and 600,000 pulses at mlz = 3 5 at 308 nm. In the 248 nm experiment, we continued recording 50,000 pulse scans at a detector angle of 30° for mlz = 35 and 51 for a total of 500,000 pulses at each mass. In the 308 nm experiment, 20 sets of50,000 pulse scans were obtained separately from the angular distribution experiments at 15° and 40° for a total of 1,000,000 pulses at each angle, and 5 sets of 100,000 pulse scans were obtained at 30° for a total of 500,000 pulses. These back-and-forth scans were obtained in order to accurately measure the relative signal intensities for determination of the branching ratio between the Cl and ClO products. The peak counts per pulse at mlz = 32, 35, and 51 at 248 nm for a detector angle of30° were 0.0048, 0.0137, and 0.0035 cts/pulse, respectively. At 308 nm, the counts per pulse at the peak channel at 30° were mlz = 35, 0.034 cts/pulse and mlz = 51, 0.002 cts/pulse. Other species in the molecular beam besides ClOOCl contributed to the photolysis signal. These included ClO and Cl20 at 248 nm and Cl2, ClO, and Cl20 at 308 nm. Photolysis experiments of Cl2 and Cl20 were performed at both wavelengths, although the Cl2 photolysis signal was negligible at 248 nm. A molecular beam of Cl2 was formed by flowing the He/SF6 buffer gas mixture over liquid Cl2 at -95°C and expanding through the nozzle, which was held at the same temperature as in the ClOOCl photolysis experiments. Data was collected at mlz = 35 at 10°, 20°, 30°, 40°, and 50° for 5 scans of20,000 pulses each, rotating the detector back and forth among the angles. The analysis of the Cl2 photolysis data provided a check on the input parameters for our fitting routine. Cl20 photolysis experiments were performed before or after each ClOOCl data taking session. The Cl20 beam was formed by turning off the source laser used to make the ClOOCl. The results 93 of the Cl2O photolysis experiments have been published elsewhere.19•20 The relative detector sensitivity for detecting Cl and ClO was required for determination of the CVClO branching ratio. The Cl2O photolysis experiment at 308 run also allowed for calibration of the relative detector sensitivities for Cl and ClO, since the only photolysis products at this wavelength are Cl and ClO in a 1: 1 ratio. Results and Analysis Analysis Ovenriew A forward convolution technique was utilized in the analysis of TOF and angular distributions. For each two-body dissociation channel, trial center-of-mass (c.m.) translational energy P(ET) and angular w(8) distributions were transformed to the lab frame of reference to obtain a predicted number density TOF distribution N(t). The predicted N(t) distributions were compared with the experimental data, and the P(ET) and w(0) distributions were adjusted iteratively until the predicted TOF distributions agreed with the observed distributions at all angles. An extension of this procedure21 was needed to model three-body dissociation. Spontaneous three-body dissociation can occur via a concerted or a stepwise mechanism. We will use the terms defined by Maul and Gericke, 22 based on terms originally developed by Dewar23 for the three body decay of a molecule ABC. A concerted decay takes place in a single kinetic event, whereas a sequential decay is characterized by two independent dissociation steps in which the decay of the primary fragment AB is not influenced by forces from the first fragment C. The kinetic event is taken to be a rotational period. Thus, a three 94 body dissociation is called sequential if the time span 6.t between the first and second bond cleavage steps is greater than a rotational period 'trot of the primary fragment AB. As 6.t becomes smaller than 'trot, the two decay processes are no longer independent of each other, and the dissociation is called concerted. As 6.t approaches zero, all bond breaking processes take place simultaneously, and the dissociation occurs by the synchronous concerted mechanism, as opposed to the asynchronous concerted mechanism in which O< 6.t ~ 'trot· The sequential dissociation of Cl00Cl Cl00Cl + hv - Cl + CI00 - Cl + Cl + 0 2 was modeled by separating the dissociation into two independent center-of-mass reference frames in which linear momentum is conserved. Thus, a primary dissociation step Cl + Cl00 was modelled, followed by secondary dissociation of the Cl00 to Cl+ 0 2 . The secondary angular distributionP(Q) is referenced relative to the primary c.m. dissociation direction (i.e. along the Cl-0 bond, assuming the geometry doesn't change from the ground state) and not the electric vector of the laser. Since the rotational excitation of the intermediate Cl00 may be broad, the secondary angular distribution is likely to be highly averaged. If the secondary angular distribution is asymmetric about Q = 90°, then the dissociation occurs within a rotational period, and is considered asynchronous concerted rather than stepwise. 24 The synchronous concerted dissociation channel Cl00Cl + hv - Cl + 0 2 + Cl was modeled with a separate analysis routine. 25 An overall translational energy distribution for the total energy released by all three fragments was determined by a forward convolution technique. The angles between the relative c.m. velocity vectors were determined for a given 95 dissociation geometry by assuming that dissociation occurs along the Cl -0 bonds, and angle width parameters were chosen which determined the width of the assumed Gaussian distribution of geometries about the most probable dissociation geometry. Given the P(ET) and the angles between the c.m. velocity vectors, conservation of linear momentum and energy were applied to give a predicted TOF spectra. The parameters were adjusted until the predicted TOF matched the observed 0 2 and Cl data at all angles. Stimulated secondary dissociation, in which a primary photofragment absorbs a second photon during the laser pulse and dissociates, was not modeled in this work. The laser power was kept low so that secondary photodissociation was minimized. Relative Scaling of Primary and Secondazy Distributions The scaling for the secondary products relative to the primary products when both only contributed signal to a single mass was determined by modeling a single primary channel and the corresponding secondary dissociation in the analysis routine. The ratio of the primary to secondary scaling factors was then set to be I: 1, in which case the analysis routine rigorously enforces the relative scaling. The ratio between the integrals of the primary and secondary products at each angle obtained by this method was then used to determine the secondary contribution from each channel when adding in all of the other channels that contribute signal to that mass. Where the primary or secondary channel contributes signal to multiple masses, but at least one of the masses has a contribution from both primary and secondary dissociations, the scaling for a given channel was more difficult to determine but a good approximation could be made. For instance, for the ClOOCl + hv ... ClO(X) + ClO(A) ... ClO(X) + Cl + 0(3P) channel, the scaling for the secondary Cl products was first 96 approximated by modeling only the primary and secondary dissociations from this channel at mlz = 35. Because 1) 13% of the ClO (X) fragments in the ionizer to mlz = 35 and 2) the relative detector efficiency for ClO is approximately equal to that for Cl, the ratio of the primary to secondary scaling factor was set to 0.13: 1. Relative Yields The total c.m. flux distribution for a given channel was assumed to be proportional to the product of the P(ET) and w(8) distributions, which were determined for each channel by simultaneously fitting of all the laboratory TOF distributions for all observed channels. The relative proportionality of the product fluxes for two channels which contributed to the signal at mlz = 3 5 was determined by fitting both primary components of the mlz = 3 5 TOF distributions. The relative detection efficiencies of ClO and Cl at mlz = 3 5 were calibrated using information from an analogous study of Cl20 photolysis. The flux of ClO for a given channel detected at mlz = 35 is equal to the product of the total flux of ClO (la0) for that channel, the total ionization cross section of ClO ( a ao), and the fraction of ClO that fragments to mlz = 35 in the ionizer lfa0 (35)]: 1 ~oC 35 ) = 1ao • 0 ao • faoC3S) (1) When two photolysis channels can lead either to ClO or Cl photoproducts, the ratio of the relative yield for the two channels is given by: /~0 (35) /~(35) . act . fa(3S) 0 ao faoC35) (2) 97 wherefcJ...35) = 1. For Cl20 photolysis at 308 run (CI 20 + hv - CIO + Cl), this ratio must be 1, and the relative ionization cross sections for Cl and CIO can be obtained. The first factor on the right-hand side of this equation is determined from the primary c.m. flux distribution scaling factors obtained by simultaneously fitting all of the lab TOF distributions at mlz = 35. The scaling factors between the primary flux distributions remained constant for all angles. The third factor in eq. 2 depends on the fragmentation pattern of CIO, which depends on the internal energy of the ClO. Eq. (2) implicitly assumes that Oao does not vary with internal excitation of ClO for a given ionizer condition. Because we used the Cl20 photolysis experiment for calibration, no assumptions about absolute detection efficiencies or the massdependent transmission of the quadrupole mass filter were required. 248nm Overview of Results Photolysis signal was observed at CIO+, 0 2+, and Ci+ masses. The signal at mlz = 51 (see figure 2) comprised a fast bimodal peak and a slow broad peak. The signal at mlz = 32, shown in figure 3, was sharp and fast, with the leading edge arriving at ~ 100 µs for a detector angle of 15 ° . This arrival time was faster than any of the observed signal at the other two masses. The largest amount of signal was observed at mlz = 35 (see figure 4), where the data consisted of a fast peak with multiple unresolved components and a slow sharp peak and a slow broad peak that were prominent at smaller angles. Clo Contribution The sharp peak in the mlz = 35 TOF distribution arriving at z 235 µs that was prominent at smaller detector angles was assigned to Cl products from photolysis of diatomic 98 ClO radicals which remained in the beam. At 248 run, ClO photolyzes to produce Cl(2P312) + O(1D) products. 26 The bond dissociation energy27 for this channel is 108.78 kcaVmol. Vibrations are not efficiently cooled in a supersonic expansion, so the vibrational temperature of the CIO will be ~ 200K, at which temperature greater than 99% of the ClO is in the ground vibrational state. The 248 nm photon energy is 115.01 kcaVmol. Since ClO is a diatomic, all of the energy left after breaking the ClO bond must appear as fragment translational energy. Thus, the translational energy distribution used to fit the CIO consisted of a spike ~ 0.1 kcal/mot wide at 6.22 kcal/mot and another spike 0.91 kcaVmol higher in energy and 19.9 times weaker corresponding to dissociation of spin-orbit excited ClO, assuming a thermal distribution at 200 K for the excited CIO. The beam velocity was adjusted until the fit to the ClO peak matched the observed data. The resulting beam velocity was then fixed and used in all subsequent analysis for that experimental run. The angular distribution for this channel was fit by an anisotropy parameter of2.0. Cl2O Contribution Because the beam contains Cl20 as well as ClOOCl, some Cl20 contribution to the mlz = 51 data is expected. The translational energy and angular distribution for Cl20 photolysis were obtained from the Cl20 photolysis experiments. Ifwe assumed that the mlz = 51 signal from the ClOOCl beam photolysis arose solely from Cl20 for a given angle, the other angles could be scaled to the Cl20 angular distribution as shown in Figure 5. The leading edge of the resulting predicted TOF spectra rose faster than the observed data, especially at the smaller angles, and fell off too quickly with increasing angle. The leading edge discrepancy is evident in Figure 6, which compares the best Cl20 fit to the data at 0LAB 99 = 30°. Furthennore, the two peaks in the TOF spectrum do not correspond well to the three components in the predicted Cl20 spectrum. Since the beam velocity is constrained by the ClO bond energy (which is well known) in order to fit the ClO photolysis signal at mlz = 35, the discrepancy with the rising edge can not be explained by an incorrect beam velocity. Thus, it seems reasonable to conclude that the signal at mlz = 51 does not all arise from Cl20 photolysis. The maximum reasonable contribution to the mlz = 51 signal from Cl20 photolysis products was determined by scaling the Cl20 contribution as high as possible without causing the rising edge of the predicted TOF to become faster than the observed data. This scaling was carried out at 15 ° ( see Figure 7), where the rising edge of the TOF is the most sensitive, and the remaining angles were scaled according to the Cl20 angular distribution. The Cl20 contribution at mlz = 35, including both Cl photofragments and ClO that fragmented in the ionizer to form Cl+, was obtained based on the Cl20 contribution determined at mlz = 51 and scaled based on the pure Cl20 experiments. From the Cl20 experiments, the extent of ClO fragmentation in the ionizer was observed to depend approximately linearly on the ClO internal energy. ClO + ClO Channels We assigned the remaining fast signal at mlz = 5 l to ClO from ClOOCl photolysis. The resulting P(Er) distribution, shown in Figure 8, was bimodal. The w(0) distribution was fit with an anisotropy parameter of P= 0.5. The data is clearly bimodal, with the two peaks becoming better resolved as more scans were averaged. The dip between the two peaks occurs at approximately the same time as the prominent middle component in the Cl20 100 distribution, and reducing the Cl2O contribution could not eliminate the bimodality of the ClO + ClO translational energy distribution. The average energy released in translation was (.ET) = 26 kcal/mol, or~ 26% of E:,;vail, indicating that most of the energy (<.Em1) = 73 .5 kcal/mol) tended to go into internal excitation of the ClO products. The ClO bond dissociation energy is 63.4 kcal/mol and the maximum internal energy in the ClO products is £int= 93.5 kcal/mol, so the energy would have to be unequally distributed between the two ClO fragments for spontaneous ClO bond breakage to occur. The signal at mlz = 51 is unlikely to have any contribution from a Clo+crack from ClOO because 1) the Cl-O bond in ClOO is very weak, and any ClOO from the primary dissociation is unlikely to survive intact, and 2) the ClOO equilibrium structure resembles a Cl loosely associated with an 0 2 molecule. Slow signal, which was more prominent at the smaller angles, was also observed at mlz = 51. We assigned this slow ClO signal to the channel ClOOCl + hv .... ClO(X) + ClO(A) .... ClO(X) + Cl + O(3P). The P(Ec) distribution fell more rapidly with increasing energy than calculated for statistical RRKM dissociation to ground state products. The ClO A 2IT state is 90.5 kcal/mol above the ground state, leaving 9 kcal/mol available to be partitioned into translational and internal energy. If one of the ClO fragments was formed in the A 2IT state (and the other in the ground state X 2II) then the resulting statistical translational energy distribution ( see Figure 6) predicted TOF spectra which matched the observed data. The ClO A 2II state predissociates rapidly to give Cl + O(3P), so the slow signal observed at mlz = 51 consists only of the ground state counterfragment. The angular distribution for this slow peak was fit by P= 0, indicating that the excited ClOOCl may live longer than a rotational period. 101 The slow signal was not present when the source laser was off (that is, for pure Cl2O photolysis); thus, it was correlated with the ClOOCl beam. The signal could also arise from photodissociation of clusters containing ClOOCl. However, when looking at the ClOOCl beam TOF spectra, the ClOOCl-correlated cluster peaks rose much more slowly than the ClOOCl peak, so at the photolysis laser delay we chose (sampling the ClOOCl peak), the cluster concentration should be small. The contribution at mlz = 35 from the ClO products from ClOOCl photolysis fragmenting in the ionizer was determined by assuming that the internal energy in the two ClO products was distributed evenly and calculating the extent of fragmentation based on the average internal energy of the ClO. The percent of signal fragmenting to mlz = 35 vs. internal energy was plotted for the three Cl2O components, and the ClOOCl fragmentation values were inferred from this plot. The average internal energy for the fast peak was 3 7 kcal/mo I, resulting in 32% of the ClO fragmenting to mlz = 35, and the average internal energy of the slow peak was ~ 3 kcal/mol, which was extrapolated to give 13% of the slow ClO fragmenting to Ci+. Spontaneous secondary dissociation of the ClO A211 state was also modeled. The ClO A state energy is 90.5 kcal/mol above the ground state, and the ClO bond dissociation energy is 63 .4 kcal/mo1, so that the minimum translational energy of the secondary Cl products will be 27.1 kcal/mol. Up to 9 kcal/mol from the primary dissociation can be partitioned into the ClO(A) internal energy, which would go into translational energy of the secondary products; the statistical translational energy distribution ranging from 27.1 to 3 6. I kcal/mol shown in Figure 8 was used to model the secondary dissociation. The secondary angular distribution was assumed to be isotropic. The signal from the secondary 102 dissociation of ClO (A) contributed to the slow shoulder of the fast peak in the mlz = 3 5 TOF spectra. Concerted Cl + 0 2 + Cl Channel A sharp, fast signal was obseived at mlz = 32, shown in figure 3. Because this signal was faster than the observed er signal, we assigned it to 0 2 recoiling from both Cl atoms from the concerted synchronous dissociation of ClOOCl to Cl + 0 2 + Cl. The translational energy distribution, dissociation geometry, and the width of the Gaussian distribution of angles about the equilibrium geometry were determined by iterating between the 0 2 peak and the corresponding Cl peak to achieve the best fit. For a concerted synchronous dissociation, the two Cl atoms will be identical and give rise to a single peak. The predicted Cl peak position was very sensitive to the assumed dissociation geometry. Since the dissociation geometry is not known, we started with the ground state geometry and then adjusted the dihedral angle. The Cl position was predicted to be around the main peak in the mlz = 3 5 TOF spectra, and the best fit was obtained for a dissociation geometry with a smaller dihedral angle of 65 ° (instead of 81 °) and angle width parameters of 4 ° . The translational energy distribution (Fig. 9) indicated that on average 65 kcal/mol, or 72% of Eavm1 ,went into translation, leaving an average of25.7 kcal/mol in internal energy, most of which must go into the 0 2. The maximum translational energy for this dissociation was 87 kcal/mol, leaving ~4 kcal/mol in internal energy of the products. The angular distribution for the mlz = 32 peak indicated a perpendicular transition with an anisotropy parameter of - 1.0 s ~ s -0.5 if the 0 2 signal was modeled as arising from a two-body process; the Cl angular distribution corresponded to O.0 s ~ s O.4. 103 Because the 0 2 and Cl products are detected at different masses with no contributions from both products at the same mass, the relative scaling of the Cl to the 0 2 signal could not be done exactly; however, an estimate can be made. Two Cl atoms are produced for each 0 2, and the relative ioniz.ation cross section for Cl is about 1.3 times that for 0 2 ( 0 02 = 2. 7 x 10- 15 cm 2 for an electron energy of 100 eV, a a= 3.5 x 10- 15 cm2 at 75 eV). 28•29 The Cl signal should thus be about a factor of 2. 6 larger than the 0 2 signal. The integrals of the 0 2 peaks were on the order of 10,000 counts, while the integrals of the corresponding Cl products were around 42,000 if scaled to the same number of shots. However, the angular distribution of the Cl products is not the same as that of the 02, and a direct comparison of integrals can only provide a rough estimate of the relative amounts of Cl and 0 2. Ifwe were to instead assign the fast 0 2 to products from the secondary dissociation of ClOO following primary dissociation of ClOOCl to Cl + Cl00, no more than 14 kcal/moI (15% of Eavmi) could be partitioned into translation in the primary dissociation step, with the remainder going into ClOO internal excitation. The ClOO internal excitation must then all become translational energy in the secondary dissociation to Cl + 0 2, with very little or no energy remaining in the 0 2 molecule. Such a dissociation mechanism seems very unlikely. Furthermore, the existence of a peak at mlz = 35 that corresponds to the Cl counterfragment from the concerted dissociation lends credence to our assignment of the 0 2 peak as arising from the concerted dissociation mechanism. No other peaks were observed at mlz = 32, but the signal-to-noise ratio was very low due to the high 0 2 background, so that the slower, broader signal from the secondary dissociation of any ClOO was lost in the noise. Another possibility is that the 0 2 arises from the Cl2 + 0 2 channel. However, no peak 104 was observed at mlz = 70, and the average background was 4.6 times lower than the background at mlz = 32. The ionization cross section27 for Cl2 is Oru = 7.0 x 10- 16 cm2, a factor of2.6 times the 0 2 ionization cross section, and the fragmentation pattern for Cl2 was measured to be Ci+: Cl2+ = 1. 0 : 1. 9 for similar ionizer conditions. Therefore, if the 0 2 were from the Cl2 + 0 2 channel, we should have observed the Cl2 counterfragment at mlz = 70. Cl + CIOO Channels The remaining signal at mlz = 35 consisted of a fast shoulder, a slow shoulder, and a very slow peak. The fast shoulder was assigned to primary Cl products from the Cl + ClOO channel and fit with the translational energy distribution shown in figure 10. The angular distribution was fit with an anisotropy parameter of0.6. The average energy in translation for this channel was (£,T) = 25 kcal/mol, or~ 26 % of £avail, leaving on average 66 kcal/mol of internal energy primarily in the ClOO fragment. The minimum internal energy in the ClOO fragment was 37.4 kcal/mol. This is significantly greater than the bond dissociation energy ofCIOO, which is only 4.7 kcal/mol; thus, the ClOO must undergo spontaneous secondary dissociation if it is formed in the ground state. Calculations30 on the four lowest energy doublet states of both A' and A" symmetry indicate that only the ground state is bound; all excited doublet states investigated were repulsive with respect to dissociation to Cl + 0 2 . Table I shows the relative energies of the eight calculated doublet states ofClOO. Even if the ClOO were formed in an excited electronic state, dissociation would still occur. Thus, the lack of signal at mlz = 67 (CIOO+) is not surprising. We modeled the spontaneous secondary dissociation ofClOO assuming formation of 105 ground state Cl+ 0 2 products. The average energy in translation was 6.4 kcal/mol, with a maximum translational energy of 19 kcal/mo I. The secondary angular distribution P( Q) was estimated by assuming that the ClOO dissociated promptly from the ground state ClOOCl geometry. The reference direction for the secondary angular distribution is the c.m. primary dissociation direction, i.e. the first Cl-O bond axis. In the ground state geometry the vectors along the two Cl-O bonds make an angle of 89°, so the secondary angular distribution was assumed to be sideways-peaked about 89° . A sideways peaked angular distribution gave the best fit to the slow shoulder of the fast peak; however, the fact that this signal consisted of multiple unresolved components precluded us from unambiguously determining the secondary angular distribution. Even if the true angular distribution is sideways-peaked, because it is symmetric about Q = 90° we can not rule out the possibility that the CIOO* lives longer than a rotational period. Table 1. The eight lowest doublet states calculated for ClOO. Taken from Ref 29. state relative energy (kcal/mol) correlates to ground or excited state products 4 2A' 65.6 e.s. 4 2A" 61.3 e.s. 3 2A' 58 .9 g.s. 3 2A" 52.0 e.s. 2 2A" 47.3 g.s. 2 2A' 45.0 g.s. 1 2A' 36.1 g.s. 1 2A" 0.0 106 Slow Signal at mlz = 3 5 The mlz = 35 TOF distributions also exhibited a broad, slow component that was more prominent at the smaller angles. This signal could arise from several sources: 1) dissociation of CIOOCI to form Cl + CIOO* which then dissociates to Cl + 0 2*, 2) dissociation of CIOOCI- or Cl 2-containing clusters, 3) secondary photodissociation of ClO photoproducts or other secondary processes, or 4) a combination of the above. Some of the low-lying excited states29 predicted for CIOO correlate to formation of electronically excited 0 2 * products (see Table I). If we modeled the slow signal at mlz = 35 as a statistical dissociation to Cl + CIOO* followed by a statistical dissociation of CIOO* to Cl + 0 2 *, a satisfactory fit at all angles could not be achieved unless the translational energy fell more rapidly than the predicted RRKM P(Er) distribution, as shown in figure 11. This was so even for the highest calculated excited state of ClOO* that correlates to excited state products, the 4 2 A' state at 65.6 kcal/mot above the ground state, and for subsequent dissociation of the ClOO* to Cl + 0 2 (11:g) 39.0 kcal/mol above the ground state. The primary angular distribution was fit with p = 0, and the secondary dissociation was fit assuming an isotropic secondary angular distribution; thus, both dissociations may occur after a rotational period. As with the slow signal at mlz = 51, the slow signal at mlz = 35 is not observed in the pure Cl2O photolysis experiments, indicating that if the signal arises from cluster dissociation, the clusters must be correlated with the CIOOCI beam. The slow signal is not likely to arise exclusively from secondary photolysis of primary ClO, again because the signal was not observed for Cl2O photolysis at the same laser power. Some slow signal was observed at mlz = 3 5 for pure Cl2O photolysis. This slow signal was small compared to the fast peak and 107 broader than the slow signal observed in the e1ooe1 photolysis. For e120, the slow signal was modeled as cluster dissociation, and was included when modeling the e120 contribution to the elOOel photolysis signal. 308nm Overview of Results Photolysis signal was observed for Clo+ and er fragments (Figure 12). A single Clo+ peak was observed, with no discernable "slow" signal. The er TOF spectrum consisted of three peaks, dominated by a sharp, fast peak which had faster and slower weak shoulders followed by two slower peaks. The signal-to-noise (SIN) ratio was worse than at 248 run because of the decreased absorption cross section. Contributions from other components in the beam We observed contributions to the photolysis signal from elO, e12, Cl20, and elOOCl. Of these, dissociation of e12 (the major component in the beam) dominated the er TOF spectrum. A separate e12 photolysis experiment was performed to obtain the P(Er) and P needed to fit the Cl2 contribution to the photolysis signal. The e12 experiment also allowed us to test our input file for the analysis routine, fixing the length of the ionizer and providing a double check on the other parameters. Figure 13 shows the e12 photolysis data at mlz = 3 5. We modeled the signal with a P(ET) that reflected dissociation to 94% CI(2P 3n} + e1(2P 312), 2% el(2P:vi) + el(2Pvi), and 4% el(2P112) + el(2P11i). The angular distribution was fit with J3 = -1 for the el{2P:vi) + el(2P312) channel; the other two channels were too small to be able to determine p values. Matsumi et a/. 31 ' 32 observed a lower yield of0.0095 for Cl{2P 1n) from 108 Cl 2 photolysis at 308 run, and anisotropy parameters P = -1 for Cl(2P3n) and p = -0. 7 for Cl(2P 112 ). Samartzis et a/.33 observed a slightly higher yield of Cl* than Matsurni et al. for photolysis at 351 run, 0.08 as opposed to 0.016. The magnitudes of the CIO and Cl2O contributions could not be determined from the TOF data alone. We therefore scaled the ClO (dot-dot-dashed lines, Fig. 12) and Cl2O (dashed line, Fig. 12) contributions to the Ci+ signal relative to the Cl2 contribution using relative abundances determined by mass spectrometer measurements of the source, as discussed in Chapter 3, using the absorption cross sections at 308 nm. Knowledge of the ClOOCl absorption cross sections or the amount of signal arising from ClOOCl was not needed for this calculation. The ClO contribution was modeled as a combination of photolysis to Cl(2P 112) + O(3P) and Cl(2P312) + O(3P) and was observed in the middle peak of the Cl+ TOF spectrum. The anisotropy parameter was assumed to be P= 0. The Cl2 O P(ET) and P as well as the relative scaling for Cl and ClO photofragments were obtained from separate Cl2O photolysis experiments. With this information we could model the Cl2O contribution for both Cl+ and Clo+ products. ClOOCl photolysis The only contributions to the mlz = 51 signal were from Cl 2O and CIOOCI. Once the Cl2 O contribution was determined, the remaining signal was assigned to the CIO + CIO channel from ClOOCl photolysis (dotted line, Fig. 12) and fit with the translational energy distribution shown in Fig. 14. The angular distribution was fit with an anisotropy parameter of p = 2.0. The average energy in translation was 27 kcaVmol, or~ 35% ofEavail, indicating that the ClO fragments were highly vibrationally excited, with /,£in1) = 50 kcal/mol distributed 109 between the two fragments. The maximum amount of energy in Cl0 internal energy was 65.3 kcal/mol, which is barely enough energy for spontaneous dissociation of the Cl0 if all the internal energy were partitioned into one of the Cl0 fragments. Thus, spontaneous secondary dissociation of Cl0 is unlikely. We also modeled the Cl0 signal as only arising from Cl0 + Cl0 (no Cl20 contribution) in order to determine the upper limit for the Cl0 branching ratio from Cl00CI photolysis. A fast shoulder, slow shoulder, and a very slow peak remained after taking into account er signal from the Cl2, Cl20, and CI0 photolysis contributions and from CI0 fragmenting in the ionizer. The fast shoulder was assigned to primary Cl from the Cl + Cl00 channel and fit with the P(ET) shown in Figure 15. The angular distribution was modeled with an anisotropy parameter of P= 1.5. The average energy in translation was I.ET)= 27 kcal/mol (37% ofEav;,;.J, leaving on average 46 kcal/mol in the Cl00 fragment, well above its dissociation energy (D 0 = 4.7 kcal/mol). As at 248 nm, no Cl02+ signal was observed. The secondary dissociation of Cl00 to Cl + 0 2 was modeled with an impulsive P(Er) and an isotropic angular distribution. The secondary Cl+ signal appeared in the slow shoulder of the fast peak. Because the Cl2 photolysis signal obscured the signal from Cl00Cl, and because the signal-to-noise ratio was low, and because no 0 2 signal was observed, it was impossible to determine the dynamics ( sequential or concerted) of the dissociation. Very slow Cl+ signal remained, which we modeled as arising from a statistical dissociation to Cl + CI00* - 2Cl + 0i{1 i::g+). This signal is larger in proportion to the er Cl00Cl photolysis signal than it was at 248 nm, and could also arise from dissociation of clusters in the beam. llO Relative Yields To obtain the overall relative yields for ClOOCl photodissociation at 248 run, we had to assume that ClOOCl only dissociates via the channels considered. Relative yields could be calculated assuming that 1) all slow signal, 2) only the slow signal at mlz = 3 5 or 51, or 3) none of the slow signal arises from ClOOCl photolysis. Including the slow signal did not greatly change the overall relative yields. Only the primary dissociation channels were used to calculate the relative yields, with the secondary contributions scaled so that the primary and secondary dissociation processes had a 1: 1 ratio for a given channel. The relative yield for the concerted channel was determined by modeling the Cl products as arising from a primary two-body channel to get a primary c.m. flux distribution. All of the calculated yields depended somewhat on the choice we made for the contribution from Cl2O photolysis. At 248 run, we determined the Cl2O contribution from the Ci+ product TOF spectrum as discussed in the Results and Analysis section. This method gave a Cl2O contribution that was comparable to that predicted from source TOF measurements, as described in Chapter 3. The ratio of Cl2O to ClOOCl photolysis signal was predicted to be Cl2O : ClOOCl = 1. 0 : 5.6 for this experiment. After the final analysis of all the photolysis signal using the CIO+ "rising edge" method to determine the Cl2O contribution, the Cl2O: ClOOCl photolysis signal ratio was 1.0 : 4.6 if the very slow Ci+ and CIO+ signal was attributed to clusters and 1.0 : 6.8 if this signal was attributed to ClOOCl photolysis. Because the ClO channel for ClOOCl photolysis is minor, changing the Cl2O contribution over this range does not greatly affect the Cl : ClO branching ratio for ClOOCl photolysis. We could obtain an absolute upper limit on the relative yield of the ClO + ClO channel by 111 assuming that all of the ClO signal was from ClOOCl photolysis with no contribution from Cl2O. This upper limit on the relative yield is 4>a0 +Clo= 0.19 for ClOOCl excited at 248 nm. The upper limit we measured would be too low if any of the primary ClO photofragments absorbed a photon. However, secondary dissociation of ClO was minimal for all of these experiments, as verified by the corresponding Cl2O photolysis experiments. At 308 nm, we observed two main product channels, ClO + ClO and Cl + ClOO .... 2Cl + 02, as well as some slow Ci+ signal. The slow signal was much higher with respect to the ClOOCl signal in the fast peak than it was at 248 nm, and thus makes more of a difference in the relative Cl and ClO yields from ClOOCl photolysis. We calculated the relative yields for both the case where with slow signal was attributed to clusters as well as the case where it arose from ClOOCl photolysis. The Cl2O contribution was determined relative to the Cl2 contribution from TOF measurements of the source and did not rely on any assumptions involving ClOOCI. We can again obtain an upper limit for the relative yield of the ClO + ClO channel by assuming that all the signal arises from ClOOCl; this upper limit for ClOOCl photoexcited at 308 nm is 4>ao+ao = 0.31. Table II indicates the observed channels and their relative yields, as well as the relative Cl: ClO product yields for ClOOCl photolysis at 248 and 308 nm. The values given were calculated assuming that the slow signal can all be attributed to cluster photolysis; the values in parentheses were calculated assuming that the slow signal arises from ClOOCl photolysis. For these calculations, all of the Cl2O2 photolysis signal was attributed to the ClOOCl isomer. The other possible isomers, ClOClO and ClCIO2, were discussed in Chapter 3. ClOOCl is the major product from the three-body recombination of ClO, on which we based our source 112 design; however, we can not preclude the presence of the other isomers in our source. In particular, it is possible (although unlikely) that some of the ClO signal came from the ClOClO isomer. Table IT. Relative yields* for the ClOOCl photolysis channels observed at 248 and 308 nm. relative yield 248nm 308nm Cl + ClOO .... Cl + Cl + 0 2 0.67 (0.55) 0.83 (0.50) Cl + 0 2 + Cl (concerted) 0.22 (0.18) ClO + ClO 0.12 (0.09) 0.17 (0.10) Cl+ ClOO*- Cl+ Cl+ 0 2 * 0.00 (0.15) 0.00 (0.40) ClO + CIO* .... ClO + Cl + 0 0.00 (0.04) Dynamical Channels Relative Product Yields 2Cl + 0 2 0.89 (0.88) 0.83 (0.90) ClO + ClO 0.12 (0.13) 0.17 (0.10) *The numbers listed assume that products with low c.m. translational energies arise from some other process such as cluster dissociation; the numbers in parentheses include those products as arising from ClOOCl dissociation. Discussion At 248 run, we observed three primary dissociation channels: ClO + ClO, a concerted channel in which both Cl -0 bonds break simultaneously to form 2Cl + 0 2, and a sequential dissociation to 2Cl + 0 2 via Cl+ CIOO. In addition, slow signal was observed for both CIO+ and er photofragments that could correspond to ClO(X) + ClO(A) and Cl+ CIOO*- 2 Cl + 0 2 channels. The average and maximum c.m. total translational energies and anisotropy 113 parameters for these channels are summarized in Table III. Table ill. Summary ofresults for each dissociation channel at 248 nm. Energies are all in kcal/mol. The ClO + ClO* and Cl + CIOO*-, 2Cl + 0 2 * channels were modeled, but this signal is likely to arise from clusters. channel <Er) £-r,max £avail <Er)/£avail ET,nwl£avail p Cl +ClOO 34 58 95.4 0.36 0.61 0.6 Cl +02 +Cl 65 87 90.7 0.72 0.96 -1.0, o.1t ClO+ CIO 26 45 99.5 0.26 0.45 0.5 Cl+ ClOO* 2.3 12 25.1 0.09 0.48 0 ClO+ ClO* 2.7 9 9.0 0.30 1.0 0 ClOO--Cl + 0 2 6.4 19 90.7- ET,1· i isotropic isotropic ClOO--Cl + 0 2 * 4 15 51.7- ET,1· i t The angular distribution for 0 2 corresponded to -1.0 ~ P ~ -0.5 and that for Cl to 0.0 ~ P ~ 0.4, if each was modeled as ifit were a two-body dissociation. The best values of-1.0 and 0.1 shown in this table are explained in the "Transition Dipole of the Concerted Channel" section later in this chapter. :The available energy for the secondary dissociation depends on the primary translational energy, and is thus a distribution. At 308 nm, two main channels for ClOOCl photolysis were observed: ClO + ClO and 2Cl + 0 2 via Cl + ClOO. The signal-to-noise ratio was worse than that at 248 nm, the Cl2 photolysis signal obscured much of the ClOOCl photolysis signal at mlz = 35, and no 0/ signal was observed; therefore, we can not determine whether a concerted process to form 2Cl + 0 2 also occurred. Rather, all of the Ci+ signal was assigned to sequential Cl+ ClOO and the spontaneous secondary dissociation of ClOO. Since slow Ct signal was observed, which most likely arose from clusters but could have also arisen from some Cl + CIOO* ... 2Cl + 0 2 * channel. This slow signal was much larger in proportion to the CIOOCl photolysis signal than it was at 248 nm. The average and maximum c.m. total translational energies and 114 anisotropy parameters for these channels are summarized in Table IV. Table IV. Summary ofresults for each dissociation channel at 308 nm. Energies are all given in kcal/mol. The Cl + CIOQ*-. 2Cl + 0 2 * channels was modeled, but this signal is likely to arise from clusters. channel (ET) Enrmx ElN1111 (ET)/EIMil Ennw/£;wail p ClO + ClO 27 31 77.3 0.35 0.40 2.0 Cl +ClOO 27 45 73 .2 0.37 0.61 1.5 Cl+ ClOO* 1.6 5.7 34.2 0.05 0.17 0 ClOO -- Cl + 0 2 ~8 20 68.5- ET,1· i isotropic ClOO*-- Cl+ 0 2* ~14 29.5 29.5- Er,1· i isotropic Clo Photolysis Signal at 248 nm The angular distribution for the Cl products from the photolysis of ClO was fit by P = 2.0, which differs from the anisotropy parameter P= 1.2 reported by Davis and Lee. 25 The ClO (A 2rr312 .... X2II~ excitation is a parallel transition (A Q = 0), for which p = 2.0. Davis and Lee explained their observation of P= 1.2 by assuming that ~70% of the dissociation was from the A state, with the remaining ~30% occurring via a perpendicular transition ( p = -1) from a state of 2 :E or 2 A symmetry (AO=± 1). Their ClO source was from photolysis of the (6,0,0) band of OClO in a jet, which produces a higher-than-thermal yield of spin-orbit excited ClO. 34 Our ClO arose from Cl20 photolysis and the ensuing chemical reactions, and resided in the quartz cell for over 100 µs after the Cl20 photolysis occurred. Thus, our ClO was more likely to have a thermal spinorbit population distribution, which, for a cell temperature of ~200K is ~95% ClO X 2II312 . It is possible that the spin-orbit excited ClO X 2TI 112 has enough extra energy to be able to 115 access the higher 2 ~ or 2 A state while the ClO X 2II312 state can not. In this case, we would only see signal from the ClO (A 2rr312 - X 2II312) transition, and the two experiments would be consistent. Transition Dipole of the Concerted Channel and Effects of Excited State Symmetzy (248 nm) The angular distribution of the 0 2 peak corresponded to an anisotropy parameter of -1.0 ~ p ~ -0.5, if we modeled the dissociation as a two-body process. The angular distribution of the Cl counterfragments can be described by 0.0 ~ P ;;; 0.4. The low signal-tonoise ratio at mlz = 32 and the multiple components at mlz = 35 made it difficult to assign more precise values for the anisotropy parameters. P has the form P = 2Pi(cosx) where P 2 is the second Legendre polynomial and x is the angle between the electronic transition moment and the c.m. recoil axis of the fragments. For a given P value, the transition moment resides on a cone with a half-angle x centered about the recoil axis. For the concerted channeL both p values must correspond to a single electronic transition, so that the transition moment must lie at the intersection of the two cones determined from the 0 2 and Cl angular distributions. Furthermore, the transition moment must be unique, so that the two cones must intersect only along one line (that is, the edges of the two cones must just touch.) For the 0 2 products, x = 90° for P= - 1.0, and x = 66° for P= - 0.5, so that X must lie between 66° and 90°. For the Cl products, x = 55° for P = 0.0 and X = 47° for P = 0.4. The angles between the c. m. recoil vectors for the ClOOCl dissociation geometry determined from the peak positions in the three-body dissociation analysis are Le1a,b-o2 = 142.4° and 116 Laa-en,= 75.2°. Because the angle between the Cl and 0 2 c.m. velocity vectors is 142.4°, for the two cones to intersect at the edges the two x's for the Cl and 0 2 angular distributions must add up to 142.4 ° . This is only possible within the bounds of the observed angular distributions if p02 = -1.0 (x = 90°), in which case Pa must be 0.1 (x = 52°). Thus, the angular distributions for the products from the concerted dissociation channel are consistent with an electronic transition moment along the 0-0 bond. The other two primary dissociation channels, ClO + ClO and Cl + CIOO, were parallel transitions, as indicated by the anisotropy parameters P = 0.5 and P = 0.6, respectively. Theory3 predicts a series of pairs of nearly degenerate states at 248 nm, formed as symmetric (A) and antisymmetric (B) linear combinations of the ClO antibonding orbitals. The transition moments (as determined from the anisotropy parameter) for these channels suggest that excitation to an antisymmetric B state (transition moment along the 0-0 bond axis) leads to both the ClO + ClO and the concerted 2Cl + 0 2 channels, whereas excitation to a symmetric A state (transition moment along the Cl-O bond) leads preferentially to dissociation via Cl + CIOO. Thus, the symmetry of the excited state appears to influence the dissociation dynamics. Comparison with Previous Experiments Our results indicate that the previous reports of the yield for Cl-atom formation by Cox and Hayman9 and Molina et. al. 10 are qualitatively correct, but somewhat high. One uncertainty in the absolute quantum yield determined by Molina et al. at 308 nm is the dependence on the ClOOCl absorption cross section. Using the most recent absorption cross section measurements12 results in a re-calculated yield for the 2Cl + 0 2 channel (7) of <1> 7 = 117 0.88, in excellent agreement with our results (<I> 7 = 0.83). Furthermore, the high ClO vibrational excitation that we observe could explain their inability to detect ClO products. Recent experiments35 show that the reaction rate of vibrationally excited ClO with NO is much slower than that for ground state ClO reacting with NO. Since Molina et. al. used titration with NO to detect ClO, and our results suggest that the ClO is vibrationally excited, it is not surprising that they did not detect ClO. Schmidt et. al. 13 reported a much lower yield for the Cl + ClOO channel at 250 nm of Cl> 7 = 0.65 ± 0.15, with an average translational energy of (ET)= 0.42 eV (~ 10 kcal/mol). At 248 nm, we observe a relative yield for Cl-producing channels of <I> 7 = 0.88, and were able to obtain an upper limit for ClO products that implies that the lower limit for Cl-producing channels must be Cl> 7 :2:: 0.81. This lower limit is right on the edge of the error bars they give for the Cl yield. Also, we observed that the average translational energy for primary and secondary Cl from the Cl + ClOO - Cl + Cl + 0 2 channel was 22 kcal/mol and 3 kcal/mol, respectively. The Cl from the concerted 2Cl + 0 2 channel had an average translational energy of 21 kcal/mo I. If we only consider our yield for the Cl + ClOO - 2Cl + 0 2 channel, Cl> = 0.67, we get reasonable agreement with their results. If the very slow products we observed arose from ClOOCl and not clusters, the average translational energy of all the Cl products was ~ 11 kcal/mol with a yield of Cl> = 0.89, and the average translational energy if we exclude the Cl products from the concerted channel was ~ 7 kcal/mol with a yield of Cl> = 0. 70. The best explanation for the discrepancy between our results and the results of Schmidt et. al. is that their technique may discriminate against the Cl products with high translational energies. 118 Stratospheric Implications From our results at 308 run, we conclude that in the stratosphere, ClOOCI + hv - 2Cl + 0 2 <P7= 0.83 (7) - ClO + ClO <1> 5 =0.17 (5), with an absolute upper limit of <I> = 0.31 for the ClO product channel. The intermediate ClOO (reaction 6) is formed withEin1 >>De and is unlikely to be collisionally stabilized under stratospheric conditions. Also, we observed highly vibrationally excited ClO and 0 2 photoproducts. The role of vibrationally excited states in the stratosphere has recently received some attention. Specifically, the reactions36•37 and 0 2 (v ~ 26) + 0 2 - 0 3 + 0 (11) ClO (v ~ 10) + N2 - Cl+ N20 (12) could potentially be of importance in the stratosphere. The 0 2 we observed from the sequential 2Cl + 0 2 channel at 248 run had an average internal energy of 50 kcal/mo~ corresponding to v = 12, which is not high enough to exceed the threshold for reaction with 0 2. However, the ClO produced from the CIO + ClO channels is highly vibrationally excited, with an average energy per ClO of37 kcal/mol at 248 run and 24 kcal/mol at 308 run. The threshold for reaction with N 2 is 24.3 kcal/mol. Thus, the production of highly excited ClO from ClOOCl photolysis could lead to a small production term for N 20 in the stratosphere. N 20 is currently used as a tracer species, with no stratospheric sources, so discovering a stratospheric source for N20 could potentially be important. The production ofN20 from ClOOCl photolysis is not likely to be a major perturbation on current models, because 1) the 119 ClOOCl photolysis mainly occurs for longer wavelengths, where less energy is available and 2) the ClO + ClO channel is minor. At 248 nm, slow signal was observed at mlz = 51 which could arise from the ClO(X) + ClO(A) - ClO + Cl + 0 channel. This channel is unlikely to be of great importance in the stratosphere, as it requires wavelengths }. :s:; 290 nm and most of the contribution to the ClOOCl photolysis rate occurs for the wavelength range 310 to 4 70 nm. 12 The ClO(A) fragment predissociates, producing a Cl-atom and an O-atom. Thus, the ClO(X) + ClO(A) channel results in a null cycle for ozone depletion. Our results confirm that the catalytic cycle involving photolysis of ClOOCl is the dominant mechanism for ozone loss under perturbed polar conditions. However, the production of ClO from ClOOCl photolysis reduces the efficiency of the ClOOCl catalytic cycle for ozone depletion and could shift the predicted partitioning of chlorine-containing species in the stratosphere. Because of the extreme importance of the ClOOCl catalytic cycle in the current models of Antarctic stratospheric ozone depletion, even the small ClO yield (<I>c10 +eta= 0.17) from ClOOCl photolysis at 308 nm reported in this paper could have a significant effect on the predicted ozone depletion. Further modeling studies are necessary to determine whether the small reduction in the Cl yield may require additional ozone loss mechanisms to balance the ozone budget. This result may be most important in the Arctic vortex, where warmer temperatures decrease ClOOCl concentrations and reduce ozone loss rates. 120 References 1. L. T. Molina and M. J. Molina, J. Phys. Chem. 91, 433 (1987). 2. T. J. Lee, C. M. Rohlfing, and J.E. Rice, J. Chem. Phys. 97, 6593 (1992). 3. J. F. Stanton, C. M. L. Rittby, R. J. Bartlett, and D. W. Toohey, J. Phys. Chem.95, 2107 (1991). 4. M. P. McGrath, K. C. Clemitshaw, F. S. Rowland, and W. H. Hehre, Geophys. Res. Lett. 15, 883 (I 988). 5. F. Jensen and J. Oddershede, J. Phys. Chem. 94, 2235 (1990). 6. M. Schwell, H.-W. Jochims, B. Wassermann, U. Rockland, R Flesch, and E. Ruhl, J. Phys. Chem. 100, 10700 (1996). 7. M Birk, R. R. Friedl, E. A. Cohen, H. M. Pickett, and S. P. Sander, J. Chem. Phys. 91, 6588 (1989). 8. Heat of formation for ClOOCl from S. L. Nickolaisen, R.R. Friedl, and S. P. Sander, J. Phys. Chem. 98, 155 (1994). 9. R. A. Cox and G.D. Hayman, Nature 332, 796 (1988). 10. M. J. Molina, A. J. Colussi, L. T. Molina, R. N. Schindler, and T.-L. Tso, Chem. Phys. Lett. 173, 310 (1990). 11. W. B. DeMore, S. P. Sander, D. M. Golden, R. F. Hampson, M. J. Kurylo, C. J. Howard, A. R. Ravishankara, C. E. Kolb, and M. J. Molina, Chemical Kinetics and Photochemical Data for Use in Stratospheric Modeling, Evaluation Number 12, JPL Publication 97-4, JPL (1997). 12. K. J. Huder and W. B. DeMore, J. Phys. Chem. 99, 3905 (1995). 13. S. Schmidt, Th. Benter, and R. N. Schindler, poster at XXIII Informal Conj on Photochemistry, May 10-15 1998, Pasadena, CA, USA. 14. Y. T. Lee, J. D. McDonald, P.R. LeBreton, and D. R. Herschbach, Rev. Sci. Inst. 40, 1402 (1969). 15. T. K. Minton, C. M. Nelson, T. A. Moore, and M. Okumura, Science 258, 1342 (1992). 16. T. K. Minton, P. Felder, R. J. Brudzynski, and Y. T. Lee, J. Chem. Phys. 81, 1759 (1984). 121 17. M. J. O'Laughlin, B. P. Reid, and R. K. Sparks, J. Chem. Phys. 83. 5647 (1985). 18. D. J. Krajnovich, Ph.D. Thesis, University of California, Berkeley (1983). 19. C. M. Nelson, T. A. Moore, M. Okumura, and T. K. Minton, J. Chem. Phys. 100, 8055 (1994). 20. T. A. Moore, M. Okumura, and T. K. Minton, J. Chem. Phys. 107, 3337 (1997), 21. P. Felder, Habilitationsschrift, University of Zurich, Switzerland (1993). 22. C. Maul and K.-H. Gericke, Int. Rev. in Phys. Chem. 16, 1 (1997). 23. M. J. S. Dewar, J. Am. Chem. Soc. 106,209 1984. 24. S. W. North, Ph.D. Thesis, University of California, Berkeley (1995). 25. CMLABl, presented in X. Zhao, Ph.D. Thesis, University of California, Berkeley (1988). 26. H.F. Davis and Y. T. Lee, J. Phys. Chem. 100, 30 (1996). 27.L. V. Gurvich, I. V. Veyts, and C. B. Alcock, Thermodynamic Properties of Individual Substances, Fourth Edition, Volume 1, Hemisphere Publishing Corp., p. 182 (1989). 28. R. E. Center and A. Mandi, J. Chem. Phys. 57, 4104 (1972). 29. T. R. Hayes, R. C. Wetzel, and R. S. Freund, Phys. Rev. A. 35, 578 (1987). 30. J. A. Jafri, B. H. Lengsfield, III, C. W. Bauschlicher, Jr., and D. H. Phillips, J. Chem. Phys. 83, 1693 (1985). 31. Y. Matsumi, M. Kawasaki, T. Sato, T. Kinugawa, and T. Arikawa, Chem. Phys. Lett. 155, 486 (1989). 32. Y. Matsumi, K. Tonokura, and M. Kawasaki, J. Chem. Phys. 97, 1065 (1992). 33. P. C. Samartzis, I. Sakellariou, T. Gougous~ and T. Kitsopoulos, J. Chem. Phys. 107, 43 (1997). 34. H.F. Davis and Y. T. Lee, J. Chem. Phys. 105, 8142 (1996). 35. G. S, Tyndall, C. S. Kegley-Owen, J. J. Orlando, and J. G. Calvert, J. Chem. Soc., Faraday Trans. 93, 2675 (1997). 36. J.M. Price, J. A. Mack, C. A. Rogas~ and A. M. Wodtke, Chem. Phys. 175, 83 (1993)~ R L. Miller, A.G. Suits, P. L. Houston, R. Toumi, J. A. Mack, and A. M. Wodtke, Science 122 265, 1831 (1994). 37. R. F. Delmdahl and K.-H. Gericke, Chem. Phys. Lett. 281,407 (1997). 123 Figure Captions Figure 1. Structure of ground state ClOOCl, from ref 7. Figure 2. TOF spectra of CIO+ signal from ClOOCl photolysis at 248 nm. The dotted lines correspond to the Cl2O contribution, which in this case was determined by examining the rising edge of the data at a detector angle of 15°. The dashed line is the remaining signal modeled as ClO + ClO from ClOOCl photolysis. The slow signal was modeled as ClO(X) + ClO(A). Figure 3. TOF spectra of Clo+ signal from ClOOCl photolysis at 248 nm. The solid line is the predicted spectra assuming that all of the signal at 30° arises from Cl2O only and scaling the other angles according to the Cl2O angular distribution. Figure 4. TOF spectra of Clo+ signal at 30° from ClOOCl photolysis at 248 nm, signal averaged for 1,000,000 laser pulses. The solid line is again assuming that all of the signal arises from Cl2O photolysis. Figure 5. TOF spectra ofClO+ signal at 15° from CIOOCl photolysis at 248 nm. The dotted lines correspond to the Cl2O contribution, which was determined by examining the rising edge of the data. 124 Figure 6. Optimized c.m. translational energy distributions for the ClO + ClO and the ClO(X) + ClO(A) channels. Figure 7. TOF spectra of the 0/ signal at 30° from ClOOCl photolysis at 248 run. The fast peak was modeled as arising from the concerted channel. The dotted line indicates where the 0 2 products from the secondary dissociation from the fast Cl + ClOO channel would appear; the dot-dashed line indicates the position of the 0 2 products from secondary dissociation from the Cl + CIOO* channel. Figure 8. Optimized c.m. translational energy distribution for the total translational energy release from the concerted Cl+ 0 1 + Cl channel. Figure 9. TOF spectra of the Ci+ signal from ClOOCl photolysis at 248 run. The shortdashed line indicates the Cl2O contribution, determined from the mlz = 51 data. The mediumdashed line is from photolyis of ClO in the beam. The long-dashed lines indicate the Ci+ signal from the ClO + ClO channels fragmenting in the ionizer. The dotted line corresponds to the primary and secondary dissociation from the fast Cl+ ClOO channel. The dot-dashed line is from the Cl+ 0 2 + Cl concerted channel. The dot-dot-dashed line indicates the signal from the primary and secondary dissociation for the slow Cl + CIOO* channel. Figure 10. Optimized c.m. translational energy distributions for the Cl+ CIOO primary channel and the ensuing ClOO secondary dissociation. 125 Figure 11. TOF spectra of the er signal at 15°. The dashed line shows the predicted fit for Cl + ClOO* if a statistical dissociation to the highest calculated CIOO* state that correlates to excited state products. The dotted line indicates the predicted fit for the secondary dissociation of CIOO* to Cl + 0 2* assuming a statistical dissociation. Figure 12. TOF spectra collected at a detector angle of 40° for (A) Clo+ and (B) Cl+ fragments and the predicted TOF distributions for ClOOCl photolysis at 308 run. The overall Clo+ fit (solid line) in (A) has two components, Cl2O (dashed line) and the ClOOCl - ClO(X) + ClO(X) channel (dotted line). The overall er fit (solid line) in (B) is the sum of contributions from photolysis of Cl2 (long-dashed line), Cl 2O (short-dashed line), ClO Cl(2P112,3n) + 0(3P) (dot-dot-dashed lines), and ClOOCI. The ClOOCl photolysis contributions to the er signal include primary and secondary dissociation from the ClOOCl - Cl + ClOO - Cl + Cl + 0 2 channel (dotted lines), primary and secondary dissociation from the ClOOCl - Cl + ClOO* - Cl + Cl + 0 2 * channel which could also arise from clusters (medium-dashed lines), and Ci+from the ClO(X) fragmenting in the ionizer (dot-dashed line). Figure 13. TOF spectra of the er signal from Cl2 photolysis at 308 nm. Figure 14. Optimized c.m. translational energy distribution for the ClO + ClO channel at 308 nm. Figure 15. Optimized c.m. translational energy distributions for the Cl+ ClOO primary channel and the ensuing ClOO secondary dissociation. 126 1.426 A 0 110.10 '1.704A - Cl Figure 1. 127 15° oO 30° 0 40° 0 0 ,:····-... . .~.····- . 50° 0 0 0 0 0 0 Cb 0 0 0 0 0 0 0 0 o0 0 0 Oo 0 o0 00 0 0 100 Figure 2. 200 300 Flight time, µs 0 400 128 15° . .. . ,.I ....... o o O 0 0 Co Cb 0 n... 00O c9o oCb o<P <P"U 30° ti) =: C: ::, .c < 0 b ·en 40° C: Jg C: 50° 0 0 0 (9 0 oo 0 0 0 100 Figure 3. 0 150 200 Flight time, µs 250 300 129 cSJ CIO+ products at 30° --... z 150 200 flight time, µs Figure 4. 250 130 15° Cb U) 0 :t:! C: ::::::> .c L.. 0 0 0 Q)O 0 0 0 0 0 0 0 0 0 <( 'b 0 0 0 >. ~ 0 U) C: 0 .! C: Cb 0 0 0 0 O oc9 100 0 150 200 Flight time, µs Figure 5. 250 131 ~ C: ::, -e <( ~ lJJ er 0 5 10 15 20 25 30 35 40 Translational Energy, kcal/mol Figure 6. 45 50 132 mlz = 32 , 30° 0 0 O 0 50 100 150 200 250 300 350 400 450 500 Flight time, µs Figure 7. 133 15 U) ~ C :J .c <C 10 ~ - LJJ it:' 5 0 L....1.....L...L.L...J.,4...J._I....L..J.....L...J...J-L...J....1....L..J.....L....L..lL..J....L.J....L...L...1....L..JL..J....L.J....L...L...1....L..JL....L...L....L.L...J.....L...L...I....L..J....L.L...J.-L...J.....1...J....J...u..L..L..J...J 30 35 40 45 50 55 60 65 Energy, kcal/mo I Figure 8. 70 75 80 85 90 134 ' •ji.;~~\J{l.~ •• -·:..:.....:--..· ·---------------- 30° 40° /· .. • . -.-,:-:;..i..:·· . ..,.__- .. 50° 100 Figure 9. 150 200 250 Flight time, µs 300 350 400 135 secondary primary "'C: :t:: ::, .ci <C ~ - ur ct' 0 4 8 12 16 20 24 28 32 36 40 44 48 52 56 60 Energy (kcal/mol) Figure 10. 136 <O Cl+ products at 15° ,g & <a ~ & 8 i f/) :::: C :::, .c i.... <( - 9 d3~o 0 0 0 ~ p z 0 0 0 '/ / 0 '/ -------- --- _ _ _ _ _ _ _...c=.;...:.=-:..:.='-'-='-1 0 •1/- _ _ _ _-----------------------------0 ~,...,..,_.-:::::....._ 100 150 200 250 300 flight time, µs Figure 11. 350 400 450 500 137 0 0 Oo 0 0 Co 0 0 en ......, ■- C: :::> 0 0 0 0 O 0 &;rYc9 0 0 0 0 0 0 0 0 0 • ... o 0 cB 0 O 0 Oo 0 00 Oo .c I- <( 0 Cl♦, 40° B ~ ......, ■- en C: Q) ......, - C: i oO 0 0 150 Figure 12. 200 250 0 300 Flight time, µs 0 350 138 10° 20° en ~ C: :::> . ..c I- 30° <( - ~ ~ en ... C: Q) C: 40° 50° Figure 13. 100 125 150 Flight time, µs 175 139 CIO + CIO at 308 nm 10 15 20 25 Energy, kcal/mol Figure 14. 30 35 140 Cl + CIOO ➔ 2 Cl + 0 2 at 308 nm primary secondary 0 5 10 15 20 25 30 Energy, kcal/mo I Figure 15. 35 40 45 50