Electronic states and photodissociation dynamics of hydroxyalkyl radicals

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Request accessible transcript

Transcript (if available)

Content ELECTRONIC STATES AND PHOTODISSOCIATION DYNAMICS OF HYDROXYALKYL RADICALS by Boris Karpichev __________________________________________________________________ A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment of the Requirements for the Degree DOCTOR OF PHILOSOPHY (CHEMISTRY) August 2009 Copyright 2009 Boris Karpichev ! ii ! Acknowledgments My transition from Physics to Chemistry, from theory to experiment from Master to Doctor would have been impossible without guidance and support of my adviser, my colleagues and friends, and of course my family. I am grateful to Professor Hanna Reisler, my adviser, for all the opportunities she has given to me to work on interesting projects, to solve challenging scientific problems and to interact with the scientific world. Also I would like to thank Hanna for the wonderful environment she created in our group that truly facilitated collaboration and research I want to thank all past and present members of Hanna’s group. Specifically, I wish to express my gratitude to Dr. Vladimir Dribinsky for being my friend and mentor, from whom I learned many valuable lessons in research and life; Dr. Lin Feng, for teaching me how to work on the TOF machine and also for all her knowledge and experience she shared with me; Dr. Jie Wei for all the work we have done together, for all the fun we had in the lab. Good old Jie, I really miss you! I wish thank to Laura Edwards, my partner in the lab for many years. Words cannot justly express my gratitude for all the help and support I got from Laura for all these years. I also want to thank Misha Ryazanov for collaborative work and fruitful discussions. I am grateful to the rest of the group: Dr.Guosheng Lee, Dr. Andrew Mollner, Dr. Jessica Parr, Blythe Casterline, Lee Chiat Ch’ng and Dr. Igor Fedorov for constant fruitful interaction and maintaining a friendly working atmosphere. iii ! My ab initio project would not succeed without support and guidance from Professor Anna Krylov, Dr. Kadir Diri. I would like to thank also Dr. Piotr Pieniazek for numerous thought-provoking discussions. I have greatly benefited from many excellent classes I have attended at USC. For well-crafted lectures and engaging discussions I would like to thank Professor Curt Wittig, Professor Stephen Bradforth and Professor Hanna Reisler. I met a lot of friends here who made USC such a fun place to be. In no particular order I want to thank Dr. Dmitry Skvortsov, Dr. Roman Rabinovich, Dr. Sergey Malyk, Dr Igor Fedorov, Dr. Miklhail and Lyuda Slipchenko, Dr. Kirill Kuyanov, Dr. Piotr Pieniazek, Dr. Nikolai Markovsky, Vadim Mozhaysky, Dr. Kadir Diri, Sergey Zakharov, Dr. Askat Jaluybekov, Dr. Danmiil Stolyarov, Dr. Elena Polyakova and Dr. Sergey Levchemko. I wish to thank Department staff members for their highly professional work: Michele Dea (Michele, you are amazing!) Heather Meunier, Danielle Hayes, Katie McKissik, Yuki Yabuta, Valerie Childress, Ross Lewis, and Frank Niertit. Finally, I must thank my family, who supported my bold undertaking and my special big thanks to my beloved wife, Natalya Razumikhina, for her unselfish support, care and understanding. iv ! TABLE OF CONTENTS Acknowledgments ..................................................................................................... ii List of Tables ............................................................................................................ vi List of Figures ......................................................................................................... vii Abstract ..................................................................................................................... x Chapter 1. Introduction ...................................................................................... 1 1.1 Overview ............................................................................................ 1 1.2 Hydroxymethyl Radical ...................................................................... 1 1.3 1-Hydroxyethyl Radical ..................................................................... 4 Chapter 1 References ...................................................................................... 8 Chapter 2. Experimental Details ...................................................................... 12 2.1 Overview .......................................................................................... 12 2.2 Radical Production ........................................................................... 14 2.3 Laser System and H/D Detection ..................................................... 17 2.4 Time-of-Flight .................................................................................. 20 Chapter 2 References .................................................................................... 28 Chapter 3. Unimolecular Processes In CH 2 OH Below The Dissociation Barrier: O–H Stretch Overtone Excitation And Dissociation .... 29 3.1 Introduction ...................................................................................... 29 3.2 Experiment ....................................................................................... 32 3.3 Results and Analysis ......................................................................... 36 3.3.1. Excitation of the Second OH-stretch Overtone, 3! 1 ............... 36 3.3.2 Excitation of the Third OH-stretch Overtone, 4! 1 ................... 40 3.3.2.1 Hydrogen Fragment Yield Spectra Of CH 2 OH And CD 2 OH .............................................................. 40 3.3.2.2 Time-Of-Flight (TOF) Analysis Of Hydrogen Fragments ................................................................. 42 3.3.3. Spectroscopic Analysis Of OH-Stretch Overtones ................. 45 3.3.3.1 Analysis Of The Second Overtone Transition .......... 45 3.3.3.2 Analysis Of The Third Overtone Transition ............. 46 3.3.4. Birge-Sponer Analysis ............................................................ 47 3.4 Discussion ......................................................................................... 49 v ! 3.4.1. Roles Of IVR, Predissociation, And Isomerization ................ 49 3.4.2. Predissociation By Tunneling ................................................. 55 3.5 Conclusions ...................................................................................... 56 Chapter 3 References .................................................................................... 58 Chapter 4. Electronic Spectroscopy and Photodissociation Dynamics of the 1-Hydroxyethyl Radical CH 3 CHOH .................................. 61 4.1 Introduction ...................................................................................... 61 4.2 Experimentsl Details ........................................................................ 64 4.3 Results .............................................................................................. 68 4.3.1 Photofragment Yield And REMPI Spectra .......................... 68 4.3.2 Time Of Flight Of H(D) Photofragments ............................. 72 4.4 Discussion ......................................................................................... 76 4.4.1 Assignment Of The 3s And 3p z States ................................. 76 4.4.2 Photodissociation Dynamics ................................................ 82 4.5 Conclusions ...................................................................................... 87 Chapter 4 References .................................................................................... 90 Chapter 5. Effect of Hyperconjugation on Ionization Energies of Hydroxyalkyl Radicals ................................................................... 93 5.1 Introduction ...................................................................................... 93 5.2 Computational Details ...................................................................... 99 5.3 Results and Discussion ................................................................... 101 5.4 Conclusions .................................................................................... 109 Chapter 5 References .................................................................................. 111 Chapter 6. Future Experiments ...................................................................... 115 6.1 Unimolecular reactions on the ground electronic state .................. 115 6.2 Electronic states of the 2-hydroxyethyl radical .............................. 119 6.3 4 th overtone of OH stretch in CH 2 OH ............................................ 124 Chapter 6 References .................................................................................. 128 Alphabetized Bibliography .................................................................................. 131 ! vi ! List of Tables Table 3.1 OH-stretch vibrational frequencies and anharmonicities of CH 2 OH in the ground and the Rydberg 3p z state .............................................. 50 Table 4.1 Comparison of energies and quantum defects of Rydberg states for CH 2 OH and CH 3 CHOH. ! is obtained from the experimental results ................................................................................................... 67 Table 5.1. Computed Vertical and Adiabatic IEs (eV) of CH 3 CHOH and CH 2 OH .............................................................................................. 101 ! vii ! List of Figures Figure 2.1 Schematic diagram of the molecular beam apparatus (side view): 1) Source chamber 2) Detection chamber 3) Piezoelectric nozzle 4) Ion optic system 5) Field-free drif region 6) MCP detector ............ 13 Figure 2.2 Schematic diagram of the piezoelectric nozzle ................................... 15 Figure 3.1 IR+UV double resonance REMPI spectrum of CH2OH obtained via the 3p z state under "IR on" (solid line) and "IR off" (dotted line) conditions. The pump laser frequency is fixed at 10 489.8 cm – 1 , the frequency of the second overtone transition in CH 2 OH ............. 37 Figure 3.2 IR spectrum of CH 2 OH in the region of the second overtone of the OH stretch obtained by DRID. The upper panel displays the experimental spectrum. The solid and dotted lines represent signals corresponding to "IR-on" and "IR-off" experiments, respectively. In the bottom panel a best-fit spectrum with a linewidth of 0.4 cm –1 is shown. The calculated rotational transitions are given by the stick spectrum. In the simulations, we used A”=6.51 cm –1 , B”=1.01 cm –1 , C”=0.88 cm –1 , A’=6.30 cm –1 , B’=1.00 cm –1 , C’=0.88 cm –1 , T rot =13 K, and " 0 =10 484.2 cm –1 .................................. 39 Figure 3.3 The solid curve depicts experimental H-atom photofragment yield spectra in the region of the third overtone of the OH stretch for (a) CH 2 OH and (b) CD 2 OH. The dashed line shows a spectral fit to the data for an a-type transition and linewidth of 1.3 cm –1 . See text for details. In the simulations, we used A”=6.51 cm –1 , B”=1.01 cm –1 , C”=0.88 cm –1 , A’=6.00 cm –1 , B’=B”, C’=C”, T rot =13 K, and ! 0 =13 597.9 cm –1 for CH 2 OH; and A”=3.87 cm –1 , B”=0.86 cm –1 , C”=0.71 cm –1 , A”=3.70, B’=B”, C’=C”, T rot =13 K, and ! 0 =13 616.6 cm –1 for CD 2 OH .............................................................. 41 Figure 3.4 Time-of-flight (TOF) spectra of H fragments produced in the dissociation of CH 2 OH by one photon excitation at (a) 27 210 cm –1 (3s Rydberg excited state) and (b) 13 603 cm –1 (4! 1 transition). Zero TOF indicates no recoil energy. The arrows show the maximum and minimum TOF values allowed by the thermochemistry for one-photon dissociation. The polarizations of the pump laser radiation used in (a) and (b) are perpendicular and parallel, respectively, to the extraction field. Each spectrum is a summation of 5000 laser firings, and background is subtracted ......... 44 viii ! Figure 3.5 Birge-Sponer plots for CH 2 OH in the ground electronic state and the 3p z Rydberg excited state. The lines are fits to the Birge-Sponer expression and yield the A and B parameters indicated in the figure .. 48 Figure 4.1 D photofragment yield spectrum in the region 19000 cm -1 – 21500 cm -1 . Background signal is subtracted, and the signal normalized to the OPO/OPA laser energy ........................................... 69 Figure 4.2 (top) 2+1 REMPI of CH 2 OH in the region of absorption to the 3p z state (adapted from ref. 30) (bottom) 2+2 REMPI of CH 3 CHOH in the region of absorption to the 3p z state. The lowest energy band of each transition is the origin band ......................................................... 71 Figure 4.3 H fragment time-of-flight spectrum from CH 3 CHOH obtained at 21,276 cm -1 excitation. The polarization of the pump laser is alternated between parallel (solid line) and perpendicular (dashed line) to the TOF axis. Background is subtracted. Zero time indicates fragment with no recoil ........................................................ 73 Figure 4.4 The c.m. E t distribution obtained by monitoring H photofragments following the 1 2 A(3s) ! 1 2 A transition. at 21,276 cm -1 .(2.63 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel .......................................... 75 Figure 4.5 The c.m. E t distribution obtained by monitoring H photofragments following excitation of CH 3 CHOH at 31,250 cm -1 (3.87 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel. .......................................................... 77 Figure 4.6. The c.m. E t distribution obtained by monitoring D photofragments following excitation of CH 3 CHOD at 31,250 cm -1 (3.87 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel ........................................................... 79 Figure 4.7 The c.m. E t distribution obtained by monitoring H photofragments following excitation of CH 3 CHOH at 35,460 cm -1 (4.39 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel ........................................................... 81 Figure 4.8. The c.m. E t distribution obtained by monitoring D photofragments following excitation of CH 3 CHOD at 35,460 cm -1 (4.39 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel ........................................................... 83 ix ! Figure 4.9 Recoil anisotropy parameter " as a function of excitation energy ...... 85 Figure 5.1 Equilibrium structures and nuclear repulsion energies (V NN ) of the CH 2 OH and CH 3 CHOH radicals and their cations optimized at the CCSD(T)/cc-pVTZ level of theory. Angles listed in parentheses correspond to the neutral radicals ........................................................ 96 Figure 5.2 Three MOs resulting from hyperconjugation of # CH , LP(O), and LP(C) in CH 3 CHOH at the cation and radical geometries. The bonding and nonbonding MOs are doubly occupied in both the cation and the neutral. The antibonding orbital is singly occupied in the radical and is unoccupied in the cation. At the neutral geometry (right), the shape of the orbitals is slightly different, but their character is preserved. For the sake of clarity, we refer to these three orbitals as SOMO, HOMO, and HOMO-1 of the radical, although energetically there are other MOs (which do not participate in hyperconjugation) in between the nonbonding and bonding MOs ....................................................................................... 98 Figure 5.3. Frontier MOs of hydroxymethyl. Oxygen lone pair, LP(O), and LP(C) form bonding and antibonding $-like orbitals hosting three electrons in the radical ....................................................................... 103 Figure 5.4 Hyperconjugation energies of the three CH bonds for various orientations of the CH 3 group in CH 3 CHOH + at the cation (upper panel) and neutral (lower panel) geometries ..................................... 107 Figure 6.1 Energies, barrier heights and product channels (kcal/mol) relevant to the OH + C 2 H 4 reaction ................................................................. 116 Figure 6.2 Equilibrium structures of CH 2 CH 2 OH, potential energies of the conformers relative to the most stable conformer and symmetries calculated at CCSD/6-311+G(d,p) level of theory ............................ 120 Figure 6.3 Vertical excitation energies for the first three electronic states of CH 2 CH 2 OH in the cation equilibrium geometries (cyclic and CH 3 CHOH + ) and neutral equilibrium geometries (A,B and C ), calculated at CCSD/6-311(2+,+)G(d,p) level of theory .................... 122 ! x ! Abstract The predissociation of vibrationally excited hydroxymethyl radical and the ionization spectroscopy and the photodissociation dynamics of 1-hydroxyethyl radicals from excited Rydberg states are described. The OH-stretch overtone spectroscopy and dynamics of the hydroxymethyl radical, CH 2 OH, are reported in the region of the second and third overtones. The second overtone spectrum at 10484 cm –1 is obtained by double resonance IR-UV resonance enhanced multiphoton ionization (REMPI) spectroscopy via the 3p z electronic state. The third overtone spectra of CH 2 OH and CD 2 OH are observed at ~13600 cm –1 by monitoring H-atom photofragments while scanning the excitation laser frequency. Dissociation via tunneling is proposed. No isomerization to methoxy is observed. The electronic spectroscopy and photodissociation dynamics of the CH 3 CHOH radical in the region 19400%37000 cm -1 were studied in a molecular beam using resonance-enhanced multiphoton ionization (REMPI), photofragment yield spectroscopy, and time-of-flight (TOF) spectra of H and D fragments. The onset of the transition to the Rydberg 3s state, the lowest excited state, is estimated at 19600. The 3s state dissociates fast, and no REMPI spectrum is observed. The origin band of the transition to the 3p z state, identified by 2+2 REMPI, lies at 32360 cm -1 By comparison of the TOF distributions of the isotopologs CH 3 CHOH, CH 3 CHOD, and CD 3 CHOH, it is concluded that two major product channels dominate xi ! the photodissociation, one leading to acetaldehyde and the other to vinyl alcohol (enol) products. There is no indication of isomerization to ethoxy. On the basis of electronic structure calculations, we offer a physical explanation of the observed large decrease (0.9 eV) in ionization energies (IE) in going from hydroxymethyl to hydroxyethyl radical. The effect is attributed to hyperconjugative interactions between the # CH orbitals of the methyl group in hydroxyethyl, the singly occupied p orbital of carbon, and the lone pair p orbital of oxygen. Analyses of vertical and adiabatic IEs and hyperconjugation energies computed by the natural bond orbital (NBO) procedure reveal that the decrease is due to the destabilization of the singly occupied molecular orbital in hydroxyethyl radical as well as structural relaxation of the cation maximizing the hyperconjugative interactions. 1 ! Chapter 1 Introduction 1.1 Overview Hydroxyalkyl radicals play an important role many atmospheric and combustion reactions such as: reaction of alcohols with OH[1] and halogen atoms [2- 9]; reaction of O( 1 D) with methane [10-14], intermediates in O + C 2 H 5 and OH + C 2 H 4 reactions [15-18]. Hydroxyethyl radicals can be found in interstellar space [19] as well as in mouse or human liver [20]. Hydroxyethyl radicals are also precursors of the enol tautomer of acetaldehyde – important intermediates formed in flames of ethanol, olefins and commercial fuels [21]. 1.2 Hydroxymethyl radical CH 2 OH can be used as a model to predict the dynamics of larger hydroxyalkyls with a radical center on the carbon next to the oxygen atom. For example, aspects of hydroxymethyl applicable to other hydroxyalkyl radicals (such as CH 3 CHOH) can be the nature of electronically excited states and dynamics on the excited state potential energy surface (PES). A short summary of hydroxymethyl properties is presented here. According to ab initio calculations, the ground state of CH 2 OH is nonplanar, belonging to the C 1 symmetry point group [22]. The singly occupied molecular orbital 2 ! (SOMO) has $ * shape with a node between C and O atoms which gives it an antibonding character [23]. Upon electronic excitation from the SOMO to a non- bonding (Rydberg) orbital or the ionization continuum, the C-O bond shortens by ~ 0.1 & and the radical becomes planar with C s symmetry [22,23]. Electronic excited states of CH 2 OH were studied extensively both experimentally [24-29] and theoretically [30-33]. The lowest electronic state is a Rydberg 3s state. It is coupled to the ground state through a conical intersection located in the region of an elongated O-H bond [33]. Conversion to the ground state with subsequent dissociation is so efficient that no resonance-enhanced multiphoton ionization (REMPI) signal is observed for this state [26,28]. The next optically accessible electronic states are Rydberg 3p x and 3p z states (here z axis is perpendicular to the C-O-H plane and x axis is lying along the C-O bond), both having access to the ground state through a seam of conical intersections. While it is possible to observe REMPI signals for both 3p states, the REMPI spectrum of 3p z has narrower lines than the similar spectrum of the 3p x state [24,25,28,29], indicating that the coupling to the 3s state and eventually to the ground state is more efficient for the 3p x state than for 3p z . The dynamics on the ground state is itself an interesting question. For example, the isomerization of CH 2 OH to CH 3 O is an important process in atmospheric chemistry as the reaction of hydroxymethyl with O 2 is almost 3000 times faster than the reaction of methoxy with O 2 [34-36]. The barrier to isomerization on the ground 3 ! state potential energy surface (PES) is calculated to be slightly lower than the barrier to O-H bond breaking. According to ab initio calculations the isomerization barrier is ~ 14000 cm -1 [36-38] while the direct dissociation barrier is ~ 16000 cm -1 [34,36]. After isomerization, CH 3 O has enough internal energy to surmount the barrier to H + CH 2 O on the ground state. Competition is possible between two dissociation pathways and it depends on the relative height of the barriers which is difficult to calculate with the required accuracy. Hence, experimental study is necessary. In Chapter 3, the predissociation of CH 2 OH on the ground state PES is studied by overtone OH-stretch excitation. A similar approach has been used to access highly vibrationally-excited levels of the ground electronic states of stable molecules such as HOOH, NH 2 OH and CH 3 OH [39-43]. For these molecules, the OH bond energy is much higher than the lowest dissociation energy and the dissociation coordinate is not involved in the excitation. Therefore, high overtone excitation of the OH-stretch can be achieved in a potential that behaves as a local-mode Morse oscillator. In CH 2 OH on the other hand, the O-H bond is the reaction coordinate to H + CH 2 O and the dissociation energy is low. Thus, it is not clear how high the OH-stretch vibration can be optically excited. In fact, I am not aware of any other study in which the dissociation coordinate was excited directly. 4 ! 1.3 1-Hydroxyethyl radical The next homolog of hydroxymethyl is the 1-hydroxyethyl radical (CH 3 CHOH), whose the radical center is on the carbon next to the oxygen, similar to CH 2 OH. In Chapters 4 and 5 the following questions are addressed: how addition of the methyl group changes the dynamics of hydroxyethyl relative to hydroxymethyl and what aspects are similar for both radicals? CH 3 CHOH has twice as many vibrational degrees of freedom as CH 2 OH (18 and 9 respectively). The added vibrations increase significantly the density of states, which facilitates internal vibrational energy redistribution (IVR). Faster IVR rates may lead to more efficient coupling between electronic states through conical intersections, if the new vibrations induce modes that promote the conical intersection. The geometry and molecular orbitals of the 1-hydroxyethyl radical are in many ways similar to the hydroxymethyl case. The ground state of 1-hydroxyethyl radical has the C 1 symmetry point group [23,44]. The SOMO has nodes between the C and O atoms as well as between the two carbon atoms making it an antibonding orbital for both C-O and C-C bonds [23]. Thus, in the cation both bonds are shorter than in the neutral radical. Similar to hydroxymethyl ion, CH 3 CHOH + has C s symmetry [23,44,45]. The ionization energy (IE) of 1-hydroxyethyl is 6.64 eV [8,46]which is almost 0.9 eV lower than the IE of CH 2 OH (7.56 eV [47]). Corresponding values for ethanol and methanol are 10.85 and 10.41 eV, respectively [48-51]. Typically, the ionization 5 ! energy is inversely proportional to molecular size. If the difference in IE for alcohols (0.5 eV) can be explained as effect of size increase, the larger reduction in IE for hydroxyalkyl radicals (0.9 eV) indicates that other effects are involved such as additional stabilization of CH 3 CHOH + compared to CH 2 OH + . Stabilization of cations by addition of methyl groups can be caused by the well-known effect of hyperconjugation. Hyperconjugation is a concept used to describe conjugation effects of #-bonds with $-bonds or p-lone pairs. Similar to conjugation between double or triple bonds, hyperconjugation involves delocalization of charge density over several atoms or groups of atoms. Hyperconjugation is responsible for the stability of secondary and tertiary radicals and cations [52], the changes in bond strength [53] and conformations [54] upon substitutions, the vibrational spectra and structures of hydrocarbon radicals[55,56], and more. In CH 3 CHOH this effect appears as interaction of one or several of its # CH orbitals with p lone pairs on C and O. The exact number of interacting orbitals depends on the orientation of the methyl group relative to the rest of the molecule [23]. The combination of these orbitals creates bonding, non-bonding and anti-bonding molecular orbitals. The anti-bonding combination lies higher in energy than any fragment orbital and is the SOMO of the radical. According to Koopmans’ theorem, the vertical IE is proportional to the energy of the HOMO (or SOMO for radicals). Destabilization of the SOMO by hyperconjugation decreases the vertical IE of hydroxyethyl relative to hydroxymethyl. 6 ! Little is known about the electronically excited states of CH 3 CHOH. There are no theoretical studies on the excited states of hydroxyethyl radical and there are only a few experimental studies. The absorption spectrum of CH 3 CHOH taken at room temperature in the region 230-300 nm is structureless [3]. An observed resonance- enhanced multiphoton ionization (REMPI) signal of the CH 3 CHOH ion recorded in the range 430-460 nm is likely the result of from 2+1 ionization via several Rydberg states [57], though particular states are observed and assigned in the reported spectra. Unlike the valence states, the positions of the Rydberg states can be estimated from the Rydberg formula: ! "#$%&'( )*!+, - ./+01 2 , where IE is the ionization energy of the molecule, n is the principal quantum number of the state, # is the quantum defect of the state and R is the Rydberg constant. Quantum defects from the corresponding states of CH 2 OH can be used to predict the positions of the respective Rydberg states of CH 3 CHOH. Due to low IE of 1- hydroxyethyl, it is reasonable to assume that the first excited electronic states will be Rydberg similar to the case of hydroxymethyl. As the CH 3 CHOH + ion is stable [23,44,45], the Rydberg states are expected to be bound. Dissociation on the Rydberg state PES would lead to products in Rydberg states. As radical dissociation products are usually closed-shell molecules, their Rydberg states lie high in energy, which makes such dissociation channels 7 ! inaccessible for low excitation energies. Thus the major dissociation pathway will be internal conversion to the ground electronic state followed by dissociation. The lowest-barrier channels calculated for the ground state PES are [18,58]: CH 3 CHOH ' CH 3 CHO + H (H = 1.04 eV ' CH 2 CHOH + H (H = 1.46 eV ' CH 3 + CHOH (H = 3.13 eV ' CH 3 COH + H (H = 3.37 eV Certain dissociation pathways like CH 3 CO + H 2 and CH 3 + CO + H 2 have high barriers and tight transition states which may make them less competitive with simple bond fission channels, as in CH 2 OH [26,27]. Isomerization to ethoxy radical with the barrier of 1.61 eV from hydroxyethyl can lead to dissociation to CH 2 O + )H 3 ((H = 0.85 eV) or CH 3 CHO + H [18]. ! 8 ! Chapter 1 References (1) Grosjean, D. J Brazil Chem Soc 1997, 8, 433. (2) Ahmed, M.; Peterka, D. S.; Suits, A. G. Physical Chemistry Chemical Physics 2000, 2, 861. (3) Anastasi, C.; Simpson, V.; Munk, J.; Pagsberg, P. Chemical Physics Letters 1989, 164, 18. (4) Anastasi, C.; Simpson, V.; Munk, J.; Pagsberg, P. Journal of Physical Chemistry 1990, 94, 6327. (5) Pagsberg, P.; Munk, J.; Anastasi, C.; Simpson, V. J. Journal of Physical Chemistry 1989, 93, 5162. (6) Pagsberg, P.; Munk, J.; Sillesen, A.; Anastasi, C. Chemical Physics Letters 1988, 146, 375. (7) Rudic, S.; Murray, C.; Ascenzi, D.; Anderson, H.; Harvey, J. N.; Orr- Ewing, A. J. Journal of Chemical Physics 2002, 117, 5692. (8) Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1994, 101, 10936. (9) Taatjes, C. A.; Christensen, L. K.; Hurley, M. D.; Wallington, T. J. Journal of Physical Chemistry A 1999, 103, 9805. (10) Fockenberg, C.; Hall, G. E.; Preses, J. M.; Sears, T. J.; Muckerman, J. T. Journal of Physical Chemistry A 1999, 103, 5722. (11) Lin, J. J.; Harich, S.; Lee, Y. T.; Yang, X. Journal of Chemical Physics 1999, 110, 10821. (12) Lin, J. J.; Shu, J.; Lee, Y. T.; Yang, X. Journal of Chemical Physics 2000, 113, 5287. (13) Marcy, T. P.; Diaz, R. R.; Heard, D.; Leone, S. R.; Harding, L. B.; Klippenstein, S. J. Journal of Physical Chemistry A 2001, 105, 8361. (14) Seakins, P. W.; Leone, S. R. Journal of Physical Chemistry 1992, 96, 4478. 9 ! (15) Cleary, P. A.; Romero, M. T. B.; Blitz, M. A.; Heard, D. E.; Pilling, M. J.; Seakins, P. W.; Wang, L. Physical Chemistry Chemical Physics 2006, 8, 5633. (16) Hoyermann, K.; Olzmann, M.; Seeba, J.; Viskolcz, B. Journal of Physical Chemistry A 1999, 103, 5692. (17) Lindner, J.; Loomis, R. A.; Klaassen, J. J.; Leone, S. R. Journal of Chemical Physics 1998, 108, 1944. (18) Senosiain, J. P.; Klippenstein, S. J.; Miller, J. A. Journal of Physical Chemistry A 2006, 110, 6960. (19) Fairley, D. A.; Scott, G. B. I.; Freeman, C. G.; Maclagan, R. G. A. R.; McEwan, M. J. Journal of the Chemical Society-Faraday Transactions 1996, 92, 1305. (20) Moncada, C.; Torres, V.; Varghese, G.; Albano, E.; Israel, Y. Mol Pharmacol 1994, 46, 786. (21) Taatjes, C. A.; Hansen, N.; McIlroy, A.; Miller, J. A.; Senosiain, J. P.; Klippenstein, S. J.; Qi, F.; Sheng, L. S.; Zhang, Y. W.; Cool, T. A.; Wang, J.; Westmoreland, P. R.; Law, M. E.; Kasper, T.; Kohse-Hoinghaus, K. Science 2005, 308, 1887. (22) Johnson, R. D.; Hudgens, J. W. Journal of Physical Chemistry 1996, 100, 19874. (23) Karpichev, B.; Reisler, H.; Krylov, A. I.; Diri, K. Journal of Physical Chemistry A 2008, 112, 9965. (24) Aristov, V.; Conroy, D.; Reisler, H. Chemical Physics Letters 2000, 318, 393. (25) Conroy, D.; Aristov, V.; Feng, L.; Reisler, H. Journal of Physical Chemistry A 2000, 104, 10288. (26) Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2003, 118, 9623. (27) Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2004, 120, 6524. (28) Feng, L.; Huang, X.; Reisler, H. Journal of Chemical Physics 2002, 117, 4820. 10 ! (29) Feng, L.; Reisler, H. Journal of Physical Chemistry A 2004, 108, 9847. (30) Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 1998, 102, 3141. (31) Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 2001, 105, 8599. (32) Hoffman, B. C.; Yarkony, D. R. Journal of Chemical Physics 2002, 116, 8300. (33) Yarkony, D. R. Journal of Chemical Physics 2005, 122. (34) Adams, G. F.; Bartlett, R. J.; Purvis, G. D. Chemical Physics Letters 1982, 87, 311. (35) Radford, H. E. Chemical Physics Letters 1980, 71, 195. (36) Saebo, S.; Radom, L.; Schaefer, H. F. Journal of Chemical Physics 1983, 78, 845. (37) Hippler, H.; Striebel, F.; Viskolcz, B. Physical Chemistry Chemical Physics 2001, 3, 2450. (38) Walch, S. P. Journal of Chemical Physics 1993, 98, 3076. (39) Boyarkin, O. V.; Rizzo, T. R.; Perry, D. S. Journal of Chemical Physics 1999, 110, 11346. (40) Kuhn, B.; Boyarkin, O. V.; Rizzo, T. R. Berichte der Bunsen- Gesellschaft für Physikalische Chemie 1997, 101, 339. (41) Kuhn, B.; Rizzo, T. R. Journal of Chemical Physics 2000, 112, 7461. (42) Phillips, J. A.; Orlando, J. J.; Tyndall, G. S.; Vaida, V. Chemical Physics Letters 1998, 296, 377. (43) Scott, J. L.; Luckhaus, D.; Brown, S. S.; Crim, F. F. Journal of Chemical Physics 1995, 102, 675. (44) Curtiss, L. A.; Lucas, D. J.; Pople, J. A. Journal of Chemical Physics 1995, 102, 3292. 11 ! (45) Nobes, R. H.; Rodwell, W. R.; Bouma, W. J.; Radom, L. Journal of the American Chemical Society 1981, 103, 1913. (46) Dyke, J. M.; Groves, A. P.; Lee, E. P. F.; Niavaran, M. H. Z. Journal of Physical Chemistry A 1997, 101, 373. (47) Dyke, J. M.; Ellis, A. R.; Jonathan, N.; Keddar, N.; Morris, A. Chemical Physics Letters 1984, 111, 207. (48) Holmes, J. L.; Lossing, F. P. Organic Mass Spectrometry 1991, 26, 537. (49) Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1991, 95, 4033. (50) Tao, W.; Klemm, R. B.; Nesbitt, F. L.; Stief, L. J. Journal of Physical Chemistry 1992, 96, 104. (51) Williams, J. M.; Hamill, W. H. Journal of Chemical Physics 1968, 49, 4467. (52) Muller, N.; Mulliken, R. S. Journal of the American Chemical Society 1958, 80, 3489. (53) Schneider, W. F.; Nance, B. I.; Wallington, T. J. Journal of the American Chemical Society 1995, 117, 478. (54) Hoffmann, R.; Hehre, W. J.; Salem, L.; Schleyer, P. V.; Pople, J. A.; Radom, L. Journal of the American Chemical Society 1972, 94, 6221. (55) Koziol, L.; Levchenko, S. V.; Krylov, A. I. Journal of Physical Chemistry A 2006, 110, 2746. (56) Haber, T.; Blair, A. C.; Nesbitt, D. J.; Schuder, M. D. Journal of Chemical Physics 2006, 124. (57) Edelbuttel-Einhaus, J.; Hoyermann, K.; Rohde, G.; Seeba, J. Proceedings of the 24th Symposium International on Combustion, Combustion Institute, Pittsburgh, PA 1992, 661. (58) Klippenstein, S, private communication. 12! ! Chapter 2 Experimental Details 2.1 Overview All the experiments described in this Dissertation were performed in the molecular beam apparatus shown schematically in Figure 2.1. The source chamber contains a piezoelectric-driven nozzle and is separated from the detection chamber by a flange with a skimmer (Beam Dynamics 1.51 mm). The detection chamber has a custom-built Wiley-McLaren time-of-flight mass spectrometer mounted vertically [1, 2]. The source and detection chambers are evacuated by Leybold TMP1000C and Leybold TMP361 turbomolecular pumps to a base pressure of 1·10 -7 Torr and 1·10 -8 Torr respectively. The pressure is measured with an ion gauge glass tube (Duniway I- 75-N) connected with a Philips-style cable to a controller (Granville Phillips 270). In a typical experiment radicals produced in the source chamber are excited to a specific state by a pump laser in the detection chamber. The outcome of the excitation is detected by ionization of the radicals or their fragments by a probe laser. Ions are accelerated and focused on a microchannel plate (MCP) detector by a charged particle optics system. The signal from the MCP goes through a fast current preamplifier (KOTA E105) to the oscilloscope (Tektronics TDS640A), which is connected by a GPIB interface to a computer running a Labview program. The same program controls the laser wavelengths via the serial port. The repetition rate of the 13! ! Figure 2.1 Schematic diagram of the molecular beam apparatus (side view): 1) Source chamber 2) Detection chamber 3) Piezoelectric nozzle 4) Ion optic system 5) Field-free drif region 6) MCP detector ! 6 5 4 gas mixture in to turbomolecular pumps to oscilloscope 1 3 2 14! ! pulsed nozzle and lasers is 10Hz. Two delay generators (Stanford Research Systems DG535) synchronize the firing of the lasers, the opening of the pulsed valve and data acquisition. Delays can be also controlled from the Labview program through a separate GPIB interface. 2.2 Radical Production Ethanol (AAPER Alcohol, used without further purification) or methanol (Aldrich, used without further purification) vapors are mixed with 5% Cl 2 /He mixture in a glass manifold outside the chamber. Typical concentrations are 3% methanol or 2.6% ethanol and 1% Cl 2 in 2 atmospheres of helium for CH 2 OH or CH 3 CHOH production, respectively. Partially deuterated alcohols (Aldrich, used without further purification) are used for CH 2 OD, CD 2 OH, CH 3 CHOD, CD 3 CHOH production. The mixture is introduced into the source chamber through a pulsed piezoelectric driven nozzle schematically shown in Figure 2.2 The initial plunger position relative to the opening as well as the voltage applied to the piezoelectric actuator controls the amount of gas administered to the chamber in each pulse. The plunger is adjusted in such a way that the pressure in the detection chamber is 2·10 -7 Torr when the voltage applied to the piezoelectric actuator is between 300V and 400V. In such a plunger position it is typical to have a small leak through the nozzle. The leak should not increase the base pressure in the detection 15! ! Figure 2.2 Schematic diagram of the piezoelectric nozzle. ! 16! ! chamber by more than 2·10 -8 Torr. In a normal regime, the nozzle requires adjustment every month and the piezoelectric actuator has a lifetime of about a year. Cl 2 molecules are dissociated in a quartz tube attached to the end of the nozzle by 355 nm laser light (3rd harmonic of Quanta-Ray GCR11 Nd:YAG laser, 5mJ, focused with cylindrical 30 cm focal length lens): Cl 2 34 5 62Cl. Chlorine atoms react with methanol or ethanol molecules producing CH 2 OH or CH 3 CHOH, respectively with 95% yield[3-5]: C 3 CH 2 OH + Cl ' CH 3 CHOH + HCl CH 3 OH + Cl ' CH 2 OH + HCl. The position of the laser beam along the quartz tube is very important. Radicals created far from the tube exit can be destroyed in secondary reactions with chlorine atoms or other radicals. The optimal position for the laser beam is at the tip of the tube where only half of the laser light goes through the tube. The radical signal is not sensitive to small movements of the photolysis laser beam or its focus so after initial setup (laser beam intersects the quartz tube as described above, focus is slightly behind the tube) no further adjustment is necessary. In supersonic expansion, radicals rapidly cool down by collisions with helium atoms. The rotational temperature for CH 2 OH after the expansion is 10-15K [1, 6]. A similar rotational temperature is expected for CH 3 CHOH as the expansion parameters 17! ! are the same. However, due to the lack of rotational line resolution it is impossible to measure the rotational temperature directly. 2.3 Laser system and product detection In the detection chamber the molecular beam is intersected at right angle by two counter-propagating laser beams. The output of a Continuum Sunlite OPO/OPA system pumped by the 3 rd harmonic of seeded Continuum Powerlite 8000 Nd:YAG laser is used to excite radicals to a specific electronic or vibrational state. After excitation either the radical itself or a product of dissociation are ionized with a probe laser. The probe laser is not required in the case of resonance enhanced multiphoton ionization (REMPI) where, upon excitation to a Rydberg state, the radical is ionized with a same photon from the pump laser. H/D detection. The most common case in this Dissertation is dissociation of the radical into a molecular fragment and hydrogen or deuterium atom. Atoms are ionized in a 1+1’ REMPI scheme where the first step is the one-photon transition to the 2s state with 121.6 nm light (Lyman-* line) followed by excitation to the ionization continuum with 365.4 nm light. The Lyman-* line is obtained by the tripling of 365.4 nm light in a gas cell. The ~728 nm output of a dye laser (Continuum ND6000) pumped by the 2 nd harmonic of Nd:YAG laser (Continuum NY81) is doubled by a KDP crystal and focused inside the tripling cell with a 25-cm focal length lens. The cell is filled with krypton (partial pressure 200 Torr) and argon 18! ! (partial pressure ~740 Torr). The final pressure ratio of two gases is adjusted to maximize the H + /D + signal. The tripling cell is attached to the detection chamber as 121.6 nm radiation is readily absorbed by air. A short focal length MgF 2 lens (9.8 cm f.l.) at the exit of the cell focuses the 121.6 nm light at the molecular beam inside the detection chamber. This setup is sensitive to the alignment of the probe laser beam and gas mixture in the tripling cell. Signal from background hydrogen is much stronger than from background deuterium, so even if the deuterium detection is required it is advisable to establish hydrogen detection first. The following procedure to find the H + signal was used: 1. Align the probe laser beam to go exactly through the center of the tripling cell and the vacuum chamber window on the opposite side. A 25 cm f.l. lens must be placed approximately 10 cm before the cell 2. Thoroughly pump the cell down to 100 mTorr. 3. Fill the cell with 200 Torr of Kr gas. 4. Slowly fill the cell with Ar. If the laser wavelength after doubling is not far from 365.4 nm (wavelength calibration with a spectrometer is required sometimes), signal from the scattered VUV light should appear on the oscilloscope. Maximize the signal by adding more argon. If a signal for the scattered VUV light is present but H + signal is not, then it is necessary to change the wavelength. The optimum Kr/Ar ratio is wavelength dependent so it must be adjusted accordingly. 19! ! 5. Adjust the probe laser alignment vertically and horizontally to maximize H + signal. 6. If deuterium detection is required then decrease wavelength of the dye laser by 0.2 nm. 7. Scan the hydrogen (deuterium) Doppler profile. Set the wavelength at the center of the profile. Overtone spectroscopy. Unlike hydroxyethyl, the hydroxymethyl radical can be selectively detected by 1+1 (or 2+2) REMPI via the origin band of the 2 2 A”(3p z )+1 2 A” transition. The overtone spectrum can be recorded by the depletion technique, where the probe laser selectively ionizes molecules in the vibrational ground state via the 3p z level at 41062 cm -1 and the pump laser excites rovibrational levels in the electronic ground state. If the wavelength of the pump laser matches a resonance in the rovibrational level, the population of the radicals in the vibrational ground state decreases leading to depletion of the REMPI signal. The advantage of this method is that no a priori knowledge of the energy position of the vibrational level is required and only the pump laser wavelength is scanned. The disadvantage is the low signal-to-noise (S/N) ratio due to instability of the REMPI signal intensity. Better S/N ratio and resolution were obtained in the double resonance ionization detection (DRID) method. Here after excitation to the rovibrational level of the ground electronic state by the pump laser, the molecule is ionized by the probe laser via resonance excitation to a selected vibrational level of the 3p z state that has a suitable 20! ! Franck-Condon factor. For example, in the study of the second overtone of the OH stretch of CH 2 OH, ionization via transition to the 3" 1 level of the 3p z state was used. The exact wavelength of the probe laser for this transition was found during a scan where the wavelength of the pump laser was fixed at the maximum depletion found using the previous method. In order to obtain the overtone spectrum using DRID, the frequencies of both lasers were adjusted during the scan so that the sum of the frequencies was kept constant. In this case the CH 2 OH + signal appeared only if the frequency of the pump laser was in resonance with a transition to the rovibrational level, and the frequency of the probe laser is too small to ionize resonantly radicals in the ground vibrational state. In a typical DRID experiment two spectra were obtained “pump-on” and “pump-off” with the resulting overtone excitation spectrum being the difference of these two. 2.4 Time-of-flight The ion optics assembly is comprised of three plates: extractor, repeller and accelerator. The assembly and time-of-flight tube with a field-free region are oriented vertically with the MCP at the top, so the TOF axis is perpendicular to the molecular beam and laser plane. Ions are created between the extractor and repeller plates. The purpose of this system is to focus ions with same speed along the TOF axis at the multichannel plate (MCP) detector. Ions formed farther from the exit will attain a higher speed than ions created closer to the exit. When the space focusing 21! ! condition is met, the difference in speed is such that ions formed at different positions arrive at the detector simultaneously. The focusing condition dictates a specific ratio between the extractor and repeller voltages. The electric field between the repeller and accelerator plates is used to move the focal plane farther from the ion optic assembly, which allows a field free region between the accelerator plate (which is grounded) and the MCP (the first plate of the MCP assembly is also grounded). In mass spectrometer mode a high voltage is applied to the extractor and repeller plates (3000V and 2000V, respectively). This creates a strong electric field in the region that rapidly accelerates ions to a speed that depends only on the ion mass- to-charge ratio and not on the initial velocity. After flying out through the drift region, ions with different masses are separated in their time of arrival to the MCP detector. If the time of arrival is known, the mass can be deduced from the following equation: 7 )89.:+: ; 1 2 (1) where m is the mass of the ion in amu, t is the time of arrival in ns, t 0 is the time of the laser pulse which corresponds to the position of the scattered laser light peak and b is a coefficient obtained from the mass spectrometer calibration. It is necessary to perform a calibration before each measurement, because small deviations in the laser beam position and voltages can give an error of ±1 amu at 40 amu. Fortunately, calibration is easy as ultraviolet laser light readily ionizes hydrocarbon molecules present in the background via multiphoton absorption. This process produces several easily recognizable progression of peaks such as C, CH, CH 2 , CH 3 and C 2 ,C 2 H, C 2 H 2 , C 2 H 3 22! ! and so on. Several peaks with known arrival times and ion masses are required in order to get an accurate a value. In the so called time-of-flight mode, the time of arrival to the detector depends on the projection of the initial ion velocity on the TOF axis. In this mode the voltage on the extractor/repeller plates is kept low (for most TOF experiments it were 320V/246V for repeller/extractor, correspondingly) so ions that initially were flying toward the detector will arrive noticeably earlier than those that were flying initially away from the detector. The distribution of ions in time can therefore be used to get information about their speed and translational energy distribution. The projection of the initial ion velocity on the TOF axis (v TOF ) is linearly dependent on its time of arrival to the detector: < =>? )@.:+: ; 1. Here a is parameter that depends on the field and the mass of the fragment and t 0 is the time of arrival of particles with zero initial velocity. Under space focusing conditions a can be calculated: @) A 7 BC D , Where q is the electric charge of the particle, m is particle mass, (V is voltage difference between repeller and extractor plates, and l is the distance between the plates which is 1 cm or 0.01 m. Unfortunately, the voltage difference measured outside of the chamber provides only approximate estimation of the electric field between the plates. Hence, for accurate measurements of the particle speed, calibration 23! ! is required. Calibration is performed by measuring the TOF spectrum from photodissociation of a molecule with a known kinetic energy distribution of the photofragments under the same conditions at which the main experiments are carried (same voltage, alignment etc) In general, diatomic molecules (for instance HBr, DBr) provide better calibration, as the excess energy in the photodissociation process is partitioned between kinetic energy of the fragments and well-defined atomic levels of one fragment. However, it is more important to select for calibration a molecule that has kinetic energy distribution of photofragments approximately in the same range as the radical under investigation. In such a case, a polyatomic molecules should be used with distinctly resolved peaks in the translational energy distribution. For example in the 1-hydroxyethyl experiments, photodissociation of the hydroxymethyl radical via excitation to the 3s state was used for calibration. The TOF trace I $ (v TOF ) in velocity space is linearly proportional to the observed TOF trace I $ (t) in time space. Here $ is the angle between the TOF axis and the pump laser polarization vector. It is an important parameter as photodissociation can be anisotropic in space with respect to the laser polarization, which means that two different laser polarizations relative to the TOF axis can produce different TOF traces from the same dissociation process. One can calculate I $ (v TOF ) if the velocity distribution is known from the following equation[7]: * E .< =>? 1),F G H IJKL.<1M 2 .NOPQ1RM.<1S< TH UVW X YH Z X H UVW (2) < ' )[ : ; \ 24! ! where "(v) is the recoil anisotropy parameter at a given velocity v, P 2 is a second Legendre polynomial, P(v) is the angle-averaged speed distribution, r is the radius of the MCP detector and % is the recoil angle with respect to the pump laser polarization vector. % can be represented as a sum of $ and % TOF , where % TOF is the angle between the ion velocity vector and the TOF axis. As a result, M 2 .NOPQ1)M 2 .NOPQ =>? 1M 2 .NOP]1. For $ = 0 (laser polarized parallel to the TOF axis): * ; .< =>? 1),F G H IJKL.<1M 2 .NOPQ =>? 1RM.<1S< TH UVW X YH Z X H UVW (3) for $ = $/2 (laser polarized perpendicular to the TOF axis): *^ X .< =>? 1),F G H _J+ `.H1 2 M 2 .NOPQ =>? 1aM.<1S< TH UVW X YH Z X H UVW (4) The anisotropy parameter " can be obtained from the TOF traces[7] (under core-sampling condition, described below): L.<1)b c d .H UVW 1ec^ X .H UVW 1 c d .H UVW 1Y2c^ X .H UVW 1 (5) " is an intrinsic parameter of the dissociation mechanism. When several dissociation pathways compete, the observed anisotropy is a weighted average of " for each pathway. If v r is small relatively to v TOF , it possibly to obtain P(v) directly from (3),(4) with the approximate equation [7]: M.<1)fg/h:,< i 2 .* j .< =>? 1Kb*^ X .< =>? 11 (6) 25! ! Empirically, the ratio of v TOF /v r = 4 can be used as the lower limit at which the approximation (6) is still valid. v r can be made smaller by either decreasing the voltage, so the electric field is weaker and t 0 is increased, or by decreasing the effective detector size r with an aperture installed in front of the MCP. The latter approach, called core-sampling or core-extraction, is used to obtain high resolution TOF traces, which can be converted straightforwardly to kinetic energy distributions. The major downside of this method is that a large fraction of the produced ions does not reach the detector, and the recorded ion signal may become unacceptably low for certain experiments. Also if the photodissociation process creates a considerable number of slow ions, equation (6) can become inapplicable even with the core installed. In general, if there is a noticeable amount of ions with v TOF < 4v r the speed distribution can be obtained only through simulation. First, some model velocity distribution is assumed. Then, the speed distribution can be obtained from the formula (for isotropic distribution): M kl&&$ .<1S< )m m M H&njopq# .<1 2r ; r ; S<,<Ss,<PtuQSQ )vw< 2 M H&njopq# .<1S< In a typical iteration, TOF traces are calculated from the speed distribution using equations (3) and (4). The calculated traces are compared with the experimental results and parameters of the model are adjusted for the next iteration cycle, until a good fit is obtained 26! ! In most experiments a translational energy distribution is more important for interpretation. The speed distribution can be converted to a translational energy distribution of the fragment using the following equation: M kl&&$ .<1S< )M q'xyk .!1S! M q'xyk I!.<1R) J < M kl&&$ .<1)vw<M H&njopq# .<1 M q'xyk .!1)vw z b! 7 M H&njopq# { z b! 7 | While the TOF traces and speed distributions have inherently different shapes, they have the same width. In certain cases only the maximum translational energy is needed and the exact distribution is not required. The maximum velocity of one fragment can be obtained directly from a TOF trace: < }x~ )@9 =>? b where T TOF is full width the TOF trace. The maximum translation energy of the fragment is 'x(}&yqG ) } Z H X 2 . The center-of-mass (c.m.) total translational energy of the system: qjqxn ) 'x(}&yqG K 'x(}&yq2 ) } Z Y} ZX } ZX 'x(}&yqG (6) TOF traces can be used to identify multiple dissociation pathways when hydrogen atoms is produced in the fission of different bonds. Each dissociation process can be studied separately by recording D + TOF spectra of partially deuterated radicals, in which certain hydrogen atoms are substituted with deuterium. The D + TOF spectra from such partially deuterated radicals will have distinctively different shapes 27! ! caused by different dynamics and different bond energies for each channel. The H + TOF spectrum from the original molecule is a weighted sum of such D + TOF traces. If only one channel is active then the H + and D + TOF traces will coincide after normalizing for the mass effect. ! 28! ! Chapter 2 References (1) Aristov, V.; Conroy D.; Reisler, H. Chemical Physics Letters 2000, 318, 393. (2) Conroy, D.; Aristov, V.; Feng, L.; Reisler, H. Journal of Physical Chemistry A 2000, 104, 10288. (3) Ahmed, M.; Peterka, D. S.; Suits, A. G. Physical Chemistry Chemical Physics 2000, 2, 861. (4) Rudic, S.; Murray, C.; Ascenzi, D.; Anderson, H.; Harvey, J. N.; Orr- Ewing, A. J. Journal of Chemical Physics 2002, 117, 5692. (5) Taatjes, C. A.; Christensen, L. K.; Hurley, M. D.; Wallington, T. J. Journal of Physical Chemistry A 1999, 103, 9805. (6) Wei, J.; Karpichev, B.; Reisler, H.; Journal of Chemical Physics 2006, 125, 034303. (7) Syage, J. A. Journal of Chemical Physics 1996,105, 1007. 29 Chapter 3 Unimolecular processes in CH2OH below the dissociation barrier: O–H stretch overtone excitation and dissociation 3.1 Introduction The hydroxymethyl (CH 2 OH) radical and its isomer, the methoxy (CH 3 O) radical, have long been implicated in diverse areas such as atmospheric chemistry [1], surface reactions [2], and combustion chemistry [3]. Unimolecular processes in CH 3 O and CH 2 OH on the ground state potential energy surface (PES) have attracted considerable experimental and theoretical attention [4-8], and it was found that both radicals have low dissociation barriers. Ab initio calculations show that the reaction barrier for the loss of hydrogen atom from CH 3 O is lower than that for isomerization to CH 2 OH [6,8]. In the case of CH 2 OH, the situation is not as clear because the calculated barriers to direct O–H bond breaking and dissociation through isomerization are comparable, both leading to H atom and formaldehyde. According to calculations, the former is slightly higher than the latter [6,8]. The quantum state resolved unimolecular dynamics for CH 3 O in highly vibrationally excited states (up to 10 000 cm –1 ) have been examined by stimulated emission pumping (SEP) spectroscopy [4,9-11]. It was concluded that competition between intramolecular vibrational energy redistribution (IVR) and unimolecular 30 decomposition was important in the vibrationally excited state, but there was no indication of isomerization. SEP experiments cannot be performed with CH 2 OH, because its electronic excited states are dissociative with lifetimes shorter than 0.5 ps [12-17], and so far no information on its unimolecular reaction on the ground state has been reported. The photodissociation of CH 2 OH radicals in excited electronic states has been examined experimentally [12-16] and theoretically [17-20]. The lowest excited electronic state, 3s, internally converts to the ground state through efficient conical intersections that take place at long O–H distances, and the subsequent dissociation at low energies above the 3s onset gives rise exclusively to O–H bond breaking. When initial excitation accesses the 3p x and 3p z states, sequential couplings to the 3s state followed by conical intersections with the ground state propel the system to dissociation along the O–H and C–H dissociation coordinates. No isomerization to methoxy has been detected [13-15]. The present study concerns the excitation and predissociation of CH 2 OH on the ground PES by overtone OH-stretch excitation. Optical excitation of overtones has been used before to access highly vibrationally excited levels in the ground electronic state of stable molecules, e.g., HOOH, NH 2 OH, and CH 3 OH [21-25]. For example, in methanol OH-stretch (! 1 ) overtones up to 5 ! 1 have been excited directly [21,24], and levels as high as 8! 1 have been reached by using double resonance overtone 31 excitation [21]. The IVR rates deduced for HOOH, NH 2 OH, and CH 3 OH were rather similar [22]. There are two main differences between these studies and the one reported here. First and most important, in the other studies the excited OH vibration was not the reaction coordinate and had a much higher bond energy than the lowest dissociation energy. Thus, high overtone excitation could be achieved in a potential that behaved as a local-mode Morse oscillator, and IVR preceded dissociation. In contrast, in CH 2 OH the O–H bond is the reaction coordinate to H+CH 2 O and therefore it is not clear how high the OH-stretch vibration can be optically excited. Second, the species excited here is a free radical, adding the challenge of working with a minor component that is reactive. Previously, we reported spectroscopic studies in the region of the CH and OH fundamental transitions and first overtone of the OH-stretch vibration in CH 2 OH, and compared the experimental results to ab initio calculations [26]. Rotationally resolved spectra were recorded by double resonance ionization detection (DRID) via the 3p z Rydberg state and showed line broadenings in the 2 ! 1 spectrum that were thought to arise from low-order resonances. Here, we extend our studies to levels up to 4 ! 1 (13 600 cm –1 ), thereby approaching the dissociation barrier. By using DRID, we recorded a rotationally resolved spectrum of the second OH-stretch overtone. In contrast, the spectra obtained for CH 2 OH and CD 2 OH by exciting the 4 ! 1 level could only be recorded by detecting hydrogen fragments and were partially rotationally 32 resolved. We conclude that this state is predissociative and discuss linewidth broadening and predissociation via tunneling. 3.2 Experiment The experimental methods for generating hydroxymethyl radicals in a molecular beam and their detection by time-of-flight (TOF) mass spectroscopy are described in detail elsewhere [12]. Reactant mixtures of Cl 2 (1%), CH 3 OH (3%), and He are prepared in a glass bulb (5.0 l) at a total pressure of 2.0 atm at room temperature. The mixture is transported to the vacuum chamber through a piezoelectrically controlled pulsed nozzle. Hydroxymethyl radicals are produced at the end of a quartz tube attachment to the nozzle aperture by the following reactions: Cl 2 34 5 62Cl CH 3 OH + Cl ' CH 2 OH + HCl Photodissociation of Cl 2 is achieved with the 355 nm tripled output of a Nd:YAG (yttrium aluminum garnet) laser (Spectra Physics GCR-11, 6 mJ) focused by a 30 cm focal length (fl) cylindrical lens at the edge of the quartz tube. In order to minimize contributions from secondary reaction products [14], CH 2 OH radicals generated at the leading edge of the molecular beam pulse are probed. The molecular beam enters the detection chamber through a skimmer (Beam Dynamics, 1.50 mm). The rotational temperature of the radicals in the molecular beam is typically 10–15 K. 33 The pressure in the detection region is ~2.0×10 –7 Torr with the nozzle operating at 10 Hz, and the base pressure is below 2.0×10 –8 Torr. Spectra are obtained by exploiting pump and probe techniques, with different experimental schemes used for detection of 3! 1 and 4! 1 levels, as described below. The experimental schemes for detection of 3! 1 are similar to those used for the fundamental and the first overtone transitions, namely, depletion and DRID [26]. The main difference is that in this work we use (2+2) resonance enhanced multiphoton ionization (REMPI) with visible radiation to probe the radical [5,12], instead of the (1+1) REMPI employed before [26]. The (2+2) scheme gives better signal-to-noise ratio (S/N), particularly in the DRID experiments. Detailed descriptions of the depletion and DRID schemes are given elsewhere [26]. Briefly, in depletion the pump wavelength is scanned with the probe wavelength fixed at a resonance for the (2+2) REMPI transition to 3p z . Because of the low depletion depth for overtone transitions and fluctuations in the (2+2) REMPI signal, this method gives lower S/N than DRID. In DRID, the pump wavelength is first fixed at the deepest depletion value, and the probe wavelength scanned in search of the J resonance for the CH 2 OH (2+2) REMPI transition via 3p z , which has a favorable Franck-Condon factor. Once the transition is found, the probe is fixed at the resonance peak wavelength and the pump is scanned to obtain the second overtone transition spectrum. To distinguish between signal and background contributions, "pump-on" and "pump-off" experiments are carried out. The pump laser is fired 8 ns before the probe laser in a typical pump-on experiment, 34 whereas the pump laser is set to fire ~1 µs later than the probe under pump-off conditions. The pump radiation required for 3! 1 excitation (around 954 nm) is obtained as the idler output of a seeded Nd:YAG laser pumped optical parametric oscillator (OPO) (Continuum, PL8000/Sunlite, 12 mJ; 0.1 cm –1 , 20 cm fl lens), and the probe radiation around 490 nm is generated by another Nd:YAG laser pumped dye laser (Continuum, ND6000, 0.1 cm –1 , 25 cm fl lens). In depletion experiments, a typical probe laser energy is 0.8 mJ, whereas in DRID experiments energies up to 4 mJ are used to increase the S/N. With third overtone (4! 1 ) excitation, no depletion or DRID signal is observed above the background, but H-atom fragments are detected by (1+1’) REMPI via the L- * hydrogen transition. Vacuum ultraviolet (vuv) radiation at 121.6 nm is obtained by frequency tripling 365 nm radiation (2 mJ) in a Kr/Ar filled cell, as described before [13,14]. The pump radiation is obtained from another Nd:YAG laser pumped dye laser (Continuum, ND6000, 35 mJ; 0.1 cm –1 , 25 cm fl lens). Action spectra in the 4 ! 1 region are obtained by monitoring hydrogen fragments while varying the pump wavelength. In order to elucidate the role of isomerization, experiments were also carried out with CD 2 OH (Aldrich, 99.5%-d) with the probe tuned, in turn, to detect H and D fragments. In our experimental arrangement, the probe and pump laser beams are counterpropagating and cross the molecular beam at a right angle at the center of the 35 repeller and accelerator plates of the vertically mounted TOF mass spectrometer. The ions are detected mass selectively. The output of a 25 mm diameter dual microchannel plate (MCP) detector (Galileo) is amplified with a wide-band preamplifier (KOA microcircuits, KE104), and the signal is digitized by a 500 MHz oscilloscope (Tektronix, TDS640A). The timing sequence is controlled by two pulse generators (Stanford Research Systems). The lasers, pulse generators, and digitized oscilloscope are controlled with LABVIEW computer programs. Laser wavelengths are calibrated with a wavemeter (Burleigh WA-4500). The TOF mass spectrometer is used in two modes. For mass resolution, the repeller and accelerator voltages are held at 3000 and 2330 V, respectively, to assure full separation of the masses of methanol and hydroxymethyl. For the TOF analysis required to obtain hydrogen translational energies, the repeller and extractor voltages are lowered to 360 and 280 V, respectively, while maintaining space-focusing conditions. At these voltages, the TOF traces have sufficient resolution to distinguish between hydrogen fragments generated by one- and two-photon dissociation. Unfortunately, at the lower voltages required to achieve full core-sampling conditions [13,14], the S/N is too low, precluding a full analysis of the kinetic energy distributions. 36 3.3 Results And Analysis 3.3.1. Excitation of the second OH-stretch overtone, 3! 1 The second overtone of the OH-stretch of CH 2 OH was observed by depletion and DRID in a similar manner to the fundamental and first overtone transitions [26]. In depletion spectroscopy, the pump laser was scanned while the probe laser was fixed to monitor the (2+2) REMPI signal of CH 2 OH via the 3p z +1A ’’ ; ; transition [5,12]. A maximum depletion signal (~10%) was recorded at 10 489.8 cm –1 . The narrow REMPI probe band obtained via the ; ; transition makes depletion a convenient method. The S/N, however, is low because of the low intensity of the overtone transition and the fluctuations in the (2+2) REMPI signal and molecular beam intensity. Nevertheless, depletion gives an initial indication of the transition frequency to be used in the DRID experiments. In the DRID method, the probe transition originates in 3! 1 and terminates in a vibrational state in 3p z for which a favorable Frank-Condon factor exists. In Fig. 3.1, the (2+2) REMPI signal of CH 2 OH + is monitored by scanning the probe wavelength while keeping the pump frequency at 10 489.8 cm –1 , i.e., the 3! 1 ( q R(2)) frequency (see assignments below). Three "hot bands", G ; , 2 ; , and 2 ; , are observed under both pump-on and pump-off conditions [5]. The peak signal of the origin band (41 068 cm –1 ; not shown in Fig. 3.1) is two orders of magnitude higher in intensity than the hot bands. ! 37 ! Figure 3.1 IR+UV double resonance REMPI spectrum of CH2OH obtained via the 3p z state under "IR on" (solid line) and "IR off" (dotted line) conditions. The pump laser frequency is fixed at 10 489.8 cm –1 , the frequency of the second overtone transition in CH 2 OH. ! 38 The band appearing at 40 320±15 cm –1 only under pump-on conditions is assigned as J . Previously, vibronic transitions originating in ! 1 and 2 ! 1 in the ground electronic state were assigned as J G G and J 2 2 [26]. The propensity for the (v = 0 sequence for OH-stretch transitions to Rydberg states is explained by the nearly equal OH bond lengths in the CH 2 OH neutral and cation [5]. For example, the DRID signal from the J G G band is three times more intense than the one from the J G ; band. The energy of the 3! 1 level in the 3p z Rydberg state is determined at 9741±15 cm –1 above the origin by adding the pump laser frequency and the two-photon probe laser frequency and subtracting the ; ; frequency. The assignment of the resonant level in 3p z to 3! 1 is confirmed by the Birge-Sponer relationship discussed below. By scanning the pump laser wavelength while keeping the probe at the two- photon peak transition, the rotationally resolved 3! 1 DRID spectrum shown in Fig. 3.2(a) is obtained. This scheme takes advantage of the broad linewidth of the J transition (60 cm –1 , see Fig. 3.1) . A simulated spectrum obtained by using the asymmetric rotor program ASYROTWIN [27] is shown in Fig. 3.2 (b) (see Sec. IV for details). In order to get a rough estimate of the lifetime of CH 2 OH in 3! 1 , the time delay between the pump and probe lasers was varied while monitoring the CH 2 OH + DRID signal. The decrease in DRID signal with time was exponential with a lifetime of 60±5 ns. The diameter of the focal spot of probe laser beam is ~0.1 mm, and that for 39 ! Figure 3.2 IR spectrum of CH 2 OH in the region of the second overtone of the OH stretch obtained by DRID. The upper panel displays the experimental spectrum. The solid and dotted lines represent signals corresponding to "IR-on" and "IR-off" experiments, respectively. In the bottom panel a best-fit spectrum with a linewidth of 0.4 cm –1 is shown. The calculated rotational transitions are given by the stick spectrum. In the simulations, we used A”=6.51 cm –1 , B”=1.01 cm –1 , C”=0.88 cm –1 , A’=6.30 cm –1 , B’=1.00 cm –1 , C’=0.88 cm –1 , T rot =13 K, and " 0 =10 484.2 cm –1 ! 40 the pump is <0.2 mm. Considering a molecular beam speed of about 1.8 mm/µs, the decay time represents mainly the fly-out time. No hydrogen fragments could be detected, consistent with the long lifetime of 3! 1 . 3.3.2. Excitation of the third OH-stretch overtone, 4! 1 3.3.2.1. Hydrogen fragment yield spectra of CH 2 OH and CD 2 OH Neither depletion nor DRID resulted in a clear signal above the background in the region of the 4! 1 transition. This is not surprising, especially with respect to depletion, because the 4! 1 transition is weaker than 3! 1 . As 4! 1 lies well above the thermochemical threshold for dissociation to H + CH 2 O, the next step was to monitor hydrogen fragments. Figure 3.3(a) shows the 4! 1 spectrum for CH 2 OH obtained by monitoring hydrogen fragments by (1+1’) REMPI. A similar experiment was carried out with CD 2 OH, in search of hydrogen and deuterium. Only hydrogen was observed and Fig. 3.3(b) shows the 4! 1 spectrum for CD 2 OH obtained by monitoring H fragments, which is blueshifted by ~19 cm –1 with respect to CH 2 OH. The pump laser was scanned over a range of ±600 cm –1 around the 13 600 cm –1 peak of the 4! 1 transition. No other transitions were detected. ! 41 ! Figure 3.3 The solid curve depicts experimental H-atom photofragment yield spectra in the region of the third overtone of the OH stretch for (a) CH 2 OH and (b) CD 2 OH. The dashed line shows a spectral fit to the data for an a-type transition and linewidth of 1.3 cm –1 . See text for details. In the simulations, we used A”=6.51 cm –1 , B”=1.01 cm –1 , C”=0.88 cm –1 , A’=6.00 cm –1 , B’=B”, C’=C”, T rot =13 K, and ! 0 =13 597.9 cm –1 for CH 2 OH; and A”=3.87 cm –1 , B”=0.86 cm –1 , C”=0.71 cm –1 , A”=3.70, B’=B”, C’=C”, T rot =13 K, and ! 0 =13 616.6 cm –1 for CD 2 OH ! 42 3.3.2.2.Time-of-flight (TOF) analysis of hydrogen fragments Previous studies of the transition to the lowest excited electronic state of CH 2 OH, 3s, revealed a broad and structureless spectrum with an onset at ~26 000 cm –1 [16]. The 3s state is dissociative, and near its onset H atoms are produced solely by O– H bond fission [14]. A two-photon transition via 4! 1 of the ground state can access 3s at energies ~1200 cm –1 above its onset. Therefore, two possible pathways for hydrogen production via 4! 1 exist: (i) one-photon unimolecular predissociation on the ground electronic state and (ii) two-photon dissociation on 3s mediated by 4! 1 excitation. The H-atom fragment signal varies linearly with laser fluence, which favors a one-photon process. In addition, analysis of the H-atom translational energies measured by TOF was carried out in order to determine the energetics involved in the dissociation. Figure 3.4 shows TOF spectra of hydrogen fragments from CH 2 OH (obtained at low repeller and extractor voltages) at two pump laser frequencies: the peak frequency of the 4! 1 transition (13 603 cm –1 ) and the doubled frequency (27 206 cm –1 ), which can excite directly the 3s dissociative state. The minimum (maximum) allowed flight times corresponding to cold CH 2 O cofragments flying along (opposite to) the direction of the extraction electric field are estimated by using D 0 (CH 2 O–H)=9600 cm – 1 [6] and are shown as the inner (outer) arrows for the two excitation energies. Clearly, the H-atom TOF distribution obtained via excitation of 4! 1 is within the one-photon limit and quite different from the result obtained by doubling the frequency. 43 As the measured TOF spectrum is a one-dimensional projection of the velocity distribution, its main value for our purpose is in determining maximum speeds rather than the complete velocity distributions. In particular, slower speeds are not resolved because of the convolution of angular distributions [28]. Some information on the translational energy distribution can still be obtained by lowering the extraction voltage, as seen in Fig. 3.4. Fast products give a better reflection of the speed distribution when the voltage is lowered, because velocity components perpendicular to the TOF axis are discriminated against. This enables us to distinguish between the energy limits associated with H fragments produced by one- and two-photon processes. In summary, both the laser fluence dependence and the H-fragment TOF analysis support the production of fragments by one-photon dissociation. Vibrationally mediated photodissociation on 3s via 4! 1 is unfavored. This may be due to unfavorable Franck-Condon factors for excitation from 4! 1 . Considering that two-photon excitation via 4! 1 reaches an energy that is only 1200 cm –1 higher than the onset of 3s, there is not enough energy to excite the Franck-Condon favored diagonal transitions to high OH-stretch levels. Experiments were also carried out by monitoring H-atom fragments from CD 2 OH, and the results were essentially the same. ! 44 ! Figure 3.4 Time-of-flight (TOF) spectra of H fragments produced in the dissociation of CH 2 OH by one photon excitation at (a) 27 210 cm –1 (3s Rydberg excited state) and (b) 13 603 cm –1 (4! 1 transition). Zero TOF indicates no recoil energy. The arrows show the maximum and minimum TOF values allowed by the thermochemistry for one-photon dissociation. The polarizations of the pump laser radiation used in (a) and (b) are perpendicular and parallel, respectively, to the extraction field. Each spectrum is a summation of 5000 laser firings, and background is subtracted 45 3.3.3. Spectroscopic analysis of OH-stretch overtones CH 2 OH is well described as a near-prolate asymmetric top and rotational energy levels are designated by , where N denotes the rotational angular momentum. Simulations were done with the ASYROTWIN program [27]. Ground-level rotational constants were derived from the calculated equilibrium structure [5,26], and the upper state constants A’, B’, and C’ were varied until a best fit to the spectra is obtained. The A rotational constants from the ground state to 4! 1 (6.51, 6.41 cm –1 , 6.39 cm –1 , 6.30 ± 0.2, and 6.00 ± 0.2 cm –1 , respectively [26]) are quite similar, with a slight decrease at high overtones, reflecting probably the extension of the OH bond. However, the simulations are rather insensitive to the value of A, because the transitions are mainly a-type (parallel) bands, and higher order effects cannot be resolved. 3.3.3.1. Analysis of the second overtone transition A simulated spectrum of the 3! 1 transition is displayed in Fig. 3.2 (b). The positions of the rotational transitions are shown by the stick spectrum. The transition is dominated by an a-type band with some b-type character (I b /I a =0.4±0.2), rather similar to the fundamental and the first overtone transitions [26]. For simplicity, only a-type transitions are assigned in Fig. 3.2 (b). The two branches, q P and q R, of the a-type transition are well resolved for the quantum number N, whereas fine rotational structures due to different K a and K c quantum numbers are not resolvable. For 46 example, q R(1) involves three transitions with (K a =0, (N=1, and N”=1:2 12 +1 11 , 2 11 +1 10 , and 2 02 +1 01 , which can be seen in the calculated stick spectrum but are not resolved. The radical's b-type transition consists of a predominant peak of the r Q branch, which is overlapped with the a-type q R(2) transition. The rotational temperature, T rot =13 ± 2 K, is determined from the rotational level population (maximum N”=5 or 6) and the best-fit linewidth is 0.4±0.1 cm –1 —smaller than that obtained for 2! 1 (0.8 ± 0.1 cm –1 ) [26], as discussed in Sec. 3.4. 3.3.3.2.Analysis of the third overtone transition Simulated 4! 1 spectra are depicted in Fig. 3.3for CH 2 OH and CD 2 OH. The rotational linewidths in the 4! 1 transition are clearly broader than those in 3! 1 , with a value of 1.3 cm –1 for the former and 0.4 cm –1 for the latter. Varying the rotational temperature and the linewidth affects the spectrum differently. Specifically, for T rot = 13 ± 2 K, rotational levels are populated up to N”=5 or 6. At higher temperatures, higher N” 's contribute more and this changes primarily the intensity distribution within the band, but has little effect on the line broadening. On the other hand, varying the linewidth changes the widths of all N”-resolved transitions, without changing their relative intensities. These two effects are quite distinct and enable us to estimate the linewidth at 1.3 cm –1 . The third overtone transition can be described fairly well as an a-type band. The three branches, q P, q Q, and q R, are separated from each other, whereas the transitions in each branch are overlapped and only barely separated. The line 47 broadening of the 4! 1 transition reflects the combined effect of IVR and unimolecular dissociation. Despite the higher state density of CD 2 OH (320/cm –1 , compared with 110/cm –1 for CH 2 OH, by the Beyer-Swinehart algorithm neglecting anharmonicity [29]), the linewidths for CD 2 OH and CH 2 OH are similar, suggesting that reaction rates give the dominant contribution to the linewidth (see Sec. 3.4). 3.3.4. Birge-Sponer analysis Additional insight into the nature of the O–H vibrational potential can be obtained by fitting the observed vibrational levels to an expression for a Morse oscillator and comparing the extracted fundamental frequency and anharmonicity with the values for analogous OH containing molecules. The energy levels of a Morse oscillator follow the Birge-Sponer expression )+ 2 (1) where ! is the frequency of the vibration, B=& e ' e is the anharmonicity, & e = A + B is the harmonic frequency, and ( is the number of quanta. The experimental values for v=1,2 are taken from Refs. [26,30]. The Birge-Sponer plots for CH 2 OH shown in Fig. 3.5give the (A,B) values for the ground state and the 3p z excited electronic state: (3766.3±2.7 cm –1 ,91.4±1.0 cm –1 ) and (3487.7±3.1 cm –1 ,80.5±1.4 cm –1 ), respectively. These values are close to those for the free hydroxyl radical (3653, 82.5) [31] and other OH containing molecules, such as CH 3 OH (3769.6, 86.1) [22], NH 2 OH (3743, 90.6) [25], and HOOH (3701, 90.5) [23,25]. The linearity of the Birge-Sponer plot is 48 ! Figure 3.5 Birge-Sponer plots for CH 2 OH in the ground electronic state and the 3p z Rydberg excited state. The lines are fits to the Birge- Sponer expression and yield the A and B parameters indicated in the figure 49 accurate for a one-dimensional (1D) Morse oscillator, and the similarity of the (A,B) parameters for all the OH containing molecules suggests that the 1D PESs along the OH coordinate are rather similar over the examined energy regions. The OH-stretch vibrational frequencies and anharmonicities derived from the Birge-Sponer expression for CH 2 OH are listed in Table 3.1. The Birge-Sponer plot of the O–H stretch overtones in the Rydberg 3p z state (Fig. 3.5) confirms our assignment of the OH-stretch levels. Despite the fact that the 3p z state is predissociative with a lifetime shorter than 0.5 ps [12-17], the OH-stretch oscillator is less anharmonic than the ground state. A strong O–H bond is expected in 3p z , because the corresponding bond in the ion is stronger [D 0 (CH 2 O–H + )=59 000 cm – 1 ] [5] than in the ground state of the neutral (9600 cm –1 ) [6]. 3.4. Discussion 3.4.1. Roles of IVR, predissociation, and isomerization The parameters characterizing the Birge-Sponer plot for the CH 2 OH OH- stretch are similar to those for hydroxyl radical and other OH containing small molecules, implying that 50 ! Ground 3p z , e /cm -1 3857.7±2.9 3526.2±3.4 , e - e /cm -1 91.4±1.0 80.5±1.4 Table 3.1 OH-stretch vibrational frequencies and anharmonicities of CH 2 OH in the ground and the Rydberg 3p z state.! 51 the observed transitions have a predominant local-mode OH-stretch character. In studies of overtone spectroscopy of the species containing XH stretch vibration (X=C,N,O) it has been found that the hydrogen stretch overtones are described well as local-mode vibrations [32-38]. We assume a similar local-mode behavior for the OH-stretch overtones of CH 2 OH. In our previous study of 1! 1 and 2! 1 , the 1! 1 linewidth was limited by the laser bandwidth (0.4 cm –1 ), whereas the 2! 1 linewidth was slightly broader (0.8 cm –1 ) [26]. This broadening was explained by the existence of low-order resonances that give rise to nearby unresolved satellite bands. The linewidth of 3! 1 (0.4 cm –1 ) is narrower than that of 2! 1 despite a factor of ~6 increase in the harmonic vibrational state density (from 5/cm –1 for 2 ! 1 to 30/cm –1 for 3! 1 ) [26,29], but is broader than the laser bandwidth in the present experiment (0.1 cm –1 ). The estimated linewidths of 4! 1 for CH 2 OH and CD 2 OH are considerably larger (1.3 cm –1 ). The decrease in linewidth from 2! 1 to 3! 1 reinforces our previous conclusion that accidental low-order resonances are responsible for the observed 2! 1 line broadening [26]. Because of their different anharmonicities, the levels resonant with 2! 1 are detuned out of resonance with 3! 1 . The long lifetime of CH 2 OH in 3! 1 indicates that unimolecular reactions, including isomerization to CH 3 O, do not contribute to the linewidth. This is not surprising, as the barriers to dissociation and isomerization are higher than the energy of 3! 1 by 3000–6000 cm –1 . Measurements of linewidth as a function of laser intensity show no power broadening. The 0.4 cm –1 linewidth of 3! 1 52 may thus reflect homogeneous broadening due to weak couplings to bath states. In the OH overtone transitions of CH 3 OH and NH 2 OH, homogeneous broadening of ~0.2 cm –1 was observed starting from the second overtone of the OH stretch, which increased for higher overtones [21,22]. In OH-stretch overtones of CH 3 OH, couplings due to low-order resonant states were shown to be sensitive to the value of the rotational quantum number K and the carbon isotope [21,35]. The increased linewidth of 4! 1 is likely the result of contributions from predissociation via tunneling. In order to distinguish between the contributions of IVR and lifetime broadening, the overtone spectrum of CD 2 OH was also recorded. The 4! 1 transition of CD 2 OH is blueshifted by 19 cm –1 from the corresponding level of CH 2 OH, consistent with the OH-stretch character of the transition. Of the nine vibrational modes of hydroxymethyl, all except ! 1 (OH stretch) and ! 6 (CO stretch) are related to the motion of H(C) atoms and therefore have significantly lower vibrational frequencies in CD 2 OH than in CH 2 OH [5]. The different frequencies of the vibrational modes in CD 2 OH and CH 2 OH should result in different low-order resonances and coupling matrix elements for IVR. The observation that the two have identical 4! 1 linewidths within experimental uncertainty suggests that low-order resonances are not responsible for the sharp increase in linewidth in going from 3! 1 to 4! 1 . Comparisons with the overtone spectroscopy of NH 2 OH, HOOH, and CH 3 OH are enlightening. The dissociation thresholds of these molecules are higher than that of CH 2 OH and therefore lifetime broadening is insignificant in their 4! 1 energy regions. 53 The thermochemical dissociation threshold in CH 2 OH is D 0 (O–H)=9600 cm –1 [6], whereas the corresponding values for NH 2 OH and HOOH are D 0 (N–O)=21 620 cm –1 and D 0 (O–O)=17 052 cm –1 , respectively [22]. The OH oscillator in HOOH, for example, follows the Birge-Sponer expression up to 6! 1 , 2042 cm –1 above D 0 (O–O), because the O–H bond energy is much higher than D 0 (O–O) and IVR is slow [23]. Even though 4! 1 of CH 2 OH is estimated to lie ~4000 cm –1 above the O–H bond energy, the values of A and B returned by the Birge-Sponer plot are similar to those in the other molecules, apparently because the reaction barrier, calculated at around 15 000 cm –1 [6,8], traps the OH oscillator. Near the top of the barrier the energy of the vibrational levels may deviate from the Birge-Sponer straight line. For example, in HOOH, the 7! 1 energy is lower by 50 cm –1 than the value predicted by extrapolation of the Birge-Sponer plot that describes well the lower overtones, indicating that the 1D OH potential curve is less steep at this energy than the Morse representation [23]. Our results suggest that in CH 2 OH the 1D OH potential curve follows a Morse function from 1! 1 to at least 4! 1 , with A and B values similar to those in the other OH containing molecules. This indicates that the barrier to O–H bond fission lies higher than the 4! 1 level, and consequently dissociation in this energy region must proceed via tunneling. In hydroxylamine and methanol, which have higher dissociation energies, 3! 1 and 4! 1 exhibit restricted IVR, and for these molecules a multitier coupling scheme, in which low-order resonances serve as gateway states to a bath of dark states, is invoked to explain the small increase in linewidth [21,25]. In contrast, in CH 2 OH, which has a 54 much lower dissociation energy, the sharp linewidth increase from 3! 1 to 4! 1 is most likely caused by the increased probability of unimolecular reaction. Last, we discuss the possible contribution of isomerization to methoxy to the linewidth. In SEP studies of CH 3 O, predissociation lifetimes of vibrational states of energies ~10 000 cm –1 were estimated to be as short as 5 ps [10,11]. No indication of decay due to isomerization to CH 2 OH was obtained [9]. The heat of formation of CH 3 O is higher by 2800–3200 cm –1 than that of CH 2 OH [5-7], and therefore this excitation energy is slightly lower than the energy that would have been achieved in methoxy had 4! 1 excitation in CH 2 OH (~13 600 cm –1 ) led to isomerization. Such isomerization would be followed by dissociation to H+CH 2 O, because the calculated barrier to CH 3 O dissociation is lower than that for CH 3 O . CH 2 OH isomerization [6,8]. The absence of deuterium fragments following excitation of CD 2 OH suggests that isomerization to CD 2 HO is not favored. Also, the similarity of the 4! 1 linewidths in CH 2 OH and CD 2 OH, which indicates similar rates for their unimolecular processes, favors direct fission of the O–H stretch. The conclusion that isomerization is not important is reinforced by the spectroscopic studies. As mentioned above, no vibrational bands other than the OH- stretch overtone were found in the region of the 4! 1 band. The geometry of the transition state for hydrogen shift from O to C has the H atom equidistant between the O and C atoms [8]. To reach this configuration, IVR to bending modes must be involved, and this should be reflected in the spectrum. The absence of such bands is 55 another indication that isomerization does not take place and the OH stretch acts as a local mode. We conclude, therefore, that H-atom photofragments are produced by OH fission via tunneling through the barrier. 3.4.2. Predissociation by tunneling As discussed above, the linear Birge-Sponer plot can be rationalized if we assume that the reaction barrier to direct OH bond fission is higher than the 4! 1 energy. To a first approximation tunneling through the barrier can be treated by a one- dimensional semiclassical model [39]. The tunneling rate constant k is given by, )M (2) where P is the tunneling probability and f is the classical vibrational frequency given by [40] )f & + & & .KJ1, (3) where c is the speed of light in vacuum. Using the & e and & e - e values in Table 3.1 for the ground state, f(4! 1 )=1.0×10 14 s –1 is obtained. If we assume that the linewidth is determined exclusively by the predissociation rate (/=1.3 cm –1 or k=2.5×10 11 s –1 ), we obtain an upper limit to the tunneling probability, P max =2.5×10 –3 . A rough estimate of the barrier height can be obtained by using an Eckart barrier [41] with three parameters: the reaction barrier height, the dissociation energy (D 0 =9600 cm –1 ) [6], and the imaginary vibrational frequency (" c =1712 cm –1 ) [42]. Assuming tunneling probabilities in the range of P=(1.0–2.5)×10 –3 , dissociation 56 barriers of 15 100–15 400 cm –1 are obtained. This simple estimation compares favorably with the calculated barrier heights of 16 000 [8] and 14 000 cm –1 [6]. Evidently, lifetime broadening is not the only contributor to the 4! 1 linewidth. Assuming that the homogeneous IVR broadening for 3! 1 and 4! 1 (0.4 cm –1 ) are comparable, the predissociation rate should be higher than the rate of IVR. In statistical rate theories, it is assumed that IVR is rapid compared to the dissociation rate. However, for small molecules, state-specific behavior is sometimes observed as a result of incomplete IVR. The situation in CH 2 OH is somewhat unique: because the reaction coordinate is excited directly, IVR is expected to hinder, rather than promote, reaction. Thus, it is plausible that reaction in the tunneling regime takes place without significant IVR. 3.5. Conclusions The 3! 1 and 4! 1 levels of CH 2 OH have been examined by overtone excitation under molecular beam conditions, and H fragments have been observed following 4! 1 excitation. The second overtone spectrum of CH 2 OH was recorded by both depletion spectroscopy and DRID. The DRID spectrum is rotationally well resolved with linewidth of 0.4 cm –1 , reflecting IVR due to weak couplings to zeroth-order dark bath states. No dissociation is observed. Third overtone spectra of CH 2 OH and CD 2 OH are obtained by monitoring H- atom fragments. They are only partially rotationally resolved, and can be simulated 57 with similar 1.3 cm –1 linewidths. Both laser intensity dependence and TOF translational energy analyses support the production of H-atom fragments by one- photon dissociation from 4! 1 on the ground electronic state. Considering the similar 4! 1 linewidths for CH 2 OH and CD 2 OH, the sharp linewidth increase from 3! 1 to 4! 1 is interpreted as having a major contribution from predissociation. The linearity of the Birge-Sponer relationship indicates that the O–H overtones behave approximately as local modes of a Morse oscillator at least up to 4! 1 . Dissociation takes place via tunneling through the barrier to direct O–H fission. The absence of signal from D atoms, equal linewidths for CH 2 OH and CD 2 OH, and the absence of any other spectral features except the OH-stretch overtone indicate that IVR and isomerization to methoxy are not important. Thus, we conclude that the H- atom fission rate is faster than IVR and reaction takes place by imparting energy directly to the O–H reaction coordinate. Calculations of the potential energy surface and tunneling and predissociation dynamics will shed further light on the unimolecular processes. An intriguing issue is what would happen when the O–H dissociation barrier is exceeded. Our attempts to excite directly the fourth OH-stretch overtone failed because the two-photon absorption cross section to the 3s Rydberg state exceeded the overtone excitation cross section. Sequential excitation in which 1! 1 or 2! 1 are first excited followed by excitation to 5! 1 may be a better way to carry out these experiments. 58 Chapter 3 References (1) Heicklen, J. Atmospheric Chemistry; Academic: New York, 1976. (2) Whitten, J. E.; Young, C. E.; Pellin, M. J.; Gruen, D. M.; Jones, P. L. Surface Science 1991, 241, 73. (3) Demerjian, K. L.; Kerr, J. A.; Calvert, J. G. THE MECHANISM OF PHOTOCHEMICAL SMOG FORMATION, 1974. (4) Dertinger, S.; Geers, A.; Kappert, J.; Wiebrecht, J.; Temps, F. Faraday Discussion. 1995, 102, 31. (5) Johnson, R. D.; Hudgens, J. W. Journal of Physical Chemistry 1996, 100, 19874. (6) Marcy, T. P.; Diaz, R. R.; Heard, D.; Leone, S. R.; Harding, L. B.; Klippenstein, S. J. Journal of Physical Chemistry A 2001, 105, 8361. (7) Petraco, N. D. K.; Allen, W. D.; Schaefer, H. F. Journal of Chemical Physics 2002, 116, 10229. (8) Saebo, S.; Radom, L.; Schaefer, H. F. Journal of Chemical Physics 1983, 78, 845. (9) Geers, A.; Kappert, J.; Temps, F.; Wiebrecht, J. W. Journal of Chemical Physics 1993, 99, 2271. (10) Geers, A.; Kappert, J.; Temps, F.; Wiebrecht, J. W. Journal of Chemical Physics 1994, 101, 3618. (11) Geers, A.; Kappert, J.; Temps, F.; Wiebrecht, J. W. Journal of Chemical Physics 1994, 101, 3634. (12) Aristov, V.; Conroy, D.; Reisler, H. Chemical Physics Letters 2000, 318, 393. (13) Conroy, D.; Aristov, V.; Feng, L.; Reisler, H. Journal of Physical Chemistry A 2000, 104, 10288. (14) Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2003, 118, 9623. 59 (15) Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2004, 120, 6524. (16) Feng, L.; Huang, X.; Reisler, H. Journal of Chemical Physics 2002, 117, 4820. (17) Hoffman, B. C.; Yarkony, D. R. Journal of Chemical Physics 2002, 116, 8300. (18) Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 1998, 102, 3141. (19) Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 2001, 105, 8599. (20) Yarkony, D. R. Journal of Chemical Physics 2005, 122. (21) Boyarkin, O. V.; Rizzo, T. R.; Perry, D. S. Journal of Chemical Physics 1999, 110, 11346. (22) Kuhn, B.; Boyarkin, O. V.; Rizzo, T. R. Berichte der Bunsen- Gesellschaft für Physikalische Chemie 1997, 101, 339. (23) Kuhn, B.; Rizzo, T. R. Journal of Chemical Physics 2000, 112, 7461. (24) Phillips, J. A.; Orlando, J. J.; Tyndall, G. S.; Vaida, V. Chemical Physics Letters 1998, 296, 377. (25) Scott, J. L.; Luckhaus, D.; Brown, S. S.; Crim, F. F. Journal of Chemical Physics 1995, 102, 675. (26) Feng, L.; Wei, J.; Reisler, H. Journal of Physical Chemistry A 2004, 108, 7903. (27) Judge, R. H.; Clouthier, D. J. Computational Physics Communications 2001, 135, 293. (28) Xu, Z.; Koplitz, B.; Wittig, C. Journal of Chemical Physics 1989, 90, 2692. (29) Bear, T.; Hase, W. Unimolecular Reaction Dynamics; Oxford University Press: New York, 1996. (30) Feng, L.; Reisler, H. Journal of Physical Chemistry A 2004, 108, 9847. 60 (31) Coxon, J. A.; Foster, S. C. Canadian Journal of Physics 1982, 60, 41. (32) Child, M. S. Accounts of Chemical Research 1985, 18, 45. (33) Chirokolava, A.; Perry, D. S.; Boyarkin, O. V.; Schmid, M.; Rizzo, T. R. Journal of Chemical Physics 2000, 113, 10068. (34) Halonen, L. Advances in Chemical Physics, Vol 104 1998, 104, 41. (35) Henry, B. R. Accounts of Chemical Research 1977, 10, 207. (36) Henry, B. R.; Kjaergaard, H. G. Canadian. Journal of Chemical Review. Can. Chim. 2002, 80, 1635. (37) Mills, I. M.; Robiette, A. G. Molecular Physics 1985, 56, 743. (38) Quack, M.; Willeke, M. Journal of Chemical Physics 1999, 110, 11958. (39) Waite, B. A.; Miller, W. H. Journal of Chemical Physics 1980, 73, 3713. (40) Herzberg, G. Molecular Spectra and Molecular Structure I: Spectra of Diatomic Molecules; Van Nostrand: New York, 1950. (41) Miller, W. H. Journal of the American Chemical Society 1979, 101, 6810. (42) Harding, L. B., unpublished. 61 Chapter 4 Electronic Spectroscopy and Photodissociation Dynamics of the 1-Hydroxyethyl Radical CH 3 CHOH 4.1 Introduction The chemistry of hydroxyalkyl radicals is important in atmospheric chemistry and combustion environments. For example, CH 2 OH is a significant product in the reaction of O( 1 D) with methane,[1-5] and of Cl atoms and OH radicals with methanol.[6-9] Of the hydroxyalkyl radicals larger than CH 2 OH, the structural isomers 1- and 2- hydroxyethyl radicals (CH 3 CHOH and CH 2 CH 2 OH, respectively),[10-14] are relevant to ethanol combustion. These isomers are also involved in the photochemistry of the ethoxy radical,[15,16] as products in the reaction of halogen atoms with ethanol,[6,9-11,13,17] and as intermediates in the O + C 2 H 5 and OH + C 2 H 4 reactions.[18-22] The role of hydroxyalkyl intermediates in combustion reactions has recently been highlighted by the discovery that the enol tautomers of aldehydes are significant combustion intermediates. For example, vinyl alcohol (ethenol, CH 2 =CHOH), the less stable tautomer of acetaldehyde (CH 3 –HC=O), is formed in flames of ethanol, olefins, and commercial fuels.[23] Both tautomers can be formed in the decomposition of ground state hydroxyethyl radicals.[18,21]. In spite of their importance very little is known about the photophysics and photochemistry of hydroxyalkyl radicals. 62 An unstructured electronic absorption spectrum of CH 3 CHOH was recorded at 300 K in the region 230-300 nm with a maximum absorption cross section of 3.6x10 - 18 cm 2 .[10] The spectrum shows that absorption at < 300 nm rises strongly towards shorter wavelengths but the measurements stopped at 230 nm, where the spectrum reached a plateau, and did not extend to wavelengths longer than 300 nm. A paper on the kinetics of CH 3 CHOH reports a resonance enhanced multiphoton ionization (REMPI) signal of the parent ion in the range 430-460 nm,[24] which likely arises from 2+1 ionization via a Rydberg state. The adiabatic ionization energy of CH 3 CHOH is 6.64 eV,[13,25] much lower than the corresponding value for CH 2 OH (7.56 eV).[26] A photoelectron spectroscopy study revealed a progression of ~ 1600 cm -1 in the CO stretch of the ion, in agreement with the expected shortening of the CO bond upon ionization.[25] The photodissociation of the hydroxyethyl radical has not been studied before. However, following UV excitation of the ethoxy isomer (CH 3 CH 2 O), Neumark and coworkers observed the unexpected C 2 H 3 + H 2 O channel, which they suggested evolved via isomerization on the excited state to hydroxyethyl.[15,16] There are no theoretical papers on the excited states of the hydroxyethyl radical. However, there are several ab initio calculations of the geometry of the ground state and the ion as well as barriers to isomerization and dissociation on the ground- state potential energy surface (PES).[12,14,18,21] The lowest barrier channels calculated for the ground state PES are:[21,27] 63 CH 3 CHOH ' CH 3 CHO + H "H = 1.04 eV (1) ' CH 2 CHOH + H "H = 1.46 eV (2) ' CH 3 + CHOH "H = 3.13 eV (3) The molecular elimination channels, CH 3 CO + H 2 and CH 3 + CO + H 2 , involve high barriers and tight TS’s and, as in CH 2 OH, may not be competitive with simple bond fission channels.[28,29] In addition, The CH 3 CH 2 O # CH 3 CHOH isomerization can lead to CH 2 O + CH 3 ("H = 0.85 eV) and CH 3 CHO + H.[21] The barrier to isomerization is calculated at 1.61 eV from the hydroxyethyl side.[21] In this paper we present the first study of the photophysics and photochemistry of the 1-hydroxyethyl radical (CH 3 CHOH) in a molecular beam in which we: (i) identify and assign its lowest electronic transitions; (ii) discuss the role of nonadiabatic transitions following electronic excitation in the region of the lowest Rydberg states; and (iii) identify dissociation pathways. In interpreting the results we rely on information gained from similar studies on the prototype hydroxymethyl radical.[28-37] For example, the energy of the 3s and 3p Rydberg states of CH 3 CHOH is predicted by using the Rydberg formula with quantum defects $ similar to those of CH 2 OH (see Section 4 and Table 4.1). Likewise, theoretical studies of CH 2 OH photodissociation have shown efficient couplings via conical intersections to the ground PES followed by fast dissociation,[36,37] and we expect that these would be important in CH 3 CHOH as well. In particular, the out-of- 64 plane modes involving the methyl group are likely to increase the efficiency of such nonadiabatic couplings relative to CH 2 OH, and this may account for the lack of structure in the 300K absorption spectrum. Many radicals have low ionization energies and consequently low-lying Rydberg states. Rydberg-Rydberg and Rydberg-valence interactions are of fundamental importance to the photochemistry of these radicals. 4.2 Experimental Details The experimental apparatus and procedures have been described in detail elsewhere and here we elaborate only on procedures that have changed.[28,30,33] The radical is produced efficiently by the reaction of Cl with ethanol, where the more stable 1-hydroxyethyl radical, CH 3 CHOH, is produced with 95% yield:[6,9,17] CH 3 CH 2 OH + Cl ' HCl + CH 3 CHOH "H = - 10.7 kcal mol -1 (4) A mixture of 2.6% CH 3 CH 2 OD, CD 3 CH 2 OH (99.5% and 99% respectively, Aldrich, used without further purification) or CH 3 CH 2 OH (99.98% AAPER Alcohol, used without further purification) and ~1% Cl 2 (Matheson Tri Gas, High Purity) in He at 2 atm total pressure is prepared in a 4-L glass bulb. A piezoelectrically driven pulsed nozzle operating at 10 Hz introduces this mixture to the source region of the differentially pumped vacuum chamber. Cl 2 is photodissociated by 355 nm radiation from a Nd:YAG laser (Spectra Physics, GCR-11; 5 mJ, focused by 30 cm f.l. cylindrical lens) directed at the edge of a 4 mm long quartz tube attached in front of 65 the nozzle orifice. Cl atoms react with CH 3 CH 2 OH (CH 3 CH 2 OD) to create CH 3 CHOH (CH 3 CHOD). Radicals generated in the quartz tube undergo cooling during the supersonic expansion. The rotational temperature is estimated at 10-15 K, as shown in our work with CH 2 OH.[30] The detection chamber is separated from the source chamber by a skimmer (Beam Dynamics, 1.51 mm diameter). In the detection chamber, the pump and probe laser beams are counterpropagating and cross the molecular beam at a right angle. The pump (dissociation) radiation is generated by a Nd:YAG pumped OPO/OPA laser system (Continuum, PL8000/Sunlite/FX-1; 0.5-2 mJ at 250-400, 14- 20 mJ at 470-520 nm; 25 cm f.l. lens). The probe laser detects H and D fragments by 1+1' REMPI via the Lyman-$ transition. The doubled output (365 nm, 2 mJ) of a Nd:YAG pumped dye laser system (Continuum, NY81/ND6000, LDS 751) is focused (20 cm f.l. lens) into a 1000 Torr mixture of Kr (25%) and Ar (75%). The tripled 121.6 nm radiation is focused in the detection chamber (MgF 2 7.5 cm f.l. lens) along with the residual 356 nm light. Ions generated in the experiment are accelerated perpendicularly to the laser and molecular beams by a static electric field. A system of two electrostatic lenses provides space focusing conditions for a range of acceleration voltages. After traversing a 18 cm field-free drift region, the ions reach a multichannel plate detector (MCP, Galileo, 25 nm) installed at the top of the TOF tube. 66 The polarization of the pump laser is controlled by a photoelastic modulator (PEM-80, HINDS International, Inc.). All TOF distributions are recorded with two polarizations (parallel and perpendicular to the TOF axis), which allows calculation of the recoil anisotropy parameter ".[38] Because the ion signals were low, proper core sampling of the ion velocity distribution was impossible to achieve. Instead, the voltages applied to the extractor and repeller plates were gradually decreased while preserving space focusing conditions. At the lowest possible voltage with an acceptable signal-to-noise ratio the MCP collected all ions with a perpendicular component of translational energy (E t ) lower than 0.6 eV, whereas increasingly better discrimination against ions with velocity components not aligned with the TOF axis was achieved at higher E t . Under these conditions, an accurate center-of-mass (c.m.) E t distribution of photofragments cannot be obtained without a deconvolution procedure. However, it is still possible to correctly determine the maximum translational energy, E t,max , and identify different dissociation pathways that have distinct E t distribution. In order to find the position and width of the peak in the region of low E t, the experimental TOF spectrum was fitted with a spectrum calculated from a model velocity distribution composed of two gaussian-like distributions. This method is described in more detail by Syage.[38] The conversion of the TOF spectra to an E t scale was calibrated using E t distributions obtained in our previous experiments with CH 2 OH photodissociation.[28] 67 State Hydroxymethyl (CH 2 OH) 1-Hydroxyethyl (CH 3 CHOH) T 0 E , cm -1 (eV) Experiment a $ T 0 E , cm -1 (eV) Estimated T 0 E , cm -1 (eV) Experiment $ 3s 25971 (3.22) 1.23 18527 (2.29) 19600 (2.43) 1.20 3p x 35004 (4.34) 0.94 27695 (3.43) --- --- 3p z 41053 (5.09) 0.65 33684 (4.18) 32360 (4.01) 0.73 Table 4.1 Comparison of energies and quantum defects of Rydberg states for CH 2 OH and CH 3 CHOH. ! is obtained from the experimental results. 68 4.3 Results Two types of experiments were carried out. First, REMPI and photofragment yield spectra of H or D were recorded in regions where transitions to Rydberg states were expected in order to assign the electronic transitions. To elucidate the photodissociation dynamics, TOF spectra were obtained at selected energies and the recoil anisotropy parameters were determined 4.3.1 Photofragment yield and REMPI spectra As stated in Section 1, our search for electronic absorption was guided by results obtained for CH 2 OH.[30,33-35] In analogy with CH 2 OH, we assumed that the lowest absorption bands involved Rydberg states and that the quantum defects did not change much relative to those in CH 2 OH. As described in Section 4, this was indeed found to be true and therefore in the material that follows we use spectroscopic assignments of the upper electronic states that correspond to those in CH 2 OH. 3s region. This is the lowest excited state and, as in CH 2 OH,[34] no ions of CH 3 CHOH (CH 3 CHOD) were detected upon excitation in this region. Therefore, H(D)-photofragment yield spectroscopy was used to monitor absorption. The onset of H (or D from CH 3 CHOD) fragment signal was determined at 19,600 ± 100 cm -1 (2.43 ± 0.01 eV). Figure 4.1 displays a background subtracted D-photofragment yield spectrum in the region where the transition to the 3s Rydberg state was predicted to be 69 19000 20000 21000 0.0 0.2 0.4 0.6 0.8 1.0 D + signal, arb. units Frequency, cm -1 Figure 4.1 D photofragment yield spectrum in the region 19000 cm -1 – 21500 cm -1 . Background signal is subtracted, and the signal normalized to the OPO/OPA laser energy. 70 located. A similar spectrum was obtained by monitoring H-fragments but the signal- to-noise ratio was slightly better with CH 3 CHOD due to lower D background. 3p x region. A distinct feature of the 3p x state in CH 2 OH was a longer lifetime than the 3s state, which has made REMPI detection possible.[34] In contrast, in CH 3 CHOH no REMPI signal was observed up to the origin band of the 3p z state (see below). Photofragment yield spectra of H atoms taken in the wavelength region where absorption to 3p x was expected did not show features that could distinguish it from absorption to the 3s state. Therefore, this transition remains unassigned. 3p z region. In contrast, a distinct 2+2 REMPI spectrum was observed in the 32000-38,000 cm -1 region expected for the transition to the 3p z state. The lowest energy peak in Figure 4.2 is assigned as the origin band of the transition, which is located at 32,360 ± 70 cm -1 (4.01 ± 0.01 eV). Similarly to the corresponding CH 2 OH spectrum shown in the top panel of Figure 4.2,[30] the two other peaks, located at 1560 ± 100 cm -1 intervals, form a vibrational progression in the C-O stretch excitation. The scan was taken with a step size of 5 cm -1 , but no additional structure was observed with a finer step size. The REMPI spectrum continues to shorter wavelengths than displayed in Figure 4.2, but no further structure is observed, probably due to a combination of spectral congestion and lifetime broadening. 71 32000 33000 34000 35000 36000 Frequency, cm -1 41000 42000 43000 44000 Figure 4.2 (top) 2+1 REMPI of CH 2 OH in the region of absorption to the 3p z state (adapted from ref. 30) (bottom) 2+2 REMPI of CH 3 CHOH in the region of absorption to the 3p z state. The lowest energy band of each transition is the origin band. 72 4.3.2 Time of flight of H(D) photofragments Figure 4.3 displays a typical background-subtracted TOF spectrum of H atoms recorded following excitation at 21,212 cm -1 (2.63 eV; 470 nm) with laser polarization parallel and perpendicular to the ion TOF axis. The corresponding E t distribution and the recoil anisotropy parameter "(E t ) are shown on Figure 4.4. The maximum available energy is determined by energy conservation: E avail = E h" – E p int – D 0 . The parent internal energy, E p int , is negligible in the supersonic expansion and we use E h" = 2.63 eV. From Figure 4.3 we obtain E t,max , the maximum observed E t : E t,max = E avail = 1.53 eV ± 0.1 eV, which leads to D 0 = 1.1 eV ± 0.1 eV, in good agreement with previous experimental and theoretical estimates (1.13 eV ± 0.04 eV and 1.04 eV, respectively).[21,25,39] The same D 0 value is obtained from E t,max at other wavelengths near the onset of the transition to the 3s state. The photofragment E t distribution exhibits a single anisotropic peak with " = –0.7 ± 0.1 Figure 4.5 displays the c.m. E t distribution obtained by monitoring H fragments from photodissociation of CH 3 CHOH at 31250 cm -1 (3.87 eV) excitation energy with parallel and perpendicular polarizations of the pump laser. For high E t the recoil distribution is anisotropic, but with a lower anisotropy parameter than was observed at lower excitation energies (near the onset of absorption to the 3s state), and " = – 0.4 ± 0.1 is obtained for both CH 3 CHOH and CH 3 CHOD. In addition, a small isotropic peak (" = –0.0 ± 0.1) appears in the low E t region. The peak of slow photofragments is absent in Figure 4.6, which shows an E t distribution for CH 3 CHOD 73 -40 -20 0 20 40 H + signal, arb. units Polarization: % || Relative TOF, ns Figure 4.3 H fragment time-of-flight spectrum from CH 3 CHOH obtained at 21,276 cm -1 excitation. The polarization of the pump laser is alternated between parallel (solid line) and perpendicular (dashed line) to the TOF axis. Background is subtracted. Zero time indicates fragment with no recoil. 74 dissociation obtained at the same excitation energy by recording the D fragment TOF spectrum. A notable feature of this distribution is the difference between E t,max and the available energy: E avail = E h" – D 0 = 3.87 – 1.2 = 2.67 ± 0.1 eV (where D 0 is adjusted for zero point energy), whereas E t,max = 2.2 ± 0.1 eV, giving, E avail – E t,max = 0.47 ± 0.14 eV as the minimum internal energy in the molecular co-fragment. In contrast, in the case of CH 3 CHOH, E t,max and E avail are equal within the error bars. H- and D-fragment TOF spectra in the 21000 – 31300 cm -1 excitation range show similar trends. Figures 7 and 8 show typical c.m. E t distributions obtained following excitation of CH 3 CHOH (CH 3 CHOD) at 35,460 cm -1 (the third peak in Figure 4.2). The two peaks in Figure 4.7 become broader than those at lower excitation energy, and the recoil anisotropy parameters of the high E t peaks are lower (" = -0.2 ± 0.1). The low E t peak grows in relative intensity but remains isotropic. The gap between E t,max and E avail increases to 0.39 eV (0.79 eV) for CH 3 CHOH (CH 3 CHOD). The E t,max - E avail gap increases gradually with excitation energy up to the highest measured excitation energy of 35,460 cm -1 , while the " parameter decreases with increasing excitation energy. The dependence of " on excitation energy is shown in Figure 4.9. A TOF distribution was also recorded by monitoring D atoms from the CD 3 CHOH isotopolog. It showed only a slow peak comparable to the low E t peak in Figure 4.7, in contrast to the D-fragment TOF spectrum from CH 3 CHOD, which displayed only the high E t peak. 75 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 P(E t ) E t , eV Polarization: % || E t,max = 1.53 eV Monitor H -1 0 & Figure 4.4 The c.m. E t distribution obtained by monitoring H photofragments following the 1 2 A(3s) ! 1 2 A transition. at 21,276 cm -1 .(2.63 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel. 76 4.4 Discussion 4.4.1 Assignment of the 3s and 3p z states The low ionization energy of CH 2 OH results in electronic Rydberg states that are lower in energy than the lowest valence states.[32,36,40] Since the ionization energy of CH 3 CHOH is almost 1 eV lower than that of CH 2 OH (6.64 eV versus 7.56 eV)[25,26] the lowest-lying electronic states of the former are also expected to be of Rydberg nature. The positions of the states can be estimated by the Rydberg formula: E = IE ) R / (n ) #) 2 , where R = 109737.316 cm -1 , n is the principal quantum number and # is the quantum defect obtained experimentally from the energies of the corresponding Rydberg states of CH 2 OH (see Table 4.1). The first excited electronic state is the 3s state. The onset of the H- photofragment yield spectrum is at 19,600 ± 100 cm -1 , which is close to the predicted value of 18,530 cm -1 . Our inability to observe a CH 3 CHOH + REMPI signal despite the stability of the ion[12,14] suggests that internal conversion from the 3s to the ground state with subsequent dissociation is faster than ionization from the Rydberg state. A similar conclusion was reached with regard to CH 2 OH in the 3s state.[29,34] Notice that CH 3 CHOH starts absorbing at 510 nm, which makes it unstable in visible light. Another feature common to CH 2 OH and CH 3 CHOH is the strong negative recoil anisotropy of the photofragment distribution in excitation to the 3s state 77 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Monitor H P(E t ) E t , eV Polarization: % || E avail = 2.77 eV -1 0 & Figure 4.5 The c.m. E t distribution obtained by monitoring H photofragments following excitation of CH 3 CHOH at 31,250 cm -1 (3.87 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel. 78 (" 3s = – 0.7 ± 0.1 for both radicals near the onset of the transition). The ground state equilibrium geometry of both CH 2 OH and CH 3 CHOH is non-planar.[12,14] In CH 2 OH, a low barrier to CH 2 inversion (140 cm -1 ) allows the two mirror image equilibrium structures of C 1 symmetry to rapidly interconvert.[41-43] As a result, the electronic wavefunction of CH 2 OH complies with the C s point group. This gives the transition from the ground state (A”) to the 3s and 3p x states, which are of A’ symmetry, a perpendicular moment. The observation of similar anisotropy in the transition to 3s for both radicals suggests similar orientation of the transition dipole moment relative to the O-H bond (the only bond fission channel observed near the onset of 3s state excitation) as well as similar symmetry of the electronic wavefunctions. Using C s symmetry for the electronic transitions of CH 3 CHOH, we assign the first two Rydberg states as 1 2 A'(3s) and 2 2 A'(3p x ). For CH 2 OH, the transition to 3p y was calculated to be very weak, and was not observed;[31,40] likewise, we do not see any evidence for a transition to this state in CH 3 CHOH. The 3p x state of CH 3 CHOH could not be identified in our experiment. No REMPI of CH 3 CHOH was seen around the predicted onset of the transition (about 27,700 cm -1 ) and no discernible change in the H-photofragment spectrum was detected. In the hydroxymethyl radical, fast predissociation of the 3p x state caused considerable broadening of the peaks in the REMPI spectrum.[34] Since CH 3 CHOH has twice as many vibrational degrees of freedom as CH 2 OH, it is likely that the 3p x state predissociates even faster, reducing its ionization efficiency 79 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Monitor D P(E t ) E t , eV Polarization: % || E avail = 2.67 eV -1 0 & Figure 4.6. The c.m. E t distribution obtained by monitoring D photofragments following excitation of CH 3 CHOD at 31,250 cm -1 (3.87 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel. 80 and broadening the peaks beyond the sensitivity of our measurements. Apparently, H- photofragment spectra in this excitation region are dominated by 3s state absorption and show no special features that indicate the onset of additional absorption to another state. The first REMPI signal of CH 3 CHOH appears at 32,360 cm -1 , near the predicted origin of absorption to the 3p z (2 2 A”) state at 33,684 cm -1 . In keeping with a model of faster predissociation than in CH 2 OH, the peaks are much broader than for the 3p z state of CH 2 OH (Figure 4.2). As in hydroxymethyl, they form a vibrational progression with peaks separated by 1560 ± 100 cm -1 , close to the frequency of the C- O stretch vibration in the ion.[25] The progression can be rationalized by the promotion of an electron from the C-O antibonding $ * orbital to a non-bonding Rydberg orbital in excitation to the 3p z state, which leads to a shortening of the C-O bond length. In contrast to the H-photofragment yield spectrum of CH 2 OH in the region of the 3p z state, which shows the same vibronic bands as the REMPI spectrum, there was no discernible structure in the H(D) yield spectra of CH 3 CHOH(D) above the onset of the transition to 3p z , indicating that absorption to the 3s state still dominates the absorption spectrum in this region. It is noteworthy that a 2+1 REMPI spectrum of CH 3 CHOH was reported in the region around 430-460 nm (2h' = 46,510- 43,480 cm -1 ),[24] which corresponds according to the Rydberg formula to one of the 4p states. Higher Rydberg states usually couple less efficiently to other Rydberg and 81 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 P(E t ) E t , eV Polarization: % || -1 0 Monitor H & E avail = 3.29 eV Figure 4.7 The c.m. E t distribution obtained by monitoring H photofragments following excitation of CH 3 CHOH at 35,460 cm -1 (4.39 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel. 82 valence states, and their dissociation is less rapid, which allows observation of the 2+1 REMPI signal. 4.4.2. Photodissociation dynamics The conclusion from the spectroscopic studies is that throughout the examined region, which encompasses absorptions to the 3s, 3p x and 3p z Rydberg states, the predominant absorption is to the 3s state. Therefore, in the discussion below, we assume that dissociation is initiated from 3s, although contributions from 3p x excitation are likely at shorter wavelengths. The ground state of the CH 3 CHOH + ion is bound, and therefore so are the Rydberg states of the neutral, provided they do not interact with other states. As Rydberg states correlate with excited state products, the 3s Rydberg state must couple to the ground state in order to predissociate to ground state products. Near the onset of the 3s state, dissociation channels (1) and (2) are energetically accessible; i.e., both the O-H (D 0 = 1.04 eV) and the C (2) -H (D 0 = 1.46 eV) bond fission channels,[21] where C (2) is the terminal carbon atom, are allowed. Both channels have similar barrier heights on the ground state PES (calculated at 1.52 eV and 1.58 eV respectively).[21] Had internal conversion from 3s to the ground state been followed by statistical redistribution of energy among the vibrational modes on the ground PES, one would expect close competition between these two simple bond fission channels above the barrier. However, the E t distribution near the onset of the absorption to 3s is clearly 83 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 P(E t ) E t , eV Polarization: % || -1 0 Monitor D & E avail = 3.19 eV Figure 4.8. The c.m. E t distribution obtained by monitoring D photofragments following excitation of CH 3 CHOD at 35,460 cm -1 (4.39 eV). The arrow indicates the available energy. The " parameter is plotted as a function of E t in the top panel. 84 non-statistical as most of E avail is deposited in translation. Also, only the O-H bond fission channel (1) is detected, even though channel (2) is energetically accessible. This follows from a comparison of the E t distributions obtained by monitoring H- and D-fragments from dissociation of CH 3 CHOH and CH 3 CHOD, respectively: each exhibits a single peak of the same shape, with E t,max = E avail corresponding to D 0 = 1.1 ± 0.1 eV. At 3.87 eV (31,250 cm -1 ) excitation, however, a second dissociation channel is open, as seen in the H-fragment E t distribution of Figure 4.5. It exhibits an additional peak in the region of low E t , which is not present in the corresponding D-fragment E t distribution (Figure 4.6). Deconvolution of the H-fragment TOF spectrum obtained by assuming two Gaussian-shape distributions of E t (Section 2) shows that the channel represented by the slow peak must have D 0 < 2.7 ± 0.1 eV. This value is larger than the calculated D 0 value of the C (2) -H bond fission (1.46 eV) but smaller than D 0 for C (1) -H bond fission (calculated at 3.37 eV).[21,27] The interpretation that the low E t peak arises from channel (2) giving rise to CH 2 CHOH product agrees with the result of the D-fragment TOF spectrum of CD 3 CHOH, which showed only the low E t peak. In other words the C (2) -H bond fission channel opens > 1 eV higher than the O-H bond fission channel, despite their similar barrier heights on the ground state PES. We conclude, therefore, that the dynamics on the 3s state PES must be responsible for this difference. 85 20000 25000 30000 35000 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 & Frequency()cm *+ Figure 4.9 Recoil anisotropy parameter " as a function of excitation energy. 86 In CH 2 OH, calculations identified an efficient (vertical) conical intersection between the 3s and the ground state, which was located in the PES region of elongated O-H bond [36,37]. This geometry favors fast and efficient dissociation of the O-H bond. A second conical intersection, located at much higher energies, led to C-H bond fission.[36,37] It is reasonable to assume that in CH 3 CHOH there are at least two conical intersections between the 3s and the ground state, each favoring one of the two exit channels. Thus, after the radical goes through the conical intersection it dissociates predominantly through the corresponding channel. Accessibility to the channel in each case is controlled by the geometry and minimum energy of the intersection on the 3s PES. Referring to the properties of the conical intersection, it appears that certain vibrations must be active for efficient surface coupling to take place. As the excitation energy increases in the ground state PES following internal conversion, an increasing fraction of the parent vibrational energy is deposited into modes that appear finally as internal modes (vibration and rotation) of the acetaldehyde fragments. This is manifested in the broadening of the high energy peak in the E t distributions and the increasing value of E avail -E t,max . The latter is especially pronounced in CH 3 CHOD. The slow vinyl alcohol product from channel (2) also has a broad internal energy distribution. It appears that some of the excited parent vibrational levels involve out-of- plane motions, which reduce the recoil anisotropy in the perpendicular transition. 87 Indeed, the recoil anisotropy parameter " becomes less negative as the excitation energy increases (Figure 4.9), concomitant with the internal energy buildup in the products. This effect is much less noticeable in CH 2 OH(D) where the " parameter is nearly constant over the entire range of perpendicular transitions to 3s and 3p x , and a sudden change in anisotropy occurs only at the onset of the parallel transition to the Rydberg 3p z state.[28,30,34] 4.5 Conclusions Electronic transitions to low-lying Rydberg states and photodissociation dynamics of the CH 3 CHOH(D) radical in the region 19,600 - 37,000 cm -1 were studied for the first time in a molecular beam. On the basis of the REMPI and H(D)- photofragment yield spectra we conclude that the predominant absorption in the observed region is to the lowest excited state -- the Rydberg 3s state -- whose onset is at 19,600 cm -1 . The origin band of the transition to the 3p z state lies at 32,360 cm -1 . The onsets of the observed absorptions agree well with predictions of the Rydberg formula with quantum defects similar to those obtained for the corresponding observed transitions in CH 2 OH. To identify the photodissociation channels, TOF spectra of D- and H- photofragments were recorded for CH 3 CHOH, CH 3 CHOD and CD 3 CHOH. At dissociation energies near the onset of absorption to the 3s state, c.m. fragment E t distributions from CH 3 CHOH show only one peak at high E t , close to the maximum 88 allowed by energy, with a recoil anisotropy parameter " typical of a perpendicular transition. At higher excitation energies, a second, low E t peak appears, which is isotropic; its relative intensity increases at higher excitation energies. The TOF-spectra obtained with the three isotopologs allow us to identify two independent dissociation channels, one leading to acetaldehyde and the other to vinyl alcohol (enol): (1) CH 3 CHO + H; (2) CH 2 CHOH + H. There is no indication of isomerization to ethoxy. The former channel appears at the onset of the 3s absorption, whereas the latter first appears only ~ 1 eV above its thermochemical threshold. We suggest that conical intersections with the ground state lead to O-H bond fission [channel (1)] from the onset of absorption to 3s, but the conical intersection leading to the enol product [channel (2)] has a minimum energy much above the thermochemical threshold for this channel. Another notable feature of the dissociation is the increasing amount of internal energy in acetaldehyde [channel (1)] as the excitation energy increases. The less negative " of the fast peak at higher excitation energies may result from out-of-plane motions during dissociation of vibrationally excited molecules generated in the nonadiabatic transition with the ground electronic state. These results reflect the dynamics in the region of the conical intersections as well as dynamics in the exit channel. From experiment and theory we conclude that the Rydberg formula should hold rather well in the homologous series of CH 2 OH, serving as a guide to the 89 absorption spectra of higher hydroxyalkyl radicals whose radical center is located on the C adjacent to OH (1-position). The structureless electronic absorption extends from the visible to the UV, making these radicals unstable to visible light. Also, nonadiabatic transitions coupling the Rydberg states to the ground state are expected to be efficient, resulting in fast dissociation to several product channels. 90 Chapter 4 References (1) Lin, J. J.; Harich, S.; Lee, Y. T.; Yang, X. Journal of Chemical Physics 1999, 110, 10821. (2) Lin, J. J.; Shu, J.; Lee, Y. T.; Yang, X. Journal of Chemical Physics 2000, 113, 5287. (3) Fockenberg, C.; Hall, G. E.; Preses, J. M.; Sears, T. J.; Muckerman, J. T. Journal of Physical Chemistry A 1999, 103, 5722. (4) Marcy, T. P.; Diaz, R. R.; Heard, D.; Leone, S. R.; Harding, L. B.; Klippenstein, S. J. Journal of Physical Chemistry A 2001, 105, 8361. (5) Seakins, P. W.; Leone, S. R. Journal of Physical Chemistry 1992, 96, 4478. (6) Ahmed, M.; Peterka, D. S.; Suits, A. G. Physical Chemistry Chemical Physics 2000, 2, 861. (7) Pagsberg, P.; Munk, J.; Anastasi, C.; Simpson, V. J. Journal of Physical Chemistry 1989, 93, 5162. (8) Pagsberg, P.; Munk, J.; Sillesen, A.; Anastasi, C. Chemical Physics Letters 1988, 146, 375. (9) Rudic, S.; Murray, C.; Ascenzi, D.; Anderson, H.; Harvey, J. N.; Orr- Ewing, A. J. Journal of Chemical Physics 2002, 117, 5692. (10) Anastasi, C.; Simpson, V.; Munk, J.; Pagsberg, P. Chemical Physics Letters 1989, 164, 18. (11) Anastasi, C.; Simpson, V.; Munk, J.; Pagsberg, P. Journal of Physical Chemistry 1990, 94, 6327. (12) Curtiss, L. A.; Lucas, D. J.; Pople, J. A. Journal of Chemical Physics 1995, 102, 3292. (13) Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1994, 101, 10936. (14) Nobes, R. H.; Rodwell, W. R.; Bouma, W. J.; Radom, L. Journal of the American Chemical Society 1981, 103, 1913. 91 (15) Choi, H.; Bise, R. T.; Neumark, D. M. Journal of Physical Chemistry A 2000, 104, 10112. (16) Faulhaber, A. E.; Szpunar, D. E.; Kautzman, K. E.; Neumark, D. M. Journal of Physical Chemistry A 2005, 109, 10239. (17) Taatjes, C. A.; Christensen, L. K.; Hurley, M. D.; Wallington, T. J. Journal of Physical Chemistry A 1999, 103, 9805. (18) Cleary, P. A.; Romero, M. T. B.; Blitz, M. A.; Heard, D. E.; Pilling, M. J.; Seakins, P. W.; Wang, L. Physical Chemistry Chemical Physics 2006, 8, 5633. (19) Hoyermann, K.; Olzmann, M.; Seeba, J.; Viskolcz, B. Journal of Physical Chemistry A 1999, 103, 5692. (20) Lindner, J.; Loomis, R. A.; Klaassen, J. J.; Leone, S. R. Journal of Chemical Physics 1998, 108, 1944. (21) Senosiain, J. P.; Klippenstein, S. J.; Miller, J. A. Journal of Physical Chemistry A 2006, 110, 6960. (22) Tully, F. P. Chemical Physics Letters 1988, 143, 510. (23) Taatjes, C. A.; Hansen, N.; McIlroy, A.; Miller, J. A.; Senosiain, J. P.; Klippenstein, S. J.; Qi, F.; Sheng, L. S.; Zhang, Y. W.; Cool, T. A.; Wang, J.; Westmoreland, P. R.; Law, M. E.; Kasper, T.; Kohse-Hoinghaus, K. Science 2005, 308, 1887. (24) Edelbuttel-Einhaus, J.; Hoyermann, K.; Rohde, G.; Seeba, J. Proceedings of the 24th Symposium International on Combustion, Combustion Institute, Pittsburgh, PA 1992, 661. (25) Dyke, J. M.; Groves, A. P.; Lee, E. P. F.; Niavaran, M. H. Z. Journal of Physical Chemistry A 1997, 101, 373. (26) Dyke, J. M.; Ellis, A. R.; Jonathan, N.; Keddar, N.; Morris, A. Chemical Physics Letters 1984, 111, 207. (27) Klippenstein, S. private communication (28) Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2003, 118, 9623. 92 (29) Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2004, 120, 6524. (30) Aristov, V.; Conroy, D.; Reisler, H. Chemical Physics Letters 2000, 318, 393. (31) Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 1998, 102, 3141. (32) Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 2001, 105, 8599. (33) Conroy, D.; Aristov, V.; Feng, L.; Reisler, H. Journal of Physical Chemistry A 2000, 104, 10288. (34) Feng, L.; Huang, X.; Reisler, H. Journal of Chemical Physics 2002, 117, 4820. (35) Feng, L.; Reisler, H. Journal of Physical Chemistry A 2004, 108, 9847. (36) Hoffman, B. C.; Yarkony, D. R. Journal of Chemical Physics 2002, 116, 8300. (37) Yarkony, D. R. Journal of Chemical Physics 2005, 122. (38) Syage, J. A. Journal of Chemical Physics 1996, 105, 1007. (39) Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. Journal of Physical and Chemical Reference Data 1988, 17, 1. (40) Rettrup, S.; Pagsberg, P.; Anastasi, C. Chemical Physics 1988, 122, 45. (41) Marenich, A. V.; Boggs, J. E. Journal of Chemical Physics 2003, 119, 10105. (42) Marenich, A. V.; Boggs, J. E. Journal of Chemical Physics 2003, 119, 3098. (43) Johnson, R. D.; Hudgens, J. W. Journal of Physical Chemistry 1996, 100, 19874. 93 ! Chapter 5. Effect of Hyperconjugation on Ionization Energies of Hydroxyalkyl Radicals 5.1 Introduction Hydroxyalkyl radicals are important intermediates in combustion and atmospheric processes. [1,2] The most studied of these radicals are the hydroxymethyl and 1-hydroxyethyl radicals. [3-12] They are produced in reactions of the corresponding alcohols with atoms (F, Cl, O) and OH by abstracting a carbon-bound H atom and are the most stable structural isomers among the CH 3 O and C 2 H 5 O species, which have been implicated as important reaction intermediates. One of the notable characteristics of hydroxyalkyl radicals, as compared for example with the corresponding alcohols or the isomeric alkoxy radicals, is their low ionization energies (IEs). The adiabatic IEs of the hydroxymethyl and hydroxyethyl radicals are 7.56 and 6.64 eV, respectively, [13,14] while the corresponding values for methanol, ethanol, methoxy, and ethoxy radicals are 10.85, 10.41, 10.72, and 9.11 eV, respectively.[15-18] What is striking is not only the large reduction in ionization energy in going from the closed- to the open-shell compounds but also the large decrease (about 0.9 eV) in going from hydroxymethyl to the hydroxyethyl radical, which is much larger than the 0.4 eV difference between the corresponding alcohols. 94 ! The goal of the present work is to offer a physical explanation for this decrease in IE by using high-level electronic structure calculations and in particular by comparing the highest occupied, singly occupied, and lowest unoccupied molecular orbitals (HOMO, SOMO, and LUMO, respectively) of the radicals and the cations. Specifically, our analysis highlights the role of hyperconjugation in stabilizing the hydroxyethyl cation and destabilizing the radical, compared to hydroxymethyl. Hyperconjugation is a concept used often in physical organic chemistry to describe conjugation effects that involve #-bonds in addition to $-bonds. Both conjugation between double or triple bonds and hyperconjugation involve delocalization of charge density over several atoms or groups of atoms. This correspondence was recognized early on by Mulliken and others, [19,20] who considered the CH 3 group in a molecule as equivalent to one triple bond with respect to electron donation. Hyperconjugation is responsible, for example, for the stability of secondary and tertiary radicals and cations, [19] the changes in bond strengths [21] and conformations [20] upon substitutions, the vibrational spectra and structures of hydrocarbon radicals,[22, 23] the trends in electronically excited states in alkylperoxy radicals, [24] and more. There is also a strong evidence that the torsional barrier in ethane is due to hyperconjugation [25-27] rather than steric repulsion, although this view on the primary cause for the staggered conformation is not uniformly accepted.[28,29] 95 ! With the advent of high-level electronic structure models, it is now possible to analyze the delocalization of electronic density in terms of electronic configuration and changes in nuclear geometry. Such insight can help understand trends in IEs and stability of ions in a homologous group. In general, a reduction in IE can originate from the stabilization of the cation and/or destabilization of the neutral, and electronic structure models can help analyze the contributions of different factors. In the case of the hydroxyethyl radical and its cation, we show below that both factors above contribute to the observed large decrease in IE in going from hydroxymethyl to hydroxyethyl. Ab initio calculations of the geometry and energy of hydroxymethyl [30,31] and hydroxyethyl [10,32] radicals and cations have been published before, and our calculations (see Figure 5.1) agree with previous results within error bars of the theoretical methods employed. The CH 2 OH radical has C 1 nuclear symmetry with strongly coupled OH torsional and CH 2 wagging motions. [31] In contrast, the cation has a planar C s symmetry and can be viewed as protonated formaldehyde. The major change upon ionization is the reduction in the C%O bond length, which acquires $- bond character upon removal of an electron from the antibonding SOMO centered on the C%O bond. Thus, ionization creates a closed-shell cation that is quite stable. Similar behavior has been observed in the cations of halogenated methyl radicals.[33,34] Adding a methyl group does not change this situation qualitatively, although the shape of the HOMO, SOMO, and LUMO change due to admixture of 96 ! ! Figure 5.1 Equilibrium structures and nuclear repulsion energies (V NN ) of the CH 2 OH and CH 3 CHOH radicals and their cations optimized at the CCSD(T)/cc-pVTZ level of theory. Angles listed in parentheses correspond to the neutral radicals. ! 97 ! # CH . The hydroxyethyl radical has an unpaired electron in an antibonding orbital and possesses C 1 symmetry, whereas the cation can be viewed as a protonated acetaldehyde having C s symmetry. As in hydroxymethyl, the main effect of ionization is a large contraction of the C%O bond, which (in hydroxyethyl) is accompanied by a small contraction of the C%C bond. Both bonds acquire additional double-bond character upon removal of the unpaired electron. In hydroxyethyl, the H atom in O%H prefers the anti position, as shown in Figure 5.1, although the gauche isomer is less than 1 kcal/mol higher in energy. [32] The orientation of the methyl hydrogens relative to the OCC plane changes upon ionization, with one of the methyl hydrogens in the cation being in the same plane as OCC, whereas none is in that plane in the neutral (Figure 5.1). Again, the anti isomer is only marginally more stable than the syn structure. [32] In agreement with the analysis of Hoffman et al., [20] the barriers to internal rotation of the methyl group are less than 2 kcal/mol both in the neutral and in the cation. Thus, the minimum-energy positions of the methyl hydrogens are determined by subtle effects related to the removal of the unpaired electron and the concomitant skeletal rearrangement, in particular, the contraction of the O%C and C%C bonds. Because of the presence of the OH group, the hyperconjugation in hydroxyethyl involves three rather than two interacting MOs, which is the case in the ethyl radical. The participating orbitals are the p-like orbitals on C and O and one # CH MO of the methyl group that has favorable overlap with the p-like orbitals. In 98 ! ! Figure 5.2 Three MOs resulting from hyperconjugation of # CH , LP(O), and LP(C) in CH 3 CHOH at the cation and radical geometries. The bonding and nonbonding MOs are doubly occupied in both the cation and the neutral. The antibonding orbital is singly occupied in the radical and is unoccupied in the cation. At the neutral geometry (right), the shape of the orbitals is slightly different, but their character is preserved. For the sake of clarity, we refer to these three orbitals as SOMO, HOMO, and HOMO-1 of the radical, although energetically there are other MOs (which do not participate in hyperconjugation) in between the nonbonding and bonding MOs. 99 ! molecular orbital language, these orbitals, which have the same symmetry, can interact, creating a new set of three allyl-like delocalized orbitals as shown in Figure 5.2. The bonding and antibonding character of these delocalized MOs and the extent to which they are filled are responsible for the low IE of 1-hydroxyalkyl radicals and for the stabilization/destabilization of the neutral and cation species relative to the case of no hyperconjugation interaction. The NBO analysis confirms this qualitative picture, and the computed hyperconjugation energies are in qualitative agreement with the estimates derived from a simple Hückel-like model. Our main conclusion is that hyperconjugation destabilizes the neutral hydroxyethyl radical due to the antibonding character of the SOMO (Figure 5.2), while stabilizing the cation by lowering the energies of the HOMO and HOMO-1 relative to hydroxymethyl. Additional stabilization is achieved owing to the more extensive charge delocalization in hydroxyethyl. 5.2 Computational details All geometries were optimized by the coupled-cluster method with single and double substitutions and perturbative account of triples [CCSD(T)] [35, 36] with the cc-pVTZ basis set [37] using the ACES II electronic structure package. [38] Optimized geometries are summarized in Figure 5.1. The optimized geometry of CH 3 CHOH deviates by 0.1° from C s symmetry. Unrestricted Hartree%Fock references were used in all radical calculations. The spin contamination was found to be small; ! e.g., S 2 eq possible effe shell referen CCSD(T) en CCSD(T) op CHEM,[41] as noted bel calculations Vert cation and th and R-CCSD the CCSD(T Table 5.1. T energy of th are directly interpretatio The by analytic on the magn the 6-31G(d quals 0.76 fo fect of spin c nce and parti nergies were ptimized geo ] again using low. R-CCSD s, all electron tical IEs wer he neutral ra D(T) levels. T)/cc-pVTZ Table 5.1 also he UHF SOM related to th on. NBO [42] c gradients an nitude of hyp d) basis set, a or both radic contaminatio ially spin-ad e calculated b ometries. Al g the cc-pVT D(T) energie ns were corr re calculated adical at the Adiabatic IE optimized g o lists Koop MO of the ne he orbital ene calculations w nd properties perconjugati assuming a r cals at their e on, we also c dapted CCSD by MOLPRO ll other calcu TZ basis set, es were com elated. d as the diffe equilibrium Es were com geometries. T mans IEs co eutral radical ergies and ar were perform s code. [43] T ive interactio rigid molecu equilibrium g omputed IEs D(T) method O [40]with t ulations were except for s mputed with f rence betwe geometry o mputed as the The calculate omputed as th l. Although re therefore u med using th The effect of ons in CH 3 C ular structure geometries. s using restr d, R-CCSD(T the same bas e performed some of the N frozen core. een the total f the latter, a e total energ ed IEs are su he absolute less accurate useful for th he CCSD den f rotation of CHOH was ex e (i.e., no geo To account f ricted open- T). [39] R- sis set, at the with Q- NBO analys In all other energies of t at CCSD(T) gy difference ummarized in value of the e, these valu he nsity compu f the CH 3 gro xamined usi ometrical 100 for e es, the es at n ues uted oup ing 101 ! relaxation of any coordinate was allowed upon rotation) using parameters from Figure 5.1. All the other NBO calculations employed the cc-pVTZ basis set. Table 5.1. Computed Vertical and Adiabatic IEs (eV) of CH 3 CHOH and CH 2 OH Method vertical adiabatic CH 2 OH CH 3 CHOH CH 2 OH CH 3 CHOH CCSD(T)/cc-pVTZ 7.897 7.292 7.285 6.516 RCCSD(T)/cc-pVTZ 7.865 7.260 7.282 6.532 Koopmans’ theorem 9.198 8.724 -- -- Expt a 8.14 7.29 7.56 6.64 a References [13]and [14]. 5.3 Results and Discussion In the discussion below, we analyze the results of electronic structure calculations of the radicals and cations in a molecular orbital framework. We also employ the NBO analysis [42] to support the qualitative conclusions obtained on the basis of the molecular orbital analysis. The NBO procedure allows one to represent the total molecular electron density (either correlated or not) in terms of contributions from core, lone pairs, localized bonding, and antibonding orbitals. The advantage of NBO is that it produces a Lewis structure from delocalized electron density, thus providing chemical insight. For well-behaved closed-shell molecules, a large fraction (more than 95%) of the electron density fits a single Lewis structure consistent with ! chemical bo excited spec found that th hydroxyethy properties (e less clear. Mor bonding and hyperconjug scheme) res CH 3 CHOH + energies of t Thes derived from emphasize t procedure a only semiqu and partly b We b its cation sh substituted m onding theor cies, the resu he NBO inte yl cations is e.g., structur reover, NBO d antibondin gation as the sulting from + , we compu the three # CH se energies c m IEs using that even tho nd uses high uantitative, p because of th begin by con hown in Figu methyl radic ies. Howeve ults of NBO erpretation o consistent w res), whereas O is capable o ng orbitals. T e interaction charge deloc uted hyperco H of CH 3 wit can be comp a simple H ough the NB hly accurate partly becaus he nonadditiv nsidering the ure 5.3. Altho cal, its electr er, in the cas decompositi of the bondin with the mole s in the case of quantifyin This allows o energies (co calization am onjugation en th the $ CO pared with es ckel-like pi O analysis i electron den se of the sim vity of differ e relevant M ough hydrox ronic structur e of open-sh ions are not ng in the hyd ecular orbita of the radic ng interactio one to compu omputed by a mong relevan nergies as th orbital. stimates of h icture, as des s based on a nsity, the res mplicity of th rent contribu MOs of the hy xymethyl can re is conside hell and elect always mea droxymethyl al analysis an cals the NBO ns between d ute energies a second-ord nt bond orbi e sum of the hyperconjuga scribed below a well-define sulting energ he underlying utions to the ydroxymethy n be describ erably differe tronically ningful. We and nd molecular O picture was different of der perturbat itals. In e interaction ation effects w. We ed mathemat gy estimates g Lewis pictu total energy yl radical an ed as an OH ent. The lon 102 r s tion tical are ure y. d H- e 103 ! ! Figure 5.3. Frontier MOs of hydroxymethyl. Oxygen lone pair, LP(O), and LP(C) form bonding and antibonding $-like orbitals hosting three electrons in the radical. 104 ! pair of oxygen is sufficiently close in energy to the carbon p z orbital such that they form bonding and antibonding $-type orbitals. Consequently, the CO bond acquires partial $-bond character, which further increases upon ionization. This interaction destabilizes the SOMO, which results in a rather low IE of 7.56 eV (8.14 eV vertically), much lower than methyl’s IE of 9.84 eV and slightly lower than the 8.12 eV IE of CH 3 CH 2 [44,45] In hydroxyethyl, the character of the SOMO changes further due to hyperconjugation interactions with the # CH orbitals of the CH 3 group. As described above, in hydroxyethyl and other radicals in which the radical center is adjacent to a lone pair (LP), one needs to consider the interactions of three orbitals, LP(O), LP(C), and # CH , yielding the bonding, nonbonding, and antibonding orbitals shown in Figure 5.2, which are remarkably similar to the allylic orbitals. This pattern differs from the more familiar hyperconjugation scenario involving two interacting orbitals, e.g., LP(C) and # CH in hydrocarbon radicals. [23] The vertical IE of CH 3 CHOH is 7.29 eV, which is 0.85 eV lower than that of CH 2 OH. The observed trends in IEs can be analyzed in terms of a simple Huckel-like model. For the sake of simplicity, consider two interacting orbitals (e.g., LP(C) and # CH ) producing bonding and antibonding combinations split by 20. The case of three interacting orbitals is essentially the same, since the nonbonding allylic orbital does not contribute to the energies discussed below. In the radical, the bonding orbital is doubly occupied, and the antibonding orbital hosts one electron. Neglecting 105 ! electron%electron interaction, the change in IE of the unpaired electron relative to the noninteracting orbitals is 0, and the stabilization energy of the cation due to delocalization is 20. Thus, in this simple model, the hyperconjugation energy in the cation is double the reduction in the IE of the radical. To evaluate the change in IE due to hyperconjugation, we take 0.85 eV (the observed drop in vertical IE between hydroxymethyl and hydroxyethyl) and subtract from it 0.32 eV, the change in vertical IEs between the two saturated compounds, ethanol (10.64 eV) and methanol (10.96 eV).[46, 47] This allows us to separate the reduction in IE due to hyperconjugation from the reduction in IE due to the increase in molecular size. Thus, our estimate based on vertical IEs is 0 = 0.53 eV. As will become evident below, this number is in remarkable agreement with simple NBO calculations of hyperconjugation energy. Consistent with the MO picture, both radicals undergo significant geometrical relaxation upon ionization. As summarized in Figure 5.1, ionization induces changes in the CO bond length, which is consistent with removing an electron from an antibonding orbital. The effect is slightly larger in CH 2 OH (0.117 Å) relative to CH 3 CHOH (0.111 Å) because the participating # CH orbitals in the SOMO of the latter dilute its CO antibonding character. In agreement with this explanation, the CC bond is also contracted in the cation (by 0.032 Å). Ionization also induces rotation of the methyl group by 64°; however, the overall energy effect of this relaxation is small. Note that the CH bonds participating in hyperconjugation change accordingly, resulting in unequal bond lengths; i.e., the CH bonds involved in hyperconjugation are 106 ! weakened due to the donation of some of their electron density into the $* CO orbital; thus they are elongated. Because of significant geometry relaxation, the adiabatic IEs of CH 2 OH and CH 3 CHOH are lower than the vertical values by 0.58 and 0.65 eV, respectively. Since the changes in CO bond lengths are very similar in both radicals, the difference between the two values, 0.07 eV, can be interpreted as the increase in 0 due to stronger hyperconjugation interactions at the cation geometry and can be compared with the increase in the hyperconjugation energy in the hydroxyethyl cation upon relaxation computed by NBO. The NBO procedure confirms the bonding pattern described above. The Lewis structure of the cation includes a double CO bond at both the radical and the cation equilibrium geometries. The NBO calculation of the hyperconjugation energy (the delocalization energy between $ CO and all # CH ) of the hydroxyethyl cation at the equilibrium geometry of the radical is 0.98 eV, yielding 0 = 0.49 eV, which is in a good agreement with the value of 0.53 eV estimated from the IE differences. At the cation equilibrium geometry, the hyperconjugation energy increases by 0.23 eV and is 1.21 eV. Thus, the change in 0 due to the more favorable overlap at the relaxed geometry is 0.12 eV, which agrees nicely with 0.07 eV estimated from IEs. Hyperconjugation energies as defined above are not equivalent to the change in the total molecular energy because of the nonadditivity of different contributions. However, the changes in the total molecular energies due to hyperconjugation are 107 ! ! Figure 5.4 Hyperconjugation energies of the three CH bonds for various orientations of the CH 3 group in CH 3 CHOH + at the cation (upper panel) and neutral (lower panel) geometries. 108 ! proportional to the NBO hyperconjugation energies. This has been demonstrated by Alabugin et al., [48] who compared NBO hyperconjugation energies with changes in the total energy due to the deletion of the corresponding blocks from the Fock matrix at DFT level for series of molecules. Hyperconjugation depends on the orientation of the interacting orbitals, which can be examined by calculating the hyperconjugation interactions at different positions of the CH 3 group produced by a rotation of the group around the CC axis. Figure 5.4 shows the individual hyperconjugation energies for each of the CH bonds, as well as their sum, along the torsional coordinate for CH 3 CHOH + at the cation and radical geometries (only the torsional angle is varied, all other degrees of freedom are frozen; see section 5.3 ). Because of the different lengths of the three CH bonds, the torsional curves in Figure 5.4 are not symmetric. If the three CH bonds were identical, the period of rotation would be 120° and the potentials will be symmetric with respect to 60°. As expected, the interaction between an individual # CH bond and the $ CO decreases as the angle between them increases; however, hyperconjugation as the sum of the individual interactions is almost constant, since as one CH bond leaves the zone of favorable overlap, another CH bond enters in. Thus, unless one of the hydrogens is appropriately substituted, the torsion potential is rather flat. [20] The total hyperconjugation energy is larger at the cation geometry because of the more favorable overlap at this geometry due to the shorter CC and CO bonds. Thus, 109 ! significant energy relaxation upon ionization is largely due to the more efficient hyperconjugation at shorter bond lengths, whereas the changes in energy due to CH 3 group orientation are minor. This is in agreement with the analysis of the nature of the torsional barrier in ethane,[26] where hyperconjugation was found to be more efficient at the staggered configuration due to the shorter CC bond length rather than more favorable orientation of the CH 3 groups. 5.4 Conclusions Hyperconjugation explains the observed large changes in the IE of the hydroxyethyl radical as compared to the hydroxymethyl radical by destabilizing the SOMO and stabilizing the HOMO-1 of the cation. At the radical geometry, the change in IE due to SOMO destabilization is estimated to be 0.53 eV, which is in excellent agreement with the hyperconjugation energy of the hydroxyethyl cation computed by NBO. Upon geometry relaxation following ionization, the hyperconjugation energy increases by about 0.1 eV, which explains the larger difference between the vertical and adiabatic IEs in hydroxyethyl relative to hydroxymethyl. Thus, we can interpret the large change between the adiabatic and vertical IE in CH 3 CHOH as the cumulative effects of the bonding interactions between the carbon’s unpaired electron, the lone pair of oxygen, and hyperconjugation interaction with # CH . The energy relaxation 110 ! upon ionization is due to more efficient hyperconjugation at shorter CC and CO bond lengths, whereas energy changes due to the CH 3 torsion are minor. ! 111 ! Chapter 5 References (1) Gardiner, W. C. , Ed. Gas-Phase Combustion Chemistry; Springer: New York, 2000. (2) Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications; Academic Press: New York, 1999. (3) Lin, J. J.; Shu, J.; Lee, Y. T.; Yang, X. Journal of Chemical Physics 2000 113 5287 (4) Marcy, T. P.; Diaz, R. R.; Heard, D.; Leone, S. R.; Harding, L. B.; Klippenstein, S. J. Journal of Physical Chemistry A 2001 105 8361 (5) Seakins, P. W.; Leone, S. R. Journal of Physical Chemistry. 1992 96 4478 (6) Ahmed, M.; Peterka, D. S.; Suits, A. G. Physical Chemistry Chemical Physics. 2000 2 861 (7) Rudic, S.; Murray, C.; Ascenzi, D.; Anderson, H.; Harve, J. N.; Orr- Ewing, A. J. Journal of Chemical Physics 2002 117 5692 (8) Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1994 101 10936 (9) Hoyermann, K.; Olzmann, M.; Seeba, J.; Viskolcz, B. Journal of Physical Chemistry A 1999 103 5692 (10) Senosiain, J. P.; Klippenstein, S. J.; Miller, J. A. Journal of Physical Chemistry A 2006 110 6960 (11) Cleary, P. A.; Romero, M. T. B.; Blitz, M. A.; Heard, D. E.; Pilling, M. J.; Seakins, P. W.; Wang, L. Physical Chemistry Chemical Physics. 2006 8 5633 (12) . Taatjes, C. A.; Hansen, N.; McIlroy, A.; Miller, J. A.; Senosiain, J. P.; Klippenstein, S. J.; Qi, F.; Sheng, L. S.; Zhang, Y. W.; Cool, T. A.; Wang, J.; Westmoreland, P. R.; Law, M. E.; Kasper, T.; Kohse-Hoinghaus, K. Science 2005 308 1887 (13) Dyke, J. M.; Ellis, A. R.; Ellis, A. R.; Jonathan, N.; Keddar, N.; Morris, A. Chemical Physics Letters 1971 111 207 112 ! (14) Dyke, J. M.; Groves, P.; Lee, E. P. F.; Niavaran, M. H. Z. J. Journal of Physical Chemistry A 1997 101 373 (15) Tao, W.; Klemm, R. B.; Nesbitt, F. L.; Stief, J. L. Journal of Physical Chemistry 1992 96 104 (16) Holmes, J. L.; Lossing, F. P. Organic Mass Spectrometry 1991 26 537 (17) Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1991 95 4033 (18) Williams, J. M.; Hamill, W. H. Journal of Chemical Physics 1968 49 4467 (19) Muller, N.; Mulliken, R. S. Journal of the American Chemical Society. 1958 80 3489 (20) Hoffman, R.; Radom, L.; Pople, J. A.; Schleyer, P.; von, R.; Here; W, J.;; Salem, L. Journal of the American Chemical Society. 1972 94 6221 (21) Schneider, W. R.; Nance, B. I.; Wallington, T. J. Journal of the American Chemical Society. 1995 117 478 (22) Koziol, L.; Levchenko, S. V.; Krylov, A. I. Journal of Physical Chemistry A 2006 110 2746 (23) Häber, T.; Blair, A. C.; Nesbitt, D. J. Journal of Chemical Physics 2006 124 054316 (24) Sharp, E. N.; Rupper, P.; Miller, T. A. Physical Chemistry Chemical Physics. 2008,in press (25) Brunck, T. K.; Weinhold, F. Journal of the American Chemical Society. 1978 101 1700 (26) Prophristic, V.; Goodman, L. Nature (London) 2001 411 565 (27) Weinhold, F. Angew. Chemistry, Int. Ed. Engl. 2003 42 4188 (28) Bickelhaupt, F. M.; Baerends, E. J. Angew. Chemistry, Int. Ed. Engl. 2003 42 4183 (29) Mo, Y.; Wu, W.; Song, L.; Lin, M.; Zhang, Q.; Gao, J. Angew. Chemistry, Int. Ed. 2004 43 1986 113 ! (30) Saebo, S.; Radom, L.; Schaefer, H. F. Journal of Chemical Physics 1983 78 845 (31) Johnson III, R. D.; Hudgens, J. F. Journal of Physical Chemistry. 1996 100 19874 (32) Curtiss, L. A.; Lucas, D. J.; Pople, J. A. Journal of Chemical Physics 1995 102 3292 (33) Levchenko, S. V.; Krylov, A. I. Journal of Chemical Physics 2001 115 7485 (34) Levchenko, S. V.; Demyanenko, A. V.; Dribinski, V.; Potter, A. B.; Reisler, H.; Krylov, A. I. Journal of Chemical Physics 2003 118 9233 (35) Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M. Chemical Physics Letters 1989 157 479 (36) Watts, J. D.; Gauss, J.; Bartlett, R. J. Journal of Chemical Physics 1993 98 8718 (37) Dunning, T. H. Journal of Chemical Physics 1989 90 1007 (38) Stanton, J. F.; Gauss, J.; Watts, J. D.; Lauderdale, W. J.; Bartlett, R. J. ACES II, 1993.The package also contains modified versions of the MOLECULE Gaussian integral program of J. Almlöf and P. R. Taylor, the ABACUS integral derivative program written by T. U. Helgaker, H. J. Aa. Jensen, P. Jørgensen, and P. R. Taylor, and the PROPS property evaluation integral code of P. R. Taylor. (39) Knowles, P. J.; Hampel, C.; Werner, H. J. Journal of Chemical Physics 1993 99 5219 (40) Werner, H.-J.; Knowles, P. J.; Lindh, R.; Schutz, M. ; et al. MOLPRO 2002, 2003. 114 ! (41) Shao, Y.; Molnar, L. F.; Jung, Y.; Kussmann, J.; Ochsenfeld, C.; Brown, S.; Gilbert, A. T. B.; Slipchenko, L. V.; Levchenko, S. V.; O’Neil, D. P.; Distasio, R. A., Jr.; Lochan, R. C.; Wang, T.; Beran, G. J. O.; Besley, N. A.; Herbert, J. M.; Lin, C. Y.; Van Voorhis, T.; Chien, S. H.; Sodt, A.; Steele, R. P.; Rassolov, V. A.; Maslen, P.; Korambath, P. P.; Adamson R.D; Austin, B.; Baker, J.; Bird, E. F. C.; Daschel, H.; Doerksen, R. J.; Drew, A.; Dunietz, B. D.; Dutoi, A. D.; Furlani, T. R.; Gwaltney, S. R.; Heyden, A.; Hirata, S.; Hsu, C.-P.; Kedziora, G. S.; Khalliulin, R. Z.; Klunziger, P.; Lee, A. M.; Liang, W. Z.; Lotan, I.; Nair, N.; Peters, B.; Proynov, E. I.; Pieniazek, P. A.; Rhee, Y. M.; Ritchie, J.; Rosta, E.; Sherrill, C. D.; Simmonett, A. C.; Subotnik, J. E.; Woodcock, H. L., III; Zhang, W.; Bell, A. T.; Chakraborty, A. K.; Chipman, D. M.; Keil, F. J.; Warshel, A.; Herhe, W. J.; Schaefer, H. F., III; Kong, J.; Krylov, A. I.; Gill, P. M. W.; Head-Gordon, M. Physical Chemistry Chemical Physics. 2006 8 3172 (42) Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Weinhold, F. NBO 5.0, Theoretical Chemistry Institute, University of Wisconsin, Madison, WI, 2001. (43) Levchenko, S. V.; Wang, T.; Krylov, A. I. Journal of Chemical Physics 2005 122 224106 (44) Berkowitz, J.; Ellison, G. B.; Gutman, D. Journal of Physical Chemistry. 1994 98 2744 (45) Ruscic, B.; Berkowitz, J.; Curtiss, L. A. Journal of Chemical Physics 1989 91 114 (46) Vorob’ev, A. S.; Furlei, I. I.; Sultanov, A. S.; Khvostenko, V. I.; Leplyanin,; G, V.; Derzhinskii, A. R.; Tolstikov, G. A. Bull. Acad. Sci. USSR, Div. Chem. Sci. 1989 1388 (47) Ohno, K.; Imai, K.; Harada, Y. Journal of the American Chemical Society. 1985 107 8078 (48) Alabugin, I. V.; Zeidan, T. A. Journal of the American Chemical Society. 2002 124 3175 115! ! Chapter 6 Future Experiments 2-hydroxyethyl radical, the less stable isomer of CH 3 CHOH, is an intermediate in the low temperature OH + C 2 H 4 reaction, which proceeds by addition and stabilization of the HOC 2 H 4 adduct [1-14]. Of the three C 2 H 5 O isomers, the best studied is ethoxy [15,16]; our study is the only one on the spectroscopy and photochemistry of CH 3 CHOH in a molecular beam[17], and little work exists on CH 2 CH 2 OH [18-23]. Therefore it is a natural target for investigation. The goal of the proposed experiments is to study the photoinitiated reactions of CH 2 CH 2 OH on the ground and the excited potential energy surfaces. 6.1 Unimolecular reactions on the ground electronic state The ground state potential energy surface (PES) of this reaction, calculated by Senossian et al,. shows that there are two low-lying simple bond fission channels for 2-hydroxymethyl (see Figure 6.1)[1]: CH 2 CH 2 OH ' OH + CH 2 CH 2 (H = 26 kcal/mol (1) ' H + CH 2 CHOH (H = 27 kcal/mol (2) Both reactions yield closed-shell products: ethylene (1) or vinyl alcohol (2) Reaction (1) is the entrance channel in the OH+C 2 H 4 reaction[1-3] and has a small reverse barrier of 0.2 kcal/mol [1]. It is considered to be the major dissociation ! 116! ! ! Figure 6.1 Energies, barrier heights and product channels (kcal/mol) relevant to the OH + C 2 H 4 reaction (Ref 1). 117! ! channel on the ground state for the radical and it is the only one observed experimentally so far [22,23]. The vinyl alcohol channel has a higher reverse barrier of ~ 6 kcal/mol and cannot compete with the ethylene channel at low energies [1]. However, as the internal energy of the radical increases, dissociation via the vinyl channel becomes viable [24]. Besides the direct bond fission, other dissociation channels are available via isomerization to ethoxy or 1-hydroxyethyl radicals [1]. However, barriers for isomerization are higher and usually have tight transition states [1], which may hinder competition of these dissociation pathways with direct bond fission channels. Unimolecular reactions on the ground state can be investigated by either state- specific excitation of the “cold” CH 2 CH 2 OH or by observing predissociation of the “hot” CH 2 CH 2 OH. For these studies 2-hydroxyethyl radicals can be generated by UV photolysis of 2-haloethanols, specifically 2-iodo- and 2-bromoethanol [22,23,25,26]. The photodissociation of 2-haloethanols was studied before by using photofragment translational spectroscopy [22], and by recording LIF signal of OH products from secondary dissociation of 2-hydroxyethyl [23]. CH 2 CH 2 OH generated by 193 nm or 202 nm photolysis of 2-bromoehtanol has high internal energy, a substantial amount of which is tied as rotational excitation[22,23]. Hence, BrCH 2 CH 2 OH can be considered as a good source of “hot” 2-hydroxyethyl radicals. The photodissociation of the 2-bromoethanol will be carried in the interaction region. Its room temperature absorption spectrum shows a maximum around 202 nm. This wavelength can be obtained by generating the third harmonic of the 606 nm output of 118! ! a dye laser. Some of the CH 2 CH 2 OH products will have sufficient internal energy to predissociate, and OH from reaction (1) has already been identified [23]. Hydrogen atoms produced by reaction (2) can be detected by 1+1’ REMPI via the Lyman-* transition. The advantage of the 2-bromoethanol as a source of CH 2 CH 2 OH is a workable vapor pressure at room temperature, which allows sample preparation without additional heating. In the case of 2-iodoethanol, heating of the sample and the nozzle is required because its vapor pressure at room temperature is too low. However, it has a broader absorption spectrum than 2-bromoethanol. 2-hydroxyethyl produced by 266 nm excitation of ICH 2 CH 2 OH shows only minor vibrational predissociation [23]. Photodissociation of 2-iodoethanol in the source chamber with subsequent cooling the molecular beam makes this precursor a good source of “cold” CH 2 CH 2 OH. Radicals generated in this process will allow controlled excitation of vibrational overtones, which would provide a great opportunity to study dynamics on the ground state PES in a manner similar to that shown in Chapter 3. For example, surpassing the barrier to reaction (2) requires 34 kcal/mol (11900 cm -1 ), and therefore exciting CH 2 CH 2 OH to the 3 rd overtone of the OH stretch (> 13000 cm -1 ) should exceed this barrier. Although other species can absorb IR radiation in the region of the OH stretch of CH 2 CH 2 OH, only CH 2 CH 2 OH will dissociate in the 3 rd overtone region. With our experimental setup the action spectrum of the radical can be obtained by monitoring H photofragments. 119! ! 6.2 Electronic states of the 2-hydroxyethyl radical To the best of my knowledge, there is no information on the excited states of CH 2 CH 2 OH, and even its published 300 K absorption spectrum has been questioned [15,16,25]. One isomer of CH 2 CH 2 OH, the 1-hydroxyethyl radical, has a low ionization energy of 6.64 eV [18,27], so its lowest electronic states are Rydberg [17]. Another isomer, the ethoxy radical, has a higher IE of 10.29 eV [18] and its lowest electronic states are all valence [15]. The ionization energy of CH 2 CH 2 OH is estimated to be less than 8.35 eV [18] . Hence, a mixture of low-lying Rydberg and valence states is possible. In order to better understand and plan the experimental study of the dynamics on the excited state PES, I performed several ab initio calculations to estimate the energies and nature of several electronic states. Geometry optimization was performed at the CCSD/6-311+G(d,p) level of theory, and electronic states excitation energies were calculated using the EOM-EE-CCSD/6-311(n+,+)G(d,p) method where n=1,2,3. There are three possible equilibrium geometries on the ground state PES, with the least stable (geometry A) lying only 1.38 kcal/mol higher in energy than the most stable (geometry C) conformer; see Figure 6.2. Geometry A has C s symmetry, while geometries B and C have C 1 . The two vibrational modes in geometry C with the lowest frequencies, the CH 2 group torsion around the C-C bond (~ 200 cm -1 ) and the OH torsion around the C-O bond (~ 370 cm -1 ), are exactly the two motions required to go from one equilibrium geometry to another. Taking into account the small 120! ! ! Figure 6.2 Equilibrium structures of CH 2 CH 2 OH, potential energies of the conformers relative to the most stable conformer and symmetries calculated at CCSD/6-311+G(d,p) level of theory. 121! ! differences in the potential energies in the three minima, it can be safely concluded that the barriers between them would be low and even with little vibrational excitation the molecule can easily convert from one conformer to another. Equation-of-motion calculations of the excited electronic state reveal a mixture of valence and Rydberg states. Valence states are formed by excitation of electrons from the doubly-occupied orbitals to the singly-occupied lone pair on the terminal carbon atom. Rydberg states are created by excitation of the electron from the SOMO to one of the Rydberg orbitals. The positions of the first three states are shown in Figure 6.3. The ordering of the valence and Rydberg states is different for different geometries. Therefore, these states can cross, especially since there is no symmetry constraint to prevent them from mixing. Calculations show that the lowest valence state is completely repulsive. Dissociation on the valence state PES is possible through at least three exit channels: ethylene plus hydroxyl radical, vinyl alcohol plus hydrogen atom, and hydroxymethyl plus methylene radicals. This can be explained by the fact that the SOMO of the CH 2 CH 2 OH is antibonding with respect to the C-C and C-O bonds. Therefore, during excitation to the valence state an electron is promoted from one of the bonding orbitals to an antibonding orbital. Thus, the stability of the radical is considerably decreased. The more intriguing case is that of the Rydberg states of CH 2 CH 2 OH. The equilibrium geometry of a Rydberg state is close to the geometry of the cation. It was shown before that CH 2 CH 2 OH + has a cyclic geometry, similar to oxirane [19]. My calculations confirm this, and also show that vertical ionization in geometries B and C 122! ! Figure 6.3 Vertical excitation energies for the first three electronic states of CH 2 CH 2 OH in the cation equilibrium geometries (cyclic and CH 3 CHOH + ) and neutral equilibrium geometries (A,B and C ), calculated at CCSD/6- 311(2+,+)G(d,p) level of theory. 123! ! opens a barrierless path to the CH 3 CHOH + isomer. There are no minima on the cation PES with an open CH 2 CH 2 OH geometry. The two equilibrium cation geometries are shown in Figure 6.3, as well as the position of the electronic states of the neutral radical calculated at these geometries. Due to the considerable changes in geometry, adiabatic transitions to the Rydberg states have small cross sections and vertical excitations should lead to fast isomerization either to CH 3 CHOH or cyclic CH 2 CH 2 OH. As was discussed in Chapter 4, efficient conical intersections of the Rydberg states with the ground state in the CH 3 CHOH geometry would lead to fast dissociation following isomerization to 1- hydroxyethyl. On the other hand, isomerization to the cyclic structure may result in a more stable molecule if there is no corresponding coupling to the ground state or if the coupling is less efficient, as in the case of the 3p z state of CH 3 CHOH. Studies of the excited states of CH 2 CH 2 OH require preparation of cold radicals generated by 266 nm photodissociation of 2-iodoethanol as described in Section 6.1. The lowest electronic state is valence in two geometries out of three, however it has a small oscillator strength (~ 0.004), whereas the oscillator strength of the 3s state is considerably larger (~ 0.027), which makes the transition to this state more probable. Excitation to the 3s state of CH 2 CH 2 OH can lead to several outcomes: a. Isomerization to CH 3 CHOH followed by dissociation on the ground state. The signature of this process is O-H bond fission, which is the major dissociation channel in CH 3 CHOH, but not for ground-state dissociation of 124! ! CH 2 CH 2 OH. This channel can be easily identified by monitoring D + signal from CH 2 CH 2 OD radicals. b. Non adiabatic transition to the valence PES. Dissociation on the repulsive surface must have clearly different dynamics than predissociation on the ground electronic state, which would manifest in distinctly different kinetic energy distribution of the products. Also, for the valence state, C-C bond fission becomes more probable (as the two electrons occupy an antibonding orbital), which may make the contribution of the hydroxymethyl plus methylene channel more noticeable. Interestingly, both products in this channel have a good REMPI detection scheme. c. Isomerization to the cyclic structure. This is the least understood and the most exciting case. According to my calculations any conical intersection with the ground state will result mostly in dissociation to OH plus ethylene. However in the absence of such coupling (or if it is not very efficient) the radical can be ionized by another photon. I expect it to be the only channel that can give a REMPI signal for CH 2 CH 2 OH. 6.3. 4 th overtone of OH stretch in CH 2 OH One of the remarkable results in the study of CH 2 OH was the observation of H fragments via tunneling through the barrier following direct (4" OH ) excitation of the O- H predissociation coordinate[28]. Excitation in the region of 5" OH should exceed the dissociation barriers to both isomerization and direct dissociation (estimated at 125! ! ~ 15000 cm -1 )[29,30], and therefore faster dissociation should ensue. It is not clear a priori which one should dominate near the barrier, and mode-selective effects can be important. The barrier for dissociation arises from an avoided crossing; in this case the character of the electronic wavefunction will change in the barrier region, creating a dynamical bottleneck that will slow dissociation. Because both direct dissociation and isomerization channels have tight TS’s[29], couplings to the dissociation continuum just above the barrier must be weak. As a result, OH-stretch excitation may be achieved. The CD 2 OH isotopolog can be used to reveal the contributions from each channel. If direct O-H fission prevails, mostly H atoms will be detected, whereas if isomerization precedes dissociation, both H and D will be detected with comparable kinetic energy distributions. If both processes are significant, differences in the KED’s of H and D will be observed as the excitation energy is changed. Also monitoring H or D fragments while scanning the pump laser near or just above the barrier to direct O-H bond fission may show different resonance structures in the photofragment yield spectra. The H spectrum will reflect preferentially states that possess a large fraction of OH stretch character, while the D spectrum will be more sensitive to states with the skeletal COH bending or deformation motions required for isomerization. As a result, the H/D ratio may fluctuate with excitation energy. One-photon excitation of 5" OH from the ground state has very low intensity as intensities for overtone pumping scale on average as 10 -0" [31,32] Unfortunately, two- photon transition to the 3s state is very efficient in this wavelength range. Thus, even 126! ! if there is a small amount of H from 5" OH excitation, it would be completely obscured by the strong H + signal from dissociation of the 3s state. However, a double resonance schemes where excitation is performed in a two-step fashion have been shown to reach overtones up to 8" OH [31,32]. Also two-photon excitation of the 3s state can be avoided. In such a scheme one laser saturates the transition to "=1 or 2 and another laser excites the molecules to the final level. The main complication of this approach is the alignment of four lasers, in particular the two pump laser beams that enter the chamber through the same window. In the 1" OH +4" OH double resonance scheme, excitation to 4" OH can generate H atoms, which simplifies initial alignment of pump and probe lasers. Also the IR OPO used for 1" OH excitation does not need to be focused, which simplifies alignment further. Excitation from the ground state to 4" OH generates H atoms only in a narrow region of ±25 cm -1 around 13600 cm -1 [28]; thus, even a small detuning from the double resonance wavelength will generate mainly H atoms from dissociation near the barrier only. In another possible scheme, 2" OH +3" OH , the advantage is that excitation of lower overtones is required, which should increase transition efficiency. While it is relatively easy to align both pump lasers with a single probe laser, the major obstacle is that in this case the probe laser will be tuned to a transition to the 3p z state for the alignment. It is then difficult to switch it to H detection by 1+1’ REMPI while preserving the established alignment. It may be easier to preserve the alignment by going with 2+1 REMPI detection of hydrogen, however it will significantly increase 127! ! the H background, as CH 2 OH also absorbs in this region. Therefore, for practical reasons, it is the best to start with the first scheme. ! 128! ! Chapter 6 References (1) Senosiain, J. P.; Klippenstein, S. J.; Miller, J. A. Journal of Physical Chemistry A 2006, 110, 6960. (2) Tully, F. P. Chemical Physics Letters 1988, 143, 510. (3) Cleary, P. A.; Romero, M. T. B.; Blitz, M. A.; Heard, D. E.; Pilling, M. J.; Seakins, P. W.; Wang, L. Physical Chemistry Chemical Physics 2006, 8, 5633. (4) Greenwald, E. E.; North, S. W.; Georgievskii, Y.; Klippenstein, S. J. Journal of Physical Chemistry A 2005, 109, 6031. (5) Zhu, R. S.; Park, J.; Lin, M. C. Chemical Physics Letters 2005, 408, 25. (6) Bhargava, A.; Westmoreland, P. R. Combustion and Flame 1998, 113, 333. (7) Hippler, H.; Viskolcz, B. Physical Chemistry Chemical Physics 2000, 2, 3591. (8) Vakhtin, A. B.; Murphy, J. E.; Leone, S. R. Journal of Physical Chemistry A 2003, 107, 10055. (9) Sosa, C.; Schlegel, H. B. Journal of the American Chemical Society 1987, 109, 7007. (10) Diau, E. W. G.; Lee, Y. P. Journal of Chemical Physics 1992, 96, 377. (11) Liu, G. X.; Ding, Y. H.; Li, Z. S.; Fu, Q.; Huang, X. R.; Sun, C. C.; Tang, A. C. Physical Chemistry Chemical Physics 2002, 4, 1021. (12) Piqueras, M. C.; Crespo, R.; Nebot-Gil, I.; Tomas, F. Journal of Molecular Structure-Theochem 2001, 537, 199. (13) Chuong, B.; Stevens, P. S. Journal of Physical Chemistry A 2000, 104, 5230. (14) Yamada, T.; Bozzelli, J. W.; Lay, T. Journal of Physical Chemistry A 1999, 103, 7646. (15) Faulhaber, A. E.; Szpunar, D. E.; Kautzman, K. E.; Neumark, D. M. Journal of Physical Chemistry A 2005, 109, 10239. 129! ! (16) Choi, H.; Bise, R. T.; Neumark, D. M. Journal of Physical Chemistry A 2000, 104, 10112. (17) Karpichev, B.; Edwards, L. W.; Wei, J.; Reisler, H. Journal of Physical Chemistry A 2008, 112, 412. (18) Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1994, 101, 10936. (19) Curtiss, L. A.; Lucas, D. J.; Pople, J. A. Journal of Chemical Physics 1995, 102, 3292. (20) Nobes, R. H.; Rodwell, W. R.; Bouma, W. J.; Radom, L. Journal of the American Chemical Society 1981, 103, 1913. (21) Anastasi, C.; Simpson, V.; Munk, J.; Pagsberg, P. Journal of Physical Chemistry 1990, 94, 6327. (22) Hintsa, E. J.; Zhao, X. S.; Lee, Y. T. Journal of Chemical Physics 1990, 92, 2280. (23) Sapers, S. P.; Hess, W. P. Journal of Chemical Physics 1992, 97, 3126. (24) Klippenstein, S. J., private communication. (25) Murrells, T. P.; Jenkin, M. E.; Shalliker, S. J.; Hayman, G. D. Journal of the Chemical Society-Faraday Transactions 1991, 87, 2351. (26) Jenkin, M. E.; Cox, R. A. Journal of Physical Chemistry 1991, 95, 3229. (27) Dyke, J. M.; Groves, A. P.; Lee, E. P. F.; Niavaran, M. H. Z. Journal of Physical Chemistry A 1997, 101, 373. (28) Wei, J.; Karpichev, B.; Reisler, H. Journal of Chemical Physics 2006, 125. (29) Saebo, S.; Radom, L.; Schaefer, H. F. Journal of Chemical Physics 1983, 78, 845. (30) Marcy, T. P.; Diaz, R. R.; Heard, D.; Leone, S. R.; Harding, L. B.; Klippenstein, S. J. Journal of Physical Chemistry A 2001, 105, 8361. (31) Boyarkin, O. V.; Rizzo, T. R.; Perry, D. S. Journal of Chemical Physics 1999, 110, 11346. 130! ! (32) Boyarkin, O. V.; Rizzo, T. R.; Perry, D. S. Journal of Chemical Physics 1999, 110, 11359. 131 ! Alphabetized Bibliography Adams, G. F.; Bartlett, R. J.; Purvis, G. D. Chemical Physics Letters 1982, 87, 311. Ahmed, M.; Peterka, D. S.; Suits, A. G. Physical Chemistry Chemical Physics 2000, 2, 861. Alabugin, I. V.; Zeidan, T. A. Journal of American Chemical Society. 2002 124 3175 Anastasi, C.; Simpson, V.; Munk, J.; Pagsberg, P. Chemical Physics Letters 1989, 164, 18. Anastasi, C.; Simpson, V.; Munk, J.; Pagsberg, P. Journal of Physical Chemistry 1990, 94, 6327. Aristov, V.; Conroy, D.; Reisler, H. Chemical Physics Letters 2000, 318, 393. Bear, T.; Hase, W. Unimolecular Reaction Dynamics; Oxford University Press: New York, 1996 Berkowitz, J.; Ellison, G. B.; Gutman, D. Journal of Physical Chemistry 1994 98 2744 Bhargava, A.; Westmoreland, P. R. Combustion and Flame 1998, 113, 333. Bickelhaupt, F. M.; Baerends, E. J. Angew. Chemistry., Int. Ed. Engl. 2003 42 4183 Boyarkin, O. V.; Rizzo, T. R.; Perry, D. S. Journal of Chemical Physics 1999, 110, 11346. Boyarkin, O. V.; Rizzo, T. R.; Perry, D. S. Journal of Chemical Physics 1999, 110, 11359. Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 1998, 102, 3141. Bruna, P. J.; Grein, F. Journal of Physical Chemistry A 2001, 105, 8599. Brunck, T. K.; Weinhold, F. Journal of American Chemical Society. 1978 101 1700 Child, M. S. Accounts of Chemical Research 1985, 18, 45 Chirokolava, A.; Perry, D. S.; Boyarkin, O. V.; Schmid, M.; Rizzo, T. R. Journal of Chemical Physics 2000, 113, 10068 132 ! Choi, H.; Bise, R. T.; Neumark, D. M. Journal of Physical Chemistry A 2000, 104, 10112. Chuong, B.; Stevens, P. S. Journal of Physical Chemistry A 2000, 104, 5230. Cleary, P. A.; Romero, M. T. B.; Blitz, M. A.; Heard, D. E.; Pilling, M. J.; Seakins, P. W.; Wang, L. Physical Chemistry Chemical Physics 2006, 8, 5633. Conroy, D.; Aristov, V.; Feng, L.; Reisler, H. Journal of Physical Chemistry A 2000, 104, 10288. Coxon, J. A.; Foster, S. C. Canadian Journal of Physics 1982, 60, 41. Curtiss, L. A.; Lucas, D. J.; Pople, J. A. Journal of Chemical Physics 1995, 102, 3292. Demerjian, K. L.; Kerr, J. A.; Calvert, J. G. THE MECHANISM OF PHOTOCHEMICAL SMOG FORMATION, 1974. Dertinger, S.; Geers, A.; Kappert, J.; Wiebrecht, J.; Temps, F. Faraday Discussion. 1995, 102, 31. Diau, E. W. G.; Lee, Y. P. Journal of Chemical Physics 1992, 96, 377. Dunning, T. H. Journal of Chemical Physics 1989 90 1007 Dyke, J. M.; Ellis, A. R.; Jonathan, N.; Keddar, N.; Morris, A. Chemical Physics Letters 1984, 111, 207. Dyke, J. M.; Groves, A. P.; Lee, E. P. F.; Niavaran, M. H. Z. Journal of Physical Chemistry A 1997, 101, 373. Edelbuttel-Einhaus, J.; Hoyermann, K.; Rohde, G.; Seeba, J. Proceedings of the 24th Symposium International on Combustion, Combustion Institute, Pittsburgh, PA 1992, 661. Fairley, D. A.; Scott, G. B. I.; Freeman, C. G.; Maclagan, R. G. A. R.; McEwan, M. J. Journal of the Chemical Society-Faraday Transactions 1996, 92, 1305. Faulhaber, A. E.; Szpunar, D. E.; Kautzman, K. E.; Neumark, D. M. Journal of Physical Chemistry A 2005, 109, 10239. Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2003, 118, 9623. 133 ! Feng, L.; Demyanenko, A. V.; Reisler, H. Journal of Chemical Physics 2004, 120, 6524. Feng, L.; Huang, X.; Reisler, H. Journal of Chemical Physics 2002, 117, 4820. Feng, L.; Huang, X.; Reisler, H. Journal of Chemical Physics 2002, 117, 4820. Feng, L.; Reisler, H. Journal of Physical Chemistry A 2004, 108, 9847. Finlayson-Pitts, B. J.; Pitts, J. N. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications; Academic Press: New York, 1999. Fockenberg, C.; Hall, G. E.; Preses, J. M.; Sears, T. J.; Muckerman, J. T. Journal of Physical Chemistry A 1999, 103, 5722. Gardiner, W. C. , Ed. Gas-Phase Combustion Chemistry; Springer: New York, 2000. Geers, A.; Kappert, J.; Temps, F.; Wiebrecht, J. W. Journal of Chemical Physics 1993, 99, 2271. Geers, A.; Kappert, J.; Temps, F.; Wiebrecht, J. W. Journal of Chemical Physics 1994, 101, 3618. Geers, A.; Kappert, J.; Temps, F.; Wiebrecht, J. W. Journal of Chemical Physics 1994, 101, 3634. Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.; Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Weinhold, F. NBO 5.0, Theoretical Chemistry Institute, University of Wisconsin, Madison, WI, 2001. Greenwald, E. E.; North, S. W.; Georgievskii, Y.; Klippenstein, S. J. Journal of Physical Chemistry A 2005, 109, 6031. Grosjean, D. Journal of Brazilian Chemical Society 1997, 8, 433. Haber, T.; Blair, A. C.; Nesbitt, D. J.; Schuder, M. D. Journal of Chemical Physics 2006, 124, 054316 Halonen, L. Advances in Chemical Physics, Vol 104 1998, 104, 41. Harding, L. B., unpublished Heicklen, J. Atmospheric Chemistry; Academic: New York, 1976. 134 ! Henry, B. R. Accounts of Chemical Research 1977, 10, 207. Henry, B. R.; Kjaergaard, H. G. Canadian. Journal of Chemical Review. Can. Chim. 2002, 80, 1635. Herzberg, G. Molecular Spectra and Molecular Structure I: Spectra of Diatomic Molecules; Van Nostrand: New York, 1950 Hintsa, E. J.; Zhao, X. S.; Lee, Y. T. Journal of Chemical Physics 1990, 92, 2280. Hippler, H.; Striebel, F.; Viskolcz, B. Physical Chemistry Chemical Physics 2001, 3, 2450. Hippler, H.; Viskolcz, B. Physical Chemistry Chemical Physics 2000, 2, 3591. Hoffman, B. C.; Yarkony, D. R. Journal of Chemical Physics 2002, 116, 8300. Hoffman, R.; Radom, L.; Pople, J. A.; Schleyer, P.; von, R.; Here; W, J.;; Salem, L. Journal of American Chemical Society. 1972 94 6221 Hoffmann, R.; Hehre, W. J.; Salem, L.; Schleyer, P. V.; Pople, J. A.; Radom, L. Journal of the American Chemical Society 1972, 94, 6221. Holmes, J. L.; Lossing, F. P. Organic Mass Spectrometry 1991, 26, 537. Hoyermann, K.; Olzmann, M.; Seeba, J.; Viskolcz, B. Journal of Physical Chemistry A 1999, 103, 5692. Jenkin, M. E.; Cox, R. A. Journal of Physical Chemistry 1991, 95, 3229. Johnson, R. D.; Hudgens, J. W. Journal of Physical Chemistry 1996, 100, 19874. Judge, R. H.; Clouthier, D. J. Computational Physics Communications 2001, 135, 293. Karpichev, B.; Edwards, L. W.; Wei, J.; Reisler, H. Journal of Physical Chemistry A 2008, 112, 412. Karpichev, B.; Reisler, H.; Krylov, A. I.; Diri, K. Journal of Physical Chemistry A 2008, 112, 9965. Klippenstein, S. J., private communication. Knowles, P. J.; Hampel, C.; Werner, H. J. Journal of Chemical Physics 1993 99 5219 135 ! Koziol, L.; Levchenko, S. V.; Krylov, A. I. Journal of Physical Chemistry A 2006 110 2746 Kuhn, B.; Boyarkin, O. V.; Rizzo, T. R. Berichte der Bunsen-Gesellschaft für Physikalische Chemie 1997, 101, 339. Kuhn, B.; Rizzo, T. R. Journal of Chemical Physics 2000, 112, 7461. Levchenko, S. V.; Demyanenko, A. V.; Dribinski, V.; Potter, A. B.; Reisler, H.; Krylov, A. I. Journal of Chemical Physics 2003 118 9233 Levchenko, S. V.; Krylov, A. I. Journal of Chemical Physics 2001 115 7485 Levchenko, S. V.; Wang, T.; Krylov, A. I. Journal of Chemical Physics 2005 122 224106 Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Mallard, W. G. Journal of Physical and Chemical Reference Data 1988, 17, 1. Lin, J. J.; Harich, S.; Lee, Y. T.; Yang, X. Journal of Chemical Physics 1999, 110, 10821. Lin, J. J.; Shu, J.; Lee, Y. T.; Yang, X. Journal of Chemical Physics 2000, 113, 5287. Lindner, J.; Loomis, R. A.; Klaassen, J. J.; Leone, S. R. Journal of Chemical Physics 1998, 108, 1944. Liu, G. X.; Ding, Y. H.; Li, Z. S.; Fu, Q.; Huang, X. R.; Sun, C. C.; Tang, A. C. Physical Chemistry Chemical Physics 2002, 4, 1021. Marcy, T. P.; Diaz, R. R.; Heard, D.; Leone, S. R.; Harding, L. B.; Klippenstein, S. J. Journal of Physical Chemistry A 2001, 105, 8361. Marenich, A. V.; Boggs, J. E. Journal of Chemical Physics 2003, 119, 10105. Marenich, A. V.; Boggs, J. E. Journal of Chemical Physics 2003, 119, 3098. Miller, W. H. Journal of the American Chemical Society 1979, 101, 6810. Mills, I. M.; Robiette, A. G. Molecular Physics 1985, 56, 743 Mo, Y.; Wu, W.; Song, L.; Lin, M.; Zhang, Q.; Gao, J. Angew. Chemistry, Int. Ed. 2004 43 1986 136 ! Moncada, C.; Torres, V.; Varghese, G.; Albano, E.; Israel, Y. Mol Pharmacol 1994, 46, 786. Muller, N.; Mulliken, R. S. Journal of the American Chemical Society 1958, 80, 3489. Murrells, T. P.; Jenkin, M. E.; Shalliker, S. J.; Hayman, G. D. Journal of the Chemical Society-Faraday Transactions 1991, 87, 2351. Nobes, R. H.; Rodwell, W. R.; Bouma, W. J.; Radom, L. Journal of the American Chemical Society 1981, 103, 1913. Ohno, K.; Imai, K.; Harada, Y. Journal of American Chemical Society. 1985 107 8078 Pagsberg, P.; Munk, J.; Anastasi, C.; Simpson, V. J. Journal of Physical Chemistry 1989, 93, 5162. Pagsberg, P.; Munk, J.; Sillesen, A.; Anastasi, C. Chemical Physics Letters 1988, 146, 375. Petraco, N. D. K.; Allen, W. D.; Schaefer, H. F. Journal of Chemical Physics 2002, 116, 10229. Phillips, J. A.; Orlando, J. J.; Tyndall, G. S.; Vaida, V. Chemical Physics Letters 1998, 296, 377. Piqueras, M. C.; Crespo, R.; Nebot-Gil, I.; Tomas, F. Journal of Molecular Structure- Theochem 2001, 537, 199. Prophristic, V.; Goodman, L. Nature (London) 2001 411 565 Quack, M.; Willeke, M. Journal of Chemical Physics 1999, 110, 11958 Radford, H. E. Chemical Physics Letters 1980, 71, 195. Raghavachari, K.; Trucks, G. W.; Pople, J. A.; Head-Gordon, M. Chemical Physics Letters 1989 157 479 Rettrup, S.; Pagsberg, P.; Anastasi, C. Chemical Physics 1988, 122, 45. Rudic, S.; Murray, C.; Ascenzi, D.; Anderson, H.; Harvey, J. N.; Orr-Ewing, A. J. Journal of Chemical Physics 2002, 117, 5692. Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1991, 95, 4033. 137 ! Ruscic, B.; Berkowitz, J. Journal of Chemical Physics 1994, 101, 10936. Ruscic, B.; Berkowitz, J.; Curtiss, L. A. Journal of Chemical Physics 1989 91 114 Saebo, S.; Radom, L.; Schaefer, H. F. Journal of Chemical Physics 1983, 78, 845. Saebo, S.; Radom, L.; Schaefer, H. F. Journal of Chemical Physics 1983, 78, 845. Sapers, S. P.; Hess, W. P. Journal of Chemical Physics 1992, 97, 3126. Schneider, W. F.; Nance, B. I.; Wallington, T. J. Journal of the American Chemical Society 1995, 117, 478. Scott, J. L.; Luckhaus, D.; Brown, S. S.; Crim, F. F. Journal of Chemical Physics 1995, 102, 675. Seakins, P. W.; Leone, S. R. Journal of Physical Chemistry 1992, 96, 4478. Senosiain, J. P.; Klippenstein, S. J.; Miller, J. A. Journal of Physical Chemistry A 2006, 110, 6960. Shao, Y.; Molnar, L. F.; Jung, Y.; Kussmann, J.; Ochsenfeld, C.; Brown, S.; Gilbert, A. T. B.; Slipchenko, L. V.; Levchenko, S. V.; O’Neil, D. P.; Distasio, R. A., Jr.; Lochan, R. C.; Wang, T.; Beran, G. J. O.; Besley, N. A.; Herbert, J. M.; Lin, C. Y.; Van Voorhis, T.; Chien, S. H.; Sodt, A.; Steele, R. P.; Rassolov, V. A.; Maslen, P.; Korambath, P. P.; Adamson R.D; Austin, B.; Baker, J.; Bird, E. F. C.; Daschel, H.; Doerksen, R. J.; Drew, A.; Dunietz, B. D.; Dutoi, A. D.; Furlani, T. R.; Gwaltney, S. R.; Heyden, A.; Hirata, S.; Hsu, C.-P.; Kedziora, G. S.; Khalliulin, R. Z.; Klunziger, P.; Lee, A. M.; Liang, W. Z.; Lotan, I.; Nair, N.; Peters, B.; Proynov, E. I.; Pieniazek, P. A.; Rhee, Y. M.; Ritchie, J.; Rosta, E.; Sherrill, C. D.; Simmonett, A. C.; Subotnik, J. E.; Woodcock, H. L., III; Zhang, W.; Bell, A. T.; Chakraborty, A. K.; Chipman, D. M.; Keil, F. J.; Warshel, A.; Herhe, W. J.; Schaefer, H. F., III; Kong, J.; Krylov, A. I.; Gill, P. M. W.; Head-Gordon, M. Physical Chemistry Chemical Physics 2006 8 3172 Sharp, E. N.; Rupper, P.; Miller, T. A. Physical Chemistry Chemical Physics 2008, Sosa, C.; Schlegel, H. B. Journal of the American Chemical Society 1987, 109, 7007. Stanton, J. F.; Gauss, J.; Watts, J. D.; Lauderdale, W. J.; Bartlett, R. J. ACES II, 1993.The package also contains modified versions of the MOLECULE Gaussian integral program of J. Almlöf and P. R. Taylor, the ABACUS integral derivative program written by T. U. Helgaker, H. J. Aa. Jensen, P. Jørgensen, and P. R. Taylor, and the PROPS property evaluation integral code of P. R. Taylor. 138 ! Syage, J. A. Journal of Chemical Physics 1996, 105, 1007. Taatjes, C. A.; Christensen, L. K.; Hurley, M. D.; Wallington, T. J. Journal of Physical Chemistry A 1999, 103, 9805. Taatjes, C. A.; Hansen, N.; McIlroy, A.; Miller, J. A.; Senosiain, J. P.; Klippenstein, S. J.; Qi, F.; Sheng, L. S.; Zhang, Y. W.; Cool, T. A.; Wang, J.; Westmoreland, P. R.; Law, M. E.; Kasper, T.; Kohse-Hoinghaus, K. Science 2005, 308, 1887. Tao, W.; Klemm, R. B.; Nesbitt, F. L.; Stief, L. J. Journal of Physical Chemistry 1992, 96, 104. Tully, F. P. Chemical Physics Letters 1988, 143, 510. Vakhtin, A. B.; Murphy, J. E.; Leone, S. R. Journal of Physical Chemistry A 2003, 107, 10055. Vorob’ev, A. S.; Furlei, I. I.; Sultanov, A. S.; Khvostenko, V. I.; Leplyanin,; G, V.; Derzhinskii, A. R.; Tolstikov, G. A. Bull. Acad. Sci. USSR, Div. Chem. Sci. 1989 1388 Waite, B. A.; Miller, W. H. Journal of Chemical Physics 1980, 73, 3713 Walch, S. P. Journal of Chemical Physics 1993, 98, 3076. Watts, J. D.; Gauss, J.; Bartlett, R. J. Journal of Chemical Physics 1993 98 8718 Wei, J.; Karpichev, B.; Reisler, H.; Journal of Chemical Physics 2006, 125, 034303. Weinhold, F. Angew. Chem., Int. Ed. Engl. 2003 42 4188 Werner, H.-J.; Knowles, P. J.; Lindh, R.; Schtz, M. ; et al. MOLPRO 2002, 2003 Whitten, J. E.; Young, C. E.; Pellin, M. J.; Gruen, D. M.; Jones, P. L. Surface Science 1991, 241, 73. Williams, J. M.; Hamill, W. H. Journal of Chemical Physics 1968, 49, 4467. Xu, Z.; Koplitz, B.; Wittig, C. Journal of Chemical Physics 1989, 90, 2692 Yamada, T.; Bozzelli, J. W.; Lay, T. Journal of Physical Chemistry A 1999, 103, 7646. Yarkony, D. R. Journal of Chemical Physics 2005, 122. 139 ! Zhu, R. S.; Park, J.; Lin, M. C. Chemical Physics Letters 2005, 408, 25.

Asset Metadata

Creator Karpichev, Boris (author)

Core Title Electronic states and photodissociation dynamics of hydroxyalkyl radicals

Contributor Electronically uploaded by the author (provenance)

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 08/07/2009

Defense Date 06/16/2009

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag CH₂OH,CH₃CHOH,hydroxyalkyl radicals,oai:digitallibrary.usc.edu:usctheses,OAI-PMH Harvest,photodissociation

Language English

Advisor Reisler, Hanna (committee chair), Bradforth, Stephen E. (committee member), Kresin, Vitaly V. (committee member)

Creator Email karpiche@usc.edu,karpichev@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m2532

Unique identifier UC1145984

Identifier etd-Karpichev-3072 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-176672 (legacy record id),usctheses-m2532 (legacy record id)

Legacy Identifier etd-Karpichev-3072.pdf

Dmrecord 176672

Document Type Dissertation

Rights Karpichev, Boris

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email uscdl@usc.edu

Abstract (if available)

Abstract The predissociation of vibrationally excited hydroxymethyl radical and the ionization spectroscopy and the photodissociation dynamics of 1-hydroxyethyl radicals from excited Rydberg states are described.