University of Southern California Dissertations and Theses

Photodissociation dynamics of atmospherically relevant small molecules in molecular beams

(USC Thesis Other)

Photodissociation dynamics of atmospherically relevant small molecules in molecular beams

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Photodissociation Dynamics of Atmospherically
Relevant Small Molecules in Molecular Beams

By

Subhasish Sutradhar

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
in Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMISTRY: CHEMICAL PHYSICS)

December 2019

Copyright 2019 Subhasish Sutradhar

Dedicated to my amazing mom, Doli Sutradhar whose unrelenting effort to get her kids higher
education from a small village brought me here.

i

Table of Contents
Table of Contents i
Acknowledgement v
List of Tables viii
List of Figures x
Abstract xvi
Chapter 1 Introduction 1
1.1 Background 2
1.1.1 CH 2OH 2
1.1.2 CO 2 8
1.1.3 Pyruvic Acid 12
1.2 Chapter Overview 15
References 17
Chapter 2 Experimental Methods 32
2.1 Molecular Beam Apparatus 33
2.1.1 Vacuum Chamber 33
2.1.2 Pulsed Nozzle 33
2.2 Time of Flight (TOF) mass spectrometer 34
2.3 Sliced Velocity Map Imaging (SVMI) 35
2.4 Tripling cell and Generation of VUV 43
2.5 Pyrolysis Nozzle 46
References 48
Chapter 3 Photodissociation of the Hydroxymethyl Radical from the 3px and 3pz Rydberg
States 51
ii

3.1 Introduction 51
3.2 Experimental Details 52
3.3 Results and Discussion 55
3.3.1 Triplet States of Formaldehyde and Hydroxymethylene 55
3.3.2 Vibrational excitation and dissociation HCOD products 62
3.3.3 Implications to conical intersections in CH 2OH(D) 72
3.4 Summary and Conclusions 74
References 76
Chapter 4 Temperature Dependence of the Photodissociation of CO2 from High
Vibrational Levels 80
4.1 Introduction 80
4.2 Experimental Details 81
4.3 Results and Discussion 83
4.3.1 REMPI spectra of CO products 83
4.3.2 Kinetic energy release distributions of CO products 88
4.3.3 Kinetic energy distributions of O(
3
P,
1
D) products 93
4.3.4 REMPI spectra of hot CO 2 96
4.3.5 Implications to CO 2 photodissociation dynamics 98
4.4 Summary and Conclusions 105
References 107
Chapter 5 Spectroscopy and Two-photon Dissociation of Jet-Cooled Pyruvic Acid 113
5.1 Introduction 113
5.2 Experimental Details 114
5.3 Results 116
5.3.1 Photofragment Yield Spectra. 116
iii

5.3.2 Reaction Thermochemistry and Fragments’ Kinetic Energy Release. 120
5.4 Discussion 127
5.4.1 Electronic Structure Calculations of Excited States of PA. 127
5.4.2 S 1 ← S 0 Spectroscopy 129
5.4.3 Two-Photon Dissociation. 132
5.4.3.A Two-Body Fragmentation Pathways. 132
5.4.3.B Three-Body Fragmentation Pathways 133
5.4.3.B.I Synchronous Three-Body Fragmentation 134
5.4.3.B.II Sequential Three-Body Fragmentation 138
5.4.3.C Mechanistic Interpretations 142
5.5 Summary and Conclusions 146
References 148
Chapter 6 Future Direction 154
6.1 Preparation of HCOH and CH 3COH at the nozzle tip 154
6.2 Direct Detection of HC and MHC via REMPI 156
6.3 Photodissociation of vibrationally excited CO 2 157
6.4 Generation of VUV Radiation to detect CO 2 160
References 164
Appendices 169
I Possible assignments of the CO 2+1 REMPI transition responsible for peak M at ~ 86,830 cm
-1
169
I.A Estimation of B(v')-X(v") transitions of CO 169
I.B k
3
П←X vibronic transitions of CO 171
I.C C
1
Σ𝑔 +

←X vibronic transitions 173
I.D E
1
П ←X vibronic transitions 175
II Synchronous and sequential three-body fragmentation 176
iv

II.A Synchronous fragmentation 176
II.B Sequential fragmentation 178
v

Acknowledgement
I sincerely thank my advisor, Professor Hanna Reisler, for giving me the opportunity to work in
her lab and providing me funding throughout my Ph.D. career. Her passion and enthusiasm
about fundamental molecular science are not only inspiring, but contagious as well. I truly
appreciate her relentless effort to keep my scientific and critical thinking on point.
Chirantha Rodrigo, our former postdoc, was an excellent mentor when I joined the group. The
Old Monk, as I called him, blessed me with his vast knowledge about almost everything while
retaining his patience for my constant bugging. I had an opportunity to work with Dr. Amit
Samanta, another smart and hardworking postdoc. His excellent intuition and attention-to-detail
helped us to solve numerous technical and scientific problems. I am grateful to our former
graduate student, Dr. Mikhail Ryazanov, who helped me with valuable discussion and technical
assistance whenever I was stuck in the lab.
I thank Dr. Ravin Fernando, our current postdoc, for his hard work and perseverance. We teamed
up pretty strongly through the difficult times in the lab. I found it remarkably helpful that he has
solid technical skills with absolutely no ego. I thank my very capable lab partner, Bibek
Samanta, for providing theoretical and technical expertise with the added bonus of an intelligent
sense of humor when needed. Though I am leaving, I am sure the project is in good hands with
these two very capable fellas. I appreciate the positive reinforcement of my friend and
colleague, Daniel Kwasniewski, throughout my Ph.D. life. I am lucky to have had Dhritiman as
my roommate for last five years. It is remarkable that he never loses his sense of humor even in
the utmost crisis. I thank my friend, Dr. David Crombecque, with whom I share so many great
vi

memories. Saptaparna and Anirban were, and continue to be, great mentors who have helped me
to navigate the academic and professional ladders. I sincerely thank Alexandra Aloia who
checked the Grammar and English of my thesis meticulously and gave me valuable suggestions
regarding my resume writing. I made a fabulous group of friends: Avik, Arnab, Sohini,
Dibyendu, Debanjan, Sayan, Surabhee, Tirthendu, Gaurav, Eric, Shayne to name a few. The time
spent with each of them definitely transformed me in a positive way.
I am grateful to USC chemistry and the staff for their constant effort to create a smooth work
environment for us. I would especially like to thank Michele Dea, Magnolia Benitez, Frank
Niertit, Jaime Avila, and Don Wiggins for helping me out on numerous occasions.
My parents, Sachin and Doli Sutradhar, showed a great deal of patience throughout the process.
They always believed in me and gave me the freedom to chose what I want to do in life. My
three amazing sisters, Sukla Ghosal, Kajal Karmakar, and Srabanti Halder, are always by my
side as my friends, philosophers, and guides. In many ways, I am just a shadow of these three
people, and they inspire me every day to move forward. I thank my fabulous brothers-in-law,
Debdoot Ghosal, Jayanta Karmakar, and Subhash Halder, for being ready to help me out in any
crisis. I would like to send my overflowing affection to my nephew, Rishabh, and nieces,
Shreemoyee, Sukriti, Aishee, Shreetama, and Shreemanti for bringing fresh air to my mundane
days of grad life. I feel happy to see my family proud of my accomplishment, and there is no
word of gratitude to thank them enough for the unconditional love I received from all of them.
Lastly, I would say that the seven years of graduate school changed me in a positive way, both
scientifically and psychologically. Living in a diverse city like Los Angeles, I met so many
people in and outside the lab with different educational, ethnic and cultural backgrounds. All of
vii

these encounters changed me in such an unprecedented way that I never before would have
imagined. I thank all these people who crossed my path for last seven years and who have
transformed me into a better person.
Okay, I am done. Mic Drop.
Subhasish Sutradhar
July 2019
Los Angeles, CA

viii

List of Tables
Table 3.1: List of fundamentals for cis-HCOD ............................................................................ 69

Table 4.1: Tentative assignments of the 3pσu
1
Пu (v'1v'2v'3)← X
1
Σ
+
g (v"1v"2v"3) of CO2. ......... 98

Table 4.2: Ground and excited electronic states of CO2. Vertical and adiabatic excitation
energies (Vvert and Vmin, respectively) are listed. The geometries at the minimum of the potentials
are given by the O-CO distance rmin and bond angle αmin. .......................................................... 101

Table 5.1: Calculated vertical excitation energies for the indicated transitions, transition dipole
moment directions, and oscillator strengths calculated using EOM-EE-CCSD/6-311(2+)G** 128
Table 5.2: Positions and tentative assignments of the H-PFY peaks. The transitions marked as
“unassigned” are probably due to splittings in internal rotor transitions of the CH3 moiety...... 130
Table 5.3: Relevant vibrational frequencies of ground state PA................................................ 131
Table 5.4: Estimated angle α and corresponding KEmax values of fragments produced via
synchronous three body-fragmentation reactions 1a-1d. The error bars for KEmax represent the
deviations for a change of ± 5° in the corresponding α value ..................................................... 137
Table 5.5: Computed KEmax values of fragments produced in the secondary reactions 5-8,
correlated with the observed KEmax of CH3CO or HOCO generated in reaction 1. ................... 141
Table A1: Band origins of vibronic B(v'=0-9)-X(v"=0) bands of CO. The band origins of the
diagonal (v'=v") vibronic transitions are estimated by subtracting the corresponding vibrational
energies v" of the ground electronic state of CO from Tv'0. ........................................................ 169
Table A2: Diagonal k-X (v'-v") vibronic transitions ................................................................. 171
Table A3: Off-diagonal K-X (v'-v") vibronic transitions ........................................................... 172
Table A4: Diagonal (v-v) C-X vibronic transitions ................................................................... 174
Table A5: Off-diagonal C-X (v'-v") vibronic transitions ........................................................... 174

ix

Table A6: E(v')-X(v") vibronic transitions ................................................................................ 175
Table A7: Heats of formation (ΔfH
0
) of the relevant species. The errors in ΔfH
0
are less than 50
cm
-1
for all the species listed. ...................................................................................................... 176

x

List of Figures
Figure 1.1: 1D potential energy curves for the lowest electronic states of CO2 along the C-O
bond length (A) and O-C-O bending angle α (B). The singlet (A, B, C) and triplet states (a, b, c)
are shown in solid and dotted lines, respectively ............................................................................ 9
Figure 1.2: The UV–VIS absorption spectrum (S1←S0) of pyruvic acid recorded at 298 K ...... 13
Figure 2.1: Schematic of the general experimental setup used in the present studies ................. 32
Figure 2.2: (A) The 3D velocity distributions of AB correlated with ABCD → AB + CD (red)
and ABCD → AB + CD* (blue) channels. (B) The SVMI setup is illustrated in three-electrode
configuration following the dissociation of ABCD by the pump laser. The AB product in a
particular quantum state is selectively ionized by the appropriate probe laser wavelength
radiation. The 3D velocity distribution of ionized AB product is stretched temporally using
suitable repeller and extractor voltages, and a thin slice around the center of the distributions (∆z
≈ 0) along the x-y plane is recorded by fast gating the detector. .................................................. 37
Figure 2.3: Ar:Kr gas cell and associated optical arrangement used for VUV generation. ......... 43
Figure 2.4: Electron signal recorded in the VUV range of 114-115 nm ...................................... 45
Figure 3.1: Energies of excited Rydberg states of CH2OH. The shaded areas correspond to the
excitation energies (Eexc) of the 3px and 3pz states investigated in this work. The scale on the
right shows the energies of singlet and triplet states of H2CO and HCOH products relative to the
ground state of H2CO. ................................................................................................................... 52
Figure 3.2: The left and right panels show, respectively, SVMI images and the corresponding
radial velocity distributions of H photofragments recorded at the following CH2OH excitation
energies (3px): (a) 35,053 cm
-1
; (b) 35,714 cm
-1
and (c) 36,036 cm
-1
. Onsets of product states are
indicated by red sticks. The images were obtained by zooming in on the low velocity region,
which is correlated mainly with HCOH products. The peak at near-zero velocities is from
hydrogen atom contamination....................................................................................................... 57
Figure 3.3: Image and radial D-fragment velocity distribution from CH2OD excitation recorded
at Eexc = 36,563 cm
-1
. The low velocity peak corresponds to
3
H2CO. The distribution was
obtained by zooming in on the 0-5000 m/s velocity region. ......................................................... 58
xi

Figure 3.4: H-fragment radial velocity distributions obtained with excitation energies 37,106
cm
1
(red) and 37,313 cm
-1
(black) overlaid to demonstrate the appearance of a new feature
assigned as the onset of
3
HCOD. The inset shows a magnified plot of the low velocity region. . 61
Figure 3.5: H and D-fragment KER plots obtained by monitoring D (upper panels) and H (lower
panels) following 3pz excitation of CH2OD to the (a,c) 00
0
(Eexc = 41,004 cm
–1
); and (b,d) 60
1

(Eexc = 42,608 cm
-1
) bands. Sticks in each panel indicate the KE onsets associated with H2CO,
cis-HCOD, and the appearance of H and D from secondary dissociation .................................... 63
Figure 3.6: H and D-fragment KER plots obtained by monitoring D (upper panel) and H (lower
panel) following 3pz excitation of CH2OD to the 60
2
band (Eexc = 44,207 cm
-1
). Sticks in each
panel indicate the KE onsets associated with H2CO, cis-HCOD, and H and D from secondary
dissociation ................................................................................................................................... 64
Figure 3.7: A schematic energy diagram showing relevant states and barrier energies (in cm
–1
)
for H2CO, HCOH and their dissociation to H + HCO, and H2 + CO. The values include ZPE and
are relative to the ground state of H2CO. ...................................................................................... 67
Figure 3.8: Kinetic energy distributions (center-of-mass) recorded at CH2OD excitation energy
41,004 cm
–1
(3pz origin) by monitoring H-photofragments, and single-mode vibrational levels of
cis-HCOD (shown by red sticks). The zero of the energy scale was set as the center of the first
cis-HCOD peak (the zero-point energy level). Higher vibrational energy levels (overtones) are
generated by multiplication of the fundamental frequencies by the number of quanta. The stick
height is adjusted to match the experimental intensities and serves only to guide the eye. ......... 70
Figure 3.9: The cis-HCOD product distribution is fit with a stick spectrum of vibrational energy
levels of the ν1–4 modes (combinations and overtones). The stick spectrum intensities are set to
match the experimental signal intensities and do not represent relative populations. .................. 71
Figure 3.10: Comparison of cis-HCOD internal state distributions obtained via 3pz excitation of
CH2OD to the 60
1
band (Eexc = 42,608 cm
-1
), and by excitation to the 3s/3px underlying
background (Eexc = 42,553 cm
-1
). Sticks indicate the KE onsets associated with cis-HCOD and H
from secondary dissociation ......................................................................................................... 73
Figure 4.1: One color 2+1 REMPI spectra of CO produced in the photodissociation of CO2
around 230 nm recorded with low (black), medium (red) and high (blue) heating of the CO2
parent, respectively. The REMPI transition from ground state vibrational level (X
1
Σ
+
, v") to
excited state vibrational level (B
1
Σ
+
, v') is denoted by v"-v'. The assignment of the band denoted
by M is uncertain. Intensities are normalized to the 0-0 bandhead. ............................................. 84
xii

Figure 4.2: Comparison of the REMPI spectrum of the CO 0-0 QJJ band (blue) to a fit obtained
with a rotational linewidth Γ = 1.9 cm
-1
, assuming a rotational temperature of 1700 K, and a
Gaussian component with a width of 5 J" levels centered at J" = 43. The individual Boltzmann
and Gaussian components are shown in red dotted lines and the total fit is shown in black. ...... 85
Figure 4.3: Center-of-mass (c.m.) KER distributions obtained by monitoring CO(X
1
Σ
+
,v"=0)
products at 230.1 nm by 2+1 REMPI at the bandhead of the B←X (0-0) transition. Results are
shown for low (black), medium (red) and high (blue) CO2 heating. The relative signal intensities
in the three plots are arbitrary ....................................................................................................... 88
Figure 4.4: c.m. KER distributions obtained with high CO2 heating by monitoring (a) the
bandhead of the B←X 1-1 transition at 230.27 nm, and (b) the peak of the M band at 230.34 nm.
Signal intensities are normalized with respect to the feature centered at 5000 cm
-1
.................... 91
Figure 4.5: Center of mass (c.m.) KER plot obtained with high heating by monitoring CO
(X
1
Σ
+
;

v"=0) products at the 2+1 REMPI bandhead (86,916 cm
-1
, red) and at J"= 43 (86,962 cm
-
1
, black) of the B-X 0-0 transition. The KER plot obtained at J" = 43 does not show the feature at
5000 cm
-1
. The relative signal intensities are scaled for easy visual comparison ......................... 92
Figure 4.6: 2+1 REMPI of the O(
3
Pj) product at 225.8-226.6 nm showing its three spin orbit
states (
3
P2,
3
P1,
3
P0) ....................................................................................................................... 93
Figure 4.7: c.m. KER distributions obtained by monitoring O(
3
P2) products at 225.654 nm with
low (black), medium (red) and high (blue) CO2 heating. The intensities of the KER distributions
obtained with different heating levels are normalized with respect to the broad feature centered at
~3000 cm
-1
.................................................................................................................................... 94
Figure 4.8: c.m. KER distribution obtained by monitoring O (
1
D) product at 205.47 nm with
high CO2 heating. .......................................................................................................................... 95
Figure 4.9: 3+1 REMPI spectra of CO2 obtained via the 3pσu
1
Пu  X
1
Σ
+
g transition.
Progressive increase in vibrational excitation of CO2 is observed upon increased heating, as
shown for low (cyan), medium (red) and high (blue) heating. The spectrum of cold CO2 is
shown in black for comparison. The peak at 91,910 cm
-1
corresponds to the origin band of the
3pσu
1
Пu (000)  X
1
Σ
+
g (000) transition of CO2. The intensities in the spectra obtained with
different heating levels are normalized with respect to the 3pσu
1
Пu (000)  X
1
Σ
+
g (000) peak of
cold CO2. ....................................................................................................................................... 96
xiii

Figure 4.10: Energy diagram showing adiabatic (Vmin, solid line) and vertical (Vvert, dotted line)
energies of the excited electronic states of CO2. A and B states are shown in red and black,
respectively ................................................................................................................................. 100
Figure 5.1: Molecular beam H-PFY spectrum of PA in He carrier gas (red) and room
temperature absorption spectrum (black).................................................................................... 117
Figure 5.2: Molecular beam H-PFY spectrum of PA recorded in Ar carrier gas ...................... 118
Figure 5.3: Tc and Tt conformers of PA. The Tc conformer is more stable than the Tt due to
intramolecular H bonding. .......................................................................................................... 119
Figure 5.4: CH3CO-PFY (a) and HOCO-PFY (b) spectra of PA recorded in He. .................... 120
Figure 5.5: KER distributions (top) and corresponding recoil anisotropy parameters β (bottom)
of the CH3CO fragments following S1 excitation in peaks 1, 3, 6, and 9 of the PFY spectrum in
Figure 5.2 are shown in red, green, black, and blue, respectively. The KER distributions are
normalized to the peak at low KER. The predicted KEmax values of CH3CO produced via
synchronous and sequential three-body fragmentation processes are marked by arrows .......... 122
Figure 5.6: KER distributions (top) and corresponding recoil anisotropy parameters β (bottom)
of H fragments following S1 excitation in peaks 1, 3, 6, and 9 of the PFY spectrum in Figure 5.2
are shown in red, green, black, and blue, respectively. The predicted KEmax values for
synchronous and sequential three-body fragmentation processes are marked by arrows .......... 123
Figure 5.7: KER distribution of CO fragments (X
1
Σ
+
, v"=0, bandhead) detected by 2+1 REMPI
at 230.1 nm following the excitation of PA in peak 1 of Figure 5.2. The predicted KEmax values
for synchronous and sequential three-body fragmentation processes are marked by arrows ..... 124
Figure 5.8: KER distribution of HOCO following S1 excitation in peak 9 of Figure 5.2. The
predicted KEmax values of HOCO produced by three-body synchronous and sequential
fragmentation processes are marked by arrows .......................................................................... 125
Figure 5.9: KER distributions of CH3 (𝑋
2
A2", 0000, black) and CH3 (𝑋
2
A2", 0100, red)
recorded at 333.77 and 329.95 nm, respectively. ....................................................................... 126
Figure 5.10: KEmax of the final products of synchronous three-body fragmentation reactions 1a-
1d as function of α. The values of α for each channel at the moment of fragmentation are
xiv

estimated from the equilibrium geometry of the S1 state. The KEmax values of the fragments are
shown by vertical dashed lines ................................................................................................... 136
Figure 5.11: KEmax of the three fragments as a function of the KER in the first step of the
sequential three-body fragmentation reactions 5 and 6 where hot CH3CO dissociates to give
CH3+CO and CH2CO+H, respectively ....................................................................................... 139
Figure 5.12: KEmax of the three fragments as a function of KER in the first step of the sequential
three-body reactions 7 (a) and 8 (b) where hot HOCO dissociates to give OH+CO and H+ CO2,
respectively ................................................................................................................................. 140
Figure 6.1: Example of the observed lumpy IR spectrum of CO2 up to ~ 10,000 cm
-1
. The cross-
sections are given in cm
2
/molecule. Each polyad includes different combinations of stretch and
bend vibrations. ........................................................................................................................... 158
Figure 6.2: UV and VUV transmittance of different materials ................................................. 161
Figure 6.3: Schematic of waveguide-tripling cell combination ................................................. 162
Figure 6.4: Schematic diagram for the Generation of VUV radiation in the range 6.9-16.0 eV
using Kr or Xe by resonant four-wave mixing. ω1 is fixed at either 249.629 nm (Xe) or 212.556
nm (Kr), while ω2 is scanned from 400 to 900 nm ..................................................................... 163
Figure A1: The thermal rotational distribution of the B-X (2-2) vibronic transition simulated
using PGOPHER (black) at 1700 K. The experimental 2+1 REMPI spectrum of the CO (X
1
Σ
+
;
v") product measured with high heating of the CO2 parent is shown in red............................... 170
Figure A2: The thermal rotational distribution of the B-X 5-5 vibronic transition (black) at 1700
K. The experimental 2+1 REMPI spectrum of the CO (X
1
Σ
+
; v") product is shown in red. ..... 171
Figure A3: The thermal rotational distributions of the k(v'=1) ← X(v"=2) vibronic transition
simulated using PGOPHER (black) at 500 K. The experimental 2+1 REMPI spectrum of the
CO(X
1
Σ
+
; v") product is shown in red. ...................................................................................... 173
Figure A4: A schematic depiction of synchronous three-body-fragmentation of a triatomic ABC
molecule where two bonds break simultaneously. The X-axis is defined in the momentum vector
space such that it lies along the central angle bisector of the planar molecule. The energy
disposal in A, B, and C depends on the angle α at the critical configuration. ............................ 177
xv

Figure A5: A schematic of sequential three-body fragmentation of ABC where the B-C bond
breaks first at time τ1 producing AB and C fragments. The AB fragment can further dissociate
after time τ2, which is longer than the mean rotational period of the AB intermediate .............. 178

xvi

Abstract
The spectroscopy and UV induced photodissociation of hydroxymethyl(CH2OH/CH2OD)
radical, hot CO2, and pyruvic acid in molecular beams are studied using time-of-flight (TOF)
mass spectrometry and sliced velocity map imaging (SVMI).
Rotational, vibrational and electronic states of formaldehyde and cis-hydroxymethylene
products are generated in the photodissociation of CH2OH following excitation of the radical to
its 3px and 3pz Rydberg states. With CH2OD precursors, formaldehyde and hydroxymethylene
products are examined separately by monitoring D and H, respectively. Whereas the main
dissociation channels lead to formaldehyde and cis-hydroxymethylene in their ground electronic
states, at higher excitation energies the kinetic energy distributions (KEDs) of the H and D
photofragments exhibit additional peaks, which are assigned as the triplet states of formaldehyde
and hydroxymethylene. The rotational excitation of cis-hydroxymethylene depends on the
excited Rydberg state of CH2OD, and is lower in dissociation via the 3pz state than via the lower
lying 3px and 3s states. Results obtained with the deuterated isotopologs of CH2OH demonstrate
that the yield of the triplet state of formaldehyde decreases upon increasing deuteration,
suggesting that the conical intersection seams that govern the dynamics depend on the degree of
deuteration. Vibrational excitation of cis-HCOD, which spans the entire allowed internal energy
range, consists mostly of the CO-stretch and CH2 in-plane bend modes. When the internal energy
of cis-HCOD exceeds the dissociation threshold to D + HCO, slow D and H photofragments
deriving from secondary dissociation are observed. The yields of these H and D fragments are
xvii

comparable, and we propose that they are generated from cis-HCOD via prior isomerization to
HDCO.
The 205-230 nm photodissociation of vibrationally excited CO2 at temperatures up to
1800 K was studied using Resonance Enhanced Multiphoton Ionization (REMPI) and SVMI.
CO2 molecules seeded in He were heated in a SiC tube attached to a pulsed valve and
supersonically expanded to create a molecular beam of rotationally cooled but vibrationally hot
CO2. Photodissociation was observed from vibrationally excited CO2 with internal energies up to
about 20,000 cm
-1
, and CO(X
1
Σ
+
), O(
3
P) and O(
1
D) products were detected by REMPI. The large
enhancement in the absorption cross-section with increasing CO2 vibrational excitation made this
investigation feasible. The internal energies of heated CO2 molecules that absorbed 230 nm
radiation were estimated from the kinetic energy release (KER) distributions of CO(X
1
Σ
+
)
products in v" = 0. CO2 internal energies in excess of 16,000 cm
-1
were confirmed by observing
O(
1
D) products. It is likely that initial absorption from levels with high bending excitation
accesses both the A
1
B2 and B
1
A2 states, explaining the nearly isotropic angular distributions of
the products. The CO product internal energy distributions change with increasing CO2
temperature, suggesting that more than one dynamical pathway is involved when the internal
energy of CO2 (and the corresponding available energy) increases. The KER distributions of
O(
1
D) and O(
3
P) show broad internal energy distributions in the CO(X
1
Σ
+
) co-fragment,
extending up to the maximum allowed by energy but peaking at low KER values. Although not
all the observations can be explained at this time, with the aid of available theoretical studies of
CO2 VUV photodissociation and O+CO recombination, it is proposed that following UV
absorption, the two lowest lying triplet states, a
3
B2 and b
3
A2, and the ground electronic state are
involved in the dynamical pathways that lead to product formation.
xviii

The UV photodissociation of pyruvic acid (PA) is studied in molecular beams using TOF
mass spectroscopy and SVMI following excitation to the first absorption band (S1←S0) at 330-
380 nm. CH3CO, HOCO, CO, CH3, and H are detected as photodissociation products. The
photofragment yield (PFY) spectrum of the H product is recorded at 350-380 nm in He and Ar
carrier gases. The spectrum shows sharp vibrational features reflecting the significant rotational
cooling achieved in the molecular beam. It matches well the broad features observed in the room
temperature absorption spectrum, and indicates that the S1 state lives longer than a picosecond.
The origin band of the S1←S0 transition is identified at 26,710 cm
-1
, and progressions in the CH3
and C-C torsional modes are tentatively assigned. KER and angular distributions of CH3CO,
HOCO, CO, CH3, and H fragments indicate that additional photon absorption from S1 to the
S2/S3 states is facile, and is followed by rapid dissociation to the observed fragments. Based on
the energetics of the different dissociation pathways and analyses of the observed KER
distributions, three-body fragmentation processes are proposed as major contributors to the
formation of the observed products
1

Chapter 1
Introduction
The studies described in this dissertation are focused on the UV photon absorption and
photodissociation dynamics of small molecules and radicals, which are important in combustion
and the atmosphere. In my PhD research, I have studied the photodissociation processes of three
chemical systems: hydroxymethyl radical (CH2OH), hot carbon dioxide (CO2), and pyruvic acid
(CH3COCOOH). The light induced photophysics of these species under combustion and/or
atmospheric conditions are often complicated by multiple collisions, and understanding the
mechanisms of the photochemistry leading to the different products is not straightforward.
Therefore, photodissociation studies under collisionless conditions are necessary to obtain the
nascent product state distributions and decipher photodissociation mechanisms. The studies
described herein were carried out in molecular beams produced by supersonic expansion inside a
high vacuum chamber under pristine collisionless condition. The supersonically cooled radicals
and molecules were electronically excited using tunable laser radiation and the photodissociation
products were monitored using time-of-flight mass spectrometry combined with the time-sliced
velocity map imaging technique. The results show that the photodissociation dynamics of these
species involve both excited and ground electronic states and the product state distributions are
controlled by the intricate coupling among multiple potential energy surfaces.

2

1.1 Background
1.1.1 CH2OH
The hydroxymethyl radical, CH2OH, plays an important role in atmospheric, combustion
and interstellar chemistry. It is a major product of the reactions between O (
1
D) and methane in
the upper atmosphere.
1,2
The H abstraction in the reaction between Cl atom and methanol
primarily yields CH2OH.
3,4
In addition, several studies indicate the presence of CH2OH in the
methanol enriched interstellar medium.
5–7
Recently, it has been proposed that the radical-radical
recombination between HCO and CH2OH may lead to the formation of the simple sugar
glycolaldehyde in interstellar ice grains.
6,8–10
Therefore, it is no surprise that the spectroscopy
and photochemistry of the CH2OH radical remain an active area of research. For the last two
decades, our group studied the spectroscopy and photodissociation of CH2OH both in the ground
and excited electronic states using several different techniques. For example, the near threshold
dissociation of ground state CH2OH was studied recently upon excitation to O-H overtone using
Sliced Velocity Map Imaging (SVMI), a study that provided detailed insights into dissociation
via direct O-H bond fission vs. prior CH2OH ↔ CH
3
O isomerization followed by dissociation.
While the ground state of CH2OH is now considered well characterized,
11–19
the
photodissociation CH2OH from excited electronic states revealed a much more complicated
picture due to the involvement of multiple potential energy surfaces (PESs) and therefore, the
photochemistry and product state distributions of excited state CH2OH are less well understood.
Ab-initio electronic structure calculations show that the lowest lying electronic states of
CH2OH are diffuse and possess significant Rydberg character.
20,21
The lowest lying electronic
3

states of CH2OH originate from the promotion of an unpaired electron from the 𝜋 𝐶𝑂
∗
molecular
orbital to the Rydberg orbitals with following configuration:
Based on the Rydberg formula, Dulcey and Hudgens predicted that the lowest electronic
states of CH2OH can predominantly be described by the n=3 Rydberg series.
20
Later, several
experimental and theoretical studies revealed that the four lowest electronic transitions of
CH2OH can be classified as π →3s, π →3px, π →3py, π →3pz.
22–26
Due to the low ionization
energy of CH2OH (IP = 7.56 eV), these states lie at much lower energies compared to the typical
Rydberg states of closed-shell molecules. These states are energetically quite close, and as a
result, they can interact with one another as well as with the ground state via non-adiabatic
coupling. Indeed, calculations show the existence of multiple conical intersections between these
states and the ground state. As a result, the photodissociation dynamics and products’ branching
ratios are dictated by the location and shape of the conical intersection seams.
27–30
Therefore, the
photodissociation studies of electronically excited CH2OH can provide important insights into
the couplings between the Rydberg and ground states and the role of conical intersections in the
photodissociation dynamics.
Previous calculations showed that the π→ 3px and π→ 3pz transitions are the strongest,
while the 3py state does not carry any oscillator strength.
21
The absorption spectrum of CH2OH
spans from ~ 26000 cm
-1
up to at least 45,000 cm
-1
with overlapping transitions to the 3s, 3px,
and 3pz states.
22–26
The onsets of the 3s, 3px, and 3pz transitions are reported to be at 25,970,
4

35,000, and 41,050 cm
-1
, respectively. It is interesting to note that the 3s, 3px, and 3pz states have
different lifetimes. Feng et al. reported that the 3s state is structureless and extends all the way to
the to 3pz state.
26
The 3px state, which lies above the background of the 3s state, exhibits a
progression in the CO stretch with band broadening of ~ 100 cm
-1
. The spectrum of the 3pz state
shows clear vibronic band structure with linewidths of about 10 cm
-1
.
25,31
Similar to the 3px, this
band structure is assigned to a progression in the CO stretch. All of these states are reported to be
short-lived with lifetimes < 0.5 ps.
25,26

The photodissociation of CH2OH in the region 28,000–41,000 cm
-1
(~357–244 nm)
following the excitation to the 3s, 3px, and 3pz states was studied previously by using core
sampling time-of-flight (TOF) spectroscopy
32–34
. The reported thresholds of the lowest channels
correlated to the ground state CH2OH are listed below:
18,35

CH2OH → CH2O (X
̃
1
A1) + H D0 = 10,160 ± 70 cm
-1
(I)
CH2OH → trans-HOCH (X
̃
1
A') + H D0 = 28,420 ± 70 cm
-1
(II)
CH2OH → cis-HCOH (X
̃
1
A') + H D0 = 29,970 ± 70 cm
-1
(III)
In these studies, the formaldehyde and hydroxymethelene channels were distinguished by
using the CH2OD isotope and detecting D and H fragments, respectively. The D and H
photofragment yield spectra were recorded as a function of excitation energy. With
27,400−30,500 cm
-1
(3s) photolysis, only O−D bond-breaking was observed, whereas both D
and H products were detected at energies above 30,500 cm
-1
. The signal from the C−H bond
fission channel increased monotonically with photolysis energy and reached D:H ∼1 when
exciting to the 3pz state at >41,000 cm
-1
. The H/D kinetic energy (KE) and angular distributions
5

combined with theoretical calculations revealed the existence of multiple surface crossings
between these states and the ground state, which eventually lead to O−H and C−H bond breaking
with formaldehyde and hydroxymethelene co-fragments generated via channels I-III.
27–30,36

Although the core sampling TOF spectroscopic technique sheds light on several
important aspects of the photodissociation of CH2OH/CH2OD radical, the rovibrational levels in
the formaldehyde and hydroxymethelene products could not be resolved due to poor KE
resolution of the core-sampling method. The conical intersection calculations by Yarkony and
Hoffman revealed several interesting aspects of the photodissociation of CH2OH upon excitation
to 3s and 3px states,
27–30,36
which could not be tested experimentally due to lack of sufficient
energy resolution. Therefore, the higher resolution SVMI technique was implemented in our
laboratory and adapted for detection of H/D photofragments.
37
SVMI has much higher KE
resolution and can serve as an excellent tool for elucidating photodissociation dynamics and
identifying even minor product quantum states and their energy thresholds.
Previously, it has been shown that SVMI is easy to implement for all photofragments
except hydrogen, whose mass is too small for effective slicing.
38,39
Realizing that H
photofragments carry most of the KE released in the photodissociation and therefore, are useful
reporters of quantum states of molecular co-fragments, our group has designed and implemented
an SVMI arrangement optimized specifically for H fragments, as discussed in Section 2.3. SVMI
makes it possible to zoom-in on low-velocity regions of the distributions without the need to
capture the full range of velocities (as must be done in full projection velocity map imaging).
This enhances resolution and signal-to-noise ratios, and helps to identify minor products close to
their energetic thresholds.
6

Ryazanov et al. demonstrated the vibrational resolution of the SVMI arrangement for
formaldehyde fragments produced in the overtone-induced dissociation of ground state CH2OH
and CD2OH.
18
Rodrigo et al. revisited the photodissociation dynamics of isotopologs of the
hydroxymethyl radical (CH2OH, CH2OD, and CD2OD) following excitation to the 3s and 3px
Rydberg states using the same experimental arrangement.
35
The well resolved features in the
kinetic energy distributions (KEDs) of H/D fragments revealed the formaldehyde and
hydroxymethelene product state distributions. Formaldehyde was the major product of the
photodissociation following excitation to the 3s state, and most of these fragments were
produced with high KE and low rotational excitation. Analysis showed a progression in the CO-
stretch with contributions from the CH(D) deformation modes, particularly the scissors mode, υ3.
On the other hand, the internal energy distribution in the HCOH product from the 3s state
showed a broad rovibrational energy distribution.
Interestingly, the photodissociation dynamics changed abruptly when exciting to the
origin band of the 3px state. The rotational excitation of the HCOH product decreased, and both
trans- and cis-HCOH products were detected in addition to formaldehyde. These experimental
observations were supported by the theoretical calculations by Yarkony and Hoffman who find
that, upon excitation to the 3s state, most of the CH2OH radicals access the conical intersection
seam between 3s and the ground state along O-H coordinate and efficiently dissociate to CH2O +
H in the repulsive part of the ground state.
27–30
On the other hand, when the radical is excited to
the 3px state, it first couples to 3s via a conical intersection seam and subsequently to the ground
state via 3s/ground state conical intersections along the O−H and C−H coordinates, producing
CH2O and HCOH products in the ground state. The calculations reveal that the shapes of the
conical intersections have a strong effect on the product state distributions and the relative
7

branching ratios, which were consistent with the experimental observations. The progressive
broadening of the spectra from 3pz to 3s was also rationalized by the shapes of the conical
intersections. Feng et al. argue that the vertical cone between 3s/ground state leads to efficient
internal conversion and fast dissociation which is responsible for the diffuse nature of the 3s
absorption band.
26
On the other hand, the titled cone between 3px and 3s states gives rise to less
efficient coupling with the lower electronic states. This may partly explain why the vibronic
features are narrower in the excitation to 3pz than to 3px.
25,26

The improved KE resolution of SVMI combined with theoretical calculations provide
important new insights into the photodissociation dynamics of CH2OH following excitation to 3s
and near the origin of the transition to 3px (28,000-30,000 cm
-1
). However, the photodissociation
at the energy range above the 3px origin, which also accesses the 3pz state (30,000-45,000 cm
-1
),
was studied only by the low-resolution core-sampling TOF technique, which could not resolve
the internal states of the products. Encouraged by the previously demonstrated energy resolution
of our SVMI setup, the photodissociation of CH2OH and CH2OD at energies 30,000-45,000 cm
-
1
, covering the 3px and 3pz states, is reexamined in the present study. At these high energies,
several new product channels become available, such as the excited singlet and triplet states of
CH2O, and the triplet state of HCOH, and the superior KE resolution achieved by SVMI is
particularly important. The CH2O and HCOH product state distributions can test the effect of
conical intersections on the photodissociation dynamics in the 3pz state which is expected to be
different than that on the 3s and 3px states.
For the sake of completeness, we mention that the geometries of the 3s, 3px, and 3pz
Rydberg states are planar, similar to the ground state CH2OH ion and can alternatively be
8

denoted as 1
2
A' (3s), 2
2
A' (3px), 2
2
A'' (3pz).
13
This nomenclature is important for spectroscopic
assignments but is less relevant to the photodissociation dynamics studies presented here.
Therefore, I use exclusively the Rydberg notations (3s, 3px, and 3pz) throughout this dissertation.
1.1.2 CO2
Carbon dioxide is one of the most important molecules in the solar system. It is a product
of combustion and flames and plays a key role in climate change in Earth’s atmosphere.
40–42

CO2 is also found to be a ubiquitous component of ices in the interstellar medium, and
constitutes a major part of the atmospheres of Venus and Mars.
43–46
It has been argued that the
photodissociation of CO2 induced by the solar radiation can be considered as a sink of CO2 in
Earth’s atmosphere. Therefore, its photoinitiated reactions have attracted much attention.
Photodissociation of ground state CO2 has been studied mainly at wavelengths < 190 nm,
where absorption cross-sections from the ground vibrational state are significant.
47–61
These
studies identified several photodissociation channels following excitation to highly excited
valence and Rydberg states. The lowest energy dissociation channels are listed below:
CO2 → CO(X
1
Σ
+
): + O(
3
P) D0 = 43,958 cm
-1
(I)
CO2 → CO(X
1
Σ
+
): + O(
1
D) D0 = 59,847 cm
-1
(II)
CO2 → CO(X
1
Σ
+
): + O(
1
S) D0 = 77,753 cm
-1
(III)
CO2 → CO(a
3
Π): + O(
3
P) D0 = 92,425 cm
-1
(IV)
While previous studies shed light on the photodissociation dynamics and product state
distribution of CO2 from highly excited electronic states, very little is known about the
photodissociation from the lowest excited electronic states, which are closer to the thresholds of
9

O(
3
P) + CO(X
1
Σ
+
) and O(
1
D) + CO(X
1
Σ
+
) channels. Theoretical calculations reveal that there are
three singlet (1
1
B2, 1
1
A2, and 2
1
A2) and three triplet (1
3
B2, 1
3
A2, and 2
3
A2) excited electronic
states of CO2 that can be accessed at λ > 150 nm in the Franck–Condon (FC) region.
62–64

According to the nomenclature used by Schmidt et al. in their theoretical paper, the singlet and
triplet states are abbreviated as A (1
1
B2), B (1
1
A2), C (2
1
A1) and a (
3
B2), b (1
3
A2), c (2
3
A2),
respectively.
Figure 1.1: 1D potential energy curves for the lowest electronic states along the C-O bond
length (A) and O-C-O bending angle α (B) calculated by Schmidt et al.
62
The singlet (A, B, C)
and triplet states (a, b, c) are shown in solid and dotted lines, respectively.
10

Schmidt et al. calculated 1D cuts along the O-C-O bending angle and one of the two C-O
bonds, as shown in Figure 1.1, which revealed that the geometries of these lowest excited
electronic states are bent.
62
The FC factors connecting the linear ground state and these low
lying bent excited states are vanishingly small. Consequently, the absorption cross-sections from
the vibrationless ground electronic state to the singlet electronic states become negligible
towards longer wavelengths (λ > 190 nm)
65,66
, and the photodissociation of CO2 near the
threshold of O(
3
P) + CO(X
1
Σ
+
) has not been studied.
Although ground state CO2 does not absorb radiation at wavelengths longer than 190 nm,
it has been reported that vibrationally excited CO2 absorbs even at 300 nm.
67–70
The highly
vibrationally excited CO2 is reported to play an important role in several high temperature
environments such as combustion, hydrocarbon flames and the hot atmospheres of Venus and
several exoplanets.
40,41,44,45,70–72
In this regard, the effect of temperature on the UV absorption of
CO2 has been studied extensively in heated shock tubes, laminar flames and furnaces in the
wavelength region 190-355 nm at temperatures as high as 6300 K.
50,67,69,70,73,74
These
temperature dependent studies showed that, while the absorption cross-sections of CO2 at 300 K
is less than 10
-22
cm
2
at λ > 200 nm, it increases almost exponentially with temperature.
62
The
long-wavelength enhancement of the absorption cross-sections at higher temperature is attributed
to absorption of vibrationally excited CO2, which has significantly large FC overlap with the low
lying bent electronic states compared to the vibrationless linear ground state.
63,69,73
Several
theoretical and experimental studies revealed that both the A and B states are involved in the
absorption.
62,63,71
However, it has been pointed out that C, a, and b states might also contribute.
62

11

Since vibrationally excited CO2 absorbs at wavelengths longer than 200 nm, its
photodissociation can be studied at energies much closer to the dissociation threshold of O(
3
P) +
CO(X
1

+
) and O(
1
D) + CO(X
1

+
). The absorption cross-sections and transition dipole moments
calculated at 150-210 nm showed that excitations to both A and B states play a major role in the
photodissociation of CO2 at energies < 8.5 eV, while the contribution of the C state is
negligible.
62
In addition, O(
3
P) + CO(X
1

+
) is a spin forbidden channel and can only be reached
through spin-orbit coupling.
75–80
Therefore, the optically excited singlet state must couple to
triplet states (a and b) in order to dissociate via this channel. Indeed, several theoretical
calculations identified multiple surface crossings among the lowest singlet (A, B) and triplet (a,
b, and c) states leading to dissociation.
62–64,81
These states are connected by conical intersections
and spin-orbit couplings at specific internuclear distances and bond angles. Bending and
asymmetric stretch modes are identified as tuning and promoting modes, respectively, in the
conical intersections among singlet states. Several theoretical studies have addressed the
recombination of O(
3
P) and CO to the ground singlet state of CO2, which also involves triplet
and singlet surfaces and spin-orbit couplings. Consequently, the couplings to the dissociation
continuum are likely to involve multiple PESs.
Similar multi-state photodissociation dynamics are found in other isoelectronic 16-electron
molecules, such as N2O, OCS, and HNCO, all of which exhibit linear to strongly bent electronic
transitions.
82
For example, previous work on HNCO identified several dissociation pathways
following initial photoexcitation to the bent S1 state and terminating in the lowest energy NH(X

3
Σ
-
) + CO(X
1
Σ
+
) channel: S1 →T1 →products; S1 →S0 →T1 →products; and S1 →T2 →T1 →
products.
83–87
These three pathways persist even at higher excitation energies when the allowed
singlet product channels are open. In addition, Crim and coworkers carried out vibrationally
12

mediated photodissociation experiments on HNCO, demonstrating vibrational state-specific
effects on surface crossings.
88,89

Therefore, the photodissociation of vibrationally “hot” CO2 is not only important in the
laser-based diagnostics for combustion, hydrocarbon flames, and modeling the atmosphere of hot
planets, but it also gives us an opportunity to unravel the dynamics and interactions among
excited electronic states closer to the threshold of the lowest dissociation channels.
1.1.3 Pyruvic Acid
Pyruvic acid (PA) is an important α-keto acid in the atmosphere. It forms as an
intermediate in the oxidation pathways of isoprene, an important volatile organic compounds
(VOC) in the atmosphere that leads to formation of secondary organic aerosols (SOA).
90–93
It has
been shown that organic acids including PA are rapidly produced in the polluted atmosphere of
urban areas and in the tropical region via degradation of several biogenic and anthropogenic
VOCs in addition to isoprene.
94–98
Whereas most of the VOCs and related organic compounds in
the atmosphere degrade by reacting with OH radicals, the rate of reaction of PA with OH is quite
slow.
99–102
Harris and co-workers studied the photochemical kinetics of PA both in the gas and
aqueous phases and concluded that the rate of direct photolysis of gaseous PA is several orders
of magnitude higher compared to OH oxidation.
102,103
Therefore, solar radiation-induced
photolysis acts as the primary sink of tropospheric PA.
91,100,103–108

The UV absorption and dissociation of gas-phase PA have been studied extensively
because the S1 ← S0 absorption band of PA overlaps significantly with the solar flux reaching
Earth’s surface.
100,103,106,108,109
The first absorption band of pyruvic acid in the gas phase lies at
380-300 nm as shown in Figure 1.2.
109
This band accesses the first excited singlet state (S1) via a
13

Figure 1.2: The UV–VIS absorption spectrum (S1←S0) of pyruvic acid recorded by Horowitz et
al. at 298 K.
109

π*← n+ transition in which an electron in the nonbonding lone pair of the carbonyl oxygen is
promoted to an antibonding orbital located primarily along the central C-C bond.
110
Yamamoto
and Back also recorded the emission spectrum of PA at the same wavelengths at 1-10 torr and
340 K, and inferred that the S1 state is relatively long lived, particularly at low energies where
the emission was more intense.
108
Although a high-resolution 298 K absorption spectrum of PA
(1.2 Torr) has been reported, the vibronic bands were too broad and overlapped for spectroscopic
assignment.
109

Several photolysis studies of PA performed at low pressures (1-10 torr) at λ < 366 nm
showed that decarboxylation leading to CO2 and CH3CHO final products is the predominant
photodissociation pathway.
106,108
Although the mechanism of the formation of these products has
260 280 300 320 340 360 380 400
0
1x10
-20
2x10
-20
3x10
-20
4x10
-20
5x10
-20
6x10
-20
absorption cross section/ cm
2
. molecule
-1
wavelength / nm
14

been proposed, the contributions of S1, T1 and highly vibrationally excited levels of S0 to the
photodissociation mechanism(s) are still under debate. For example, while early studies ruled out
the participation of T1 and vibrationally “hot” S0 levels in the formation of CO2 and CH3CHO,
theoretical calculation by Chang et al. showed that decarboxylation happens on S0 after an
efficient conical intersection between S1 and S0.
108,110
They also found that dissociation on T1 can
be an additional source of CO2 and CH3CHO products.
In addition to these low-pressure studies, the photodissociation of PA via S1 has been
investigated at atmospheric pressures in the presence of O2, N2 and air.
103–105,107
These studies
not only revealed new photodissociation pathways, but also showed that the photodissociation
quantum yield depended on the initial concentration of PA and the pressure and nature of the
buffer gas. For example, Berges and Warneck suggested that photolysis of PA at 350 nm at
atmospheric pressure produces significant amounts of CH3CO and acetic acid.
107
Vaida and co-
workers investigated the photodissociation of PA (2-12 ppm) in air and N2 using broadband
radiation similar to the solar flux and found acetic acid as one of the photoproducts.
103

Whereas previous studies shed light on the final products of PA photodissociation under
atmospherically relevant conditions, most of these experiments have been carried out in static
gas cells where nascent products are likely to undergo secondary reactions due to multiple
collisions. Not surprisingly, under these conditions, the final products and their yields found in
the different studies were depended on the concentration of PA, the pressure in the reaction
chamber, and the specific experimental arrangement used in each study. These studies do not
always agree with one another on the relative abundances of the final products.
103–105,107
The
results indicate that the relaxation of the S1 state and its coupling to T1 and/or S0 may be sensitive
15

to pressure. Clearly, information on the photophysics and photochemistry of PA in the pristine
collisionless environment of molecular beams is desirable.
1.2 Chapter Overview
The content of this dissertation is organized in the following chapters:
Chapter 2 describes the experimental principles and methods used to study the
spectroscopy and photodissociation processes. The molecular beam apparatus, which consists of
a high vacuum chamber and pulse nozzle system, is described. TOF mass spectrometry
combined with SVMI technique is used to unravel the photodissociation dynamics. The
generation of vacuum ultraviolet (VUV) radiation in a tripling cell is discussed in detail. The
characteristics of the pyrolysis nozzle used to create vibrationally hot CO2 is also described.

Chapter 3 presents the photodissociation studies of CH2OH and CH2OD radical from the
3px and 3pz Rydberg states at wavelengths 225-285 nm. The formaldehyde and
hydroxymethylene product channels are distinguished by selectively detecting H and D products
from CH2OD. Both singlet ground and triplet excited states of formaldehyde and
hydroxymethylene are observed by monitoring the H and D fragments’ kinetic energy release
(KER). The results indicate that internally hot hydroxymethylene undergoes secondary
dissociation. The formaldehyde and the hydroxymethylene product state distributions reveal the
effects of multiple conical intersections among the Rydberg states and the ground state on the
photodissociation dynamics. The results agree remarkably well with theoretical predictions.

16

Chapter 4 investigates the photodissociation of vibrationally hot CO2 at λ > 200 nm. The
large enhancement in the absorption cross-section with increasing CO2 vibrational excitation
makes this investigation feasible. The hot CO2 is electronically excited using 230, 225 and 205
nm radiation and the two lowest dissociation channels, CO(X
1
Σ
+
) + O(
3
P) and CO(X
1
Σ
+
) +
O(
1
D), are observed. CO(X
1
Σ
+
) product internal energies were estimated from resonance
enhanced multiphoton ionization (REMPI) spectroscopy, and the KER distributions of the
CO(X
1
Σ
+
), O(
3
P) and O(
1
D) products were obtained by SVMI. From the KEDs of the CO(X
1
Σ
+
,
v" = 0) products, the maximum internal energy of hot CO2 is estimated to be about 20,000 cm
-
1
. With the aid of available theoretical studies of CO2 VUV photodissociation and O+CO
recombination, it is proposed that following UV absorption to the A
1
B2 and B
1
A2 states, the two
lowest lying triplet states, a
3
B2 and b
3
A2, and the ground electronic state are involved in the
dynamical pathways that lead to product formation.

Chapter 5 describes the photodissociation of pyruvic acid (PA) following excitation to
the first absorption band (S1←S0) at 330-380 nm. CH3CO, HOCO, CO, CH3, and H are detected
as photodissociation products. The photofragment yield (PFY) spectrum of the H product is
recorded at 350-380 nm in He and Ar carrier gases and matches well with the room temperature
absorption spectrum. KER and angular distributions of CH3CO, HOCO, CO, CH3, and H
fragments indicate a two-photon absorption to S2 or S3 state via S1 state followed by rapid
dissociation to the observed fragments. Based on the energetics of the different dissociation
pathways and analyses of the observed KEDs, three-body fragmentation mechanisms are
proposed as the major contributors to the formation of the observed products.

17

Chapter 6 presents research ideas for the future based on the present work. As discussed
in Chapters 3 and 5, the transient species HCOH and CH3COH can be produced in the molecular
beam by photolysis of CH2OH and pyruvic acid, respectively. In this chapter, I discuss other
possible ways to generate and stabilize HCOH and CH3COH in the molecular beam. Direct
REMPI detection schemes of HCOH and CH3COH are also proposed.
References
(1) Lin, J. J.; Harich, S.; Lee, Y. T.; Yang, X. Dynamics of the O(
1
D)+CH4 Reaction: Atomic
Hydrogen Channel vs Molecular Hydrogen Channel. The Journal of Chemical Physics 1999, 110
(22), 10821–10829.
(2) Lin, J. J.; Shu, J.; Lee, Y. T.; Yang, X. Multiple Dynamical Pathways in the O(
1
D)+CH4
Reaction: A Comprehensive Crossed Beam Study. The Journal of Chemical Physics 2000, 113
(13), 5287.
(3) Ahmed, M.; Peterka, D. S.; Suits, A. G. Imaging H Abstraction Dynamics in Crossed
Molecular Beams: Cl+ROH Reactions. Physical Chemistry Chemical Physics 2000, 2 (4), 861–
868.
(4) Rudić, S.; Murray, C.; Ascenzi, D.; Anderson, H.; Harvey, J. N.; Orr-Ewing, A. J. The
Dynamics of Formation of HCl Products from the Reaction of Cl Atoms with Methanol, Ethanol,
and Dimethyl Ether. The Journal of Chemical Physics 2002, 117 (12), 5692–5706.
(5) Bennett, C. J.; Chen, S.-H.; Sun, B.-J.; Chang, A. H. H.; Kaiser, R. I. Mechanistical
Studies on the Irradiation of Methanol in Extraterrestrial Ices. The Astrophysical Journal 2007,
660 (2), 1588–1608.
18

(6) Bennett, C. J.; Kaiser, R. I. On the Formation of Glycolaldehyde (HCOCH2OH) and
Methyl Formate (HCOOCH3) in Interstellar Ice Analogs. The Astrophysical Journal 2007, 661
(2), 899–909.
(7) Garrod, R. T.; Weaver, S. L. W.; Herbst, E. Complex Chemistry in Star‐forming Regions:
An Expanded Gas‐Grain Warm‐up Chemical Model. The Astrophysical Journal 2008, 682 (1),
283–302.
(8) Butscher, T.; Duvernay, F.; Theule, P.; Danger, G.; Carissan, Y.; Hagebaum-Reignier, D.;
Chiavassa, T. Formation Mechanism of Glycolaldehyde and Ethylene Glycol in Astrophysical
Ices from HCO
•
and
•
CH2OH Recombination: An Experimental Study. Monthly Notices of the
Royal Astronomical Society 2015, 453 (2), 1587–1596.
(9) Wang, T.; Bowie, J. H. Radical Routes to Interstellar Glycolaldehyde. The Possibility of
Stereoselectivity in Gas-Phase Polymerization Reactions Involving CH2O and ˙CH2OH. Organic
& Biomolecular Chemistry 2010, 8 (20), 4757.
(10) Eckhardt, A. K.; Linden, M. M.; Wende, R. C.; Bernhardt, B.; Schreiner, P. R. Gas-Phase
Sugar Formation Using Hydroxymethylene as the Reactive Formaldehyde Isomer. Nature
Chemistry 2018, 10 (11), 1141–1147.
(11) Bauschlicher, C. W.; Partridge, H. Heat of Formation of CH2OH. J. Phys. Chem. 1994, 98
(7), 1826–1829.
(12) Curtiss, L. A.; Kock, L. D.; Pople, J. A. Energies of CH2OH, CH3O, and Related
Compounds. The Journal of Chemical Physics 1991, 95 (6), 4040–4043.
(13) Russell D. Johnson; Hudgens, J. W. Structural and Thermochemical Properties of
Hydroxymethyl (CH2OH) Radicals and Cations Derived from Observations of B ̃
2
A'(3P) ← X ̃
19

2
A" Electronic Spectra and from Ab Initio Calculations. Journal Physical Chemistry 1996, 100
(51), 19874–19890.
(14) Kamarchik, E.; Rodrigo, C.; Bowman, J. M.; Reisler, H.; Krylov, A. I. Overtone-Induced
Dissociation and Isomerization Dynamics of the Hydroxymethyl Radical (CH2OH and CD2OH).
I. A Theoretical Study. The Journal of Chemical Physics 2012, 136 (8), 084304.
(15) Marenich, A. V.; Boggs, J. E. Structural and Thermochemical Properties of the
Hydroxymethyl (CH2OH) Radical: A High Precision Ab Initio Study. The Journal of Chemical
Physics 2003, 119 (19), 10105–10114.
(16) Walch, S. P. Computed Barrier Heights for H+CH2O↔CH3O↔CH2OH. The Journal of
Chemical Physics 1993, 98 (4), 3076–3077.
(17) Jacox, M. E. The Reaction of Excited Argon Atoms and of F Atoms with Methanol.
Vibrational Spectrum of CH2OH Isolated in Solid Argon. Chemical Physics 1981, 59 (1–2),
213–230.
(18) Ryazanov, M.; Rodrigo, C.; Reisler, H. Overtone-Induced Dissociation and Isomerization
Dynamics of the Hydroxymethyl Radical (CH2OH and CD2OH). II. Velocity Map Imaging
Studies. The Journal of Chemical Physics 2012, 136 (8), 084305.
(19) Wei, J.; Karpichev, B.; Reisler, H. Unimolecular Processes in CH2OH below the
Dissociation Barrier: O–H Stretch Overtone Excitation and Dissociation. The Journal of
Chemical Physics 2006, 125 (3), 034303.
(20) Dulcey, C. S.; Hudgens, J. W. Multiphoton Ionization Spectroscopy and Vibrational
Analysis of a 3p Rydberg State of the Hydroxymethyl Radical. The Journal of Chemical Physics
1986, 84 (10), 5262–5270.
20

(21) Rettrup, S.; Pagsberg, P.; Anastasi, C. Ab Initio Configuration Interaction Study of the
Rydberg States of the Hydroxymethyl Radical CH2OH. Chemical Physics 1988, 122 (1), 45–51.
(22) Bruna, P. J.; Grein, F. Stability, Properties, and Electronic g Tensors of the H2COH
Radical. J. Phys. Chem. A 1998, 102 (18), 3141–3150.
(23) Bruna, P. J.; Grein, F. The Electronic Spectrum of H2COH Revisited. J. Phys. Chem. A
2001, 105 (37), 8599–8603.
(24) Chen, F.; Davidson, E. R. Theoretical Study of the Electronic Spectrum and ESR of the
CH2OH Radical. J. Phys. Chem. A 2001, 105 (18), 4558–4562.
(25) Aristov, V.; Conroy, D.; Reisler, H. Symmetry and Lifetime of the Hydroxymethyl
Radical in the 3p Rydberg State. Chemical Physics Letters 2000, 318 (4–5), 393–401.
(26) Feng, L.; Huang, X.; Reisler, H. Photodissociative Spectroscopy of the Hydroxymethyl
Radical (CH2OH) in the 3s and 3px States. The Journal of Chemical Physics 2002, 117 (10),
4820–4824.
(27) Xie, C.; Malbon, C.; Yarkony, D. R.; Guo, H. Nonadiabatic Photodissociation Dynamics
of the Hydroxymethyl Radical via the 2
2
A(3s) Rydberg State: A Four-Dimensional Quantum
Study. The Journal of Chemical Physics 2017, 146 (22), 224306.
(28) Hoffman, B. C.; Yarkony, D. R. Photodissociation of the Hydroxymethyl Radical. I. The
Role of Conical Intersections in Line Broadening and Decomposition Pathways. J. Chem. Phys.
2002, 116 (19), 8300.
(29) Yarkony, D. R. Statistical and Nonstatistical Nonadiabatic Photodissociation from the
First Excited State of the Hydroxymethyl Radical. The Journal of Chemical Physics 2005, 122
(8), 084316.
21

(30) Malbon, C. L.; Yarkony, D. R. Multistate, Multichannel Coupled Diabatic State
Representations of Adiabatic States Coupled by Conical Intersections. CH2OH
Photodissociation. The Journal of Chemical Physics 2017, 146 (13), 134302.
(31) Conroy, D.; Aristov, V.; Feng, L.; Reisler, H. Predissociation of the Hydroxymethyl
Radical in the 3pz Rydberg State: Formaldehyde + Hydrogen Atom Channel . J. Phys. Chem. A
2000, 104 (45), 10288–10292.
(32) Feng, L.; Reisler, H. Photodissociation of the Hydroxymethyl Radical from the 2
2
A''(3pz
) State: H 2 CO and HCOH Products. J. Phys. Chem. A 2004, 108 (45), 9847–9852.
(33) Feng, L.; Demyanenko, A. V.; Reisler, H. O–D Bond Dissociation from the 3s State of
Deuterated Hydroxymethyl Radical (CH2OD). The Journal of Chemical Physics 2003, 118 (21),
9623–9628.
(34) Feng, L.; Demyanenko, A. V.; Reisler, H. Competitive C–H and O–D Bond Fission
Channels in the UV Photodissociation of the Deuterated Hydroxymethyl Radical CH2OD. The
Journal of Chemical Physics 2004, 120 (14), 6524–6530.
(35) Rodrigo, C. P.; Zhou, C.; Reisler, H. Accessing Multiple Conical Intersections in the 3s
and 3px Photodissociation of the Hydroxymethyl Radical. J. Phys. Chem. A 2013, 117 (46),
12049–12059.
(36) Xie, C.; Malbon, C. L.; Yarkony, D. R.; Xie, D.; Guo, H. Signatures of a Conical
Intersection in Adiabatic Dissociation on the Ground Electronic State. Journal of the American
Chemical Society 2018, 140 (6), 1986–1989.
(37) Ryazanov, M.; Reisler, H. Improved Sliced Velocity Map Imaging Apparatus Optimized
for H Photofragments. The Journal of Chemical Physics 2013, 138 (14), 144201.
22

(38) Townsend, D.; Minitti, M. P.; Suits, A. G. Direct Current Slice Imaging. Review of
Scientific Instruments 2003, 74 (4), 2530–2539.
(39) Toomes, R. L.; Samartzis, P. C.; Rakitzis, T. P.; Kitsopoulos, T. N. Slice Imaging of H-
Atom Photofragments: Effects of the REMPI Detection Process on the Observed Velocity
Distribution. Chemical Physics 2004, 301 (2–3), 209–212.
(40) Jeffries, J. B.; Schulz, C.; Mattison, D. W.; Oehlschlaeger, M. A.; Bessler, W. G.; Lee, T.;
Davidson, D. F.; Hanson, R. K. UV Absorption of CO2 for Temperature Diagnostics of
Hydrocarbon Combustion Applications. Proc. Combust. Inst. 2005, 30 (1), 1591–1599.
(41) Schulz, C.; Jeffries, J. B.; Davidson, D. F.; Koch, J. D.; Wolfrum, J.; Hanson, R. K.
Impact of UV Absorption by CO2 and H2O on No Lif Inhigh-Pressure Combustion Applications.
Proc. Combust. Inst. 2002, 29 (2), 2735–2742.
(42) Settersten, T. B.; Dreizler, A.; Patterson, B. D.; Schrader, P. E.; Farrow, R. L. Photolytic
Interference Affecting Two-Photon Laser-Induced Fluorescence Detection of Atomic Oxygen in
Hydrocarbon Flames. Appl. Phys. B 2003, 76 (4), 479–482.
(43) Anbar, A. D.; Allen, M.; Nair, H. A. Photodissociation in the Atmosphere of Mars: Impact
of High Resolution, Temperature‐dependent CO2 Cross‐section Measurements. J. Geophys. Res.:
Planets 1993, 98 (E6), 10925–10931.
(44) McElroy, M. B.; Dak Sze, N.; Ling Yung, Y. Photochemistry of the Venus Atmosphere. J.
Atmos. Sci. 1973, 30 (7), 1437–1447.
(45) Venot, O.; Fray, N.; Bénilan, Y.; Gazeau, M.-C.; Hébrard, E.; Larcher, G.; Schwell, M.;
Dobrijevic, M.; Selsis, F. High-Temperature Measurements of VUV-Absorption Cross Sections
of CO2 and Their Application to Exoplanets. Astron. Astrophys. 2013, 551, A131.
23

(46) Yung, Y. L.; DeMore, W. B. Photochemistry of Planetary Atmospheres; Oxford
University Press on Demand, 1999.
(47) Chen, Z.; Liu, F.; Jiang, B.; Yang, X.; Parker, D. H. Imaging CO2 Photodissociation at 157
nm: State-to-State Correlations between CO (ν) and O (
3
PJ = 0, 1, 2). J. Phys. Chem. Lett. 2010,
1 (12), 1861–1865.
(48) Gao, H.; Song, Y.; Jackson, W. M.; Ng, C. Y. Communication: State-to-State
Photodissociation Study by the Two-Color VUV-VUV Laser Pump-Probe Time-Slice Velocity-
Map-Imaging-Photoion Method. J. Chem. Phys. 2013, 138 (19), 191102.
(49) Gao, H.; Song, Y.; Yang, L.; Shi, X.; Yin, Q.-Z.; Ng, C. Y.; Jackson, W. M. Branching
Ratio Measurements of the Predissociation of
12
C
16
O by Time-Slice Velocity-Map Ion Imaging
in the Energy Region from 108 000 to 110 500 cm
−1
. J. Chem. Phys. 2012, 137 (3), 034305.
(50) Koshi, M.; Yoshimura, M.; Matsui, H. Photodissociation of O2 and CO2 from
Vibrationally Excited States at High Temperatures. Chem. Phys. Lett. 1991, 176 (6), 519–525.
(51) Koyano, I.; Wauchop, T. S.; Welge, K. H. Relative Efficiencies of O(
1
S) Production from
Photodissociation of CO2 between 1080 and 1160 Å. J. Chem. Phys. 1975, 63 (1), 110–113.
(52) Lu, Z.; Chang, Y. C.; Benitez, Y.; Luo, Z.; Houria, A. B.; Ayari, T.; Al Mogren, M. M.;
Hochlaf, M.; Jackson, W. M.; Ng, C. Y. State-to-State Vacuum Ultraviolet Photodissociation
Study of CO2 on the Formation of State-Correlated CO (X
1
Σ
+
; v) with O(
1
D) and O(
1
S)
Photoproducts at 11.95–12.22 eV. Phys. Chem. Chem. Phys. 2015, 17 (17), 11752–11762.
(53) Lu, Z.; Chang, Y. C.; Gao, H.; Benitez, Y.; Song, Y.; Ng, C. Y.; Jackson, W. M.
Communication: Direct Measurements of Nascent O(
3
P0, 1, 2) Fine-Structure Distributions and
Branching Ratios of Correlated Spin-Orbit Resolved Product Channels CO (Ã
3
Π; v)+ O(
3
P0, 1, 2)
24

and CO (X ̃ 1
Σ
+
; v) + O (
3
P0, 1, 2) in VUV Photodissociation of CO2. J. Chem. Phys. 2014, 140
(23), 231101.
(54) Lu, Z.; Chang, Y. C.; Yin, Q.-Z.; Ng, C. Y.; Jackson, W. M. Evidence for Direct
Molecular Oxygen Production in CO2 Photodissociation. Science 2014, 346 (6205), 61–64.
(55) Miller, R. L.; Kable, S. H.; Houston, P. L.; Burak, I. Product Distributions in the 157 nm
Photodissociation of CO2. J. Chem. Phys. 1992, 96 (1), 332–338.
(56) Pan, Y.; Gao, H.; Yang, L.; Zhou, J.; Ng, C. Y.; Jackson, W. M. Communication: Vacuum
Ultraviolet Laser Photodissociation Studies of Small Molecules by the Vacuum Ultraviolet Laser
Photoionization Time-Sliced Velocity-Mapped Ion Imaging Method. J. Chem. Phys. 2011, 135
(7), 071101.
(57) Slanger, T. G.; Sharpless, R. L.; Black, G. CO2 Photodissociation, 1060–1175 Å. J. Chem.
Phys. 1977, 67 (11), 5317–5323.
(58) Song, Y.; Gao, H.; Chang, Y. C.; Lu, Z.; Ng, C. Y.; Jackson, W. M. Photodissociation of
CO2 between 13.540 eV and 13.678 eV. Phys. Chem. Chem. Phys. 2014, 16 (2), 563–569.
(59) Stolow, A.; Lee, Y. T. Photodissociation Dynamics of CO2 at 157.6 nm by
Photofragment‐translational Spectroscopy. J. Chem. Phys. 1993, 98 (3), 2066–2076.
(60) Zhu, Y.-F.; Gordon, R. J. The Production of O (
3
P) in the 157 nm Photodissociation of
CO2. J. Chem. Phys. 1990, 92 (5), 2897–2901.
(61) Matsumi, Y.; Shafer, N.; Tonokura, K.; Kawasaki, M.; Huang, Y.-L.; Gordon, R. J.
Doppler Profiles and Fine‐structure Branching Ratios of O (
3
Pj) from Photodissociation of
Carbon Dioxide at 157 nm. J. Chem. Phys. 1991, 95 (10), 7311–7316.
25

(62) Schmidt, J. A.; Johnson, M. S.; Schinke, R. Carbon Dioxide Photolysis from 150 to 210
nm: Singlet and Triplet Channel Dynamics, UV-Spectrum, and Isotope Effects. Proc. Natl. Acad.
Sci. 2013, 110 (44), 17691–17696.
(63) Spielfiedel, A.; Feautrier, N.; Cossart‐Magos, C.; Chambaud, G.; Rosmus, P.; Werner, H.-
J.; Botschwina, P. Bent Valence Excited States of CO2. J. Chem. Phys. 1992, 97 (11), 8382–
8388.
(64) Grebenshchikov, S. Y. Photodissociation of Carbon Dioxide in Singlet Valence Electronic
States. II. Five State Absorption Spectrum and Vibronic Assignment. J. Chem. Phys. 2013, 138
(22), 224107.
(65) Parkinson, W. H.; Rufus, J.; Yoshino, K. Absolute Absorption Cross Section
Measurements of CO2 in the Wavelength Region 163–200 nm and the Temperature Dependence.
Chem. Phys. 2003, 290 (2), 251–256.
(66) Yoshino, K.; Esmond, J. R.; Sun, Y.; Parkinson, W. H.; Ito, K.; Matsui, T. Absorption
Cross Section Measurements of Carbon Dioxide in the Wavelength Region 118.7–175.5 nm and
the Temperature Dependence. J. Quant. Spectrosc. Radiat. Transfer 1996, 55 (1), 53–60.
(67) Jensen, R. J.; Guettler, R. D.; Lyman, J. L. The Ultraviolet Absorption Spectrum of Hot
Carbon Dioxide. Chem. Phys. Lett. 1997, 277 (4), 356–360.
(68) Oehlschlaeger, M. A.; Davidson, D. F.; Jeffries, J. B. Temperature Measurement Using
Ultraviolet Laser Absorption of Carbon Dioxide behind Shock Waves. Appl. Opt. 2005, 44 (31),
6599–6605.
(69) Oehlschlaeger, M. A.; Davidson, D. F.; Jeffries, J. B.; Hanson, R. K. Ultraviolet
Absorption Cross-Sections of Hot Carbon Dioxide. Chemical Physics Letters 2004, 399 (4–6),
490–495.
26

(70) Schulz, C.; Koch, J. D.; Davidson, D. F.; Jeffries, J. B.; Hanson, R. K. Ultraviolet
Absorption Spectra of Shock-Heated Carbon Dioxide and Water between 900 and 3050 K.
Chem. Phys. Lett. 2002, 355 (1), 82–88.
(71) Dixon, R. N. The Carbon Monoxide Flame Bands; The Royal Society, 1963; Vol. 275, pp
431–446.
(72) Lee, T.; Bessler, W. G.; Yoo, J.; Schulz, C.; Jeffries, J. B.; Hanson, R. K. Fluorescence
Quantum Yield of Carbon Dioxide for Quantitative UV Laser-Induced Fluorescence in High-
Pressure Flames. Appl. Phys. B 2008, 93 (2–3), 677–685.
(73) Hartinger, K. T.; Nord, S.; Monkhouse, P. B. Temperature- and Pressure-Dependent
Absorption Coefficients for CO2 and O2 at 193 nm. Applied Physics B: Lasers and Optics 2000,
70 (1), 133–137.
(74) Joutsenoja, T.; D’Anna, A.; D’Alessio, A.; Nazzaro, M. I. Ultraviolet Absorption Spectra
of Carbon Dioxide and Oxygen at Elevated Temperatures. Appl. Spectrosc. 2001, 55 (2), 130–
135.
(75) Ibraguimova, L. B.; Minaev, B. F.; Irgibaeva, I. S. Spin-Orbit Coupling and Dissociation
of CO2 Molecules. Opt. Spectrosc. 2014, 117 (5), 695–702.
(76) Hwang, D.-Y.; Mebel, A. M. Ab Initio Study of Spin-Forbidden Unimolecular
Decomposition of Carbon Dioxide. Chem. Phys. 2000, 256 (2), 169–176.
(77) Ibraguimova, L. B.; Minaev, B. F. Analysis of Dissociation—Recombination Processes
for the CO2 Molecule with the Spin—Orbit Coupling Taken into Account. Opt. Spectrosc. 2016,
120 (3), 345–351.
27

(78) Jasper, A. W. Multidimensional Effects in Nonadiabatic Statistical Theories of Spin-
Forbidden Kinetics: A Case Study of
3
O+ CO→ CO2. J. Phys. Chem. A 2015, 119 (28), 7339–
7351.
(79) Jasper, A. W.; Dawes, R. Non-Born–Oppenheimer Molecular Dynamics of the Spin-
Forbidden Reaction O (
3
P)+ CO (X
1
Σ
+
)→ CO2 (X ̃ 1
Σg
+
). J. Chem. Phys. 2013, 139 (15), 154313.
(80) Ma, Y.; Peng, L.; Zhang, H.; Yu, J.-G. The Potential Energy Surfaces of the Ground and
Excited States of Carbon Dioxide Molecule. Russ. J. Phys. Chem. A 2014, 88 (13), 2339–2347.
(81) Zhou, B.; Zhu, C.; Wen, Z.; Jiang, Z.; Yu, J.; Lee, Y.-P.; Lin, S. H. Topology of
Conical/Surface Intersections among Five Low-Lying Electronic States of CO2: Multireference
Configuration Interaction Calculations. J. Chem. Phys. 2013, 139 (15), 154302.
(82) McGlynn, S. P.; Rabalais, J. W.; McDonald, J. R.; Scherr, V. M. Electronic Spectoscopy
of Isoelectronic Molecules. II. Linear Triatomic Groupings Containing Sixteen Valence
Electrons. Chem. Rev. 1971, 71 (1), 73–108.
(83) Berghout, H. L.; Crim, F. F.; Zyrianov, M.; Reisler, H. The Electronic Origin and
Vibrational Levels of the First Excited Singlet State of Isocyanic Acid (HNCO). J. Chem. Phys.
2000, 112 (15), 6678–6688.
(84) Droz-Georget, T.; Zyrianov, M.; Reisler, H.; Chandler, D. W. Correlated Distributions in
the Photodissociation of HNCO to NH (X
3
Σ
−
, a
1
Δ)+ CO (X
1
Σ
+
) near the Barrier on S1. Chem.
Phys. Lett. 1997, 276 (5), 316–324.
(85) Droz‐Georget, T. H.; Zyrianov, M.; Sanov, A.; Reisler, H. Photodissociation of HNCO:
Three Competing Pathways. Berichte der Bunsengesellschaft für physikalische Chemie 1997,
101 (3), 469–477.
28

(86) Zyrianov, M.; Droz-Georget, T.; Reisler, H. Fragment Recoil Anisotropies in the
Photoinitiated Decomposition of HNCO. J. Chem. Phys. 1999, 110 (4).
(87) Zyrianov, M.; Droz‐Georget, T.; Sanov, A.; Reisler, H. Competitive Photodissociation
Channels in Jet‐cooled HNCO: Thermochemistry and Near‐threshold Predissociation. J. Chem.
Phys. 1996, 105 (18), 8111–8116.
(88) Brown, S. S.; Berghout, H. L.; Crim, F. F. Vibrational State Controlled Bond Cleavage in
the Photodissociation of Isocyanic Acid (HNCO). J. Chem. Phys. 1995, 102 (21), 8440–8447.
(89) Brown, S. S.; Metz, R. B.; Berghout, H. L.; Crim, F. F. Vibrationally Mediated
Photodissociation of Isocyanic Acid (HNCO): Preferential N–H Bond Fission by Excitation of
the Reaction Coordinate. J. Chem. Phys. 1996, 105 (15), 6293–6303.
(90) Carlton, A. G.; Turpin, B. J.; Lim, H.-J.; Altieri, K. E.; Seitzinger, S. Link between
Isoprene and Secondary Organic Aerosol (SOA): Pyruvic Acid Oxidation Yields Low Volatility
Organic Acids in Clouds. Geophys. Res. Lett. 2006, 33 (6), L06822.
(91) Grosjean, D. Atmospheric Reactions of Pyruvic Acid. Atmospheric Environment (1967)
1983, 17 (11), 2379–2382.
(92) Grosjean, D.; Williams, E. L.; Grosjean, E. Atmospheric Chemistry of Isoprene and of Its
Carbonyl Products. Environ. Sci. Technol. 1993, 27 (5), 830–840.
(93) Nguyen, T. B.; Bateman, A. P.; Bones, D. L.; Nizkorodov, S. A.; Laskin, J.; Laskin, A.
High-Resolution Mass Spectrometry Analysis of Secondary Organic Aerosol Generated by
Ozonolysis of Isoprene. Atmospheric Environment 2010, 44 (8), 1032–1042.
(94) Bao, L.; Matsumoto, M.; Kubota, T.; Sekiguchi, K.; Wang, Q.; Sakamoto, K. Gas/Particle
Partitioning of Low-Molecular-Weight Dicarboxylic Acids at a Suburban Site in Saitama, Japan.
Atmospheric Environment 2012, 47, 546–553.
29

(95) Veres, P.; Roberts, J. M.; Burling, I. R.; Warneke, C.; de Gouw, J.; Yokelson, R. J.
Measurements of Gas-Phase Inorganic and Organic Acids from Biomass Fires by Negative-Ion
Proton-Transfer Chemical-Ionization Mass Spectrometry. J. Geophys. Res. 2010, 115 (D23),
D23302.
(96) Kawamura, K.; Tachibana, E.; Okuzawa, K.; Aggarwal, S. G.; Kanaya, Y.; Wang, Z. F.
High Abundances of Water-Soluble Dicarboxylic Acids, Ketocarboxylic Acids and α-
Dicarbonyls in the Mountaintop Aerosols over the North China Plain during Wheat Burning
Season. Atmos. Chem. Phys. 2013, 13 (16), 8285–8302.
(97) Veres, P. R.; Roberts, J. M.; Cochran, A. K.; Gilman, J. B.; Kuster, W. C.; Holloway, J.
S.; Graus, M.; Flynn, J.; Lefer, B.; Warneke, C.; et al. Evidence of Rapid Production of Organic
Acids in an Urban Air Mass: PHOTOCHEMICAL PRODUCTION OF ORGANIC ACID.
Geophys. Res. Lett. 2011, 38 (17), L17807.
(98) Mattila, J. M.; Brophy, P.; Kirkland, J.; Hall, S.; Ullmann, K.; Fischer, E. V.; Brown, S.;
McDuffie, E.; Tevlin, A.; Farmer, D. K. Tropospheric Sources and Sinks of Gas-Phase Acids in
the Colorado Front Range. Atmos. Chem. Phys. 2018, 18 (16), 12315–12327.
(99) Schaefer, T.; Schindelka, J.; Hoffmann, D.; Herrmann, H. Laboratory Kinetic and
Mechanistic Studies on the OH-Initiated Oxidation of Acetone in Aqueous Solution. J. Phys.
Chem. A 2012, 116 (24), 6317–6326.
(100) Mellouki, A.; Mu, Y. On the Atmospheric Degradation of Pyruvic Acid in the Gas Phase.
Journal of Photochemistry and Photobiology A: Chemistry 2003, 157 (2–3), 295–300.
(101) Epstein, S. A.; Nizkorodov, S. A. A Comparison of the Chemical Sinks of Atmospheric
Organics in the Gas and Aqueous Phase. Atmos. Chem. Phys. 2012, 12 (17), 8205–8222.
30

(102) Reed Harris, A. E.; Ervens, B.; Shoemaker, R. K.; Kroll, J. A.; Rapf, R. J.; Griffith, E. C.;
Monod, A.; Vaida, V. Photochemical Kinetics of Pyruvic Acid in Aqueous Solution. J. Phys.
Chem. A 2014, 118 (37), 8505–8516.
(103) Reed Harris, A. E.; Cazaunau, M.; Gratien, A.; Pangui, E.; Doussin, J.-F.; Vaida, V.
Atmospheric Simulation Chamber Studies of the Gas-Phase Photolysis of Pyruvic Acid. The
Journal of Physical Chemistry A 2017, 121 (44), 8348–8358.
(104) Reed Harris, A. E.; Doussin, J.-F.; Carpenter, B. K.; Vaida, V. Gas-Phase Photolysis of
Pyruvic Acid: The Effect of Pressure on Reaction Rates and Products. The Journal of Physical
Chemistry A 2016, 120 (51), 10123–10133.
(105) Reed Harris, A. E.; Pajunoja, A.; Cazaunau, M.; Gratien, A.; Pangui, E.; Monod, A.;
Griffith, E. C.; Virtanen, A.; Doussin, J.-F.; Vaida, V. Multiphase Photochemistry of Pyruvic
Acid under Atmospheric Conditions. The Journal of Physical Chemistry A 2017, 121 (18), 3327–
3339.
(106) Vesley, G. F.; Leermakers, P. A. The Photochemistry of α-Keto Acids and α-Keto Esters.
III. Photolysis of Pyruvic Acid in the Vapor Phase. The Journal of Physical Chemistry 1964, 68
(8), 2364–2366.
(107) Berges, M. G. M.; Warneck, P. Product Quantum Yields for the 350 nm
Photodecomposition of Pyruvic Acid in Air. Berichte der Bunsengesellschaft für physikalische
Chemie 1992, 96 (3), 413–416.
(108) Yamamoto, S.; Back, R. The Photolysis and Thermal Decomposition of Pyruvic Acid in
the Gas Phase. Canadian journal of chemistry 1985, 63 (2), 549–554.
31

(109) Horowitz, A.; Meller, R.; Moortgat, G. K. The UV–VIS Absorption Cross Sections of the
α-Dicarbonyl Compounds: Pyruvic Acid, Biacetyl and Glyoxal. Journal of Photochemistry and
Photobiology A: Chemistry 2001, 146 (1–2), 19–27.
(110) Chang, X.-P.; Fang, Q.; Cui, G. Mechanistic Photodecarboxylation of Pyruvic Acid:
Excited-State Proton Transfer and Three-State Intersection. The Journal of chemical physics
2014, 141 (15), 154311.

32

Chapter 2
Experimental Methods
The experimental setup used for the present photodissociation studies consists of several
complex components and apparatus which are shown schematically in Figure 2.1. Different parts
of the experimental setup have been built by a generation of previous group members and the
detailed design philosophy and performance of each of these experimental setups and techniques
can be found elsewhere.
1–4
While the specific experimental setups used in different
photodissociation studies are discussed in the pertinent chapters later, a brief overview of the
relevant components is presented below. The different experimental parameters and optimized
conditions that are used for the present photodissociation studies are also discussed.
Figure 2.1: Schematic of the general experimental setup used in the present studies.
Pulsed Nozzle
Time-of-Flight mass spectrometer
MCP/
Phosphor
detector
CCD Camera
Probe laser
Pump laser
Oscilloscope
Vacuum chamber
Time delay generator
LabView Control
Data
33

2.1 Molecular Beam Apparatus
2.1.1 Vacuum Chamber
In order to create cold molecular beams, the experiments are performed inside a high
vacuum chamber. The design of the vacuum chamber is discussed in detail by David Conroy and
only described briefly here.
2
The vacuum chamber is divided into two regions: source and
detection. The two regions are separated by a flange onto which a molecular beam skimmer
(Beam Dynamics, 1.51 mm) is mounted (Figure 2.1). A base pressure of 2 x 10
-7
torr is
maintained in the source region by using a turbomolecular pump (Leybold TMP1000C). The
detection region is evacuated by a smaller turbopump and the base pressure of 2 x 10
-8
torr is
maintained. Both the source and detection turbopumps are further connected to oil-based
roughing pumps.
The major source of complication in the vacuum system arises from the deposition of less
volatile organic reactants (for example pyruvic acid) on the chamber walls causing excessive
background signal. Another source of contamination can arise from oil back-streaming from the
roughing pump into the chamber resulting an increase of base pressure. In these scenarios, an
extensive differential heating of the chamber is recommended. Usually the source and detection
regions are baked at 80
0
and 90
0
C, whereas the detector is heated at 100
0
C. This differential
heating prohibits the volatile contaminations to flow towards the detector.
2.1.2 Pulsed Nozzle
The sample mixture is introduced into the vacuum chamber using a pulsed nozzle. The
basic design of the nozzle is described in detail by David Conroy.
2
The sample mixture (usually
using He or Ar as the carrier gas) is flown into the reactant chamber of the nozzle head using a
34

1/4 '' stainless steel or Teflon tubing. The reactant chamber is attached to a piezo housing, which
consists of a piezoelectric actuator and a plunger. The plunger has a Kalrez O-ring at the tip
which seals the nozzle orifice towards the vacuum. A quartz or silicon carbide (SiC) tube is
attached to the nozzle orifice to produce appropriate precursors for each experiment, as discussed
in later chapters. When applying an appropriate voltage, the piezoelectric actuator moves the
plunger to open the nozzle orifice for 100-300 μs time duration allowing the reactant gas to flow
inside the source region through the quartz or SiC tube. The supersonic expansion of the reaction
mixture from high pressure at the reactant chamber to the low pressure source region through the
small nozzle orifice leads to a significant rotational cooling of the molecules. The supersonically
expanded molecules then pass through the skimmer where they form a molecular beam. The
extent of rotational cooling in the molecular beam is quantified by recording 2+1 CO REMPI
using X ← B transition around 230 nm. The distance between the tip of the tube and the
skimmer is optimized at 2-3 cm where the molecular rotational temperature of 10-15 K is
achieved using He as the carrier gas.
2.2 Time of Flight (TOF) mass spectrometer
After passing through the skimmer, the molecular beam enters the detection chamber
where the reactant molecules are excited by the pump laser and the products of the
photodissociation are ionized by suitable probe laser radiation. The ionized products are then
detected by a Wiley-McLaren TOF mass spectrometer mounted inside the vacuum chamber
(Figure 2.1). The basic operation of the Wiley-McLaren TOF mass spectrometer requires a two-
stage ion acceleration region (repeller, extractor, and ground), a fixed-length field-free drift
region, and a multichannel plate detector. The interaction region lies approximately in the middle
of the repeller and extractor. The spatial focusing of the ions is determined by the repeller and
35

extractor voltage ratio. An excellent intuitive description of the two-stage TOF mass
spectrometer can be found in the original paper by Wiley and McLaren.
5
The current Wiley-
McLaren TOF mass spectrometer, which is mounted inside the vacuum chamber is modified
from the traditional 3 electrodes arrangement originally designed by David Conroy.
2
The current
setup built by Mikhail Ryazanov contains 14 electrodes where the repeller plate (L0) is fixed but
the accelerator (L1) and the ground electrodes can be varied.
4,6
An additional electrostatic lens
(Einzel lens, L) is mounted in the drift region. The Einzel lens and the flexibility of choosing the
different accelerator and the ground electrodes’ voltages with respect to the fixed repeller
position allow us to capture ions with a large range of kinetic energies (KEs). This aspect is
important in implementing and optimizing the sliced velocity map imaging arrangement for H
fragments, as discussed in the next section.
2.3 Sliced Velocity Map Imaging (SVMI)
The sliced velocity map imaging (SVMI) is a widely used technique that provides
correlated product state distributions in molecular photodissociation processes by recording
state-selected velocities of different products. SVMI was first developed by Gebhardt et al. as an
alternative method of ion imaging to overcome the technical drawbacks of the conventional
velocity map imaging (VMI).
7
Later, an improved version of SVMI technique based on
electrostatic ion optics was introduced resulting high sensitivity and resolution.
8,9

The principle of SVMI is illustrated in Figure 2.2 for photodissociation of an ABCD
molecule with internal energy E that dissociates to AB + CD. For the sake of simplicity, we
assume that the AB product is formed in a particular rovibrational state, whereas the CD product
36

is formed in two different rovibrational states, CD and CD*, respectively. As shown in Figure
2.2, conservation of energy leads to:
E = ∆Hrxn + Eavail (2.1)
where ∆Hrxn is the heat of reaction for ABCD → AB + CD fragmentation, Eavail is the energy
available to the AB and CD fragments following the dissociation. Eavail will distribute among all
degrees of freedoms of the AB and CD products as follows:
Eavail = Ekin (AB + BC) + Eint (AB) + Eint (BC) (2.2)
Where Ekin and Eint represent the kinetic and the internal (rovibrational) energies of the products,
respectively.
37

Figure 2.2: (A) The 3D velocity distributions of AB correlated with ABCD → AB + CD (red)
and ABCD → AB + CD* (blue) channels. (B) The SVMI setup is illustrated in three-electrode
configuration following the dissociation of ABCD by the pump laser. The AB product in a
particular quantum state is selectively ionized by the appropriate probe laser wavelength
radiation. The 3D velocity distribution of ionized AB product is stretched temporally using
suitable repeller and extractor voltages, and a thin slice around the center of the distributions (∆z
≈ 0) along the x-y plane is recorded by fast gating the detector.
38

If the AB product is detected in a particular quantum state, then Eint(AB) is known.
Therefore, equation 2.2 contains two unknown parameters, namely, Ekin (AB+BC) and Eint(BC)
for a given Eint (AB) and Eavail. In this scenario, the ABCD → AB + CD channel in Figure 2.2
corresponds to the maximum kinetic energy release (KER). If BC is formed in excited state, the
KER of the fragments decreases (ABCD → AB + CD* in Figure 2.2 ). Therefore, the
experimental measurement of KER of the fragments can provide the quantum state distributions
of the BC fragment correlated with the monitored quantum state of AB. SVMI enables us to
record the velocity distributions (hence the KE) of the products. Photodissociation of molecules
and radicals often leads to a 3D spherical products’ distribution (Newton sphere, Figure 2.2)
where the radius of the sphere is proportional to the velocity of the product. Therefore, it is
evident that the dissociation channel with higher KER will represent a larger sphere (red sphere
for ABCD → AB + CD) and that with lower KER will have a smaller sphere (blue sphere for
ABCD → AB + CD*). Since the Newton sphere possesses a cylindrical symmetry along the
axis parallel to the laser polarization (shown in Figure 2.2), a thin slice around the center of the
Newton sphere along the symmetry axis would retain all the information about the original 3D
distributions.
The slicing variant gives a huge advantage over traditional VMI technique where the
entire 3D distribution is recorded on the detector.
10
In the latter method, concentric Newton
spheres with different velocities overlap each other and the resolution along the flight axis is lost.
Image reconstruction (inverse Abel transform) is needed to retrieve the original 3D velocity
distribution. This procedure often results in artificial noise in the reconstructed image mainly in
the lower velocity regions.
11
In SVMI, the two initial velocity components (vx and v y in Figure
2.2) that lie in the slicing place are mapped to arrival positions at the detector. The third
39

component (vz), perpendicular to the detector plane (or parallel to the TOF axis), is mapped as
the arrival time. The slicing along the x-y plane of the Newton sphere is performed by fast gating
the detector using an appropriate pulser, producing a 2D image on the detector, as shown in
Figure 2.2.
It should be noted that the image represents the position of the AB product in the pixel
space. Radial distribution (I(R)) plots of product intensities are obtained by integrating the signal
at each radius in pixel space over 360°.
𝐼 (𝑅 )= ∫ 𝐼 (𝑅 ,𝜃 )𝑅 2
|sin𝜃 |𝑑𝜃 2𝜋 0
2.3
Due to linear 𝑣 → 𝑅 mapping, speed distributions can be obtained as,

𝐼 (𝑣 )~𝐼 (𝑅 ) 2.4
The speed (I(v)) can be subsequently converted to the KE distributions (I(K)) as follows,
𝐼 (𝐾 )𝑑𝐾 = 𝐼 (𝑣 )𝑑𝑣 and 𝑑𝑘 = 𝑚𝑣𝑑𝑣 ,
𝐼 (𝐾 )=
1
𝑣 𝐼 (𝑣 ), 𝑣 =
√
2𝐾 𝑚 ⁄ 2.5
Alternatively,
𝐼 (𝐾 )=
1
𝑅 𝐼 (𝑅 )=
1
𝑅 ∫ 𝐼 (𝑅 ,𝜃 )𝑅 2
|sin𝜃 |𝑑𝜃 2𝜋 0
2.6
Once KE of AB is known, KE of the co-fragment CD can be calculated using
conservation of linear momentum to obtain the center-of-mass KER, Ekin (AB+BC) in the
process. The larger (red) and smaller (blue) KE rings represent, respectively the CD and CD*
products, correlated with the monitored quantum state of the AB.
40

Achieving optimal velocity resolution in the SVMI requires temporal stretching of the
Newton sphere. This is usually done by using suitable voltages in the repeller and extractor
electrodes and introducing an appropriate field-free region as illustrated in Figure 2.2. Ideally,
the thin slicing (≈ ∆z = 0) should be done with the narrowest gating pulse available. However,
the signal-to-noise ratio decreases as the slice thickness is lowered and the gating pulse cannot be
arbitrarily small. For the temporal stretching of a few tens of nanoseconds, the optimum central
slice thickness should be a tenth of the stretching size to maintain a good signal-to-noise ratio as
well as the resolution. This can easily be achieved using 20–30 ns gating pulses from commercial
pulse generators for all the ionic fragments except H.
8,12–14

The slicing of H remains difficult for several reasons. Since H is the lightest product in
any photodissociation process, it carries away almost all the KE. Due to its high velocity, the
slicing of H atoms needs a faster gating pulse. Recalling that the TOF stretching is proportional
to the square root of the mass of the fragment, the TOF stretch for H will be much smaller
compared to that of heavier fragments and cannot be sliced sufficiently using commercially
available pulsers.
6,8
To overcome this difficulty, an improved SVMI setup optimized for H
fragment was developed in our group by previous graduate student Mikhail Ryazanov.
4,6
In
addition, a high voltage (~ 2 kV) 5 ns pulser is also designed to effectively slice the H ion cloud
which has TOF stretch as small as 50 ns.
4,6

Another general requirement of SVMI setup is to accommodate images with large KE
range within the finite detector size while keeping the optimum TOF stretch. This poses a
complication in design since the TOF stretch is proportional to √𝐾 0
/𝐸 , whereas the image size is
proportional to √𝐾 0
/E (K0 is the initial KE of the product, E is the electric field strength).
41

Therefore, while one needs to increase the voltage to fit the higher KEs on the detector, it can
make the TOF stretching too small for effective slicing. The situation is equally bad for the lower
KEs where TOF stretching is so large that the thin slice results in poor signal-to-noise ratio.
While this is a general aspect of the fixed electric field configuration of TOFMS combined with
SVMI, the situation is even worse for H fragments which are already difficult to slice. Since the
KE of H fragments can range from a fraction of an electronvolt to a few electronvolts, the
optimized slicing conditions and the appropriate image size at the detector for all the KE region
cannot be achieved simultaneously in a fixed field geometry. Therefore, an additional
electrostatic lens was introduced in the drift region of our SVMI setup to control image
magnification without altering TOF stretching. By choosing the appropriate repeller, extractor
and the lens voltages, it is possible to obtain effective slicing for any imaged KEs ranging from
0.2 to 6 eV.
For a particular ion optics voltage setting, when the products span a broad range of KEs,
the effective slice thickness varies within the image. Since the TOF stretching depends on the KE
of the species, different optimum slice thicknesses are required for H fragments with different
KEs. The slice thickness (1/10 of the TOF stretch) optimized for higher KEs (larger TOF stretch)
is often too thick for slicing the lower KE (smaller temporal stretch) regions. On the other hand,
optimized slicing for the lower KEs corresponds to excessively thin slice for the higher KEs
leading to poor signal-to-noise ratio. For example, the effective slice thickness increases from
9.6% to 25% as the H-fragment KE decreases from 1.0 to 0.15 eV, which corresponds to a
change in KE resolution, ΔKE/KE, from <2% to up to ∼12% for the thick slices. In this scenario,
the fragments with lower KEs are acquired more compared to that with higher KEs and some
correction must be included to normalize the different KE regions before converting the KE
42

distributions into the population. Nevertheless, the lower KE resolution at the lowest KEs does
not hinder the determination of product channel onsets and in fact enables detection of low-KE
minor products in the presence of other predominant product channels. It should be noted that the
thick slices give the low KE features more The variable accelerator region with multiple
electrodes also allows us to zoom into low KE regions to obtain thin slices for better KE
resolution.
As the KEs of the fragments become higher, the higher lens voltage is needed to fit the
image in the ∅ 40 mm detector. A practical problem of using higher lens voltage is the increased
production of stray electrons in the chamber probably due to field emission or arcing. In the
present experiments, the maximum lens voltage of -2200-2500 V can be used without causing
serious disruption from background electrons. However, when the signal strength is very strong,
a much higher lens voltage (~ 4200 V) can be used in spite of the background electron signal
partially reduces the actual signal. A broad range of ion optics parameters optimized for different
KEs and TOF stretching are listed by Mikhail Ryazanov.
4

Although a strong signal can result in faster image acquisition, one must be careful about
the effects of detector saturation and space charge broadening, which lead to deterioration of
image resolution. The overwhelming signal might also decrease the detector lifetime. A good
indication of the detector saturation is the appearance of bright red or blue dots on the camera
during image acquisition. This situation should be avoided by either lowering the MCP voltage
or the probe laser intensity.

43

2.4 Tripling cell and Generation of VUV
VUV radiation of wavelengths as short as 114 nm is generated using Kr:Ar mixture in a
gas cell where the incident UV or near-visible pump wavelength is tripled in a process called
third harmonic generation (THG). The details of the THG generation of VUV radiation in an
inert gas mixture has been discussed in detail before.
15,16
The tripling efficiency depends on
phase matching between the fundamental and the generated third harmonic radiation. The phase
matching condition is sensitive to the relative pressures of Ar and Kr and the total pressure. As
discussed in subsequent chapters, the photodissociation studies reported here require VUV probe
radiation at wavelengths ~121 and ~110 nm to detect H/D and CO2 products respectively.
Figure 2.3: Ar:Kr gas cell and associated optical arrangement used for VUV generation.
44

Mahon et al. discussed the regions of VUV that can be generated via Ar:Kr gas mixtures
using the refractive indices of Ar and Kr.
15
They found that the phase matching condition can be
achieved in the above region by varying the relative pressures of Ar and Kr and the total pressure
in the gas cell. In general, tripling efficiency increases with increasing total pressure. However, it
has been observed that total pressure above 900 Torr can cause optical breakdown of the gas
mixture producing an avalanche of electrons that disrupt the VUV production in the present
study. Therefore, the maximum total pressure used in the present study is 900 Torr. A schematic
of the tripling cell is shown in Figure 2.3.
VUV radiation at 121.5 nm is generated by focusing ~ 365 nm laser radiation in the
tripling cell with a 25 cm lens. The produced 121.5 nm and the residual 365 nm light beams are
focused inside the vacuum chamber using a MgF2 lens (7.5 cm f.l.). The Ar:Kr ratio in the
mixture is optimized by maximizing the H/D signal intensities at the peak of the H/D Lyman-α
transition. Optimal 200 and 590 torr pressures of Kr and Ar, respectively, are used in the present
study to generate the 121.6 nm radiation. A range of VUV radiation below 120 nm is generated
using varying incident UV radiation, the relative ratio of the Ar:Kr mixture and the total pressure
of the cell.
45

Figure 2.4: Electron signal recorded in the VUV range of 114-115 nm. The production of VUV
radiation is observed when the phase matching condition is achieved by varying the Ar:Kr
pressure ratio.
The generated VUV radiation is monitored by recording the stray electron signal in the
chamber. Figure 2.4 shows the generation of VUV radiation at different Ar:Kr pressure ratios as
a function of incident UV wavelength where the phase matching condition is achieved. As
shown in Figure 2.4, 345-nm radiation can be tripled using 550 Torr of Ar and 200 Torr of Kr to
produce VUV ~ 115 nm. Wavelengths longer and shorter than the 115 nm can be generated by
increasing or decreasing the Ar and/or Kr pressures (which in turn changes the Ar:Kr ratio).
Figure 2.4 shows that the VUV profile shifts towards shorter wavelengths with decreasing Ar:Kr
ratio. VUV wavelengths down to 114 nm were generated in the present study. The sharp drop of
114.0 114.4 114.8 115.2
Ar:Kr (torr)
400:200
500:200
550:200
electron signal (arb. unit)
VUV wavelength / nm
46

electron signal at 114 nm can be either due to inefficiency of VUV production because of
decreasing total pressure and/or poor transmission of the MgF2 lens at wavelength below 115
nm.
2.5 Pyrolysis Nozzle
A heated pulsed nozzle is used to produce vibrationally hot CO2. The basic design of the
heated nozzle was originally developed by Kohn et al. for pyrolysis purposes and later modified
by several other groups.
17–21
The pyrolysis nozzle used in the present studies is adopted from
Barney Ellison’s group at the University of Colorado, Boulder.
19
The design philosophy and the
usage of the pulsed pyrolysis source described elsewhere
22
and only the relevant optimizations
are discussed here. The pyrolysis nozzle is built by modifying an existing pulsed nozzle, which is
described in Section 2.1.2. While the basic architecture of the nozzle remains the same before,
the quartz tube was replaced by a SiC tube. Two copper electrodes are attached to a small part
(1-2 cm) at the tip of the SiC tube via graphite split disks and the tube is resistively heated by a
Variac. The voltage drops and current across the electrodes at a given Variac power are
measured using an ammeter and a voltmeter, respectively. Since SiC shows a negative electrical
resistance, three 60 W incandescent bulbs are connected in series in order to maintain a constant
current throughout the experiment.
22

Due to the unavailability of an optical pyrometer to measure the temperature of the SiC
wall, the temperature of the SiC tube is estimated by the color of its glow
23
. Assuming that the
hot SiC tube is a blackbody radiator, red, orange, and yellow colors are correlated (according to
the International Commission on Illumination (CIE) chromaticity diagram
24
) with temperatures
of 800-1000, 1200-1400, and 1600-1800 K, respectively. The corresponding temperatures are
47

referred as low (red), medium (orange) and high (yellow) in this work. As will be discussed in
Chapter 4, each temperature range corresponds to a specific value of current in our heating
arrangement and leads to reproducible KE distributions of CO and O products generated via
photodissociation of hot CO2. Therefore, the measured current can be a good indicator of the
temperature of the SiC tube.
The major challenge of working with the hot nozzle is the substantial heat transfer from
SiC tube to the piezo housing. When the piezo head becomes significantly hot, the working
pressure in the source chamber increases rapidly. This puts a limit on the duration of the
experiment. Then the heating needs to stop, and the nozzle cooled down to room temperature
before the experiment can be resumed. To reduce the heating, the piezo head of the nozzle is
connected to external water and air lines. The length of the SiC tube is also optimized to make
the heated length as far from nozzle head as possible. However, the length of the tube cannot be
made arbitrarily long because the tip of the tube and the skimmer orifice must be 2-3 cm apart to
achieve maximum cooling in the supersonic expansion. After several trials, the length of the tube
is optimized to 3.5-3.8 cm under the current architecture of the nozzle and the dimension of the
chamber. The heating length of the tube was also varied to optimize two compromising factors.
While shorter heating length reduces the heat transfer to the nozzle head, it also diminishes the
signal intensity originated from hot molecule. The heating length is optimized to ~ 1.5 cm after
several trials. With these optimized tube and heating lengths, the nozzle can be operated for up to
one hour at the highest heating condition (~ 1800 K).

48

References
(1) Aristov, V. Spectroscopy, Unimolecular and Bimolecular Reactions of Methyl and
Hydroxymethyl Radicals. Ph.D. dissertation, University of Southern California, 2000.
(2) Conroy, D. Rydberg State of an Open Shell Species: Characterization and Photophysics of
the 3pz State of CH2OH. Ph.D. dissertation, University of Southern California, 2000.
(3) Feng, L. Spectroscopy and Photodissociation Dynamics of the Hydroxymethyl Radical
(CH2OH). Ph.D. dissertation, University of Southern California, 2004.
(4) Ryazanov, M. Development and Implementation of Methods for Sliced Velocity Map
Imaging. Studies of Overtone-Induced Dissociation and Isomerization Dynamics of
Hydroxymethyl Radical (CH2OH and CD2OH). Ph.D. dissertation, University of Southern
California, 2012.
(5) Wiley, W.; McLaren, I. H. Time‐of‐flight Mass Spectrometer with Improved Resolution.
Rev. Sci. Instrum. 1955, 26 (12), 1150–1157.
(6) Ryazanov, M.; Reisler, H. Improved Sliced Velocity Map Imaging Apparatus Optimized
for H Photofragments. J. Chem. Phys. 2013, 138 (14), 144201.
(7) Gebhardt, C. R.; Rakitzis, T. P.; Samartzis, P. C.; Ladopoulos, V.; Kitsopoulos, T. N. Slice
Imaging: A New Approach to Ion Imaging and Velocity Mapping. Rev. Sci. Instrum. 2001, 72
(10), 3848–3853.
(8) Townsend, D.; Minitti, M. P.; Suits, A. G. Direct Current Slice Imaging. Rev. Sci. Instrum.
2003, 74 (4), 2530–2539.
(9) Lin, J. J.; Zhou, J.; Shiu, W.; Liu, K. Application of Time-Sliced Ion Velocity Imaging to
Crossed Molecular Beam Experiments. Rev. Sci. Instrum. 2003, 74 (4), 2495–2500.
49

(10) Eppink, A. T. J. B.; Parker, D. H. Velocity Map Imaging of Ions and Electrons Using
Electrostatic Lenses: Application in Photoelectron and Photofragment Ion Imaging of Molecular
Oxygen. Rev. Sci. Instrum. 1997, 68 (9), 3477–3484.
(11) Imaging in Molecular Dynamics: Technology and Applications: A User’s Guide; Whitaker,
B. J., Ed.; Cambridge University Press: Cambridge ; New York, 2003.
(12) Li, W.; Chambreau, S. D.; Lahankar, S. A.; Suits, A. G. Megapixel Ion Imaging with
Standard Video. Rev. Sci. Instrum. 2005, 76 (6), 063106.
(13) Toomes, R. L.; Samartzis, P. C.; Rakitzis, T. P.; Kitsopoulos, T. N. Slice Imaging of H-
Atom Photofragments: Effects of the REMPI Detection Process on the Observed Velocity
Distribution. Chem. Phys. 2004, 301 (2–3), 209–212.
(14) Papadakis, V.; Kitsopoulos, T. N. Slice Imaging and Velocity Mapping Using a Single
Field. Rev. Sci. Instrum. 2006, 77 (8), 083101.
(15) Mahon, R.; J. McIlrath, T.; P. Myerscough, V.; W. Koopman, D. Third-Harmonic
Generation in Argon, Krypton, and Xenon: Bandwidth Limitations in the Vicinity of Lyman-α.
Quantum Electron. IEEE J. Of 1979, 15, 444–451.
(16) Delone, N. B.; Kraĭnov, V. P. Fundamentals of Nonlinear Optics of Atomic Gasses; Wiley
series in pure and applied optics; Wiley: New York, 1988.
(17) Kohn, D. W.; Clauberg, H.; Chen, P. Flash Pyrolysis Nozzle for Generation of Radicals in a
Supersonic Jet Expansion. Rev. Sci. Instrum. 1992, 63 (8), 4003–4005.
(18) Blush, J. A.; Clauberg, H.; Kohn, D. W.; Minsek, D. W.; Zhang, X.; Chen, P.
Photoionization Mass and Photoelectron Spectroscopy of Radicals, Carbenes, and Biradicals.
Acc. Chem. Res. 1992, 25 (9), 385–392.
50

(19) Zhang, X.; Friderichsen, A. V.; Nandi, S.; Ellison, G. B.; David, D. E.; McKinnon, J. T.;
Lindeman, T. G.; Dayton, D. C.; Nimlos, M. R. Intense, Hyperthermal Source of Organic
Radicals for Matrix-Isolation Spectroscopy. Rev. Sci. Instrum. 2003, 74 (6), 3077–3086.
(20) Rohrs, H. W.; Wickham‐Jones, C. T.; Ellison, G. B.; Berry, D.; Argrow, B. M. Fourier
Transform Infrared Absorption Spectroscopy of Jet‐cooled Radicals. Rev. Sci. Instrum. 1995, 66
(3), 2430–2441.
(21) Guan, Q.; Urness, K. N.; Ormond, T. K.; David, D. E.; Barney Ellison, G.; Daily, J. W. The
Properties of a Micro-Reactor for the Study of the Unimolecular Decomposition of Large
Molecules. Int. Rev. Phys. Chem. 2014, 33 (4), 447–487.
(22) Potter, A. Ion Imaging Studies of the Spectroscopy and Photodissociation Dynamics of
Chloromethyl Radical and Nitric Oxide Dimer. Ph.D. dissertation, University of Southern
California, 2005.
(23) Billmeyer, F. W. Color Science: Concepts and Methods, Quantitative Data and Formulae,
2nd Ed., by Gunter Wyszecki and W. S. Stiles, John Wiley and Sons, New York, 1982, 950 Pp.
Color Res. Appl. 1983, 8 (4), 262–263.
(24) Fortner, B.; Meyer, T. E. Number by Colors A Guide to Using Color to Understand
Technical Data; 1997.

51

Chapter 3
Photodissociation of the Hydroxymethyl
Radical from the 3p
x
and 3p
z
Rydberg States
3.1 Introduction
As discussed in Section 1.1.1, the photodissociation of CH2OH upon excitation to the 3s
state and the origin band of the 3px states in the energy range 27,400 - 35,053 cm
-1
were studied
previously using core-sampling TOF and SVMI techniques.
1–4
In this chapter, the
photodissociation of CH2OH is investigated in the energy range 35,000 – 41,000 cm
-1
covering
the 3px and 3pz states where high vibrational levels and excited electronic states of the products
become energetically accessible. The study is focused on excitation energies near the energy
onsets of (i)
3
H2CO (𝑎̃
3
A2) and
1
H2CO (𝐴 ̃
1
A2); (ii)
3
HCOD (𝑎̃
3
A); and (iii) secondary dissociation
of vibrationally excited H2CO and HCOH(D) products, as shown by the shaded areas in Figure
3.1. In previous studies at these excitation energies, the low-resolution core-sampling TOF
technique was used for H/D detection, which could not resolve internal states of the products.
5,6

The present study is carried out at much higher kinetic energy (KE) resolution, which allows
characterization of product vibrational and electronic states. As mentioned previously, CH2OD is
used to distinguish between the H2CO and HCOD product channels by monitoring D and H
photofragments.

52

Figure 3.1: Energies of excited Rydberg states of CH2OH are shown on the left scale. The
shaded areas correspond to the excitation energies (Eexc) of the 3px and 3pz states investigated in
this work. The scale on the right shows the energies of singlet and triplet states of H2CO and
HCOH products relative to the ground state of H2CO.
3.2 Experimental Details
CH2OH radicals are produced in a quartz tube (length = 7 mm, Ø = 1 mm) attached to a
pulsed nozzle (Section 2.1.2) by the reactions Cl2 → 2Cl and CH3OH + Cl → CH2OH + HCl.
Mixtures of CH3OH (3%; Mallinckrodt Chemicals, 99.8% purity), Cl2 (0.5%; Matheson AirGas,
99.5% purity) and He (Gilmore, Liquid-Air Company, 99.999% purity) are prepared in a glass
bulb at 2.0 atm total pressure, and are expanded into the source region of the vacuum chamber
through a piezoelectrically driven pulsed-nozzle operating at 10 Hz. Cl2 photodissociation is
53

achieved with 355 nm radiation (~12 mJ; 3
rd
harmonic of Quanta-Ray GCR-11 Nd:YAG laser)
focused with a cylindrical lens (f.l. = 15 cm) at the tip of the quartz tube. The desired isotopologs
of CH2OH are generated by selecting deuterated methanol precursors (CH 3OD; Aldrich 99.5
atom % D and CD3OD; Aldrich 99.8 atom % D).
After passing through a skimmer (Beam Dynamics, Ø = 1 mm), the molecular beam is
intersected at right angles by two counter-propagating laser beams. The radicals are first excited
by UV laser radiation (225 − 285 nm) from the frequency doubled output of a dye laser
(Continuum ND6000, Coumarin Dyes) pumped by the 3
rd
harmonic output of a pulsed Nd:YAG
laser (Continuum, PL8000). Approximately 2 ns later H and D photofragments are probed by
1+1′ two-color REMPI via Lyman-α transition. The required vacuum ultraviolet (VUV) laser
radiation at ~121.6 nm is generated by frequency tripling of ~365 nm (2-3 mJ) radiation focused
into a gas mixture of Kr:Ar 200: 590 Torr in the tripling cell, as discussed in Section 2.4. The
VUV radiation is focused into the chamber using a MgF2 lens (f.l. = 7.5 cm). The residual ~365
nm radiation ionizes the excited H(D) fragments. The 365 nm radiation is generated by
frequency doubling the output of a dye laser (Continuum, ND6000, LDS 722 dye) pumped with
532 nm laser radiation (2
nd
harmonic of a Nd:YAG laser; Continuum, PL8000).
H
+
(D
+
) ions generated in the source region are accelerated through the flight tube towards a
position sensitive detector (a phosphor screen coupled to a double-stack Ø = 40 mm
microchannel plate; Galileo Electro-Optics 3040FM series). Image magnification is controlled
by changing the voltage of the Einzel lens (Section 2.3). The degree of magnification is selected
based on the maximum H(D) fragment kinetic energy of interest. A digital video-camera
(PixeLINK PL-B741F), located behind the phosphor screen of the detector, captures ion hit
events produced in each laser shot, and the signal is transferred to a computer for further
54

analysis. All SVMI images are recorded by tuning the H(D) detection laser through the Doppler
profile of the H or D fragment.
When operating in SVMI mode, fast gating of the detector (back plate of the MCP) is
achieved by using the home-built high voltage pulser (~5ns FWHM, 2 kV peak) as discussed in
section 2.3.
7–9
In order to attain the desired slice thickness (<10% of the whole ion cloud
stretched to ~50 ns) and KE resolution, optimal voltage settings of the ion-optics are selected by
performing numerical ion trajectory simulations using the SIMION program.
8
The center slice of
the ion cloud is located by recording several slice images around the center and comparing them
to select (i) the largest image radius, and (ii) the sharpest ring structure − the two features
indicative of the center slice. The optimal time delay between the detection laser pulse and the
detector gating pulse is determined by using the above procedure. To maximize signal
collection, images recorded at excitation energies above the 3pz origin transition are acquired
with excitation laser polarization parallel to the detector surface. H(D) fragments originating in
3s/3px excitation (recoil anisotropy parameter β < 0)
3
are imaged with excitation laser
polarization perpendicular to the detector.
The sliced images are transformed from cartesian to polar coordinates before further
analysis. Slight image distortions due to inhomogeneous electric fields along the radial directions
are corrected as described before.
7
Background signals originating from the detection laser are
small compared to the pump-probe signal. Such background images are collected under identical
experimental conditions and acquisition times in the absence of pump (excitation) laser, and are
then subtracted from the pump-probe images before analysis. Images recorded by exciting to
specific 3pz vibronic levels include contributions from underlying 3s/3px states, whose
55

absorptions are structureless. These background signals are estimated by acquiring images at
laser frequencies slightly detuned from the 3pz resonances (< 100 cm
–1
), and subtracted out from
the on-resonance 3pz images before plotting.
3.3 Results and Discussion
3.3.1 Triplet States of Formaldehyde and Hydroxymethylene
The lowest excited electronic states of the products are the triplet states of formaldehyde
and hydroxymethylene, which are separated by ~ 2000 cm
-1
(see Figure 3.1). Figure 3.2 shows
H-fragment images and the corresponding radial velocity distributions recorded with CH2OH
precursor at excitation energies 35,053-36,036 cm
-1
, which are in the proximity of the
3
H2CO
product threshold. The ion optics settings (ion acceleration and lens voltages) are optimized to
“zoom-in” on H-fragments with KE <8000 cm
-1
. In other words, at these excitation energies, H-
fragments that correspond to ground state H2CO co-fragments and have KEs above this value are
not detected. The results are presented as radial velocity distributions, rather than kinetic energy
release (KER) plots, because such plots give greater weight to low velocities, and thus are useful
for identifying appearance thresholds of products that have low yields.
The singlet-triplet energy gap in formaldehyde (∆ES-T = 25,194 cm
-1
)
10,11
and the
dissociation energy of CH2OH to generate ground state H2CO (10,160 cm
-1
) are well
established.
9
These values are used to determine that the energy threshold of
3
H2CO

with respect
to the CH2OH ground state is 35,354 cm
-1
. The velocity distribution shown in Figure 3.2a was
recorded at Eexc = 35,053 cm
-1
, just below the
3
H2CO onset, and it is correlated only with
rovibrational levels of cis- and trans-HCOH co-fragments in their ground electronic states.
11
The
partially resolved peaks at higher velocities (> 6000 ms
-1
) correspond to the ground and low
56

vibrational states of HCOH products.
4
As shown in Figure 3.2, the signal corresponding to
higher rovibrational product levels (lower H velocities) is unresolved and its intensity decreases
gradually with decreasing H fragment KE.
When Eexc is increased to 35,714 cm
–1
, i.e. ~360 cm
–1
above the expected
3
H2CO
threshold, the emergence of a new low KE product peak is evident (Figure 3.2b). The energetic
onset of this peak is at 25,243 ± 140 cm
-1
above the formaldehyde ground state, in good
agreement with ∆ES-T.
10,11
The two lowest vibrational levels of triplet-H2CO are the fundamental
and first overtone of the inversion mode (ν
4
) which are at 36 and 538 cm
-1
above the zero-point
energy (ZPE).
11
Due to significant rotational excitation in the ground vibrational level and 4
1
,
these two energy levels are not resolved. However, with further increase in excitation energy, a
clear emergence of a second peak at energy corresponding to the 4
2
vibrational level (538 cm
-1
)
is observed and is labeled by a stick in Figure 3.2c. With further increase in excitation energy,
the vibrational levels get broadened by increased rotational excitation and blend into the broad
background of high rovibrational levels of ground state products.

57

Figure 3.2: The left and right panels show, respectively, SVMI images and the corresponding
radial velocity distributions of H photofragments recorded at the following CH2OH excitation
energies (3px): (a) 35,053 cm
-1
; (b) 35,714 cm
-1
and (c) 36,036 cm
-1
. Onsets of product states are
indicated by red sticks. The images were obtained by zooming in on the low velocity region,
which is correlated mainly with HCOH products. The higher velocity region corresponding to
H2CO co-fragments in the ground electronic state
4
is not shown. The peak at near-zero velocities
is from hydrogen atom contamination.

58

The assignment of the peaks to
3
H2CO is strengthened by carrying out corresponding
measurements with CH2OD. Figure 3.3 shows the “zoomed-in” portion of the D-fragment
velocity distribution obtained by exciting CH2OD at 36,563 cm
-1
. This image is acquired under
SVMI conditions that produce thicker slices (25% slice thickness) in the low KE region.
Nevertheless, the peak in the low velocity part is clear, and its appearance threshold conforms
with the formation of
3
H2CO. The rising signal intensity toward higher velocities is associated
with rovibrational levels of H2CO in the ground electronic states, as discussed elsewhere.
4
The
onset of
3
H2CO is observed at 36,017 ± 70 cm
-1
with respect to CH2OD. After taking into
account the harmonic ZPE difference between CH2OD and CH2OH (𝐸 𝐶 𝐻 2
𝑂𝐷
𝑍𝑃𝐿
− 𝐸 𝐶 𝐻 2
𝑂𝐻
𝑍𝑃𝐿
= 671 cm
-
1
),
12
formaldehyde S-T gap is estimated to be 25,186 ± 140 cm
-1
. Within experimental error, this
value agrees well with previous spectroscopic measurements.
10–12
The onset of the 4
2
vibrational
level is observed at slightly higher excitation energy (Eexc = 36,630 cm
-1
).
Figure 3.3: Image and radial D-fragment velocity distribution from CH2OD excitation recorded
at Eexc = 36,563 cm
-1
. The low velocity peak corresponds to
3
H2CO. The distribution was
obtained by zooming in on the 0-5000 m/s velocity region.
59

These observations imply that the CH2 wag mode of
3
H2CO is populated. This can be
rationalized by considering the geometries of reactants and products. The difference between the
equilibrium molecular geometries of ground-state CH2OH, CH2OH (3px) and the
3
H2CO product
are significant. The geometry of CH2OH (3px) resembles that of the CH2OH
+
ion, which is
planar (Cs symmetry), whereas ground state CH2OH is non-planar (C1 symmetry).
3
H2CO has a
pyramidal geometry around the carbon center (see Figure 3.1). This implies that a major
geometry change occurs along the CH2 wag (inversion) coordinate, in accordance with a
pyramidal geometry with a large out-of-plane angle. It is therefore likely that O-H bond fission
occurs in the non-planar geometry, and the CH2 wag levels of CH2OD convert to υ4 excitation in
3
H2CO. Analogous geometry change considerations were used to explain the observation of C-O
stretch and in-plane CH2 scissors vibrational levels in the planar H2CO(S0) products.
4

The yield of
3
H2CO is much lower in photodissociation of CH2OD than CH2OH and
attempts to record D-fragment images by exciting CD2OD failed to reveal the
3
D2CO product
onset above the background of ground state formaldehyde. We conclude that the yield of triplet
formaldehyde depends sensitively on the isotopolog precursor used, and decreases with
increasing degree of deuteration. This may reflect the dependence of the conical intersection
seams that lead to triplet formaldehyde formation on the degree of deuteration of the
hydroxymethyl precursor.
Encouraged by the ability to detect the onset of the minor
3
H2CO product, we searched
for the triplet state of hydroxymethylene, which lies ~2000 cm
-1
higher in energy but has never
been detected experimentally. Our search has been guided by the available theoretical
predictions, which place ∆ES-T of HCOH around 25 kcal/mol (~ 8740 cm
-1
).
13–17
In order to
monitor only the C-H bond breaking channel, CH2OD is as a precursor and measured the
60

velocity distributions of H products correlated with HCOD co-fragments . Although the isotopic
difference between the calculated ∆ES-T for HCOH and our measurements for HCOD were not
taken into account, this difference is expected to be small compared with the uncertainties in the
experimental measurements described below.
Figure 3.4 shows overlaid velocity distributions recorded at CH2OD excitation energies
37,106 cm
-1
and 37,313 cm
-1
, near the threshold for triplet HCOD formation. In excitation at
37,106 cm
-1
, which is 9132 cm
-1
above the threshold for production of ground state trans-HCOD,
the H
+
signal intensity decreases gradually with decreasing velocity. The H-fragment velocity
distribution obtained at slightly higher excitation energy (37,313 cm
-1
) shows a very small but
reproducible new feature emerging at the lowest velocities above the background of ground state
HCOD fragments. The high velocity part remains nearly the same with only small intensity
fluctuations. Comparing the two plots, the new feature is assigned as the predicted onset of
3
HCOD. From Figure 3.4 we determine ∆ES-T = 9230 ± 400 cm
-1
(26.39 ± 1.1 kcal/mol) for
HCOD. However, determining the exact threshold is challenging because of the small size and
large width of the new feature, the large background signal, and our inability to observe
additional distinct structures in the velocity distributions at higher energies. Recalling that
HCOD products in their ground electronic state are born with high degree of rovibrational
excitation,
4
it is likely that the same is true for
3
HCOD. Thus, when the dissociation energy is
increased, the signal associated with the triplet state blends into the high density of rovibrational
61

levels of the ground state. We, therefore, consider the assignment of the triplet state of HCOD as
tentative.
Figure 3.4: H-fragment radial velocity distributions obtained with excitation energies 37,106
cm
-1
(red) and 37,313 cm
-1
(black) overlaid to demonstrate the appearance of a new feature
assigned as the onset of
3
HCOD. The inset shows a magnified plot of the low velocity region.
The stability of
3
HCOD is unknown. Previous results in Ar matrix showed that ground state
HCOH converts to H2CO by tunneling, but HCOD is stable for a long time. However, this
situation may change for
3
HCOD, which lies at higher energies and is close to the dissociation
energy to CO + H2. This state may isomerize faster to HDCO by tunneling or undergo efficient
coupling to the ground state.
The excited singlet state of formaldehyde is located 28,188 cm
-1
above its ground state or
2994 cm
-1
above the triplet state (Figure 3.1).
10,11
D-fragment images from CH2OD
photodissociation near the energetic onset of this product were recorded but no new distinct
62

features above the broad background of high rovibrational states of ground and triplet state
H2CO fragments were observed. Thus, if H2CO(S1) is produced at all, its yield must be very
small.
3.3.2 Vibrational excitation and dissociation HCOD products
Feng et al. reported before that with excitation of CH2OD to higher vibronic levels in 3pz,
a new, low-KE feature appeared in the H and D fragments’ kinetic energy distributions (KEDs),
whose width increased with increasing excitation energy.
6
The appearance thresholds of H
coincided with the dissociation energy of formaldehyde to H + HCO.
6,18
However, the question
whether formaldehyde or hydroxymethylene (or both) is the source of the secondary dissociation
of products remained unresolved due to the low KE resolution of the experimental arrangement.
Since SVMI has demonstrated excellent vibrational level resolution in the formaldehyde and
hydroxymethylene products,
4
here the photodissociation studies of CH2OH (CH2OD) using
SVMI were extended to higher excitation energies – up to and beyond the dissociation threshold
of H2CO and HCOH(D).
Figures 3.5 and 3.6 display H and D KER plots obtained by SVMI following excitation
of CH2OD to the 00
0
, 60
1
, and 60
2
transitions of 3pz. The D-fragment KEDs are correlated with
H2CO co-fragments, except for the small peak at low KE, which originates in secondary
dissociation of HCOD.
6
The H-photofragment KEDs, on the other hand, are correlated with
HCOD co-fragments, and their partially resolved peaks reflect vibrational excitation in HCOD.
The increased signals at the lowest KEs in the 60
1
and 60
2
plots correspond to H fragments
generated by secondary dissociation of either “hot” HCOD or H2CO products (or both).

63

Figure 3.5: H and D-fragment KER plots obtained by monitoring D (upper panels) and H (lower
panels) following 3pz excitation of CH2OD to the (a,c) 00
0
(Eexc = 41,004 cm
–1
); and (b,d) 60
1

(Eexc = 42,608 cm
-1
) bands. The H-fragment KER plots were obtained by zooming in on the 0-
20,000 cm
–1
KE region to enhance resolution. Sticks in each panel indicate the KE onsets
associated with H2CO, cis-HCOD, and the appearance of H and D from secondary dissociation.
See the text for details.
64

As shown before,
6
when exciting to the 3pz origin band, which is right at the threshold for
secondary dissociation, the low KE peaks are absent. Low-KE peaks have been detected also for
the other 3pz bands located between the 00
0
and 60
2
bands.
6
Both the previous and the present
Figure 3.6: H and D-fragment KER plots obtained by monitoring D (upper panel) and H (lower
panel) following 3pz excitation of CH2OD to the 60
2
band (Eexc = 44,207 cm
-1
). The H- and D-
fragment KER plots were obtained by recording the entire allowed KER region. Sticks in each
panel indicate the KE onsets associated with H2CO, cis-HCOD, and H and D from secondary
dissociation. See the text for details.
65

studies show that as the excitation energy increases, the secondary dissociation peak broadens
and increases in integrated intensity, but its threshold energy (shown by sticks in Figures 3.5 and
3.6) always corresponds to the H(D) + HCO channel threshold.
With the higher resolution afforded by SVMI, better analysis of the H and D KEDs is
possible. The D-fragment KED recorded at excitation energy below the threshold of secondary
dissociation shows that only a very small fraction of the H2CO products are born with the high
internal energies required for secondary dissociation, and this fraction does not appear to
increase at higher excitation energies. Therefore, H2CO is not an important source of H
fragments. In contrast, the H-fragment KEDs show that a significant fraction of HCOD co-
fragments possesses sufficient internal energies to dissociate to H(D) + D(H)CO at high
excitation energies. Thus, HCOD is the only source of the observed secondary D atoms and the
main source of secondary H fragments. It can also be concluded that the observed slow H and D
products are generated predominantly by dissociation of vibrationally excited HCOD products.
This is the first report of the dissociation of hydroxymethylene. The comparable yield of H and
D products combined with the available theoretical calculations suggests that the major
dissociation pathway involves isomerization to HDCO followed by dissociation as discussed in
the next section.
The reported high quality potential energy surfaces (PESs) of H2CO/HCOH system
include the ground state of formaldehyde and hydroxymethylene, as well as formaldehyde
dissociation to H2 + CO and H + HCO.
19–22
The T1 and S1 states of the two tautomers and their
couplings to the S0 state of formaldehyde have been characterized as well.
13,20,21
It is established
by theory and experiment that the barrierless dissociation threshold of H2CO to H + HCO lies
30,328 cm
-1
above the H2CO ZPE.
18,19
This energy is higher than the dissociation barrier of
66

H2CO to H2 + CO (27,720 cm
-1
)
23
, the isomerization barrier of H2CO to trans-HCOH (28,540
cm
-1
)
24
, the trans-HCOH - cis-HCOH isomerization barrier (27,526 cm
-1
)
24
, and the T1 levels of
H2CO (25,194 cm
-1
)
10,11
and HCOH (27,281 cm
-1
)
13
– all relative to the H2CO ground state
(including ZPE). The relevant states and barriers are shown schematically in Figure 3.7, which
illustrates that it is energetically possible to reach the H + HCO products from HCOH via
isomerization to H2CO. Small changes due to ZPE differences need to be made for HCOD, but
the order of states does not change.
In contrast to the wealth of calculations available on dissociation from the H2CO global
minimum, relatively little is known about the dissociation dynamics of HCOH. Reid et al.
examined the dissociation of HCOH to H + HCO but did not include isomerization to
formaldehyde.
25
They compared calculations for HCOH performed with CAS(10,10), CAS(8,8)
and MRCI, using the cc-pVDZ basis set, and reported that the minima and barriers depended
sensitively on the method used.
Nevertheless, their analysis provides qualitative insights into the nature of the electronic
states along the O-H reaction coordinate. They find an avoided crossing along O-H between the
diabatic S0 and S1 states of cis- and trans-HCOH, which changes the electronic character of the
adiabatic S0 state at fairly short O-H separations (< 2Å). As a result of this avoided crossing
there can be either a small barrier (for the trans isomer) or a shelf-like region (for the cis isomer)
along the O-H coordinate. In the language of transition state theory this means that the transition
states associated with dissociation of HCOH to H + HCO are much tighter than the loose
transition state associated with the barrierless dissociation of H2CO to H + HCO, which does not
require a change in electronic character. The substantial electronic reorganization required for
67

Figure 3.7: A schematic energy diagram showing relevant states and barrier energies (in cm
–1
)
for H2CO, HCOH and their dissociation to H + HCO, and H2 + CO. The values include ZPE and
are relative to the ground state of H2CO. The corresponding references are: a: ref. 18; b: ref. 13;
c
4
: ref. 4; d: ref. 24 (similar values given also in ref. 19); e: refs. 11 and 12; f: ref. 12; g: ref.
23.
HCOH along the O-H coordinate can create a bottleneck to direct dissociation via this channel,
which can enable the dominance of the isomerization channel. This isomerization mechanism
explains the observation of comparable yields of secondary H and D fragments, as seen in
Figures 3.5 and 3.6.
Hopefully these experimental results would inspire additional theoretical work on the
H2CO/HCOH PES that would include trajectory calculations initiated from the global minimum
of HCOH. Up to now most of the trajectory calculations started in the H2CO minimum. It is
worth noting that in one calculation the trajectories were initiated at the T1/S0 minimum energy
68

crossing configuration.
26
This crossing occurs at nuclear configurations similar to the transition
state separating the ground-state cis- and trans-HCOH isomers. Trajectories were run at total
energies of 35,000 – 38,000 cm
-1
above the global H2CO minimum, starting from different points
corresponding to the cis- and trans-T1/S0 crossings. Shepler et al. report that trajectories
terminating in CO + H2 products sample the cis- and trans-HCOH wells before finding the much
deeper H2CO well, from which they eventually proceed to products. Regarding the H + HCO
channel, they find that when including ZPE constraints only ~6% of the HCO fragments are
produced from the HCOH well, while the majority dissociate via H2CO. It appears therefore that
the existence of the deep formaldehyde well guides the trajectories along this route. Finally, in
regard to the CO + H2 channel, we note that we have already reported the observation of CO
products in v=0 and 1 in dissociation of hydroxymethyl via the 3pz state at energies > 33,000 cm
-
1
above the H2CO minimum.
6
These products can originate in secondary dissociation of
formaldehyde either directly or following HCOD isomerization.
The H-fragment KEDs obtained by SVMI following excitation to 3pz show distinct peaks
correlated with vibrational levels of cis-HCOD. It is noteworthy that these peaks are better
resolved than in the corresponding KEDs obtained following 3s and 3px excitation.
4
To enhance
the KE resolution, the KEDs in Figure 3.5 were obtained by “zoomed in” SVMI settings to
record only H fragments with KE < 20,000 cm
-1
(the total available KE for the formaldehyde
product is > 30,000 cm
-1
). They were converted to center-of-mass KEDs in order to fit them
with vibrational levels of cis-HCOD. The calculated vibrational frequencies of cis-HCOD are
listed in Table 3.1.

69

Table 3.1: List of fundamentals for cis-HCOD taken from reference [27]

Vibrational description Symmetry

cis-HCOD
o.p. twist  (a″) 847
i.p. bend  (a′) 921
i.p. bend  (a′) 1414
C−O stretch  (a′) 1288
C−H stretch  (a′) 2516
O−D stretch  (a′) 2584

HCOD peaks shown in Figures 3.5 and 3.6 are fitted reasonably well using the four
highest frequency vibrations, including overtones and combination bands. However, the
observed peaks are quite broad, and the fits are not unique, especially since rotational excitation
may shift the peaks. Examples of fits are shown in Figures 3.8 and 3.9. It appears that the CO-
stretch and in-plane CH2 bend modes (1288 and 1414 cm
-1
, respectively) provide the major
excitations, with possible contributions from the C-H and O-D stretch modes.
It should be noted, however, that the peaks remain quite distinct even at excitation
energies close to the isomerization and dissociation barriers of HCOD. For example, the H-
fragment KED shown in Figure 3.6, which was obtained at lower KE resolution (no “zooming
in” on the HCOD region), exhibits distinct vibrational structures that extend close to ~ 10,000
cm
-1
above the cis-isomer ground state; i.e. near the dissociation and isomerization barriers.

70

Figure 3.8: Kinetic energy distributions (center-of-mass) recorded at CH2OD excitation energy
41,004 cm
–1
(3pz origin) by monitoring H-photofragments, and single-mode vibrational levels of
cis-HCOD (shown by red sticks). The zero of the energy scale was set as the center of the first
cis-HCOD peak (the zero-point energy level). The fundamental vibrational frequencies are taken
from VCI calculations carried out at CCSD(T)/cc-pVTZ level of theory.
27
Higher vibrational
energy levels (overtones) are generated by multiplication of the fundamental frequencies by the
number of quanta. The stick height is adjusted to match the experimental intensities and serves
only to guide the eye.

-2000 0 2000 4000 6000 8000 -2000 0 2000 4000 6000 8000
-2000 0 2000 4000 6000 8000 -2000 0 2000 4000 6000 8000
H
+
Signal (arb. units)
(b) 

= 1414 cm
−
(a) 

= 1228 cm
−
(d) 

= 2584 cm
−
(c) 

= 2516 cm
−
H
+
Signal (arb. units)
Energy
c.m.
/ cm
−
Energy
c.m.
/ cm
−
71

Figure 3.9: The cis-HCOD product distribution (same as in Figure 3.8) is fit with a stick
spectrum of vibrational energy levels of the ν1–4 modes (combinations and overtones). Note the
grouping of vibrational levels near the experimentally observed peaks. The stick spectrum
intensities are set to match the experimental signal intensities and do not represent relative
populations.
The present SVMI results reassess the dissociation energy for the CH2OD → cis-HCOD
+ H. Taking into account uncertainties in fragment rotational energy and checking for
consistency with ZPE changes between the fully hydrogenated and partially deuterated reactants
and products, the current estimate is 29,534 ± 200 cm
-1
. This value supersedes the previous
value of 29,945 cm
-1
. The corresponding value for CH2OD → trans -HCOD + H is lower by
1560 cm
-1
.

-2000 0 2000 4000 6000 8000
H
+
Signal (arb. units)
Energy / cm
−
72

3.3.3 Implications to conical intersections in CH2OH(D)
Finally, it is important to comment on the effect of conical intersections leading to
dissociation of CH2OH(D) on fragment state distributions. About a decade ago, Yarkony carried
out conical intersection calculations on the hydroxymethyl radical and identified conical
intersection seams along the O-H and C-H coordinates following excitation to the 3s and 3px
states.
28,29
He predicted that in conical intersections from the 3s state, formaldehyde products
will have high translational energies and fairly low internal energies but a small fraction would
sample the global minimum and would have a broad, statistical-like, internal energy distribution.
He also predicted that cis- rather than trans-HCOH would be the predominant hydroxymethylene
product. These predictions were confirmed in previous work,
4
except that the situation regarding
HCOH production is found to be more complicated. In excitation to the 3s state, the yield of
HCOH(D) was fairly small, and because of its high rotational excitation it was impossible to
distinguish between the trans- and cis-isomers. However, upon excitation to the 3px origin band,
the rotational excitation in the HCOH(D) isomers was reduced considerably and vibrational
structure was observed. This allowed the authors to show that the cis-isomer was preferentially
produced and its relative population increased with increasing excitation energy. Nonetheless,
the rotational excitation was still fairly high and obscured vibrational structure at high internal
HCOH(D) energies.
The present work shows that another abrupt reduction in the HCOH(D) rotational
excitation occurs upon excitation to the 3pz state, which in turn allows the observation of distinct
vibrational structure in the HCOD fragment even at high internal energies. The vibrational
excitation spans the full range of internal energies, reaching up and above the dissociation limit
of HCOH(D). The loss of structure observed at very high internal energies can signify either
73

increasing IVR leading to a higher density of states and/or a greater contribution of
3
HCOH(D)
at > 7570 cm
-1
above the ground vibrational state of cis-HCOH(D).
Figure 3.10: Comparison of cis-HCOD internal state distributions obtained via 3pz excitation of
CH2OD to the 60
1
band (Eexc = 42,608 cm
-1
), and by excitation to the 3s/3px underlying
background (Eexc = 42,553 cm
-1
). The background contribution is already subtracted in the 60
1

plot. Sticks indicate the KE onsets associated with cis-HCOD and H from secondary
dissociation. Note the more prominent vibrational structures in the 60
1
plot than in the 3s/3px
plot.
As reported previously, the H/D photofragment yield spectra of CH2OH(D) show that
underlying the sharp 3pz vibronic bands there is a structureless background of absorption to 3s
and 3px. This gives an opportunity to observe the differences between the HCOD rovibrational
state distributions at similar energies. Figure 3.10 presents a comparison of the H-fragment
KEDs obtained at the 60
1
peak of 3pz (after background subtraction) and the one recorded ~70
74

cm
-1
to the red, which derives from 3s/ 3px background absorptions. Indeed, the latter displays a
much greater rotational excitation that obscures much of the vibrational structure
These results suggest that the dissociation dynamics is not controlled solely by the 3s PES and its
coupling to the ground state. Rather, different dynamics ensues by going through different
conical intersections. In all cases the dissociation is fast; even the 3p z state, which shows fairly
sharp vibronic structure, has a lifetime < 0.5 ps.
30
Yet, the resulting dynamics is quite different.
It is not clear whether the 3pz and 3px states couple directly to the ground state, or whether
sequential conical intersections link the upper states to the ground state and guide the
dissociating flux in ways that affects significantly its rotational excitation. New theoretical
calculations on a global PES that includes all the participating Rydberg states as well as wave-
packet propagation may shed further light on the dissociation mechanisms.
3.4 Summary and Conclusions
The SVMI results presented here provide additional information on the elusive carbene,
hydroxymethylene, while also revealing fingerprints of the conical intersections that control the
photodissociation dynamics of the hydroxymethyl radical precursor following excitation to its
lowest lying Rydberg states. The main photodissociation products of O-H(D) and C-H bond
breaking are formaldehyde and hydroxymethylene, respectively, and their internal energies are
estimated from the KEDs of H and D co-fragments.
The HCOD fragment generated by excitation of CH2OD to the 3px and 3pz Rydberg
states is mainly the cis-isomer that lies ~ 1550 cm
-1
above the trans-isomer. It is born with high
rotational and vibrational excitation that encompasses the full range of allowed energies. Its
internal excitation depends on the excited Rydberg state of CH2OD. Specifically, the rotational
75

excitation of HCOD is much lower when it is accessed via the higher-lying 3pz state than via the
3px state. The lower rotational excitation allows characterization of the excited vibrational
modes, and the CO-stretch and in-plane CH2 bending modes carry much of the HCOD
vibrational excitation. Distinct vibrational structure can be observed even at energies as high as
~ 9,000 cm
-1
above the cis-isomer ground state, quite close to the dissociation threshold to D +
HCO and the barrier to isomerization to formaldehyde. When the internal energy of HCOD
exceeds its dissociation threshold to D + HCO, slow D and H fragments from secondary
dissociation are observed. As the yields of these H and D fragments are comparable, it is likely
that the isomerization to HDCO precedes dissociation. The results also showed evidence for
production of HCOD in the triplet state at ~ 9200 cm
-1
above trans-HCOD, a slightly higher
value (~480 cm
-1
) compared to the calculated S-T gap for HCOH. The triplet peak is small and
due to its high level of rovibrational excitation it blends into the large background of
rovibrational levels of the ground state at higher excitation energies. Therefore, this assignment
is tentative.
The formaldehyde product generated by O-H bond breaking is characterized by the KEDs
of H- and D-photofragment obtained in photodissociation of CH2OH and CH2OD. A large
fraction of the available energy is released in translation, although the extent of internal
excitation increases and its range broadens with increasing dissociation energy. In addition, the
results indicate clearly the formation threshold of the triplet state of formaldehyde, even though
it is a minor channel. The observed distinct peaks in the H/D KEDs are correlated with its ground
state and excited ν4 vibrations. The triplet state onset agrees well with the spectroscopically
determined S-T gap in formaldehyde. The relative yield of triplet formaldehyde decreases with
76

increasing deuteration of the precursor, suggesting subtle changes in the conical intersection
seams that govern its formation.
The rotational, vibrational and electronic excitations in the formaldehyde and
hydroxymethylene products depend on the excited Rydberg state of the hydroxymethyl precursor
and the degree of deuteration, and therefore they provide fingerprints of the conical intersection
seams that lead from the precursor’s Rydberg states to products.
References
(1) Feng, L. Spectroscopy and Photodissociation Dynamics of the Hydroxymethyl Radical
(CH2OH). Ph.D. dissertation, University of Southern California, 2004.
(2) Feng, L.; Demyanenko, A. V.; Reisler, H. O–D Bond Dissociation from the 3s State of
Deuterated Hydroxymethyl Radical (CH2OD). J. Chem. Phys. 2003, 118 (21), 9623–9628.
(3) Feng, L.; Demyanenko, A. V.; Reisler, H. Competitive C–H and O–D Bond Fission
Channels in the UV Photodissociation of the Deuterated Hydroxymethyl Radical CH2OD. J.
Chem. Phys. 2004, 120 (14), 6524–6530.
(4) Rodrigo, C. P.; Zhou, C.; Reisler, H. Accessing Multiple Conical Intersections in the 3s and
3px Photodissociation of the Hydroxymethyl Radical. J. Phys. Chem. A 2013, 117 (46), 12049–
12059.
(5) Conroy, D.; Aristov, V.; Feng, L.; Reisler, H. Predissociation of the Hydroxymethyl
Radical in the 3pz Rydberg State: Formaldehyde + Hydrogen Atom Channel. J. Phys. Chem. A
2000, 104 (45), 10288–10292.
(6) Feng, L.; Reisler, H. Photodissociation of the Hydroxymethyl Radical from the 2
2
A'' (3pz)
State: H 2 CO and HCOH Products
†
. J. Phys. Chem. A 2004, 108 (45), 9847–9852.
77

(7) Ryazanov, M. Development and Implementation of Methods for Sliced Velocity Map
Imaging. Studies of Overtone-Induced Dissociation and Isomerization Dynamics of
Hydroxymethyl Radical (CH2OH and CD2OH). Ph.D. dissertation, University of Southern
California, 2012.
(8) Ryazanov, M.; Reisler, H. Improved Sliced Velocity Map Imaging Apparatus Optimized
for H Photofragments. J. Chem. Phys. 2013, 138 (14), 144201.
(9) Ryazanov, M.; Rodrigo, C.; Reisler, H. Overtone-Induced Dissociation and Isomerization
Dynamics of the Hydroxymethyl Radical (CH2OH and CD2OH). II. Velocity Map Imaging
Studies. J. Chem. Phys. 2012, 136 (8), 084305.
(10) Herzberg, G. The Spectra and Structures of Simple Free Radicals: An Introduction to
Molecular Spectroscopy; The George Fisher Baker non-resident lectureship in chemistry at
Cornell University; Cornell University Press: Ithaca, 1971.
(11) Clouthier, D. J.; Ramsay, D. A. The Spectroscopy of Formaldehyde and Thioformaldehyde.
Annu. Rev. Phys. Chem. 1983, 34 (1), 31–58.
(12) Birss, F. W.; Ramsay, D. A.; Till, S. M. Further High Resolution Studies of the System of
Formaldehyde. Can. J. Phys. 1978, 56 (6), 781–785.
(13) Zhang, P.; Maeda, S.; Morokuma, K.; Braams, B. J. Photochemical Reactions of the Low-
Lying Excited States of Formaldehyde: T1/S0 Intersystem Crossings, Characteristics of the S1
and T1 Potential Energy Surfaces, and a Global T1 Potential Energy Surface. J. Chem. Phys.
2009, 130 (11), 114304.
(14) Gronert, S.; Keeffe, J. R.; More O’Ferrall, R. A. Stabilities of Carbenes: Independent
Measures for Singlets and Triplets. J. Am. Chem. Soc. 2011, 133 (10), 3381–3389.
78

(15) Kiselev, V. G.; Swinnen, S.; Nguyen, V. S.; Gritsan, N. P.; Nguyen, M. T. Fast Reactions
of Hydroxycarbenes: Tunneling Effect versus Bimolecular Processes. J. Phys. Chem. A 2010,
114 (17), 5573–5579.
(16) Matus, M. H.; Nguyen, M. T.; Dixon, D. A. Heats of Formation and Singlet−Triplet
Separations of Hydroxymethylene and 1-Hydroxyethylidene. J. Phys. Chem. A 2006, 110 (28),
8864–8871.
(17) Shen, J.; Fang, T.; Li, S. Singlet-Triplet Gaps in Substituted Carbenes Predicted from
Block-Correlated Coupled Cluster Method. Sci. China Ser. B Chem. 2008, 51 (12), 1197–1202.
(18) Terentis, A. C.; Kable, S. H. Near Threshold Dynamics and Dissociation Energy of the
Reaction H2CO → HCO + H. Chem. Phys. Lett. 1996, 258 (5–6), 626–632.
(19) Zhang, X.; Zou, S.; Harding, L. B.; Bowman, J. M. A Global Ab Initio Potential Energy
Surface for Formaldehyde
†
. J. Phys. Chem. A 2004, 108 (41), 8980–8986.
(20) Rheinecker, J. L.; Zhang, X.; Bowman, J. M. Quasiclassical Trajectory Studies of the
Dynamics of H2CO on a Global Ab Initio -Based Potential Energy Surface. Mol. Phys. 2005, 103
(6–8), 1067–1074.
(21) Fu, B.; Shepler, B. C.; Bowman, J. M. Three-State Trajectory Surface Hopping Studies of
the Photodissociation Dynamics of Formaldehyde on Ab Initio Potential Energy Surfaces. J. Am.
Chem. Soc. 2011, 133 (20), 7957–7968.
(22) Zhang, X.; Rheinecker, J. L.; Bowman, J. M. Quasiclassical Trajectory Study of
Formaldehyde Unimolecular Dissociation: H2CO→H2+CO, H+HCO. J. Chem. Phys. 2005, 122
(11), 114313.
(23) van Zee, R. D.; Foltz, M. F.; Moore, C. B. Evidence for a Second Molecular Channel in the
Fragmentation of Formaldehyde. J. Chem. Phys. 1993, 99 (3), 1664–1673.
79

(24) Schreiner, P. R.; Reisenauer, H. P.; Pickard IV, F. C.; Simmonett, A. C.; Allen, W. D.;
Mátyus, E.; Császár, A. G. Capture of Hydroxymethylene and Its Fast Disappearance through
Tunnelling. Nature 2008, 453, 906.
(25) Reid, D. L.; Hernández-Trujillo, J.; Warkentin, J. A Theoretical Study of Hydroxycarbene
as a Model for the Homolysis of Oxy- and Dioxycarbenes. J. Phys. Chem. A 2000, 104 (15),
3398–3405.
(26) Shepler, B. C.; Epifanovsky, E.; Zhang, P.; Bowman, J. M.; Krylov, A. I.; Morokuma, K.
Photodissociation Dynamics of Formaldehyde Initiated at the T1 /S0 Minimum Energy Crossing
Configurations. J. Phys. Chem. A 2008, 112 (51), 13267–13270.
(27) Koziol, L.; Wang, Y.; Braams, B. J.; Bowman, J. M.; Krylov, A. I. The Theoretical
Prediction of Infrared Spectra of Trans - and Cis -Hydroxycarbene Calculated Using Full
Dimensional Ab Initio Potential Energy and Dipole Moment Surfaces. J. Chem. Phys. 2008, 128
(20), 204310.
(28) Hoffman, B. C.; Yarkony, D. R. Photodissociation of the Hydroxymethyl Radical. I. The
Role of Conical Intersections in Line Broadening and Decomposition Pathways. J. Chem. Phys.
2002, 116 (19), 8300.
(29) Yarkony, D. R. Statistical and Nonstatistical Nonadiabatic Photodissociation from the First
Excited State of the Hydroxymethyl Radical. J. Chem. Phys. 2005, 122 (8), 084316.
(30) Aristov, V.; Conroy, D.; Reisler, H. Symmetry and Lifetime of the Hydroxymethyl Radical
in the 3p Rydberg State. Chem. Phys. Lett. 2000, 318 (4–5), 393–401.

80

Chapter 4
Temperature Dependence of the
Photodissociation of CO
2
from High
Vibrational Levels
4.1 Introduction
The effect of vibrational excitation on the photodissociation dynamics of CO2 at 205–230
nm is investigated using SVMI. While O and CO fragments from photodissociation of hot CO2
have been detected before
1,2
, the detailed product state distributions and the influence of “hot
band” absorptions on the photodissociation dynamics have not been explored.
Heating CO2 up to ~1800 K is achieved in a SiC tube attached to a pulsed valve, as discussed
in Sections 2.5 and 4.2, and the supersonically expanded hot molecules are then photodissociated
by pulsed laser irradiation. Whereas the rotational temperature is reduced in the supersonic
expansion to 50-150 K,
3,4
vibrational levels are cooled much less effectively. CO2 molecules that
remain in high vibrational levels absorb the radiation and dissociate. Depending on CO2
temperature, one or both of the following channels was observed (D0 denotes the bond
dissociation energy for each channel):
5

CO2 → CO(X
1
Σ
+
) + O(
3
P) D0 = 43,954 cm
-1
(I)
CO2 → CO(X
1
Σ
+
) + O(
1
D) D0 = 59,847 cm
-1
(II)
81

The kinetic energy release (KER) distributions obtained when monitoring CO(X
1
Σ
+
) products
in v"=0 provide information on the initial internal energy of the parent CO2. The extent of
internal excitation in the CO product is revealed from images of O(
3
P) and O(
1
D) products.
Previously, several investigators estimated the average internal energy that CO2 must possess in
order to absorb at specific wavelengths. At 230 nm, for example, this was estimated to be 0.5-1.2
eV.
6–8
These studies are consistent with the fact that the photodissociation is observed only
when the minimum vibrational excitation in CO2 required for absorption at each excitation
wavelength is reached.
4.2 Experimental Details
Vibrationally excited CO2 molecules are generated by passing a mixture of CO2 (1-5%)
seeded in He at a total pressure of 1.7 atm through the piezoelectrically driven pulsed valve
operating at 10 Hz onto a resistively heated SiC tube (3.7 cm long, 1 mm id), as discussed in
Section 2.5. The red, orange and yellow colors of the SiC tube are correlated with temperatures
of 800-1000 (low), 1200-1400 (medium), and 1600-1800 K (high), respectively. The heated gas
mixture is supersonically expanded into the source region of the vacuum chamber and passes
through a skimmer (Beam Dynamics, 1.0 mm orifice diameter) to form a molecular beam that
reaches the interaction region. As a result, the amount of internal energy of CO2 in the molecular
beam depends both on the extent of heating in the SiC tube and the subsequent cooling by
supersonic expansion. A detailed description of the gas dynamics in a pyrolysis nozzle and the
properties of the subsequent supersonic expansion were discussed by several previous studies.
3,4,9

The rotational temperature is estimated from REMPI spectra of CO molecules under the same
expansion conditions, and it increases from 50 K to 150 K as the temperature is increased up to
82

1800 K, in agreement with previous results.
3,4
The extent of vibrational excitation is discussed in
Section 4.3.
The molecular beam is intersected at a right angle by the laser beam at the center of the
interaction region, where internally excited CO2 molecules absorb UV radiation. Pulsed radiation
at 225-230 nm (0.5 mJ/pulse, focused by a 30 cm f.l. lens) is generated by the frequency doubled
output of a dye laser (Continuum ND6000, Coumarin dyes) pumped by the third harmonic
output of a pulsed Nd:YAG laser (Continuum, PL8000). Laser radiation at 205.47 nm (0.1-0.2
mJ/pulse, focused by a 30 cm f.l. lens) is generated by mixing the fundamental (616.42 nm) of a
dye laser (Continuum ND6000, DCM dye) and its frequency doubled output (308.21 nm).
CO(X
1
Σ
+
,v"=0, 1, 2) products are ionized by 2+1 REMPI at 230.1-230.5 nm via the B
1
Σ
+
←
X
1
Σ
+
transition.
10
O(
3
Pj) products are ionized by 2+1 REMPI at 225.6-226.4 nm via the 3p
3
PJ′
← 2p
3
PJ'' transition,
11
and O(
1
D) products are ionized at 205.47 nm via the 3p
1
P1 ← 2p
1
D2
transition.
12
Excitation of CO2 and product detection are achieved at the same wavelength (one
color experiment). Vibrationally excited CO2 is detected by using 3+1 REMPI via the 3pσu
1
Пu
 X
1
Σ
+
g transition at 326-329 nm (2 mJ/pulse, focused by a 30 cm f.l. lens) generated from the
frequency doubled output of a dye laser (Continuum ND6000, DCM/LDS 698) pumped by the
second harmonic output of a pulsed Nd:YAG laser.
13
The dependencies of the CO and O
REMPI signals on laser fluence were investigated in order to assess possible contributions from
multiphoton dissociation. The CO and O signals appear only with heated CO2, and they persist
even with laser energies as low as 0.2-0.3 mJ/pulse. Also, the KER distributions can only be
explained consistently by assuming one-photon dissociation. We thus believe that multiphoton
dissociation does not contribute significantly to the observed REMPI signals.
83

The ionized products are accelerated through a series of ion optics lenses in the flight
tube toward a position sensitive detector (a phosphor screen coupled to a double-stack (40 mm
dia) microchannel plates; Galileo Electro-Optics 3040FM series) as discussed in Section 2.2. A
digital video-camera (PixeLINK PL-B741F) located behind the phosphor screen of the detector
captures ion hit events produced by each laser firing, and the signal is transferred to a computer
for further analysis. Sliced images of the ion cloud are obtained by fast gating of the detector
(back plate of the MCP) using the home-built high-voltage pulser discussed in Section 2.3. (5 ns
fwhm, 2 kV peak).
4.3 Results and Discussion
4.3.1 REMPI spectra of CO products
CO products from photodissociation of hot CO2 at ~ 230 nm under different heating
conditions were detected by 2+1 REMPI at the same wavelength. Figure 4.1 shows REMPI
spectra obtained when the laser frequency is scanned in the region of the 2-photon B
1

+
- X
1

+

transition of CO.
The bandheads of the 0-0 (86,916 cm
-1
) and 1-1 (86,856 cm
-1
) vibronic peaks are assigned
based on previous spectroscopic and photodissociation studies.
14,15
The CO photofragments have
significant rotational excitation in both v"=0 and 1. Taking into account the calculated relative
FC factors for the 0-0 and 1-1 transitions (1.00 and 0.91, respectively), the relative population of
v" = 1 is estimated to be low at all temperatures. Fit spectra of the 0-0 band, obtained using
PGOPHER,
16
show that signal intensities drop significantly starting at J" > 14, and signals are
observed up to at least J" = 23-24 with low heating. Rotational excitation in the 0-0 band
increases with increased heating and the distributions become clearly bimodal, displaying a
84

weaker additional feature peaking at J" = 43-44 and extending to at least J" > 47. With
increased heating, the second feature becomes more prominent. The total signal intensity
increases strongly with increased heating.
Figure 4.1: One color 2+1 REMPI spectra of CO produced in the photodissociation of CO2
around 230 nm recorded with low (black), medium (red) and high (blue) heating of the CO2
parent, respectively. The REMPI transition from ground state vibrational level (X
1
Σ
+
, v") to
excited state vibrational level (B
1
Σ
+
, v') is denoted by v"-v'. The assignment of the band denoted
by M is uncertain (see the text for discussion). Intensities are normalized to the 0-0 bandhead.
To fit the rotational distribution of the 0-0 band, we used the expression:
𝐼 (𝐸 )= 𝐴 ∑ (2𝐽 + 1).Γ(𝐸 − 𝐸 𝐽 ).𝐹 (
𝐽 𝐸 𝐽 ) ( 4.1)
where I is the observed line intensity at the 2-photon energy E, A is a scaling constant, Γ(𝐸 − 𝐸 𝐽 )
is the linewidth function (assumed Gaussian) for each rotational transition QJJ with energy 𝐸 𝐽 ,
86800 86850 86900 86950
CO
+
signal (arb. unit)
2-photon energy / cm
-1
High heating
Medium heating
Low heating
M 1-1 0-0
85

and 𝐹 (𝐸 𝐽 ) is a fitting function. The Hönl-London factors for the Q-branch have been assumed to
be constant.
17
In the above simulation, 𝐹 is assumed to be a sum of two functions: a Boltzmann
function 𝐵 (𝐸 𝐽 ,𝑇 ) and a Gaussian distribution 𝐺 (𝐸 𝐽 0 − 𝐸 𝐽 ) centered at a specific rotational level
J
0
.
Figure 4.2: Comparison of the REMPI spectrum of the CO 0-0 QJJ band (blue) to a fit obtained
with a rotational linewidth 𝚪 = 1.9 cm
-1
, assuming a rotational temperature of 1700 K, and a
Gaussian component with a width of 5 J" levels centered at J" = 43. The individual Boltzmann
and Gaussian components are shown in red dotted lines and the total fit is shown in black.
The FWHM of the two Gaussian functions Γ and 𝐺 , the rotational temperature 𝑇 and scaling
constant 𝐴 are varied to match the experimental distributions. An example is shown in Figure 4.2
for the distribution obtained with high CO2 heating.
86900 86920 86940 86960 86980
Experiment
Fit
CO
+
signal (arb. unit)
2 photon energy (cm
-1
)
86

The fits confirm that the observed bimodal rotational distributions for the 0–0 band consist of
a statistical-like (Boltzmann) component and a Gaussian-shaped rotational component; their
relative contributions vary with heating of the CO2 parent molecule. Rotational excitation in
v"=1 rises up to J" = 17-18 at low temperatures and extends up to J" =24-25 with increased
heating. The 1-1 REMPI spectrum could not be fitted because of its partial overlap with the band
starting at 86,800 cm
-1
.
The origin of the broad spectral feature with a maximum at ~ 86,830 cm
-1
(denoted
henceforth by M) is puzzling. This feature appears only at mass 28 in the TOF mass spectrum,
and only when irradiating hot CO2. A similar feature has been observed by Spiglanin et al. in the
REMPI spectrum of the CO product from the193 nm photodissociation of HNCO,
18
and
tentatively assigned as the 2-2 vibronic band of this transition. However, this feature could not be
reproduced using the known spectroscopic constants of the B←X transition. Even though the
B(v'=0-9)-X(v"=0) vibronic bands of CO were investigated before,
10,14,19–22
due to crossings with
repulsive states, the B state is severely predissociative at v'>1.
14,19,23,24
As a result, the
assignments of the band origins of the B←X vibronic transitions involving v'>1 are quite
uncertain, having large error bars. Using the known spectroscopic constants of the ground state,
81
and the experimental values for the v' levels in the B state, the band origin frequencies of the
diagonal transitions from 2-2 to 9-9 are estimated and listed in Table A1 in Appendices.
Boltzmann rotational distributions of the 2-2 and 5-5 vibronic transitions, whose spectral
origins are closest to the M band, were simulated using PGOPHER.
16
Representative
simulations are shown in Figures A1 and A2 in Appendices. It is evident that the M band does
not match well any of the estimated B←X diagonal vibronic transitions. It is possible that this
feature originates in a different electronic transition than B←X. The three electronic states closest
87

to this energy region are k
3
П, C
1
Σ
𝑔 +
and E
1
П.
25–27
The different vibronic transitions estimated
from the known spectroscopic constants of these states are listed in Tables A2-A6 in
Appendices. The only transition close to the M band is k (v'=1) ← X (v"=2). The 500 K thermal
rotational distribution of this vibronic transition is simulated (using PGOPHER) and shown in
Figure A3 in Appendices. The simulated REMPI spectrum spans frequencies extending from
near the bandhead of the 0-0 peak of the B←X transition to the low frequency end of the
observed M band. The peak of the M band corresponds to J" = 5-10 of v"=2 of the ground state.
This transition may contribute to the REMPI signal originating in v"=2 of CO (see also Section
4.3.2), but other (yet unknown) transitions may contribute as well.
88

4.3.2 Kinetic energy release distributions of CO products
In order to obtain pair-correlated distributions, time-sliced images of CO(X
1
Σ
+
, v"=0) were
recorded at 230.1 nm (43,480 cm
-1
) with different CO2 heating levels, and the corresponding
center of mass (c.m.) KER distributions are shown in Figure 4.3.
Figure 4.3: Center-of-mass (c.m.) KER distributions obtained by monitoring CO(X
1
Σ
+
,v"=0)
products at 230.1 nm by 2+1 REMPI at the bandhead of the B←X (0-0) transition. Results are
shown for low (black), medium (red) and high (blue) CO2 heating. The relative signal intensities
in the three plots are arbitrary.
Although the excitation energy (43,480 cm
-1
) is slightly below the energy threshold of
channel I (43,954 cm
-1
), heating in the SiC tube imparts significant vibrational excitation to CO2,
and much of this excitation survives during supersonic expansion. Recall that only vibrationally
hot CO2 molecules can absorb at 230.1 nm. Since the O(
3
P) co-fragments do not possess internal
0 10000 20000 30000
0
0
CO
+
signal (arb. unit)
Kinetic energy release / cm
-1
High heating
Medium heating
Low heating
0
89

energy (except in their
3
P1 and
3
P0 spin-orbit states), the c.m. KER distributions obtained by
monitoring the 0-0 bandhead, which is centered at low J" levels, reflect the distribution of
available energies in the dissociation and, therefore, also the initial rovibrational excitation of the
heated parent CO2 molecules.
It is evident from the KER distribution obtained with the lowest heating that the observed
KER signal starts only at ~ 4000 cm
-1
, corresponding to the minimum internal excitation that
CO2 must possess in order to absorb at 230.1 nm. This KER distribution extends up to about
10,000 cm
-1
. With higher CO2 heating, the vibrational excitation in the parent CO2 molecules
increases and reaches a maximum value of 17,000 -18,000 cm
-1
, as shown in the middle and
upper traces of Figure 4.3. By energy conservation, KER values ranging between 4,000 and
~16,000 cm
-1
must be correlated only with the CO + O(
3
P) channel (channel I).
Referring to Figure 4.3, the KER signal drops when the CO2 internal excitation exceeds
17,000 – 18,000 cm
-1
even with high level of heating. At the same time, an additional feature
appears at low KER (0-2000 cm
-1
). This feature corresponds to the opening of the CO + O(
1
D)
channel (channel II), which lies 59,847 cm
-1
above the ground state CO2, or ~ 16,000 cm
-1
above
the CO + O(
3
P) channel threshold. In other words, channel II becomes energetically accessible at
230.1 nm when the available energy exceeds ~16,000 cm
-1
.
It should be noted that the observed CO
+
signal intensities reflect two opposing factors: (1)
The thermal population in the excited vibrational levels of CO2, which decreases approximately
exponentially with increasing vibrational energy, and (2) The absorption cross section at a
specific temperature, which increases approximately exponentially with CO 2 internal energy at
the long wavelength tail of the absorption curve.
6,28–30
Although the population of high
90

vibrational levels is very small even with high heating (only 15% of the population exists in the
energy region 4000-20000 cm
-1
at a temperature of 1800 K, assuming a Boltzmann distribution),
the increased absorption cross sections make these levels visible in the current experiments. As
mentioned above, the final CO2 internal energy distributions are determined by the extent of
heating in the SiC tube and the subsequent cooling in the expansion, which evidently preserves
much of the vibrational excitation. The structure in the KER distribution, which is most
pronounced at higher temperatures, reflects the well-known clumpy nature of the CO2 vibrational
states, which is also manifest in the polyad structure of its vibrational spectrum.
31

Time-sliced images of CO v"=1 are also recorded by monitoring the bandhead of the B←X
(1-1) transition at 230.27 nm, and the KER distribution is shown in Figure 4.4(a). The KER
distribution corresponding to CO(X
1
Σ
+
,v"=1) shows a sharp feature centered at 5000 cm
-1
with a
FWHM of ~ 1000 cm
-1
and a low intensity broad feature extending up to about 18,000 cm
-1
.
However, analysis of the KER distribution of CO(X
1
Σ
+
,v"=1) is complicated due to spectral
overlap of the 1-1 bandhead with the M band. In order to separate the contribution of the M band
to the 1-1 bandhead signal, the KER distribution in the M band region was obtained by recording
sliced images at 230.34 nm (86,830 cm
-1
) with different heating levels. An example, obtained
with high heating, is displayed in Figure 4.4(b), which shows that the narrow peak centered
around 5000 cm
-1
in KER distribution is associated with the M band.
91

Figure 4.4: c.m. KER distributions obtained with high CO2 heating by monitoring (a) the
bandhead of the B←X 1-1 transition at 230.27 nm, and (b) the peak of the M band at 230.34 nm.
Signal intensities are normalized with respect to the feature centered at 5000 cm
-1
.
Since both KER distributions shown in Figure 4.4 exhibit the same sharp feature centered at
5000 cm
-1
, it is likely that this feature in the plot of CO v"=1 belongs to the underlying
background of the M band. Indeed, the intensity of the sharp feature in Figure 4.4(a) increases
with increased heating as does the intensity of the M band. In contrast, the broad features at
higher KER values (Figure 4.4(a)) do not increase appreciably with heating, in accordance with
the behavior observed in the CO REMPI spectrum (Figure 4.1). The background of the M band
extends up to the bandhead of the 0-0 transition, where the images shown in Figure 4.3 were
recorded. With low heating, where the intensity of the M band is much smaller compared to the
0-0 bandhead, there is almost no feature at 5000 cm
-1
in the KER distributions shown in Figure
4.3. With increased heating, the KER distributions show the emergence of the 5000 cm
-1
feature
92

as the contribution from the M band increases. However, the M band contribution does not
extend to higher rotational levels of the 0-0 band, and the KER distributions obtained by
monitoring higher rotational levels (J" > 43) of the 0-0 band do not show the 5000 cm
-1
feature
even with high heating as shown in Figure 4.5. Evidently, a different dissociation mechanism,
presently unknown, gives rise to the feature centered at ~5000 cm
-1
.
Figure 4.5: Center of mass (c.m.) KER plot obtained with high heating by monitoring CO (X

1
Σ
+
;

v"=0) products at the 2+1 REMPI bandhead (86,916 cm
-1
, red) and at J"= 43 (86,962 cm
-1
,
black) of the B-X 0-0 transition. The KER plot obtained at J" = 43 does not show the feature at
5000 cm
-1
. This observation suggests that the contribution of the M band does not extend to the
higher frequencies, where absorption from high rotational levels of the 0-0 transition takes place.
The relative signal intensities are scaled for easy visual comparison.
93

4.3.3 Kinetic energy distributions of O(
3
P,
1
D) products
Due to the uncertainty in the spectral assignment of the M band in the REMPI spectrum, it is
best to assess the extent of CO internal excitation from the KER distributions of the atomic
O(
3
Pj) and O(
1
D2) photofragments. O(
3
Pj) fragments (channel I) are detected by 2+1 REMPI at
225.6-226.6 nm, and the three spin-orbit states (
3
P2,
3
P1, and
3
P0) produced in the
photodissociation are shown in Figure 4.6. The KER distributions obtained by monitoring O(
3
P2)
with different heating levels of CO2 are displayed in Figure 4.7.
Figure 4.6: 2+1 REMPI of the O(
3
Pj) product at 225.8-226.6 nm showing its three spin-orbit
states (
3
P2,
3
P1,
3
P0).
With low heating, the KER distribution exhibits a single broad feature extending from 0
to10,000 cm
-1
. An additional feature at 12,000-16,000 cm
-1
appears with increased heating,
indicating that two different dynamical pathways lead to channel I. Similar results (not shown)
94

are obtained when monitoring O(
3
P1) and O(
3
P0). It is important to note that the KER
distributions include contributions from CO(v, J) co-fragments generated by photolysis of all
CO2 molecules whose internal energies exceed ~4000 cm
-1
. They indicate that the dissociating
CO2 molecules, which have a broad range of vibrational energies, give rise to CO products with
a broad distribution of internal energies.
Figure 4.7: C.m. KER distributions obtained by monitoring O(
3
P2) products at 225.654 nm with
low (black), medium (red) and high (blue) CO2 heating. The intensities of the KER distributions
obtained with different heating levels are normalized with respect to the broad feature centered at
~3000 cm
-1
.
O(
1
D) photofragments produced by channel II were detected at 205.47 nm (48,669 cm
-1
) by
using 2+1 REMPI via the
1
P1←
1
D2 2-photon transition. The c.m. KER distribution obtained by
monitoring O(
1
D) photofragments with high CO2 heating is shown in Figure 4.8.
0 5000 10000 15000 20000
O
+
signal (arb. unit)
Kinetic energy release / cm
-1
High heating
Medium heating
Low heating
95

Figure 4.8: C.m. KER distribution obtained by monitoring O (
1
D) product at 205.47 nm with
high CO2 heating.
The observed KER distribution peaks at ~300-400 cm
-1
and extends to nearly 6000 cm
-1
. As
discussed above, with 230.1 nm photolysis the available energy must exceed 16,000 cm
-1
in
order to reach the threshold of channel II. This is evident in the KER distributions of CO v"=0
(Figure 4.3) where a new feature appears at 0-2000 cm
-1
. With 205.47 nm (48,669 cm
-1
)
photolysis, CO2 requires only ~11,000 cm
-1
of internal energy to reach the threshold of Channel
II. Hence, the c.m. KER of O(
1
D) photofragments can extend up to 6000-7000 cm
-1
when the
internal energy of CO2 is 17,000-18,000 cm
-1
, as seen in Figure 4.8. Clearly, the peak in the KER
distribution is shifted to lower energies than the allowed maximum, indicating that the CO co-
fragments possess significant rovibrational excitation. Parent CO2 molecules with different
rovibrational levels have different absorption cross sections and might also exhibit different
0 2500 5000 7500
O
+
signal (arb. unit)
Kinetic energy release / cm
-1
96

photodissociation dynamics. All of these determine the final shape and maximum in the KER
distribution obtained by monitoring O(
1
D).
4.3.4 REMPI spectra of hot CO2
In order to confirm that the vibrationally excited CO2 heated in the SiC tube of the pulsed
valve survives supersonic expansion, 3+1 REMPI spectra of CO2 were obtained using 331-325
nm excitation via the 3pσu
1
Пu←X
1
Σ
+
g transition with different levels of heating as shown in
Figure 4.9.
Figure 4.9: 3+1 REMPI spectra of CO2 obtained via the 3pσu
1
Пu  X
1
Σ
+
g transition.
Progressive increase in vibrational excitation of CO2 is observed upon increased heating, as
shown for low (cyan), medium (red) and high (blue) heating. The spectrum of cold CO2 is
shown in black for comparison. The peak at 91,910 cm
-1
corresponds to the origin band of the
3pσu
1
Пu (000)  X
1
Σ
+
g (000) transition of CO2. The intensities in the spectra obtained with
97

different heating levels are normalized with respect to the 3pσu
1
Пu (000)  X
1
Σ
+
g (000) peak of
cold CO2.
The origin band of the 3pσu
1
Пu (000) ← X
1
Σ
+
g (000) transition of cold CO2 is assigned
based on previous spectroscopic studies described in Ref 13.
13
A tentative assignment of the
vibronic peaks of hot CO2 is done using harmonic frequencies of the ground
32
and excited states.
Since the 3pσu
1
Пu state is a Rydberg state, its harmonic vibrational frequencies are assumed to
be the same as those of ground state CO2
+
ion.
33,34
The estimated vibrational frequencies and the
diagonal transitions (v'1v'2v'3←v"1v"2v"3) are listed in Table 4.1. The higher vibrational levels
seen in Figure 4.3 have not been observed in the REMPI spectrum because of their very low
population even with high heating. These levels show up in the KER distributions of CO (Figure
4.3) because of the increased UV absorption cross-sections of the high CO2 vibrational levels. A
tentative assignment of the vibrational progression is listed in Table 4.1. State specific dynamics
in the photodissociation of hot CO2 is suggested, among others, by the KER distributions
observed when monitoring the M REMPI band (Figure 4.4(b)), which show only a narrow range
of KER distributions centered at ~5000 cm
-1
. However, the dynamics responsible for this state-
specific KER distribution remain unexplained, as is the exact assignment of the M band (Figure
4.1).

98

Table 4.1: Tentative assignments of the 3pσu
1
Пu (v'1v'2v'3)← X
1
Σ
+
g (v"1v"2v"3) transitions using
harmonic frequencies of the ground states of CO2 and CO2
+
.

4.3.5 Implications to CO2 photodissociation dynamics
As discussed before, the absorption cross-section of CO2 at λ > 190 nm increases sharply
with temperature, and absorption can be significant up to 300 nm. Our study, which examines the
205-230 nm photodissociation of CO2 at temperatures up to ~ 1800 K, extends the
photodissociation studies of CO2 to more realistic environments where thermally excited CO2 is
dissociated at much longer wavelengths than cold CO2. It shows that several mechanisms are
likely involved in dissociation via channels I and II. The KER distributions shown in Figure 4.3,
which are obtained at different CO2 temperatures, give an idea of the extent of vibrational
excitation from which dissociation occurs at ~230 nm.
Normal
modes
CO2/
cm
-1

CO2
+
/
cm
-1

Transition
3pσ u
1
П u ← X
1
Σ
+
g
Position /
cm
-1

Symm str.

1388

1241

000-000 91870
100-100 91723

Bending

667

513

010-010 91716
110-110 91569
020-020 91562

Asymm str.

2349

1421
001-001 90942
101-101 90795
99

The REMPI spectra and KER distributions show that the CO products are born with
significant rovibrational excitation. A broad rovibrational state distribution in CO was observed
also in the 157 nm photodissociation of cold CO2, which is near the FC maximum of absorption
to the
1
B2 state. At 157 nm, channel II was dominant,
35
but a minor channel I (about 6%) was
also detected.
36–38
Understanding the dissociation dynamics is not easy even with supersonically
cooled samples, because the CO product state distributions span a broad energy range and the
rotational populations exhibit fluctuations.
35
This work demonstrates the increased complexity of
the dissociation dynamics at higher temperatures, and suggests that bending and stretch
excitation of ground state CO2 have a profound effect on the dissociation efficiency and
dynamics.
Several theoretical studies have focused on the excited electronic states of CO2 at 4-9 eV,
39–44

and on the interactions among states that are involved in the dissociation.
39,42
Schmidt et al.
computed 1D potential energy curves for the electronic states of CO2 as a function of OC-O
internuclear distance (Cs) and OCO bond angle (C2v).
39
Data on the five lowest singlet and triplet
states that can be accessed below 8.5 eV ( ̴ 68,500 cm
-1
) are summarized in Table 4.2, which lists
both the vertical and adiabatic excitation energies. Figure 4.10 presents the corresponding
schematic energy level diagram illustrating these states.
100

Figure 4.10: Energy diagram showing adiabatic (Vmin, solid line) and vertical (Vvert, dotted line)
energies of the excited electronic states of CO2 listed in Table 4.2. A and B states are shown in
red and black, respectively.

0
4

Excitation energy / eV

CO
2
5
6
7
8
3
B
2

3
A
2

1
A
2
,
1
B
2

1
A
1
(X)
CO(X
1
Σ
+
) + O(
3
P)
CO(X
1
Σ
+
) + O(
1
D)
9
3
B
2

3
A
2

1
A
2
,
1
B
2

V
min
values
V
vert
values
101

Table 4.2: Ground and excited electronic states of CO2 taken from ref. 39. Vertical and adiabatic
excitation energies (Vvert and Vmin, respectively) are listed. The geometries at the minimum of the
potentials are given by the O-CO distance rmin and bond angle αmin.
Abbreviation D∞h C2v Cs Vvert /eV Vmin /eV rmin / a0 αmin / degree
X
1
Σ
𝑔 +
1
1
A1 1
1
A
'
0 0 2.197 180
a
3
Σ
𝑢 +
1
3
B2 1
3
A' 8.228 4.642 2.355 117.5
b
3
Δu 1
3
A2 1
3
A" 8.713 5.329 2.370 127.4
B
1
Δu 1
1
A2 1
1
A" 8.950 5.528 2.370 127.2
A
1
Δu 1
1
B2 2
1
A' 8.938 5.532 2.363 117.8

The calculated absorption spectrum at long wavelengths matches the experimental
observations, and the photodissociation dynamics to channels I and II at 157 nm are also
reproduced satisfactorily in the calculations.
39
It is concluded that whereas absorption is mainly
to the A
1
B2 state at the vertical maximum, at longer wavelengths the B
1
A2 state can be excited as
well (see below), and the long wavelength spectrum reflects also mixing between the A
1
B2 and
the ground state.
39

As listed in Table 4.2, the energy differences between the ground and excited states decrease
with decreasing OCO bond angle, e.g. with increased bending excitation in the ground electronic
state. The present study shows that heating CO2 in the SiC tube imparts significant vibrational
excitation that survives supersonic expansion. Due to more favorable FC factors, these internally
hot CO2 molecules absorb 205-230 nm radiation much more efficiently, giving an opportunity to
102

study the dissociation of CO2 closer to the origins of the lowest excited states and to the
dissociation threshold, where channel I dominates.
It should be noted that although the fraction of highly excited vibrational states is low (at
1800 K, the population of molecules with 20,000 cm
-1
energy is only ~10
-4
relative to the
maximum), the steep increase in the absorption cross section with increasing vibrational
excitation allows us to observe dissociation from these levels. As discussed in section II, the
vibrational population cannot be described by a temperature, due to the supersonic expansion.
However, Figure 4.3 shows that, as expected, the CO2 vibrational levels appear in clumps, in
accordance with the vibrational spectrum of CO2 that displays a polyad structure even at high
temperatures.
31

Notably, the images recorded in the above experiments are nearly isotropic (β = 0). A
common reason for an isotropic product angular distribution is a dissociation lifetime that is
longer than the rotational period of the molecule. In this case, the dissociation lifetime is
unknown, and the average classical rotational period of CO2 molecules at 100 K is roughly 0.9
ps. This may cause a decrease from the maximum recoil anisotropy value of 2.0 for a parallel
transition. Previous studies of cold CO2 photodissociation at 157 nm (63,690 cm
-1
) showed that
the fragments’ angular distributions associated with channel I are anisotropic (β =1.25), as
expected for fast dissociation involving initial excitation via a parallel transition to the
1
B2
state.
45,46
In contrast, the angular distributions of CO fragments produced via channel II at the
same wavelength are nearly isotropic,
35
with β decreasing strongly with decreasing kinetic
energy and CO rovibrational excitation, even though the initially excited state is short-lived.
45
It
was proposed that the initially excited state may have crossed to a predissociative state whose
lifetime is longer than the rotational period of CO2.
35
Another plausible argument put forward is
103

that dissociation occurs from bent vibrational configurations or high rotational levels that reduce
the anisotropy by changing the fragments’ recoil direction. In the present study, CO2 molecules
with even higher vibrational excitation are electronically excited. If dissociation takes place from
vibrational configurations with high bending excitation, the anisotropy would be reduced.
Another likely reason for the reduced anisotropy is the concurrent excitation to the
1
B2 and
1
A2 states, which are reached via parallel and perpendicular transitions, respectively. Schmidt
and coworkers showed that at long wavelengths there are reasonable transition dipole moments
(TDMs) to both of these states, and the bending and asymmetric stretch excitations increase the
TDM to
1
A2.
39
Spielfiedel et al. assigned a long bending progression of perpendicular bands to
transitions to the B
1
A2 state at the long wavelength absorption tail, in addition to the main bands
to the A
1
B2 excited state.
47
The nearly isotropic KER distributions of the products measured in
our experiments suggest that both these excited states participate in the initial electronic
excitation, and this may be a major cause for the reduction in anisotropy.
The main photodissociation products observed here are derived from the spin-forbidden
channel I. With 157-nm photodissociation of CO2, this was the minor channel, accounting for
merely a few percents relative to channel II.
35,36
Evidently in dissociation via channel I, spin-
orbit coupling must be involved, and therefore the reverse spin-forbidden recombination
reaction, O(
3
P) + CO, is also relevant to this work. It has been established long ago that the so
called “CO flame bands”, which derive from the recombination of O(
3
P) and CO, result in also
weak emission from the
1
B2 excited state of CO2.
48
Several recent theoretical papers have
examined the recombination reaction, identifying spin-orbit couplings and conical
intersections.
7,41–44,49
Even though these calculations do not include the highly excited CO2
vibrational levels that are necessary to describe fully our results, they indicate that more than one
104

reaction mechanism must be involved. The calculations show the participation of a direct (or
non-statistical) mechanism, which is too fast to allow for randomization of energy, and an
indirect (or statistical-like) route, in which at least some energy randomization takes place. The
exact nature of the dynamics via these two pathways is not settled yet, but the outcome is clear.
Jasper and Dawes, for example, find that both the a
3
B2 and b
3
A2 triplet states are involved in the
spin-forbidden reaction O(
1
D) + CO → CO + O(
3
P).
43
In addition, the authors highlight the
importance of including geometry dependence in the spin-orbit coupling surface. Referring to the
coupling between the electronic ground state and the lowest triplet a
3
B2 state, the dynamical
pictures invoked by Jasper and Dawes
43
and Hwang and Mebel
42
are similar in that in both direct
and indirect mechanisms are involved. However, Jasper and Dawes suggest that the relative
importance of these two mechanisms depends on the depth of the geometry-dependent entrance
channel well, in addition to the properties of the saddle point and the energy minimum in the
spin-orbit crossing seam. In our studies, a broad range of geometries is sampled, which is likely
to affect the dynamics to an even greater extent. The existence of multiple dynamical channels is
amply evidenced in the KER distributions of the CO and O products, and also in the CO REMPI
spectra obtained as a function of CO2 heating. In dissociation events that start closer to the
reaction threshold, as is the case in our work, it is not hard to imagine that following initial
excitation to either the A
1
B2 or B
1
A2 state or both, the a
3
B2 and b
3
A2 triplet states and even the
ground electronic state are involved, giving rise to multiple crossings and dissociation pathways
that depend sensitively on initial vibrational excitation of parent CO2.

105

4.4 Summary and Conclusions
The 205-230 nm photodissociation of vibrationally excited CO2 at temperatures up to 1800 K
was studied in a molecular beam by using REMPI spectroscopy and SVMI. Upon heating,
dissociation from vibrational levels with energy up to nearly 20,000 cm
-1
was detected, giving
rise to CO, O(
3
P) and O(
1
D) products (channel I and II). The large enhancement of the
absorption cross section with increasing CO2 vibrational excitation made the investigation of
photodissociation from highly vibrationally excited molecules feasible. It is likely that initial
absorption is to both the A
1
B2 and B
1
A2 states.
We show that CO2 must have a minimum internal energy of ~4000 cm
-1
in order to absorb
230 nm radiation, in good agreement with previous estimates.
6–8
2+1 REMPI spectra of the CO
products show that v"=0 and 1 are generated in a broad range of rotational states, but direct
REMPI identification of higher vibrational levels was hampered by uncertainties in the
spectroscopic assignments. Nevertheless, fits of the rotational distributions of CO v"=0 show that
at higher CO2 temperatures the rotational distributions are bimodal; they can be fitted with a
~1700 K Boltzmann component at low to medium J" levels, and a narrow Gaussian-shaped
component centered at J" >40. The fragments KER distributions obtained by monitoring the CO
B←X (0-0) bandhead reveal the extent of vibrational excitation in the parent CO2, which
increases with heating.
The extent of rovibrational excitation in the CO product is assessed from KER distributions
obtained by monitoring the O(
3
P) and O(
1
D) fragments. All spin-orbit states of O(
3
P) are
produced in the dissociation and have similar KER distributions. With low CO2 heating, the KER
distribution obtained by monitoring O(
3
P) is broad and unstructured, corresponding to highly
106

rovibrationally excited CO co-fragments. At higher temperatures, a second unstructured
component centered at high KER appears, indicating the opening of another, more direct,
dissociation pathway. Following 205 nm dissociation, the KER distributions obtained by
monitoring the O(
1
D) fragment, which has 16,000 cm
-1
of internal energy, peak at lower KER
values but extend up to about 6000 cm
-1
. These distributions show that CO fragments associated
with channel II are also born with a broad distribution of rovibrational levels.
While detailed theoretical descriptions of the dissociation dynamics are unavailable at this
time (and maybe unfeasible), the observation that several dynamical pathways contribute to the
observed fragments’ REMPI spectra and KER distributions is hardly surprising. Theoretical
calculations of the absorption spectrum of CO2, especially in the long wavelength tail, show that
optical excitation can reach both the A
1
B2 or B
1
A2 excited singlet states, and that vibrational
excitation can strongly enhance absorption and change the relative contributions of these two
absorption systems. Electronic structure calculations of the conical intersections among states of
the same multiplicity and spin-orbit couplings between states of different multiplicities, which
are still incomplete, show that the low-lying excited electronic states can interact either directly
or via the ground electronic state, and that both the a
3
B2 and b
3
A2 triplet states are likely
involved in dissociation via channel I. Theory shows that reaction pathways can be direct,
without energy randomization in the shallow wells in the excited states, or more statistical-like
involving partial or total randomization of energy in the excited states during conical intersection
or in reaching the spin-orbit crossing seams.
Because the photodissociation involves intermediate states with shallow wells, it is expected
that state-specific effects with respect to initial vibrational excitation of the parent CO2 molecule
would be important. A more detailed view of the dissociation mechanism may be obtained by
107

exploiting vibrationally mediated photodissociation to reach specific vibrational levels of CO2
directly and explore their photodissociation. Excitation of combination bands with stretch and
bend components should be particularly fruitful, as discussed in Chapter 6.
References
(1) Nefedov, A. P.; Sinel’shchikov, V. A.; Usachev, A. D.; Zobnin, A. V. Photochemical
Effect in Two-Photon Laser-Induced Fluorescence Detection of Carbon Monoxide in
Hydrocarbon Flames. Appl. Opt. 1998, 37 (33), 7729.
(2) Settersten, T. B.; Dreizler, A.; Patterson, B. D.; Schrader, P. E.; Farrow, R. L. Photolytic
Interference Affecting Two-Photon Laser-Induced Fluorescence Detection of Atomic
Oxygen in Hydrocarbon Flames. Appl. Phys. B 2003, 76 (4), 479–482.
(3) Rohrs, H. W.; Wickham‐Jones, C. T.; Ellison, G. B.; Berry, D.; Argrow, B. M. Fourier
Transform Infrared Absorption Spectroscopy of Jet‐cooled Radicals. Rev. Sci. Instrum.
1995, 66 (3), 2430–2441.
(4) Guan, Q.; Urness, K. N.; Ormond, T. K.; David, D. E.; Barney Ellison, G.; Daily, J. W.
The Properties of a Micro-Reactor for the Study of the Unimolecular Decomposition of
Large Molecules. Int. Rev. Phys. Chem. 2014, 33 (4), 447–487.
(5) Song, Y.; Gao, H.; Chang, Y. C.; Lu, Z.; Ng, C. Y.; Jackson, W. M. Photodissociation of
CO2 between 13.540 eV and 13.678 eV. Phys. Chem. Chem. Phys. 2014, 16 (2), 563–569.
(6) Oehlschlaeger, M. A.; Davidson, D. F.; Jeffries, J. B.; Hanson, R. K. Ultraviolet
Absorption Cross-Sections of Hot Carbon Dioxide. Chem. Phys. Lett. 2004, 399 (4–6),
490–495.
(7) Ibraguimova, L. B.; Minaev, B. F.; Irgibaeva, I. S. Spin-Orbit Coupling and Dissociation
of CO2 Molecules. Opt. Spectrosc. 2014, 117 (5), 695–702.
108

(8) Lee, T.; Bessler, W. G.; Yoo, J.; Schulz, C.; Jeffries, J. B.; Hanson, R. K. Fluorescence
Quantum Yield of Carbon Dioxide for Quantitative UV Laser-Induced Fluorescence in
High-Pressure Flames. Appl. Phys. B 2008, 93 (2–3), 677–685.
(9) Zhang, X.; Friderichsen, A. V.; Nandi, S.; Ellison, G. B.; David, D. E.; McKinnon, J. T.;
Lindeman, T. G.; Dayton, D. C.; Nimlos, M. R. Intense, Hyperthermal Source of Organic
Radicals for Matrix-Isolation Spectroscopy. Rev. Sci. Instrum. 2003, 74 (6), 3077–3086.
(10) Tjossem, P. J. H.; Smyth, K. C. Multiphoton Excitation Spectroscopy of the B
1
Σ
+
and
C
1
Σ
+
Rydberg States of CO. J. Chem. Phys. 1989, 91 (4), 2041–2048.
(11) Bamford, D. J.; Dyer, M. J.; Bischel, W. K. Single-Frequency Laser Measurements of
Two-Photon Cross Sections and Doppler-Free Spectra for Atomic Oxygen. Phys. Rev. A
1987, 36 (7), 3497.
(12) Pratt, S. T.; Dehmer, P. M.; Dehmer, J. L. Photoionization of an Autoionizing Level of
Atomic Oxygen. Phys. Rev. A 1991, 43 (1), 282.
(13) Wu, M.; Johnson, P. M. A Study of Some Rydberg States of CO2 by (3+1) Multiphoton
Ionization Spectroscopy. J. Chem. Phys. 1989, 91 (12), 7399–7407.
(14) Eidelsberg, M.; Roncin, J.-Y.; Le Floch, A.; Launay, F.; Letzelter, C.; Rostas, J.
Reinvestigation of the Vacuum Ultraviolet Spectrum of CO and Isotopic Species: The
B
1
Σ
+
↔ X
1
Σ
+
Transition. J. Mol. Spectrosc. 1987, 121 (2), 309–336.
(15) Rijs, A. M.; Backus, E. H.; de Lange, C. A.; Janssen, M. H.; Westwood, N. P.; Wang, K.;
McKoy, V. Rotationally Resolved Photoionization Dynamics of Hot CO Fragmented from
OCS. J. Chem. Phys. 2002, 116 (7), 2776–2782.
(16) Western, C. M. A Program for Simulating Rotational Structure; 2010.
109

(17) Seitzman, J. M.; Haumann, J.; Hanson, R. K. Quantitative Two-Photon LIF Imaging of
Carbon Monoxide in Combustion Gases. Appl. Opt. 1987, 26 (14), 2892–2899.
(18) Spiglanin, T. A.; Perry, R. A.; Chandler, D. W. Internal State Distributions of CO from
HNCO Photodissociation. J. Chem. Phys. 1987, 87 (3), 1568–1576.
(19) Baker, J. The Diffuse v= 4 and 5 Vibrational Levels of the B
1
Σ
+
Rydberg State of Carbon
Monoxide. Chem. Phys. Lett. 2005, 408 (4), 312–316.
(20) Federman, S. R.; Fritts, M.; Cheng, S.; Menningen, K. M.; Knauth, D. C.; Fulk, K.
Oscillator Strengths for BX, CX, and EX Transitions in Carbon Monoxide. Astrophys. J.
Suppl. Ser. 2001, 134 (1), 133.
(21) Letzelter, C.; Eidelsberg, M.; Rostas, F.; Breton, J.; Thieblemont, B. Photoabsorption and
Photodissociation Cross Sections of CO between 88.5 and 115 nm. Chem. Phys. 1987, 114
(2), 273–288.
(22) Baker, J.; Tchang‐Brillet, W.-Ü. L.; Julienne, P. S. First Observation of the v= 3 Level of
the B
1
Σ
+
Rydberg State of CO. J. Chem. Phys. 1995, 102 (10), 3956–3961.
(23) Eidelsberg, M.; Launay, F.; Ito, K.; Matsui, T.; Hinnen, P. C.; Reinhold, E.; Ubachs, W.;
Huber, K. P. Rydberg-Valence Interactions of CO, and Spectroscopic Evidence
Characterizing the C′
1
Σ
+
Valence State. J. Chem. Phys. 2004, 121 (1), 292–308.
(24) Tchang‐Brillet, W.-Ü. L.; Julienne, P. S.; Robbe, J.-M.; Letzelter, C.; Rostas, F. A Model
of the B
1
Σ
+
–D′
1
Σ
+
Rydberg–Valence Predissociating Interaction in the CO Molecule. J.
Chem. Phys. 1992, 96 (9), 6735–6745.
(25) Baker, J.; Launay, F. The k
3
Π Valence State of CO. J. Mol. Spectrosc. 1994, 165 (1), 75–
87.
110

(26) Berden, G.; Jongma, R. T.; Meijer, G. A Reanalysis of the k
3
Π State of CO. J. Chem.
Phys. 1997, 107 (20), 8303–8310.
(27) Eidelsberg, M.; Rostas, F. Spectroscopic, Absorption and Photodissociation Data for CO
and Isotopic Species between 91 and 115 nm. Astron. Astrophys. 1990, 235, 472–489.
(28) Jeffries, J. B.; Schulz, C.; Mattison, D. W.; Oehlschlaeger, M. A.; Bessler, W. G.; Lee, T.;
Davidson, D. F.; Hanson, R. K. UV Absorption of CO2 for Temperature Diagnostics of
Hydrocarbon Combustion Applications. Proc. Combust. Inst. 2005, 30 (1), 1591–1599.
(29) Venot, O.; Fray, N.; Bénilan, Y.; Gazeau, M.-C.; Hébrard, E.; Larcher, G.; Schwell, M.;
Dobrijevic, M.; Selsis, F. High-Temperature Measurements of VUV-Absorption Cross
Sections of CO2 and Their Application to Exoplanets. Astron. Astrophys. 2013, 551, A131.
(30) Jensen, R. J.; Guettler, R. D.; Lyman, J. L. The Ultraviolet Absorption Spectrum of Hot
Carbon Dioxide. Chem. Phys. Lett. 1997, 277 (4), 356–360.
(31) Tashkun, S. A.; Perevalov, V. I.; Teffo, J.-L.; Bykov, A. D.; Lavrentieva, N. N. CDSD-
1000, the High-Temperature Carbon Dioxide Spectroscopic Databank. J. Quant.
Spectrosc. Radiat. Transf. 2003, 82 (1), 165–196.
(32) Shimanouchi, T. Tables of Molecular Vibrational Frequencies, Consolidated Volume
INational Bureau of Standards, Washington, DC, 1972).
(33) Jacox, M. E.; Thompson, W. E. The Vibrational Spectra of Molecular Ions Isolated in
Solid Neon. I. CO2
+
and CO2
−
. J. Chem. Phys. 1989, 91 (3), 1410–1416.
(34) Varfalvy, N.; Larzillière §, M.; Frigon, C.; Bacis, R.; Ross, A. J. Fourier Transform
Spectra of Laser-Induced Fluorescence from the Ion: Reinvestigation of Fermi Resonance
Effects. Mol. Phys. 2007, 105 (5–7), 627–637.
111

(35) Miller, R. L.; Kable, S. H.; Houston, P. L.; Burak, I. Product Distributions in the 157 nm
Photodissociation of CO2. J. Chem. Phys. 1992, 96 (1), 332–338.
(36) Chen, Z.; Liu, F.; Jiang, B.; Yang, X.; Parker, D. H. Imaging CO2 Photodissociation at 157
nm: State-to-State Correlations between CO (ν) and O (
3
PJ = 0, 1, 2). J. Phys. Chem. Lett.
2010, 1 (12), 1861–1865.
(37) Zhu, Y.-F.; Gordon, R. J. The Production of O (
3
P) in the 157 nm Photodissociation of
CO2. J. Chem. Phys. 1990, 92 (5), 2897–2901.
(38) Matsumi, Y.; Shafer, N.; Tonokura, K.; Kawasaki, M.; Huang, Y.-L.; Gordon, R. J.
Doppler Profiles and Fine‐structure Branching Ratios of O (
3
Pj) from Photodissociation of
Carbon Dioxide at 157 nm. J. Chem. Phys. 1991, 95 (10), 7311–7316.
(39) Schmidt, J. A.; Johnson, M. S.; Schinke, R. Carbon Dioxide Photolysis from 150 to 210
nm: Singlet and Triplet Channel Dynamics, UV-Spectrum, and Isotope Effects. Proc.
Natl. Acad. Sci. 2013, 110 (44), 17691–17696.
(40) Grebenshchikov, S. Y. Photodissociation of Carbon Dioxide in Singlet Valence Electronic
States. II. Five State Absorption Spectrum and Vibronic Assignment. J. Chem. Phys.
2013, 138 (22), 224107.
(41) Zhou, B.; Zhu, C.; Wen, Z.; Jiang, Z.; Yu, J.; Lee, Y.-P.; Lin, S. H. Topology of
Conical/Surface Intersections among Five Low-Lying Electronic States of CO2:
Multireference Configuration Interaction Calculations. J. Chem. Phys. 2013, 139 (15),
154302.
(42) Hwang, D.-Y.; Mebel, A. M. Ab Initio Study of Spin-Forbidden Unimolecular
Decomposition of Carbon Dioxide. Chem. Phys. 2000, 256 (2), 169–176.
112

(43) Jasper, A. W.; Dawes, R. Non-Born–Oppenheimer Molecular Dynamics of the Spin-
Forbidden Reaction O (
3
P)+ CO (X
1
Σ
+
)→ CO2 (X ̃ 1
Σg
+
). J. Chem. Phys. 2013, 139 (15),
154313.
(44) Ma, Y.; Peng, L.; Zhang, H.; Yu, J.-G. The Potential Energy Surfaces of the Ground and
Excited States of Carbon Dioxide Molecule. Russ. J. Phys. Chem. A 2014, 88 (13), 2339–
2347.
(45) Lu, I.-C.; Lin, J. J.; Lee, S.-H.; Lee, Y. T.; Yang, X. Product Angular Anisotropy in CO2
Photodissociation at 157 nm. Chem. Phys. Lett. 2003, 382 (5), 665–670.
(46) Chen, P.; Ng, C. Y.; Baer, T.; Powis, I. Unimolecular and Bimolecular Reaction
Dynamics. John Wiley & Sons, New York 1994.
(47) Spielfiedel, A.; Feautrier, N.; Cossart‐Magos, C.; Chambaud, G.; Rosmus, P.; Werner, H.-
J.; Botschwina, P. Bent Valence Excited States of CO2. J. Chem. Phys. 1992, 97 (11),
8382–8388.
(48) Dixon, R. N. The Carbon Monoxide Flame Bands; The Royal Society, 1963; Vol. 275, pp
431–446.
(49) Jasper, A. W. Multidimensional Effects in Nonadiabatic Statistical Theories of Spin-
Forbidden Kinetics: A Case Study of
3
O+ CO→ CO2. J. Phys. Chem. A 2015, 119 (28),
7339–7351.

113

Chapter 5
Spectroscopy and Two-photon Dissociation of
Jet-Cooled Pyruvic Acid
5.1 Introduction
As discussed in Section 1.1.3, photodissociation of pyruvic acid at low- and high-
pressures were investigated before and several photoproducts including CO2, CH3CHO, CH4,
and acetic acid were observed.
1–7
However, the photodissociation dynamics remains unclear
since the lifetime of the S1 state and the branching ratios of different products are found to be
dependent on collisions. To elucidate the photochemistry of PA under collisionless condition, the
spectroscopy and photodissociation of PA in molecular beams are studied here for the first time.
The molecular beam absorption spectrum of PA has not been measured before because
the S1←S0 absorption system is fairly weak,
1,8
and the proposed nascent products, CO2 and
methylhydroxycarbene (CH3COH), are not easily detected. However, we discovered that due to
its relatively long lifetime, a second photon absorption by S1 is facile, leading to several
photoproducts that can be detected by ionization. In this chapter, the molecular beam
photofragment yield (PFY) spectrum and multiphoton dissociation of PA at 350 −380 nm, using
He and Ar carrier gases are reported. CH3CO, HOCO, CO, CH3, and H products were detected
by TOF mass spectroscopy and VMI. The PFY spectra of the products mapped the broad
absorption peaks of PA observed at room temperature, but the rovibronic bands exhibited
significant rotational cooling and were well resolved. The two-photon nature of formation of the
114

fragments was confirmed by their KER distributions measured by monitoring CH3CO, HOCO,
CO, CH3, and H products. Three-body fragmentation is found to be a major dissociation
mechanism leading to the observed products.
The chapter is organized as follows. In Sections 5.2 and 5.3 we describe the experimental
details and results, and list the energetically allowed pathways. Electronic structure calculations
of the excited states of PA is presented in Section 5.4, followed by discussion of the S1 ← S0
spectroscopy and the probability of absorption of a second photon by the S 1 state. The two
photon dissociation reactions and possible two- and three-body fragmentation pathways are also
discussed. Section 5.5 provides summary and conclusions.
5.2 Experimental Details
Pyruvic acid (PA) (98%, Sigma-Aldrich) is purified by vacuum distillation at 50−60°C
twice, and the distilled colorless sample is then degassed using several freeze−pump−thaw
cycles.
5
PA vapor is introduced into the source region of the vacuum chamber by bubbling a
carrier gas (He or Ar) through the sample at 1300 Torr and supersonically expanding the
mixture through a piezoelectrically driven pulsed nozzle operating at 10 Hz. After passing
through a skimmer (Beam Dynamics, 1.0 mm diameter), PA is excited by UV laser
radiation (330−380 nm, 2−3 mJ/pulse, focused by a 30 cm focal-length (f.l.) lens to about 0.1
mm) from the frequency doubled output of a dye laser (Continuum ND6000, DCM
dye) pumped by the second harmonic output of a pulsed Nd:YAG laser (Continuum, PL8000).
The fluence in the 10 ns pulse used in this experiment is about 6 × 1019 photons cm
−2
.
H, CH3CO, HOCO, CO, and CH3 photofragments are detected by TOF mass spectroscopy and
VMI using the appropriate probe laser radiation.
115

H photofragments are probed by 1+1′ two-color resonance enhanced multiphoton
ionization (REMPI) via the Lyman-α transition. The required vacuum ultraviolet (VUV) laser
radiation at 121.567 nm is generated by frequency tripling of ∼365 nm (2−3 mJ) radiation
focused into a gas mixture of Kr:Ar = 200:590 Torr in the tripling cell. The 365 nm radiation is
generated by frequency doubling the output of a dye laser (Continuum, ND6000, LDS 722 dye)
pumped with 532 nm laser radiation (second harmonic of a Nd:YAG laser; Continuum, NY-
81C). The VUV and residual 365 nm beams are focused into the chamber using a MgF2 lens (f.l.
= 7.5 cm).
Because there are no known REMPI detection schemes for CH3CO and HOCO, these
photofragments are detected by one-color multiphoton nonresonant ionization, in which the
pump radiation also ionizes the photofragments. The non-state selective detection can encompass
all the internal states of these species that are below their barriers to dissociation (6000 cm
−1
for
CH3CO and 9000 cm
−1
for HOCO). CO (X
1
Σ
+
, v"=0, 1) products are detected by 2+1 REMPI at
230.1-230.5 nm via the B
1
Σ
+
← X
1
Σ
+
transition.
9
The UV radiation is generated by the
frequency doubled output of a dye laser (Continuum ND6000, Coumarin dyes, 0.5-1 mJ/pulse)
pumped by the third harmonic output of a pulsed Nd:YAG laser (Continuum, PL8000). CH3
(𝑋
̃
2
A2", v1" v2" v3" v4" ) products are detected by 2+1 REMPI at 325-336 nm (~1 mJ/pulse) via
the 3p
2
A2" ← 𝑋
̃
2
A2" transition.
10,11

The ions produced in the detection region are accelerated in the TOF tube through a
series of electrostatic lenses toward a position sensitive detector (40 mm dia double-stack
microchannel plates) coupled to a phosphor screen (Galileo Electro- Optics 3040FM series). Ion
hit events on the detector are recorded by a 1-megapixel digital video-camera (PixeLINK
PLB741F) located behind the phosphor screen, and the data are transferred to a computer for
116

further analysis. Time-sliced images of the ion sphere are obtained by fast gating of the detector
(back plate of the MCP) using the home-built high voltage pulser (5 ns fwhm, 2 kV peak), as
discussed in Section 2.3.
12

5.3 Results
5.3.1 Photofragment Yield Spectra.
Atomic hydrogen photofragments were produced by two-photon dissociation via the S1
state (see below). H-photofragment yield (H-PFY) spectra were recorded in He or Ar as a carrier
gas by scanning the pump laser radiation from 330 to 380 nm while the probe laser radiation was
fixed to detect H photofragments. The recorded spectra show significant rovibrational cooling of
PA in the molecular beam. As seen in Figure 5.1, the vibronic peaks in the H-PFY spectrum
recorded in He match well the broad S1 ← S0 absorption features recorded at 298 K. The
corresponding spectrum recorded in Ar and shown in Figure 5.2 achieves even greater rotational
cooling. Both spectra become progressively congested toward shorter wavelengths.
It is important to note the difference between the absorption and the H-PFY spectra
shown in Figures 5.1 and 5.2. The spectral intensities and vibrational progressions observed in
the 298 K absorption spectrum depend on the rovibrational populations of S0 and the Franck −
Condon factors for the S1 ← S0 transition. The cold molecular beam PFY spectrum originates
mainly in the ground vibrational state of S0. Also, the peak intensities depend on the dissociation
efficiency from each vibrational level of S1 to the monitored photofragment. In other words, the
117

Figure 5.1: Molecular beam H-PFY spectrum of PA in He carrier gas (red) and room
temperature absorption spectrum (black) taken from ref 8.
relative intensities of the peaks in the PFY spectra will not mimic the absorption spectrum if the
photodissociation cross-section leading to the monitored product changes with excitation energy.
In order to assign the H-PFY spectrum, the contributions of the different conformers of
PA must be assessed. The two relevant conformers of ground state PA are shown in Figure 5.3.
The most stable conformer, Tc, which comprises 87% of the population at room temperature, lies
730 cm
−1
below the Tt conformer (the barrier height for Tc to Tt isomerization is unknown).
13

Although gas-phase IR studies show the presence of both Tc and Tt conformers at room
temperature, the contribution of the Tt conformer to the recorded spectrum is reported to be
118

significantly lower than the Tc conformer, reflecting the large population difference of the two
conformers.
14

Figure 5.2: Molecular beam H-PFY spectrum of PA recorded in Ar carrier gas.
In order to investigate the contribution of the Tt conformer in our molecular beam
experiments, the H-PFY spectrum was recorded with different heating conditions of the PA
sample prior to expansion. The H-PFY spectra recorded when PA was heated to 298, 318, and
338 K did not exhibit new peaks or differences in peak intensities. On the basis of the large
difference in population between the Tc and the Tt conformers of S0 and the apparent
temperature independence of the HPFY spectrum, we conclude that the H-PFY spectrum
recorded in our study represents primarily absorption by the Tc conformer of PA.
119

Figure 5.3: Tc and Tt conformers of PA identified in previous studies.
13,14
The Tc conformer is
more stable than the Tt due to intramolecular H bonding.
To confirm that the H-PFY spectrum is a fingerprint of the molecular beam S1 ← S0
absorption spectrum of PA, CH3CO (mass 43) and HOCO (mass 45) products were detected in
one-color multiphoton experiments (section 5.2), and as seen in Figure 5.4, their PFY spectra
match well the H-PFY spectrum.

120

Figure 5.4: CH3CO-PFY (a) and HOCO-PFY (b) spectra of PA recorded in He.
5.3.2 Reaction Thermochemistry and Fragments’ Kinetic Energy Release.
In order to obtain a comprehensive view of the photodissociation dynamics of PA, time-
sliced images of the fragments were recorded at excitation energies corresponding to several
peaks in the H-PFY spectrum. As seen below, the high KER of fragments observed following
excitation to the ground vibrational level of S1 (26 710 cm
-1
) indicates that the detected products
are generated by two photon dissociation. Therefore, we need to estimate the thermochemical
heats of reaction (ΔHrxn) of the reactions that may give rise to the observed products. As
discussed in Section 5.4 , both two- and three-body fragmentation processes are energetically
121

allowed. ΔHrxn values for the relevant reactions, which are accurate to within 400 cm
-1
, are listed
in cm
−1
below.
CH3COCOOH → CH3CO + HOCO ∆Hrxn = 28,600 (1)
CH3COCOOH → CH3CO + H + CO2 ∆Hrxn = 28,900 (1a)
CH3COCOOH → CH3CO + OH + CO ∆Hrxn = 37,600 (1b)
CH3COCOOH → CH3 + CO + HOCO ∆Hrxn = 31,600 (1c)
CH3COCOOH → CH2CO + H + HOCO ∆Hrxn = 42,600 (1d)
CH3COCOOH → CH3COCOO + H ∆Hrxn = 39,000 (2)
CH3COCOOH → CH2COCOOH + H ∆Hrxn = 33,000 (3)
CH3COCOOH → CH3COCO +OH ∆Hrxn = 36,700 (4)
ΔHrxn for reaction 1 has been calculated by Gabriel da Silva (private communication) and
for reactions 2 and 3 by David Osborn (private communication). ΔHrxn for reactions 1a −1c are
estimated, respectively, by adding the reported ΔHrxn values of HOCO → H + CO2, HOCO →
OH + CO, and CH3CO → CH3 + CO reactions to ΔHrxn of reaction 1.
15,16
Alternatively, ΔHrxn
for these reactions can be estimated with similar results (within 400 cm
−1
or better) from the
heats of formation (ΔHf
0
) of reactants and products (see Table A7 in the Appendices). ΔHrxn for
reaction 1d is estimated from Table A7. ΔHrxn for reaction 4 is estimated from ref 17.
122

The KER distributions of CH3CO and H fragments and their corresponding recoil
anisotropy parameters, β, are shown in Figures 5.5 and 5.6. Center-of-mass (c.m.) KER values
are not given, because each product can be formed via more than one pathway.
Figure 5.5: KER distributions (top) and corresponding recoil anisotropy parameters β (bottom)
of the CH3CO fragments following S1 excitation in peaks 1, 3, 6, and 9 of the PFY spectrum in
Figure 5.2 are shown in red, green, black, and blue, respectively. The KER distributions are
normalized to the peak at low KER. The predicted KEmax values of CH3CO produced via
synchronous and sequential three-body fragmentation processes are marked by arrows, as
discussed in Section 5.4.3.
123

Figure 5.6: KER distributions (top) and corresponding recoil anisotropy parameters β (bottom)
of H fragments following S1 excitation in peaks 1, 3, 6, and 9 of the PFY spectrum in Figure 5.2
are shown in red, green, black, and blue, respectively. The predicted KEmax values for
synchronous and sequential three-body fragmentation processes are marked by arrows, as
discussed in Section 5.4.3.
The KER distributions of CH3CO (Figure 5.5) exhibit two major components: a slow
component peaking at around 200 cm
−1
, and a fast component with a maximum around 2500
cm
−1
. In addition, there is a tail that extends up to about 6500 cm
−1
. The relative contribution of
the high KER component increases with excitation energy, and the distribution becomes clearly
bimodal. The β parameter calculated for each of the distributions is shown as a function of KER
in the lower panel of Figure 5.5. At low kinetic energies (KEs), the average value of β increases
124

from 0.2 to 0.5 with increasing excitation energy. However, once the high KER component
becomes dominant, its β value stays constant at 1.4 −1.5 for all excitation energies.
Similar to CH3CO, the H-photofragment KER distributions are broad and multimodal, as
shown in Figure 5.6. They extend up to ∼25 000 cm
−1
with three apparent features centered at
about 1500, 5000, and 13 000 cm
−1
. The angular distribution of the lowest KER component is
nearly isotropic (β = 0 to −0.2) but it progressively becomes more anisotropic at higher KEs (β =
−0.4 to −0.6).
Figure 5.7: KER distribution of CO fragments (X
1
Σ
+
, v"=0, bandhead) detected by 2+1 REMPI
at 230.1 nm following the excitation of PA in peak 1 of Figure 5.2. The predicted KEmax values
for synchronous and sequential three-body fragmentation processes are marked by arrows, as
discussed in Section 5.4.3.
125

Other detected products were CO, HOCO, and CH3. The KER distribution of ground
state CO (X
1
Σ
+
, v" = 0, bandhead) was recorded following excitation of PA in peak 1 of Figure
5.2 and is shown in Figure 5.7. A small signal at mass 45 was observed in a one-color
experiment and assigned to HOCO. Its measured KER distribution is shown in Figure 5.8.
Figure 5.8: KER distribution of HOCO following S1 excitation in peak 9 of Figure 5.2. The
predicted KEmax values of HOCO produced by three-body synchronous and sequential
fragmentation processes are marked by arrows, as discussed in Section 5.4.3.
We were able to detect CH3 radicals in one-color experiments. The 2+1 REMPI spectra
of CH3 indicate the formation of CH3 products in the ground (v1" v2" v3" v4" = 0000) and excited
(v1" v2" v3" v4" = 0100) vibrational levels. Images of CH3 (𝑋
̃
2
A2", 0000) and CH3 (𝑋
̃
2
A2", 0100)
were recorded at 333.77 (0
0
0
) and 329.95 (2
1
1
) nm, respectively, and the corresponding KER
distributions are shown in Figure 5.9.
126

Figure 5.9: KER distributions of CH3 (𝑿
̃
2
A2", 0000, black) and CH3 (𝑿
̃
2
A2", 0100, red)
recorded at 333.77 and 329.95 nm, respectively.
Attempts to detect CH3 in a two-color pump-probe experiment were unsuccessful
because the large probe-only signal masks the pump −probe signal of the CH3 product. This is
because the absorption cross-section of PA at the probe wavelength (325 −336 nm) is much
greater than that at the 369 −374 nm pump wavelength.
8

All the detected fragments exhibit high KEs, much higher than allowed in one-photon
absorption, while their PFY spectra have features characteristic of absorption to the S1 state. We
conclude, therefore, that the products are formed by two-photon absorption via the S1 state. This
interpretation is supported by the quadratic dependence of the H-photofragment signal on pump
127

laser fluence. The excitation and dissociation mechanisms leading to the observed products are
discussed in detail in the next section.
5.4 Discussion
5.4.1 Electronic Structure Calculations of Excited States of PA.
As discussed above, the observed products are generated by two-photon absorption
followed by dissociation of PA. In order to characterize the excited electronic states accessed by
two-photon absorption, the vertical excitation energies from S0 to the S1-S3 states, the oscillator
strengths of the relevant transitions, and the directions of their transition dipole moments have
been calculated at the EOM-EE-CCSD/6-311(2+)G** level of theory, and are listed in Table 5.1.
The vertical excitation energies from S0 to S1 −S3 are in good agreement with previous
calculations.
17

128

Table 5.1: Calculated vertical excitation energies for the indicated transitions, transition dipole
moment directions, and oscillator strengths calculated using EOM-EE-CCSD/6-311(2+)G**.
The X, Y, and Z axes are defined with respect to the molecular plane, as shown in Figure 5.3.

The two-photon excitation energy reached via the origin (0 −0) band of S1 is 53,420 cm
−1

(see Section 5.3.2), which is higher than the calculated vertical S2 ← S0 excitation energy
(45,160 cm
−1
); thus, the S2 state can be accessed by two photon absorption via S1 at all the
excitation energies. The oscillator strength for the S2 ← S1 transition is much larger than that for
S3 ← S1, making this the most likely excitation process. Although the calculated S3 ← S0 vertical
excitation energy is higher than the lowest two-photon excitation energy employed in this study
(53,420 cm
−1
), Dhanya et al. observed the formation of OH products in the photodissociation of
PA at 193 nm (51,813 cm
−1
) and concluded that both the S2 and S3 states can be reached at this
energy.
17
In addition, we note that the S2 ← S1 transition is parallel whereas the S3 ← S1
Initial
state
Final
state
Transition
type
Vertical Excitation
Energy/cm
-1

Transition dipole
direction
Oscillator
Strength
S0 S1 π*←n+ 29,850 Perpendicular (Z) 6.0x10
-5

S0 S2 π*←n- 45,160 Perpendicular (Z) 1.6x10
-4

S0 S3 π*←π 56,500 Parallel (X,Y) 7.0x10
-2

S1 S2 11,600 Parallel (X,Y) 4.3x10
-2

S1 S3 26,050 Perpendicular (Z) 2.8x10
-5

129

transition is perpendicular. In the case of fast dissociation and axial recoil, the former should
give rise to a recoil anisotropy parameter βmax = 2, whereas for the latter βmax = −1 is expected. In
Section 5.4.3 we examine in detail the observed KER distributions and the recoil anisotropy in
the photofragment images.
5.4.2 S1 ← S0 Spectroscopy
In order to identify the origin band of the S1 ← S0 transition, H-PFY spectra were
recorded at different sample temperatures, carrier gases, and backing pressures. The lowest
energy peak (marked as 1 in Figure 5.2) appears consistently at 26,710 cm
−1
under all conditions.
We therefore assign this peak as the origin (0 −0) band of the S1 ← S0 transition of the Tc
conformer. Pressure and temperature dependence studies do not reveal hot bands in the H-PFY
spectrum. The other resolved transitions and their positions relative to the 0 −0 band are listed in
Table 5.2. Because the vibrational frequencies in the S1 state of PA are unknown, only tentative
assignments of the peaks in the H-PFY spectrum can be made at this time (see Figure 5.2). They
are based on vibrational frequencies in the S0 state and the geometries of S0 and S1. The relevant
normal modes of S0 are listed in Table 5.3.

130

Table 5.2: Positions and tentative assignments of the H-PFY peaks marked in Figure 5.2. The
notations of the normal modes are taken from from ref. 14. The transitions marked as
“unassigned” are probably due to splittings in internal rotor transitions of the CH3 moiety.

The frequencies of the C5-C7 (υ24) and CH3 (υ23) torsional modes in S0 are 90 and 134
cm
-1
, respectively.
18
The electronic structure calculations of Chang et al. show that the S1 ← S0
transition can be classified as π*←n+, with the electron density in the S1 state residing mainly
along the central C5 −C7 bond.
19
In addition, the calculated geometries of the S0 and S1 minima
Peaks Positions / cm
-1
Difference from 0-0 / cm
-1
Tentative assignment
1 26,710 0 0-0
2 26,832 122 υ23 (CH3 torsion)
3 26,847 137 υ24 (C-C torsion)
4 26,944 234 unassigned
5 26,952 242 2υ23
6 26,960 250 unassigned
7 26,979 269 2υ24
8 27,006 296 υ16 (CCC bending)
9 27,050 340 unassigned
10 27,063 353 unassigned
11 27,072 362 unassigned
12 27,078 368 3υ23
13 27,084 374 unassigned
14 27,094 384 CH3 torsion/CCO
bending
131

reveal that the C5 −C7 bond is shortened significantly in S1 compared to S0. Therefore, it is
reasonable to expect that the frequency of the torsional motion along the C5 −C7 bond will
increase in the S1 state. A similar situation has been seen in glyoxal and methylglyoxal where the
fundamental frequency of the corresponding torsional mode in S1 is nearly double that in S0.
20

On the basis of these observations, we tentatively assign peak 3 at 137 cm
−1
relative to the 0-0
band as the C5 −C7 torsion, υ24, and peak 7 at 269 cm
−1
relative to the 0 −0 band as its overtone.
The peak at 122 cm
−1
relative to the 0 −0 band (peak 2) is tentatively assigned as the fundamental
of the CH3 torsional mode, υ23.
Table 5.3: Relevant vibrational frequencies of ground state PA taken from ref. 14 (and
references therein).
14

Vibrational modes Frequencies / cm
-1

C-C torsion 90
CH3 torsion 134
CCC bend 246
CCO bend 384
C=O ketone rock 392
C=O acid bend 521
Tentative assignments of additional peaks marked in Figure 5.2 are listed in Table 5.2. It
is evident from Figure 5.2 that the two low frequency torsional modes cannot account for all the
observed peaks in the PFY spectrum up to 400 cm
−1
above the 0 −0 peak. The unassigned peaks
may correspond to internal rotations of the methyl group, which cause additional splittings of the
bands. The role of methyl internal rotors in molecular spectroscopy has been studied extensively
132

both experimentally and theoretically.
20–23
It is shown that the nondegenerate internal rotational
modes of the methyl group give rise to split peaks that complicate spectral assignments. For
example, peaks 4 and 6 (at 26,944 and 26,960 cm
−1
, respectively) might represent additional
splittings of 2υ23 (peak 5; 26,952 cm
−1
) caused by internal rotation of CH3. The same might be
true for peaks 10, 11, 13, and 14, which are adjacent to 3υ23 (peak 12 at 27,078 cm
−1
) with
almost equal spacings. Peak 14 at 27,094 cm
−1
, however, can be assigned also as the transition
terminating in the CCO bending fundamental of S1 (υ15). High-level electronic structure
calculations that include anharmonic vibrational frequencies and internal rotation of the methyl
group are required in order to offer definite assignments.
5.4.3 Two-Photon Dissociation.
The ΔHrxn values listed in Section 5.3.2 indicate that both two- and three-body
fragmentation processes are energetically accessible following two-photon absorption. In order
to distinguish between the contributions of these processes, we turn our attention to the KER
plots shown in Figures 5.5 − 5.8, focusing on the observed maximum KER values, KEmax, for
each fragment.
5.4.3.A Two-Body Fragmentation Pathways.
The maximum kinetic energy value, KEmax, of each fragment in the two body
fragmentation reactions 1 −4 will be observed when the fragments have no internal energy. In the
case of reaction 1, the maximum allowed c.m. kinetic energy is 53,400 – 28,600 cm
−1
= 24,800
cm
−1
for excitation via the ground vibrational level of S1. In accordance with momentum
conservation, the CH3CO fragment can have KEmax = 12,700 cm
−1
. Recalling that CH3CO is not
monitored state selectively, if CH3CO is generated with internal energy close to its dissociation
133

barrier of ∼6000 cm
−1
, then its KE will reach only about 9600 cm
−1
. The observed KEmax of
CH3CO, however, is only 6500 ± 500 cm
−1
(Figure 5.5), indicating that the HOCO co-fragment
possesses significant internal energy. In the complementary KER measurement of HOCO shown
in Figure 5.8, we find that the observed KEmax is 5000 ± 500 cm
−1
(recorded at peak 4), which is,
again, much smaller than the 12,100 cm
−1
value expected for cold CH3CO co-fragments . This
confirms that most CH3CO fragments are internally hot. HOCO and CH3CO fragments with
internal energies exceeding their dissociation barriers will further dissociate leading to three-
body fragmentation and these processes are discussed in the next section.
Primary H photofragments generated by reactions 2 and 3 have expected KEmax values of
14,000 ± 400 and 20,000 ± 400 cm
−1
, respectively. Therefore, whereas C −H and O −H bond
fission reactions might contribute to the observed KER distribution of H fragments (Figure 5.6),
they are not likely to explain the observed KEmax of 23,000 cm
−1
. We conclude that two-body
fragmentation processes alone cannot explain the observed KEmax of CH3CO, HOCO, and H
products, and we turn our attention to the contribution of three-body fragmentation processes.
5.4.3.B Three-Body Fragmentation Pathways
Synchronous and Sequential. Three-body fragmentation processes have been discussed in
the literature extensively, and here we follow the treatment of Maul and Gericke, who distinguish
among what they term synchronous, sequential, and asynchronous processes.
24,25
Synchronous
three-body fragmentation refers to the simultaneous breaking of two bonds to produce three
fragments. Sequential dissociation involves the initial breaking of one bond that results in two
primary products, one of which has enough internal excitation to further dissociate. The
asynchronous dissociation process is less well-defined; it involves breaking of the two bonds on
134

time scales that differ by less than a rotational period of the molecule. This process will only be
mentioned briefly in our discussion. Because we follow closely the treatment of Maul and
Gericke, the equations describing the KE released in these processes are listed in the Appendices
(Section II, equations AI-AXIV) and the plots generated by using these equations are shown in
Figures 5.10 − 5.12. Our goal is to estimate KEmax for the different pathways and compare the
values with the observed KER plot of each fragment.
5.4.3.B.I Synchronous Three-Body Fragmentation
Obviously, KEmax will be observed when the three fragments are born with no internal
energy. Therefore, we discuss here the fragmentation of an ABC molecule, where fragments A,
B, and C have only KE. For synchronous dissociation, the fragments’ KER distributions depend
on the central A −B −C angle at the critical molecular geometry at the moment of fragmentation.
In the case of PA, this geometry is unknown, but the anisotropy of the angular distributions
(Figures 5.5 and 5.6) suggests that the final dissociation on S2 (or S3) is fast and direct. We
therefore assume that the critical angle α during dissociation is similar to the corresponding bond
angle in the S1 state. Energy and linear momentum conservation dictate that
𝑝 𝐴 ⃗⃗⃗⃗ + 𝑝 𝐵 ⃗⃗⃗⃗ + 𝑝 𝐶 ⃗⃗⃗⃗ = 0 (5.1)
𝐸 𝐴 𝑘𝑖𝑛
+ 𝐸 𝐵 𝑘𝑖𝑛
+ 𝐸 𝐶 𝑘𝑖𝑛
= ℎ𝜈 − 𝛥 𝐻 𝑟𝑥𝑛 − ∑ 𝐸 𝑖𝑛𝑡 𝐴 ,𝐵 ,𝐶 = 𝜖 , (5.2)
where ℎ𝜈 refers to the two photon energy; ΔH
rxn
is the heat of reaction of the ABC molecule
fragmenting into A, B and C and leading to KER of 𝐸 A
kin
, 𝐸 B
kin
, and 𝐸 C
kin
, respectively; Eint is the
internal energy of the fragments, assumed here to be zero; and 𝜖 is the total available KE. The
above equations can be solved to obtain the fragments’ KEsas a function of the angle  projected
135

on the central atom The analytical expressions for 𝐸 𝐴 𝑘𝑖𝑛
,𝐸 𝐵 𝑘𝑖𝑛
, and 𝐸 𝐶 𝑘𝑖𝑛
with respect to  are
listed in the Appendices (equations AI-AIII).
For reactions 1b and 1c, α is defined as the angle C4C5C7 and C5C7O9, respectively,
whereas α for reaction 1a can be estimated from the two angles C5C7O9 and C7O9H10 (see
Figure 5.10). A previous theoretical study reported that the C4C5C7 and C5C7O9 angles in S 1
are 119° and 115°, respectively, and these define the values of α for reactions 1b and 1c.
19
The
C7O9H10 angle is 107°, leading to α = 42° for reaction 1a (see Figure 5.10) Similarly, α for
reaction 1d is calculated to be 49° considering the H1C4C5 (109°) and C4C5C7 (120°) angles
and breaking of the C4 −H1 bond, which lies in the molecular plane (Figure 5.10). We note that
reaction 1d can also proceed by breaking the out-of-plane C4 −H2 and C4 −H3 bonds, which will
result in different α values. However, we predict that KEmax of the H1, H2, and H3 fragments
would be similar, because these values change only slightly with α for reaction 1d (Figure 5.10).
The KEmax value for each fragment is obtained in this model by using ΔH
rxn
for the appropriate
reaction producing the fragment. These values are marked by arrows in Figures 5.5 −5.8, listed in
Table 5.4, and also shown in vertical dotted lines in Figure 5.10

136

Figure 5.10: KEmax of the final products of synchronous three-body fragmentation reactions 1a-
1d as function of α. The values of α for each channel at the moment of fragmentation are
estimated from the equilibrium geometry of the S1 state. The KEmax values of the fragments are
shown by vertical dashed lines.

137

Table 5.4: Estimated angle α and corresponding KEmax values of fragments produced via
synchronous three body-fragmentation reactions 1a-1d. The error bars for KEmax represent the
deviations for a change of ± 5° in the corresponding α value

The estimated KEmax value of the H fragment generated in reaction 1a is close to the
observed one, as shown in Figure 5.6, and so is the corresponding value for HOCO (Figure 5.8)
generated in reaction 1c. The KEmax values for H and HOCO corresponding to reaction 1d are
indicated in Figures 5.6 and 5.8, respectively. The observed values for CO (Figure 5.7) and
CH3CO (Figure 5.5), however, are lower than the predicted KEmax and might signify that either
the fragments possess significant internal energy or they are generated by sequential processes.
Reaction α Fragments KEmax / cm
-1

1a 42° CH3CO 529 ± 1
CO2 1802 ± 50
H 22,753 ± 50
1b 115° CH3CO 3138 ± 140
CO 5565 ± 500
OH 7937 ± 360
1c 119° HOCO 4965 ± 50
CO 8221 ± 860
CH3 14,894 ± 50
1d 49° HOCO 229 ± 1
CH2CO 812 ± 29
H 10300 ± 30

138

5.4.3.B.II Sequential Three-Body Fragmentation
As discussed above, the KER plots displayed in Figures 5.5 and 5.8 indicate that CH3CO
and HOCO fragments produced via reaction 1 have significant internal energies. The observed
KEmax values (6500 and 5000 cm
−1
, respectively) leave up to ∼13,300 cm
−1
to distribute in
internal energy of these fragments. In addition, noting that most CH3CO and HOCO fragments
have fairly low KEs, some of their co-fragments must have internal energies exceeding their
dissociation barriers. ΔHrxn for CH3CO and HOCO dissociation reactions are listed below in
cm
−1
, with barrier heights listed in parentheses, if known.
15,16

CH3CO → CH3 +CO ∆Hrxn = 3000 (6000) (5)
CH3CO → CH2CO + H ∆Hrxn = 14,500 (6)
HOCO → OH + CO ∆Hrxn = 9000 (barrierless) (7)
HOCO → H + CO2 ∆Hrxn = 300 (12,000) (8)
Considering again a triatomic ABC molecular analogue and assuming that, to achieve
KEmax for each fragment, the final A, B, and C fragments must have no internal energy, we now
look at the process: ABC → AB + C → A + B + C. We also take into account that KEmax will be
achieved for collinear velocity vectors of the dissociating fragments, such that
|𝑣 𝐴 | = |𝑣 𝐴 𝐶𝑀
+ 𝑣 𝐴𝐵
| and |𝑣 𝐵 | = |𝑣 𝐵 𝐶𝑀
+ 𝑣 𝐴𝐵
|
The explicit equations, again following Maul and Gericke,
24
are given in the Appendices
(equations AIV −AXIV). The internal energy of AB (𝐸 AB
int
) can be estimated from the c.m. KER in
the first step. Therefore, the predicted KEmax value of each final product is plotted as a function
of the c.m. KER in the first step by using equations AXII −AXIV in the Appendices and shown in
139

Figures 5.11 and 5.12. The maximum possible KER in the first step refers to a scenario where
the nondissociating fragment has zero internal energy and the dissociating fragment has internal
energy just above its dissociation barrier. As the dissociating fragment internal energy increases
above this value, the KE in the first step decreases concomitantly.
Figure 5.11: KEmax of the three fragments as a function of the KER in the first step of the
sequential three-body fragmentation reactions 5 and 6 where hot CH3CO dissociates to give
CH3+CO and CH2CO+H, respectively.

140

Figure 5.12: KEmax of the three fragments as a function of KER in the first step of the sequential
three-body reactions 7 (a) and 8 (b) where hot HOCO dissociates to give OH+CO and H+ CO2,
respectively.
Figures 5.11 and 5.12 show the dependencies of KEmax of the nondissociating fragment
and the KEmax values expected for the products of the secondary dissociation of the HOCO co-
fragment to CO + OH and CO2 + H (reactions 7 and 8). The KEmax values of the products of the
secondary dissociation of CH3CO (reactions 5 and 6), which are correlated with the observed
∼5000 cm
−1
KEmax value of the HOCO fragment (Figure 5.8), are also listed in Table 5.5. The
predicted KEmax values are indicated by arrows in the KER plots of the CH3CO, HOCO, CO, and
H photofragments (Figure 5.5 −5.8). Clearly, the sequential mechanism is consistent with the
observed KEmax values of CH3CO, HOCO, and possibly also CO. products of the second
dissociation step on the c.m. kinetic energy (or the internal energy of the dissociating product)
generated in the first step. In our case, we determined whether a particular three-body
fragmentation reaction is acceptable by comparing the observed and computed values of KEmax
141

for the fragments in the first and second steps. For example, the observed KEmax values of the
CH3CO fragment is ∼6500 cm
−1
(Figure 4). For this value, we list in Table 5.5 the correlated
KEmax values expected for the products of the secondary dissociation of the HOCO co-fragment
to CO + OH and CO2 + H (reactions 7 and 8). The KEmax values of the products of the secondary
dissociation of CH3CO (reactions 5 and 6), which are correlated with the observed ∼5000 cm
−1

KEmax value of the HOCO fragment (Figure 5.8), are also listed in Table 5.5. The predicted
KEmax values are indicated by arrows in the KER plots of the CH3CO, HOCO, CO, and H
photofragments (Figures 5.5 −5.8). Clearly, the sequential mechanism is consistent with the
observed KEmax values of CH3CO, HOCO, and possibly also CO.
Table 5.5: Computed KEmax values of fragments produced in the secondary reactions 5-8,
correlated with the observed KEmax of CH3CO or HOCO generated in reaction 1.
Reaction Observed KEmax of non-
dissociating fragment /cm
-1

Correlated KEmax of fragments from
secondary dissociation step / cm
-1

5 5000 (HOCO) 15,000 (CO)
17,000 (CH3)

6 5000 (HOCO) 5860 (CH2CO)
1940 (H)

7 6500 (CH3CO) 10,000 (CO)
9500 (OH)
8 6500 (CH3CO) 9000 (CO2)
15,000 (H)
142

With these KEmax values of the synchronous and sequential fragmentation reactions in
hand, we can now discuss the likely dissociation pathways following two-photon absorption of
PA.
5.4.3.C Mechanistic Interpretations
The multimodal appearance of the KER distributions of the H and CH3CO fragments
suggests that more than one dissociation pathway contributes to the observed products. As
discussed above, the fwhm of the H-PFY rovibronic bands (∼4 cm
−1
) indicates (taking into
account that each band includes an unresolved rotational envelope) that the lifetime of the S1
state is significantly longer than a picosecond. This lifetime is also consistent with the
observation of light emission on the red side of the PA absorption spectrum. Consequently, PA
in the S1 state may undergo rotational and vibrational motions before absorbing another photon,
and these motions will affect the product recoil anisotropy. Even though the interpretation of
product angular distributions in multiple-photon dissociation is not straightforward, it is likely
that the observed anisotropy is indicative of fast dissociation on S2/S3.
As shown in Table 5.1, the direction of the transition dipole moment for the S2 ← S1
transition is parallel with respect to the molecular plane whereas it is perpendicular for S3 ← S1.
The high KER component of the CH3CO distribution has a positive angular anisotropy
parameter, β = 1.4 −1.5, which is consistent with fast dissociation from the S2 state. The slow
KER component with smaller anisotropy, β = 0.2 −0.5, may derive from dissociation of
vibrationally excited PA in the S1 or T1 states reached via nonradiative transitions from S2. The
vibrational motions of PA involved in the couplings to lower electronic state(s) prior to
dissociation can lead to reduction of the angular anisotropy of products. Internal conversion and
143

intersystem crossings prior to dissociation are common in photodissociation of radicals and large
polyatomic molecules, which often exhibit fast and slow KE components that are anisotropic and
isotropic, respectively.
26–28
This scenario is also consistent with the increase in the fraction of
the fast and anisotropic CH3CO component with increasing excitation energy; at higher
excitation energies, direct dissociation on S2 may compete effectively with couplings to the
lower electronic states.
Analysis of the angular distributions of the H fragment, which shows an overall negative
anisotropy (β = −0.2 to −0.6), is more complicated because several two- and three-body
fragmentation processes can contribute to H formation. For example, the angular distributions of
H fragments produced via two-body O −H bond fission (reaction 2) and synchronous three-body
fragmentation (reaction 1a) will depend on the angle between the transition dipole moment
vector and the dissociating bond. Similarly, the angular distribution of H fragments produced via
reaction 3 will be determined by the orientation of the dissociating methyl C −H bond relative to
the transition dipole moment vector. However, the dissociation of rovibrationally hot HOCO to
H + CO2 (reaction 7) is likely to lead to H products with an isotropic angular distribution and a
broad range of KEs. We also cannot rule out that dissociation via the S3 state (a perpendicular
transition) may contribute to the observed value of β.
Taking into account the available energies for the different reactions, the KEmax values of
the fragments, and the observed KER and angular distributions, we conclude that three-body
fragmentation processes, synchronous and sequential, are major contributors to the CH3CO,
HOCO, H, and CO products generated by two-photon dissociation. At the two-photon energy of
∼54,000 cm
−1
, all the listed three-body fragmentation pathways are energetically allowed. While
144

we have restricted the above discussion of the predicted KEmax values to synchronous and
sequential three-body fragmentation processes, we note that asynchronous three-body
fragmentation might also contribute to the observed KER distributions, but this mechanism is
more difficult to evaluate without theoretical calculations. Nevertheless, analyses of the observed
KEmax values show that three-body fragmentation pathways, both synchronous and sequential,
can explain the observed KER plots of all the products most consistently.
Photoinduced three-body fragmentation has been observed previously in the
photodissociation of molecules such as alkanes, alkenes, carbonyls, alcohols, and carboxylic
acids.
29,29,30,30–34
In fact, it is quite common in molecules of the general formula CH3COR, where
R includes H, alkyl, halogens, cyanide, hydroxyl, and alkoxy.
11,24,25,30–40
For example,
photodissociation studies of acetone reveal a sequential three-body fragmentation pathway,
where nascent CH3CO fragments further dissociate to CH3 and CO.
11,31,41,42
Similar pathways
have been observed in organic aldehydes such as formaldehyde, acetaldehyde, and higher
analogues,
33,34,43
as well as in acetic acid and dihydroxybenzoic acid.
30,38
Synchronous three-
body fragmentation has also been invoked in photodissociation of glyoxal, which is structurally
similar to PA.
35,36,44
In the case of PA, R = COOH and both CO and CH3 fragments can be
observed in a sequential mechanism when the internal energy of the CH3CO co-fragment is
above the barrier. In addition, as seen in Figure 5.5, synchronous processes can contribute to the
lower KE components of the CH3CO KER distribution.
It should also be borne in mind that the predicted KEmax values refer to cases when all the
final dissociation fragments have zero internal energy. It is more realistic that many of the
CH3CO and HOCO products of reaction 1 would have nonzero internal energies. When this
145

internal energy is lower than their dissociation barriers, stable HOCO and CH3CO fragments will
be observed. In general, the observed KER decreases when the reaction products have internal
energy; therefore, internally hot fragments are correlated with KER values lower than the
predicted KEmax.
Although we conclude that the S2 state plays a major role in the formation of the
observed products, the S3 state might also contribute to the photodissociation. However, this
contribution should be much smaller given the small oscillator strength and the perpendicular
direction of the transition dipole moment of the S3 ← S1 transition.
In closing, we address briefly the issue of dissociation following S1 ← S0 excitation.
1–6

In previous studies, CO2 was identified as a final dissociation product following one-photon
excitation. Chang et al. discuss the mechanism of decarboxylation following S1 excitation in
their electronic structure calculations (carried out at the CASSCF and MSCASPT2 levels), which
include conical intersections between the S1, T1, and S0 states.
19
They find that from the
minimum energy geometry on S1, which preserves the internal hydrogen bonding, hydrogen
transfer from the carboxylic OH to the carbonyl O takes place on S1 with a small barrier of 7.2
kcal/mol, resulting in a molecular structure that they call S1-HTMIN. This transient species
accesses efficiently an S1/S0 conical intersection that leads to decarboxylation on S0. However,
they also find that some molecules on S1 reach a conical intersection with T1 prior to H-transfer.
Thus, they envision two minima on S1 separated by a small barrier. It is likely, therefore, that the
second photon absorption observed in the present work takes place from the first minimum,
before H-transfer. The fast optical pumping rate achieved with the high fluence of pulsed and
146

focused laser radiation (see Section 5.2 ) combined with the large oscillator strength of the S2 ←
S1 transition makes this process competitive with H-transfer on S1 followed by decarboxylation.
Our attempts to detect CO2 using 3+1 REMPI schemes were unsuccessful for several
reasons. First, CO2 may be formed in many internal states, and the REMPI detection schemes
that we tried were not sensitive enough to detect it state-selectively.
45
Second, the lifetime of S1
may be long such that only a few molecules dissociate during the observation time window.
Taking into account a molecular beam velocity of ∼1400 m/s and a spot size of the focused
probe laser of about 0.1 mm, we estimate the residence time of the molecules in the interaction
region to be on the order of 100 ns. While the lifetime of the S1 state is unknown, we note that
light emission was observed upon excitation of room temperature samples,1 and the linewidths
observed in this study indicate a lifetime longer than a picosecond. While all these factors may
contribute, it is most likely that the insensitivity of our REMPI detection scheme is the main
reason we are unable to detect CO2. In ongoing flow tube photodissociation experiments of room
temperature PA following S1 ← S0 excitation, we have indeed succeeded in detecting nascent
CO2 products and the isomers of methylhydroxycarbene by using tunable VUV ionization. These
results will be reported separately.
5.5 Summary and Conclusions
The first study of the photodissociation of PA in molecular beams is described providing
information on the S1 ← S0 electronic spectroscopy of the Tc conformer, and the
photodissociation products generated by two-photon dissociation via the S1 state. The main
results and conclusions are summarized below.
147

H-photofragment yield spectra at 330 −380 nm reveal narrow rovibronic bands, which are
well separated up to 400 cm
−1
above the 26, 710 cm
−1
band origin. Progressions in the C −C and
CH3 torsional modes are tentatively assigned. The spectrum is complicated by splittings, which
are probably due to the methyl group, and theoretical calculations are needed to confirm the
assignments. From the linewidths of the vibronic bands, we conclude that the S1 state lives
longer than a picosecond, a result that is supported by the previous observation of light emission
following S1 ← S0 excitation.
Electronic structure calculations show that the oscillator strength of the S2 ← S1
transition is more than 2 orders of magnitude larger than that of the S1 ← S0 transition, while the
S3 ← S1 oscillator strength is much smaller (Table 5.1). We conclude that two-photon transitions
via S1 (mainly to S2) are facile and, in fact, are difficult to suppress in focused laser experiments.
This explains the generation of H, HOCO, CH3CO, CO, and CH3 photodissociation products in
our study.
H, CO, and CH3 fragments were detected by state-selective 2+1 REMPI, and HOCO and
CH3CO, by nonresonant multiphoton ionization. KER distributions of these fragments were
determined, confirming the two-photon nature of the dissociation. The angular distributions of
the H and CH3CO products are anisotropic but show a dependence on KER. The anisotropy
indicates that dissociation on the S2 surface is fast.
Analyses of the energetically allowed dissociation pathways show that both two- and
three-body fragmentation processes are feasible. While we cannot determine branching ratios,
we have gained insight into the dissociation mechanisms by estimating the maximum allowed
148

kinetic energy release, KEmax, of the fragments for synchronous and sequential three-body
fragmentation pathways. These are compared to the computed KEmax values for two-body
fragmentation processes and the observed KER distributions of the fragments. The analyses
show that several three-body fragmentation processes, both synchronous and sequential,
contribute significantly to the observed products. Also, most of the CH3CO and HOCO
fragments are generated with significant internal energy and many of them further dissociate.
These results are hardly surprising in view of the high excess energy available for all the
dissociation pathways.
References
(1) Yamamoto, S.; Back, R. The Photolysis and Thermal Decomposition of Pyruvic Acid in the
Gas Phase. Can. J. Chem. 1985, 63 (2), 549–554.
(2) Vesley, G. F.; Leermakers, P. A. The Photochemistry of α-Keto Acids and α-Keto Esters.
III. Photolysis of Pyruvic Acid in the Vapor Phase. J. Phys. Chem. 1964, 68 (8), 2364–
2366.
(3) Rosenfeld, R. N.; Weiner, B. Energy Disposal in the Photofragmentation of Pyruvic Acid in
the Gas Phase. J. Am. Chem. Soc. 1983, 105 (11), 3485–3488.
(4) Reed Harris, A. E.; Cazaunau, M.; Gratien, A.; Pangui, E.; Doussin, J.-F.; Vaida, V.
Atmospheric Simulation Chamber Studies of the Gas-Phase Photolysis of Pyruvic Acid. J.
Phys. Chem. A 2017, 121 (44), 8348–8358.
(5) Reed Harris, A. E.; Doussin, J.-F.; Carpenter, B. K.; Vaida, V. Gas-Phase Photolysis of
Pyruvic Acid: The Effect of Pressure on Reaction Rates and Products. J. Phys. Chem. A
2016, 120 (51), 10123–10133.
149

(6) Reed Harris, A. E.; Pajunoja, A.; Cazaunau, M.; Gratien, A.; Pangui, E.; Monod, A.;
Griffith, E. C.; Virtanen, A.; Doussin, J.-F.; Vaida, V. Multiphase Photochemistry of
Pyruvic Acid under Atmospheric Conditions. J. Phys. Chem. A 2017, 121 (18), 3327–3339.
(7) Berges, M. G. M.; Warneck, P. Product Quantum Yields for the 350 nm
Photodecomposition of Pyruvic Acid in Air. Berichte Bunsenges. Für Phys. Chem. 1992,
96 (3), 413–416.
(8) Horowitz, A.; Meller, R.; Moortgat, G. K. The UV–VIS Absorption Cross Sections of the
α-Dicarbonyl Compounds: Pyruvic Acid, Biacetyl and Glyoxal. J. Photochem. Photobiol.
Chem. 2001, 146 (1–2), 19–27.
(9) Tjossem, P. J. H.; Smyth, K. C. Multiphoton Excitation Spectroscopy of the B
1
Σ
+
and C
1
Σ
+

Rydberg States of CO. J. Chem. Phys. 1989, 91 (4), 2041–2048.
(10) Hudgens, J. W.; DiGiuseppe, T. G.; Lin, M. C. Two Photon Resonance Enhanced
Multiphoton Ionization Spectroscopy and State Assignments of the Methyl Radical. J.
Chem. Phys. 1983, 79 (2), 571–582.
(11) Trentelman, K. A.; Kable, S. H.; Moss, D. B.; Houston, P. L. Photodissociation Dynamics
of Acetone at 193 nm: Photofragment Internal and Translational Energy Distributions. J.
Chem. Phys. 1989, 91 (12), 7498–7513.
(12) Ryazanov, M.; Reisler, H. Improved Sliced Velocity Map Imaging Apparatus Optimized
for H Photofragments. J. Chem. Phys. 2013, 138 (14), 144201.
(13) Reva, I. D.; Stepanian, S. G.; Adamowicz, L.; Fausto, R. Combined FTIR Matrix Isolation
and Ab Initio Studies of Pyruvic Acid: Proof for Existence of the Second Conformer. J.
Phys. Chem. A 2001, 105 (19), 4773–4780.
150

(14) Plath, K. L.; Takahashi, K.; Skodje, R. T.; Vaida, V. Fundamental and Overtone
Vibrational Spectra of Gas-Phase Pyruvic Acid
†
. J. Phys. Chem. A 2009, 113 (26), 7294–
7303.
(15) Johnson, C. J.; Otto, R.; Continetti, R. E. Spectroscopy and Dynamics of the HOCO
Radical: Insights into the OH + CO → H + CO2 Reaction. Phys Chem Chem Phys 2014, 16
(36), 19091–19105.
(16) Mordaunt, D. H.; Osborn, D. L.; Neumark, D. M. Nonstatistical Unimolecular Dissociation
over a Barrier. J. Chem. Phys. 1998, 108 (6), 2448–2457.
(17) Dhanya, S.; Maity, D. K.; Upadhyaya, H. P.; Kumar, A.; Naik, P. D.; Saini, R. D.
Dynamics of OH Formation in Photodissociation of Pyruvic Acid at 193 nm. J. Chem.
Phys. 2003, 118 (22), 10093–10100.
(18) Hollenstein, H.; Akermann, F.; Günthard, Hs. H. Vibrational Analysis of Pyruvic Acid and
D−, 13C− and 18O-Labelled Species: Matrix Spectra, Assignments, Valence Force Field
and Normal Coordinate Analysis. Spectrochim. Acta Part Mol. Spectrosc. 1978, 34 (11),
1041–1063.
(19) Chang, X.-P.; Fang, Q.; Cui, G. Mechanistic Photodecarboxylation of Pyruvic Acid:
Excited-State Proton Transfer and Three-State Intersection. J. Chem. Phys. 2014, 141 (15),
154311.
(20) Kamei, S.; Okuyama, K.; Abe, H.; Mikami, N.; Ito, M. Mode Selectivity in Intersystem
Crossing: Glyoxal, Methylglyoxal, and Biacetyl. J. Phys. Chem. 1986, 90 (1), 93–100.
(21) Dawadi, M. B.; Perry, D. S. Communication: Conical Intersections between Vibrationally
Adiabatic Surfaces in Methanol. J. Chem. Phys. 2014, 140 (16), 161101.
151

(22) Wang, X.; Perry, D. S. An Internal Coordinate Model of Coupling between the Torsion and
C–H Vibrations in Methanol. J. Chem. Phys. 1998, 109 (24), 10795–10805.
(23) Gurnick, M.; Chaiken, J.; Benson, T.; McDonald, J. D. Vibrational and Rotational
Spectroscopy of the First Electronically Allowed Transition of Α‐dicarbonyls. J. Chem.
Phys. 1981, 74 (1), 99–105.
(24) Maul, C.; Gericke, K.-H. Photo Induced Three Body Decay. Int. Rev. Phys. Chem. 1997, 16
(1), 1–79. https://doi.org/10.1080/014423597230307.
(25) Maul, C.; Gericke, K.-H. Aspects of Photoinduced Molecular Three-Body Decay. J. Phys.
Chem. A 2000, 104 (12), 2531–2541.
(26) Amaral, G.; Xu, K.; Zhang, J. UV Photodissociation Dynamics of Ethyl Radical via the
Ã
2
A′(3s) State. J. Chem. Phys. 2001, 114 (12), 5164–5169.
(27) Song, Y.; Zheng, X.; Zhou, W.; Lucas, M.; Zhang, J. Ultraviolet Photodissociation
Dynamics of the n -Propyl and i -Propyl Radicals. J. Chem. Phys. 2015, 142 (22), 224306.
(28) Sun, G.; Song, Y.; Zhang, J. Ultraviolet Photodissociation Dynamics of 1-Pentyl Radical.
Chin. J. Chem. Phys. 2018, 31 (4), 439–445.
(29) Maul, C.; Haas, T.; Gericke, K.-H. Photoinduced Near Ultraviolet Three Body Decay of
Phosgene. J. Phys. Chem. A 1997, 101 (36), 6619–6632.
(30) Hunnicutt, S. S.; Waits, L. D.; Guest, J. A.
1
(n-π*)-Photochemistry of Acetic Acid at 200
nm: Further Evidence for an Exit Channel Barrier and Reaction Selectivity. J. Phys. Chem.
1991, 95 (2), 562–570.
(31) North, S. W.; Blank, D. A.; Gezelter, J. D.; Longfellow, C. A.; Lee, Y. T. Evidence for
Stepwise Dissociation Dynamics in Acetone at 248 and 193 nm. J. Chem. Phys. 1995, 102
(11), 4447–4460.
152

(32) Tsai, P.-Y.; Chao, M.-H.; Kasai, T.; Lin, K.-C.; Lombardi, A.; Palazzetti, F.; Aquilanti, V.
Roads Leading to Roam. Role of Triple Fragmentation and of Conical Intersections in
Photochemical Reactions: Experiments and Theory on Methyl Formate. Phys Chem Chem
Phys 2014, 16 (7), 2854–2865.
(33) Hobday, N.; Quinn, M. S.; Nauta, K.; Andrews, D. U.; Jordan, M. J. T.; Kable, S. H.
Experimental and Theoretical Investigation of Triple Fragmentation in the
Photodissociation Dynamics of H2CO. J. Phys. Chem. A 2013, 117 (46), 12091–12103.
(34) Morajkar, P.; Bossolasco, A.; Schoemaecker, C.; Fittschen, C. Photolysis of CH3CHO at
248 Nm: Evidence of Triple Fragmentation from Primary Quantum Yield of CH3 and HCO
Radicals and H Atoms. J. Chem. Phys. 2014, 140 (21), 214308.
(35) Osamura, Y.; Schaefer, H. F.; Dupuis, M.; Lester, W. A. A Unimolecular Reaction
ABC→A+B+C Involving Three Product Molecules and a Single Transition State.
Photodissociation of Glyoxal: HCOHCO→H2+CO+CO. J. Chem. Phys. 1981, 75 (12),
5828–5836.
(36) Scuseria, G. E.; Schaefer, H. F. The Unimolecular Triple Dissociation of Glyoxal:
Transition-State Structures Optimized by Configuration Interaction and Coupled Cluster
Methods. J. Am. Chem. Soc. 1989, 111 (20), 7761–7765.
(37) North, S. W.; Marr, A. J.; Furlan, A.; Hall, G. E. Nonintuitive Asymmetry in the Three-
Body Photodissociation of CH3COCN. J. Phys. Chem. A 1997, 101 (49), 9224–9232.
(38) Bagchi, A.; Dyakov, Y. A.; Ni, C.-K. Photodissociation and Photoionization of 2,5-
Dihydroxybenzoic Acid at 193 and 355 nm. J. Chem. Phys. 2010, 133 (24), 244309.
(39) Deshmukh, S.; Hess, W. P. Photodissociation of Acetyl Chloride: Cl and CH3 Quantum
Yields and Energy Distributions. J. Chem. Phys. 1994, 100 (9), 6429–6433.
153

(40) de Wit, G.; Heazlewood, B. R.; Quinn, M. S.; Maccarone, A. T.; Nauta, K.; Reid, S. A.;
Jordan, M. J. T.; Kable, S. H. Product State and Speed Distributions in Photochemical
Triple Fragmentations. Faraday Discuss. 2012, 157, 227.
(41) Kim, S. K.; Pedersen, S.; Zewail, A. H. Direct Femtosecond Observation of the Transient
Intermediate in the Α‐cleavage Reaction of (CH3)2 CO to 2CH3 +CO: Resolving the Issue
of Concertedness. J. Chem. Phys. 1995, 103 (1), 477–480.
(42) Zhong, Q.; Poth, L.; Castleman, A. W. Ultrafast Dissociation Dynamics of Acetone: A
Revisit to the S1 State and 3s Rydberg State. J. Chem. Phys. 1999, 110 (1), 192–196.
(43) Chin, C.-H.; Lee, S.-H. Comparison of Two-Body and Three-Body Decomposition of
Ethanedial, Propanal, Propenal, n -Butane, 1-Butene, and 1,3-Butadiene. J. Chem. Phys.
2012, 136 (2), 024308.
(44) Hepburn, J. W.; Buss, R. J.; Butler, L. J.; Lee, Y. T. Molecular Beam Study of the
Photochemistry of S1 Glyoxal. J. Phys. Chem. 1983, 87 (19), 3638–3641.
(45) Wu, M.; Johnson, P. M. A Study of Some Rydberg States of CO2 by (3+1) Multiphoton
Ionization Spectroscopy. J. Chem. Phys. 1989, 91 (12), 7399–7407.

154

Chapter 6
Future Direction
While each of the photodissociation studies reported in the present work provided a
wealth of information about the photodissociation dynamics and product state distributions of
CH2OH, CO2, and pyruvic acid (PA), the insights gained from these studies open up also several
prospects for future research. This final chapter focuses on the scientific and technical
improvements that can be implemented in the future.
6.1 Preparation of HCOH and CH3COH at the nozzle tip
As discussed in Chapter 3, UV dissociation from the 3px state can generate both trans-
and cis-HCOH (HC). which can be characterized indirectly by monitoring H co-fragments by
SVMI. The H KER distributions reveal that both the isomers of HCOH are produced with
significant rovibrational excitation. In this experiment, it is difficult to capture and detect HCOH
directly because HCOH with significant internal energy would probably isomerize efficiently to
formaldehyde. I also discussed briefly CH3COH (MHC), the methyl homolog of HC, in Chapter
5 in the context of the photodissociation of pyruvic acid from the S1 state. HC and MHC are
inherently metastable due to their carbene character. As a result, direct spectroscopic detection
and photochemical studies of these species remain challenging.
Schreiner and co-workers produced trans-HC and MHC directly for the first time using
high-vacuum flash pyrolysis of glyoxylic and pyruvic acids, respectively, and captured them in
Ar matrix (11 K).
1,2
They also reported the UV-VIS and IR absorption spectra of both of these
155

species. The theoretical calculations combined with experimental studies suggest that HC and
MHC disappear by tunneling to form more the stable aldehyde isomers. The half-lives of the
vibrationless ground states of HC and MHC in Ar matrix (11 K) are estimated to be 2 and 1
hours, respectively.
In our experimental setup, the stabilization of HC can be achieved by dissociating
CH2OH (and its isotopologs) near the nozzle tip and subsequently cooling down the products in
the supersonic expansion. CH2OH radicals can be produced as before by the Cl + methanol
reaction in the quartz tube extension of the pulsed nozzle. A delayed 266 nm laser beam can be
focused into the effluent at the exit from the tube resulting dissociation of CH2OH to produce
HCOH and CH2O. The nascent HCOH will be immediately cooled down supersonically before
entering the interaction chamber. HCOH does not absorb 266 nm radiation and the HCOD/
CH2O ratio generated from CH2OD is about 0.8. Schreiner and coworkers report a diffuse
absorption spectrum of HCOH(D) at 390-500 nm with a peak at 427 nm.
1
Samanta et al. showed
that the higher excited electronic states of HCOH are predominantly Rydberg in character.
3
The
adiabatic excitation energies of the 3s and 3p cluster of Rydberg states of trans- and cis-HOCH
are estimated to be in the range 4.9-6.8 eV. Therefore, HCOH is not expected to absorb the 266
nm (4.66 eV) pump radiation.
While it is possible to generate MHC via same method by using CH3CHOH, the next
higher homolog of CH2OH, it has been shown that the photodissociation of CH3CHOH leads to
production of acetaldehyde and vinyl-alcohol as the major products instead of MHC.
4

HC and MHC can also be produced by photodissociation of glyoxylic and pyruvic acids,
respectively, at the nozzle tip. Currently, we are pursuing the ambitious goal of generating the
156

transient hydroxycarbene species in the molecular beam. The aim is to develop REMPI detection
schemes for these elusive carbenes and finally investigate their unimolecular photodissociation
dynamics, as discussed in the next section.
6.2 Direct Detection of HC and MHC via REMPI
Once we are able to produce and stabilize HC and MHC, they can be spectroscopically
characterized using suitable REMPI schemes. There are several requirements for a molecular
species to have an efficient REMPI detection scheme. First, the ion of each species should be
stable enough to detect. This requires excitation wavelength that can cause near threshold
ionization where fragmentation of the ion is smaller. However, if the ion is inherently unstable
then it would be difficult to have an efficient REMPI scheme for that species.
The second requirement is that the intermediate state(s) used in the scheme must not
predissociate too much prior to the absorption of consecutive photons to reach the ionization
continuum. Since valence states are often predissociative in nature, they are not usually good
candidates to serve as intermediate states. Instead, we use Rydberg states, which are far from the
molecular core and therefore less predissociative. The geometries of the Rydberg states are often
similar to the geometry of the ground state ion of the species and, therefore, the Frank-Condon
factors of the transitions connecting the Rydberg states and the ground state of the ions are
usually much larger. This makes the REMPI scheme involving the intermediate Rydberg states
efficient. In addition, the wavelengths of visible and UV radiations needed to access the Rydberg
states and subsequent ionization thresholds can be generated easily using Nd YAG pumped dye
lasers.
157

Keeping these factors in mind, Samanta et al. calculated the vertical and adiabatic
excitation energies and oscillator strengths for valence and Rydberg states of cis and trans HC
and MHC using EOM-CCSD method and the aug-cc-pVTZ basis set.
3
The study identified four
Rydberg states, namely, 3s, 3px, 3py, and 3pz, with vertical energies 5-8 eV. In addition, there are
no valence states close in energy to the 3p Rydberg states. The calculated oscillator strengths of
the transitions to the Rydberg states showed that excitation to the 3py state has a much larger
oscillator strength for trans HC and both isomers of MHC. In cis-HC, however, the oscillator
strengths for transitions to 3py and 3px are comparable. Therefore, Rydberg 3p states of HC and
MHC can be accessed via multiphoton excitation schemes and used for REMPI detection. The
study also suggested that the HC and MHC cations are fairly stable to dissociation and
isomerization. Therefore, trans and cis HC can be detected by one-color (2 + 1) REMPI schemes
via the 3py state using 330 −360 nm radiation. For MHC, one-color (1 +1+1) REMPI scheme via
the S1 and 3py states can be achieved at 364 −406 nm. If necessary, two-color schemes can also
be employed with several different combinations of wavelengths.
6.3 Photodissociation of vibrationally excited CO2
In Chapter 4, I discussed the effect of vibrational excitation on the photodissociation of
CO2 at λ >200 nm. The results suggest that the photodissociation of hot CO2 is complicated,
involving multiple surfaces, and the CO product state distributions are correlated with specific
parent vibrational excitation of parent CO2 molecule. However, the previous study cannot
achieve state-specific vibrational excitation of CO2 because the heating in SiC creates a broad
distribution of vibrational excitation up to nearly 20,000 cm
-1
.
158

Figure 6.1: Example of the observed lumpy IR spectrum of CO2 up to ~ 10,000 cm
-1
.
5
The
cross-sections are given in cm
2
/molecule. Each polyad includes different combinations of stretch
and bend vibrations.
Because the photodissociation involves intermediate states with shallow wells and
coupled surfaces, it is expected that state-specific effects with respect to initial vibrational
excitation of the parent CO2 molecule would be important. A more detailed view of the
dissociation mechanism may be obtained by exploiting vibrationally mediated photodissociation
to reach specific vibrational levels of CO2 directly and explore their photodissociation. In this
regard, excitation of combination bands with stretch and bend components should be particularly
fruitful. The IR absorption spectrum of CO2, which has been recorded up to 12000 cm
-1
, shows a
lumpy polyad structure, as shown in Figure 6.1. The vibrational transitions of CO2 are well
known and can be found in the literature.
5–10
A useful list of transitions at 3000-8000 cm
-1
is
159

given, for example, by Miller and Brown.
9
Because of the coupling between the bending and
symmetric stretch modes, vibrational levels appear in clumps or polyads, as seen clearly in the
IR absorption spectrum (Figure 6.1).
Direct IR excitation of overtones and combination bands in the range 3000-8000 cm
-1

can be achieved by using our Laser Vision OPO system (7.5-22 mJ). The vibrationless ground
state of CO2 does not absorb 193 nm wavelengths. However, excitation to the cluster of peaks
that includes the (101) vibrational level (3632 cm
-1
) followed by absorption of 193 nm radiation
should lead to enhancement in the photodissociation yield. With this level of vibrational
excitation, CO2 cannot absorb at λ > 200 nm, because it requires ~4000 cm
-1
of CO2 internal
energy to observe weak absorption at these wavelengths (Chapter 4). Therefore, we can separate
the effect of lower vibrational excitations in translational and pair correlated product state
distributions from higher vibrational excitation. This will provide a much clearer picture of the
vibrationally mediated photodissociation dynamics of CO2. At 193 nm, the available energy is
11,400 cm
-1
with respect to the CO + O(
3
P2) channel, which is sufficient to produce CO (v=0-5)
but not CO + O(
1
D). O(
3
P) and CO (v) products can be detected using established 2+1 REMPI
schemes.
It is also possible to excite CO2 to vibrational levels between 4000 and 7500 cm
-1
and
study the UV dissociation at 230 and 225.6 nm. In this region, v=3 in the CO2 asymmetric
stretch can be accessed.
5,6,9
In two-color experiments, the excitation laser wavelength can be
scanned from 205 to 225 nm, while monitoring CO and O to determine KEDs and internal state
distributions. A challenge will be to deal with background signals from one-color dissociation
and detection, but the absorption cross-section for wavelengths < 225 nm will be much larger, as
160

absorption increases by orders of magnitude with decreasing wavelength.
11–15
In other words,
absorption at 205-225 nm will dwarf absorption at > 225 nm. An insight about the vibrational
state specificity in the photodissociation dynamics can be achieved by monitoring the O(
3
P) and
CO(v) products while scanning the IR excitation laser wavelength.
Bending and asymmetric stretching motions of CO2 are identified as promoting and
tuning modes for conical intersections. The vibrationally medicated UV photodissociation of
CO2 will allow us to study the effect of these modes on product state distributions.
Photodissociation studies involving close-lying levels within a polyad that have different
amounts of asymmetric stretch and bend excitation will be helpful to determine the mode
specificity.
6.4 Generation of VUV Radiation to detect CO2
As discussed in Chapter 5, in previous low- and high-pressure studies CO2 was identified
as a major photodissociation product of PA following excitation to the S1 state.
16–22
However,
our attempts to detect CO2 using published 3+1 REMPI schemes via the 3pσu
1
Пu intermediate
Rydberg state were unsuccessful. This is most likely due to the insensitivity of our REMPI
detection scheme which involves four photons. Therefore, a more efficient REMPI detection
scheme is needed.
The multiphoton ionization (MPI) study of CO2 by Wu and Johnson revealed the
existence of several Rydberg states in the range 275 to 338 nm.
23
The two lowest Rydberg states,
3pσu
1
Пu and 3pπu
1
Δu, can be accessed at 111 and 108.8 nm, respectively. Given that the
ionization energy of CO2 is 13.77 eV, 1+1′ REMPI can be achieved by first exciting the 3pσu
1
Пu
161

and 3pπu
1
Δu Rydberg states using 111 nm and 108.8 nm VUV radiation followed by ionization
using visible radiation.
A major practical challenge in the implementation of the proposed 1+1′ REMPI scheme
is the generation of the required VUV radiation. As discussed in Section 2.4, VUV wavelengths
below 114 nm can not be generated efficiently using the existing tripling cell setup in our lab.
This is either due to a decrease in total pressure of the Ar:Kr gas mixture and/or the poor
transmission of the MgF2 lens at wavelength below 115 nm. One way to circumvent the poor
transmission of the MgF2 lens below 115 nm is to replace it with a LiF lens, which has somewhat
better (20-40%) transmittance in this region, as shown in Figure 6.2. However, using a LiF lens
requires special care because of issues with the quality,
Figure 6.2: UV and VUV transmittance of different materials.
162

orientation, and purity of the LiF material; in addition, the polishing technique can greatly affect
the LiF lens transmission at shorter VUV wavelengths. Also, the LiF lens is quite brittle and
expensive. At this moment, no vendor in the USA supplies VUV grade LiF lens.
Another way to transmit the VUV inside the chamber is to use a waveguide setup. A
waveguide is a borosilicate glass tube cut to a length that takes the VUV radiation into the
ionization zone. In this setup, the MgF2 lens can be replaced by an inexpensive LiF window and
mounted onto the chamber. The waveguide will be attached to the vacuum side of the LiF
window to ensure all the light enters the tube. Total internal reflection conditions ensure that the
light travels down the tube with little loss. The tripling cell and the waveguide can be attached to
each other with a conflat bellows that allows the tripling cell/waveguide assembly to be adjusted
so that the VUV light signal exiting the waveguide is optimized. The basic design of the tripling
cell/waveguide assembly, obtained from Mark Blitz at the University of Leeds (personal
communication), is shown in Figure 6.3.
Figure 6.3: Schematic of waveguide-tripling cell combination. 1. Vacuum chamber 2.
Borosilicate glass tube 3. Conflat bellows 4. LiF window 5. Tripling cell.
A disadvantage of the waveguide setup is that, while simple to implement, it generates
VUV radiation only at the specific wavelength that meets the phase-matching condition, and the
163

radiation is not tunable. However, REMPI schemes often require scanning the excitation
wavelengths, especially when the products are formed in many different internal states. For
example, Figure 4.9 shows that the vibrational excitation of hot CO2 is observed by scanning the
UV radiation in the in 3+1 REMPI scheme at wavelengths around 326 nm. Similarly, a tunable
VUV source is required to successfully implement 1+1′ REMPI schemes.
Recently, Ng and coworkers developed an excellent technique to generate tunable VUV
using Kr or Xe via resonant four-wave mixing,
24–26
as shown schematically in Figure 6.4. In this
technique, two laser beams of frequencies ω1 (UV) and ω2 (visible) are focused into a T-shaped
chamber. A pulsed nozzle operating at 10 Hz fills this chamber with Kr or Xe which acts as the
nonlinear medium.
Figure 6.4: Schematic diagram for the Generation of VUV radiation in the range 6.9-16.0 eV
using Kr or Xe by resonant four-wave mixing. ω1 is fixed at either 249.629 nm (Xe) or 212.556
nm (Kr), while ω2 is scanned from 400 to 900 nm.
During the experiment, ω1 fixed at the appropriate 2ω1 resonant transition frequencies of the
nonlinear medium (249.629 and 212.556 nm for Xe and Kr, respectively). The visible radiation
ω2 is tuned from 400 to 900 nm to generate VUV radiation in the ranges 6.9-11.5 eV and 11.3-
16.0 eV via different-frequency (2ω1 − ω2) and sum-frequency (2ω1 + ω2) generation. ω2=400-
164

900 nm can easily be generated using a Nd:YAG pumped dye laser (Continuum ND6000).
ω1 =249.629 nm can be generated by doubling the frequency of the output of a dye laser
(Continuum ND6000, Coumarin dyes) pumped by the third harmonic output of a pulsed
Nd:YAG laser (Continuum, PL8000). ω1 =212.556 nm can be produced by mixing the
fundamental (637.668 nm) of a dye laser (Continuum ND6000, DCM dye) and its frequency-
doubled output (318.834 nm) in a BBO crystal.
The advantage of the four-wave mixing technique over traditional tripling in a rare gas is that
the output intensity is one or two orders of magnitude higher.
27
The generated 111 and 108.8 nm
can be used to excite CO2 in the 3pσu
1
Пu and 3pπu
1
Δu states, respectively, followed by ionization
using the residual 212.557 nm. By scanning the VUV radiation, we can hopefully monitor the
CO2 fragments which are produced in higher vibrational levels.
References
(1) Schreiner, P. R.; Reisenauer, H. P.; Pickard IV, F. C.; Simmonett, A. C.; Allen, W. D.;
Mátyus, E.; Császár, A. G. Capture of Hydroxymethylene and Its Fast Disappearance
through Tunnelling. Nature 2008, 453, 906.
(2) Schreiner, P. R.; Reisenauer, H. P.; Ley, D.; Gerbig, D.; Wu, C.-H.; Allen, W. D.
Methylhydroxycarbene: Tunneling Control of a Chemical Reaction. Science 2011, 332
(6035), 1300–1303.
(3) Samanta, B. R.; Sutradhar, S.; Fernando, R.; Krylov, A. I.; Reisler, H. Electronic Structure
and Rydberg–Core Interactions in Hydroxycarbene and Methylhydroxycarbene. The
Journal of Physical Chemistry A 2018, 122 (30), 6176–6182.
165

(4) Karpichev, B.; Edwards, L. W.; Wei, J.; Reisler, H. Electronic Spectroscopy and
Photodissociation Dynamics of the 1-Hydroxyethyl Radical CH3CHOH. J. Phys. Chem. A
2008, 112 (3), 412–418.
(5) Rothman, L. S.; Gordon, I. E.; Barber, R. J.; Dothe, H.; Gamache, R. R.; Goldman, A.;
Perevalov, V. I.; Tashkun, S. A.; Tennyson, J. HITEMP, the High-Temperature Molecular
Spectroscopic Database. J. Quant. Spectrosc. Radiat. Transfer 2010, 111 (15), 2139–2150.
(6) Toth, R. A.; Brown, L. R.; Miller, C. E.; Devi, V. M.; Benner, D. C. Spectroscopic
Database of CO2 Line Parameters: 4300–7000 cm
− 1
. J. Quant. Spectrosc. Radiat. Transfer
2008, 109 (6), 906–921.
(7) Régalia-Jarlot, L.; Zéninari, V.; Parvitte, B.; Grossel, A.; Thomas, X.; Von der Heyden,
P.; Durry, G. A Complete Study of the Line Intensities of Four Bands of CO2 around 1.6
and 2.0 μm: A Comparison between Fourier Transform and Diode Laser Measurements. J.
Quant. Spectrosc. Radiat. Transfer 2006, 101 (2), 325–338.
(8) Song, K.-F.; Lu, Y.; Tan, Y.; Gao, B.; Liu, A.-W.; Hu, S.-M. High Sensitivity Cavity Ring
down Spectroscopy of CO2 Overtone Bands near 790 nm. J. Quant. Spectrosc. Radiat.
Transfer 2011, 112 (5), 761–768.
(9) Miller, C. E.; Brown, L. R. Near Infrared Spectroscopy of Carbon Dioxide I.
16
O
12
C
16
O
Line Positions. J. Mol. Spectrosc. 2004, 228 (2), 329–354.
(10) Rothman, L. S.; Hawkins, R. L.; Wattson, R. B.; Gamache, R. R. Energy Levels,
Intensities, and Linewidths of Atmospheric Carbon Dioxide Bands. J. Quant. Spectrosc.
Radiat. Transfer 1992, 48 (5–6), 537–566.
(11) Huestis, D. L.; Berkowitz, J. Critical Evaluation of the Photoabsorption Cross Section of
CO2 from 0.125 to 201.6 nm at Room Temperature; 2010; Vol. 42, p 972.
166

(12) Yoshino, K.; Esmond, J. R.; Sun, Y.; Parkinson, W. H.; Ito, K.; Matsui, T. Absorption
Cross Section Measurements of Carbon Dioxide in the Wavelength Region 118.7–175.5
nm and the Temperature Dependence. J. Quant. Spectrosc. Radiat. Transfer 1996, 55 (1),
53–60.
(13) Parkinson, W. H.; Rufus, J.; Yoshino, K. Absolute Absorption Cross Section
Measurements of CO2 in the Wavelength Region 163–200 nm and the Temperature
Dependence. Chem. Phys. 2003, 290 (2), 251–256.
(14) Jensen, R. J.; Guettler, R. D.; Lyman, J. L. The Ultraviolet Absorption Spectrum of Hot
Carbon Dioxide. Chem. Phys. Lett. 1997, 277 (4), 356–360.
(15) Oehlschlaeger, M. A.; Davidson, D. F.; Jeffries, J. B.; Hanson, R. K. Ultraviolet
Absorption Cross-Sections of Hot Carbon Dioxide. Chemical Physics Letters 2004, 399
(4–6), 490–495.
(16) Yamamoto, S.; Back, R. The Photolysis and Thermal Decomposition of Pyruvic Acid in
the Gas Phase. Canadian journal of chemistry 1985, 63 (2), 549–554.
(17) Vesley, G. F.; Leermakers, P. A. The Photochemistry of α-Keto Acids and α-Keto Esters.
III. Photolysis of Pyruvic Acid in the Vapor Phase. The Journal of Physical Chemistry
1964, 68 (8), 2364–2366.
(18) Rosenfeld, R. N.; Weiner, B. Energy Disposal in the Photofragmentation of Pyruvic Acid
in the Gas Phase. Journal of the American Chemical Society 1983, 105 (11), 3485–3488.
(19) Reed Harris, A. E.; Cazaunau, M.; Gratien, A.; Pangui, E.; Doussin, J.-F.; Vaida, V.
Atmospheric Simulation Chamber Studies of the Gas-Phase Photolysis of Pyruvic Acid.
The Journal of Physical Chemistry A 2017, 121 (44), 8348–8358.
167

(20) Reed Harris, A. E.; Doussin, J.-F.; Carpenter, B. K.; Vaida, V. Gas-Phase Photolysis of
Pyruvic Acid: The Effect of Pressure on Reaction Rates and Products. The Journal of
Physical Chemistry A 2016, 120 (51), 10123–10133.
(21) Reed Harris, A. E.; Pajunoja, A.; Cazaunau, M.; Gratien, A.; Pangui, E.; Monod, A.;
Griffith, E. C.; Virtanen, A.; Doussin, J.-F.; Vaida, V. Multiphase Photochemistry of
Pyruvic Acid under Atmospheric Conditions. The Journal of Physical Chemistry A 2017,
121 (18), 3327–3339. h
(22) Berges, M. G. M.; Warneck, P. Product Quantum Yields for the 350 nm
Photodecomposition of Pyruvic Acid in Air. Berichte der Bunsengesellschaft für
physikalische Chemie 1992, 96 (3), 413–416.
(23) Wu, M.; Johnson, P. M. A Study of Some Rydberg States of CO2 by (3+1) Multiphoton
Ionization Spectroscopy. The Journal of Chemical Physics 1989, 91 (12), 7399–7407.
(24) Gao, H.; Song, Y.; Jackson, W. M.; Ng, C. Y. Communication: State-to-State
Photodissociation Study by the Two-Color VUV-VUV Laser Pump-Probe Time-Slice
Velocity-Map-Imaging-Photoion Method. J. Chem. Phys. 2013, 138 (19), 191102.
(25) Jones, B.; Zhou, J.; Yang, L.; Ng, C. Y. High-Resolution Rydberg Tagging Time-of-Flight
Measurements of Atomic Photofragments by Single-Photon Vacuum Ultraviolet Laser
Excitation. Review of Scientific Instruments 2008, 79 (12), 123106.
(26) Pan, Y.; Gao, H.; Yang, L.; Zhou, J.; Ng, C. Y.; Jackson, W. M. Communication: Vacuum
Ultraviolet Laser Photodissociation Studies of Small Molecules by the Vacuum Ultraviolet
Laser Photoionization Time-Sliced Velocity-Mapped Ion Imaging Method. J. Chem. Phys.
2011, 135 (7), 071101.
168

(27) Hilber, G.; Lago, A.; Wallenstein, R. Broadly Tunable Vacuum-Ultraviolet/Extreme-
Ultraviolet Radiation Generated by Resonant Third-Order Frequency Conversion in
Krypton. J. Opt. Soc. Am. B 1987, 4 (11), 1753.

169

Appendices
I Possible assignments of the CO 2+1 REMPI transition responsible for peak
M at ~ 86,830 cm
-1

I.A Estimation of B(v')-X(v") transitions of CO
Table A1: Band origins of vibronic B(v'=0-9)-X(v"=0) bands of CO. The band origins of the
diagonal (v'=v") vibronic transitions are estimated by subtracting the corresponding vibrational
energies v" of the ground electronic state of CO from Tv'0.
B-X
transitions
(v'-0)
Experimental
Tv'0 / cm
-1

Ground state
vibrational
level, v"
Ground state
vibrational
energies, Ev"
/cm
-1

B-X diagonal
transitions,
B(v')-X(v")
Estimated
band origin,
Tv'0 - Ev"
/cm
-1

2-0 90988 2 4259 2-2 86729
3-0 92792 ± 40 3 6350 3-3 86402-
86482
4-0 94700 ± 120 4 8414 4-4 86166-
86406
5-0 97050 ± 150 5 10452 5-5 86448-
86748
6-0 99680-99745 6 12463 6-6 86916-
87282
7-0 101556 7 15530 7-7 86026
9-0 104996 9 19422 9-9 85554
• The 8-0 band has not been assigned in spectroscopic studies and hence the 8-8 band cannot be
estimated.
170

• The B(v'=2)-X(v"=0) frequency is taken from refs. 1 and 2. B(v'=3,4,5)-X(v"=0) values are taken
from refs. 3 and 4. B(v'=6,7,9)-X(v"=0) values are from ref. 5. E v" values are estimated from the
known spectroscopic constants of the ground state of CO listed in ref. 6
The Figures A1 and A2 present representative 2+1 REMPI simulations (using
PGOPHER) of Boltzmann distributions of specific B-X (v-v) transitions closest to the M band.
Rotational constants of the X and B states are taken from refs. 2 and 6, respectively.

Figure A1: The thermal rotational distribution of the B-X (2-2) vibronic transition simulated
using PGOPHER (black) at 1700 K. The experimental 2+1 REMPI spectrum of the CO (X
1
Σ
+
;
v") product measured with high heating of the CO2 parent is shown in red (see Figure 4.1 in
chapter 4)

171

Figure A2: The thermal rotational distribution of the B-X 5-5 vibronic transition (black) at 1700
K. The experimental 2+1 REMPI spectrum of the CO (X
1
Σ
+
; v") product is shown in red.
I.B k
3
П←X vibronic transitions of CO
In the Tables A2 and A3, the frequencies of the T
v'0
= k(v') ← X(v"=0) transitions (with ZPE
corrected Tv'0 values) were taken from ref. 7. Tv'v" = k(v') ← X(v") = Tv'0 –Ev" values were
estimated using the Ev" vibrational frequencies of CO(X) listed in Table A1.
Table A2: Diagonal k-X (v'-v") vibronic transitions

v" T
v'0
/ cm
-1
T
v'v"
/ cm
-1

0 90333 90333
1 91145 89002
2 91961 87702
3 92784 86434
4 93588 85174
5 94439 83987
6 95160 82697
172

Table A3: Off-diagonal K-X (v'-v") vibronic transitions

The Figure A3 presents a representative simulation (using PGOPHER) of a 500 K
Boltzmann distribution of the k-X (1-2) transition closest to the M band. Rotational constants of
the k state are taken from ref. 7. Note that this is a spin forbidden
3
П-
1
Σ transition, which has
different selection rules than the spin allowed Σ- Σ transition. According to the 2-photon selection
rules, the rotational distributions of П-Σ transitions have O, P, Q, R, S branches. The peak of the
M feature includes mainly J"=5-10. The simulated rotational distributions broaden towards lower
2-photon energies with increasing temperature.
v'-v" T
v'v"
/ cm
-1

0-1 88190
0-2 86074
0-3 83983
1-2 86886
1-3 84795
2-3 85611
3-4 84370
4-5 83136
173

Figure A3: The thermal rotational distributions of the k(v'=1) ← X(v"=2) vibronic transition
simulated using PGOPHER (black) at 500 K. The experimental 2+1 REMPI spectrum of the
CO(X
1
Σ
+
; v") product is shown in red.
I.C C
1
𝚺 𝒈 +
←X vibronic transitions
In the Tables A4 and A5, the Tv'0 values for the C state were taken from ref. 2 and the Tvv
and Tv'v” values were estimated as described for the other transitions.

174

Table A4: Diagonal (v-v) C-X vibronic transitions
v T
v0
/ cm
-1
T
vv
/ cm
-1

0 91919 91919
1 94065 91922
2 96176 91917
3 98244 91894
4 100391 91977

Table A5: Off-diagonal C-X (v'-v") vibronic transitions

v'-v" T
v'v"
/ cm
-1

0-1 89776"
0-2 87660
0-3 85569
1-2 89806
1-3 87715
1-4 85651
175

I.D E
1
П ←X vibronic transitions
The Tv'0 values for the E state are taken from ref. 2.
Table A6: E(v')-X(v") vibronic transitions

v'-v" T
v'v"
/ cm
-1

0-0 92929
1-0 95082
2-0 97153
1-1 92939
0-1 90786
0-2 88670
0-3 86579
1-2 90823
1-3 88732
2-2 92894
176

Table A7: Heats of formation (ΔfH
0
) of the relevant species. Most of the heat of formation
(Hf
0
) values used in the thermochemical estimations are taken from ref. 8, except for the heat of
formation of PA, Hf
0
= -43,550 cm
-1
, which is taken from ref. 9 The errors in ΔfH
0
are less than
50 cm
-1
for all the species listed below.

II Synchronous and sequential three-body fragmentation
Following the procedure given by Maul and Gericke,
10
three-body fragmentation
processes have been simulated for both synchronous and sequential three-body fragmentation of
an ABC triatomic molecule, as described below.
II.A Synchronous fragmentation
For synchronous fragmentation, the energy distribution in the fragments depends on the
central angle A-B-C at the moment of fragmentation (see Figure A4). Using energy and linear
momentum conservation of the final products A, B, and C, as described in equations 5.1 and 5.2
Species ΔfH
0
(0 K) / cm
-1
H 18,055
OH 3100
CO -9500
CO2 -32,900
CH3 12,500
CH3CO -282
CH2CO -3800
HOCO -15,150
CH3COOH -34,980
PA -43,550
177

of the main text, the fragments’ kinetic energies 𝐸 𝐴 𝑘𝑖𝑛
,𝐸 𝐵 𝑘𝑖𝑛
, and 𝐸 𝐶 𝑘𝑖𝑛
are expressed as a function
of  and the total available kinetic energy 𝜖 .
𝐸 A
kin
=
𝜖 [(𝑚 A
+𝑚 C
)/𝑚 C
]+4(𝑚 A
/𝑚 B
)cos
2
(α/2)
(AI)
𝐸 𝐵 𝑘𝑖𝑛
=
𝜖 [{𝑚 𝐵 (𝑚 𝐴 +𝑚 𝐶 )/4𝑚 𝐴 𝑚 𝐶 }𝑠𝑒𝑐 2
(𝛼 /2)]+1
(AII)
𝐸 C
kin
=
𝜖 [(𝑚 A
+𝑚 C
)/𝑚 A
]+4(𝑚 C
/𝑚 B
)cos
2
(α/2)
(AIII)
These equations are used to compute the KEmax values listed in Table 5.4 of the main text.
Figure A4: A schematic depiction of synchronous three-body-fragmentation of a triatomic ABC
molecule where two bonds break simultaneously. The X-axis is defined in the momentum vector
space such that it lies along the central angle bisector of the planar molecule. The energy
disposal in A, B, and C depends on the angle α at the critical configuration.
178

II.B Sequential fragmentation
Unlike the synchronous three-body fragmentation case, where the calculated KEmax
values of the products are determined solely by the angle α, the sequential case depends on the
energy partitioning between the two steps and the fragment velocity vectors. The kinematics of a
sequential three-body fragmentation are shown in Figure A5.

Figure A5: A schematic of sequential three-body fragmentation of ABC where the B-C bond
breaks first at time τ1 producing AB and C fragments. The AB fragment can further dissociate
after time τ2, which is longer than the mean rotational period of the AB intermediate. The
observed laboratory-frame velocities of the A and B fragments depend on the velocity of AB.
Referring to Figure A5, the maximum kinetic energy release in the final products of a
sequential process are derived by using,

179

ℎ𝜈 − 𝛥 H
rxn
𝑠𝑡𝑒𝑝 1
= 𝐸 C
kin
+ 𝐸 C
int
+ 𝐸 AB
kin
+ 𝐸 AB
int
(AIV)
𝐸 AB
int
− 𝛥 H
rxn
𝑠𝑡𝑒𝑝 2
= 𝐸 A
int
+ 𝐸 A
kin,CM
+ 𝐸 B
int
+ 𝐸 B
kin,CM
, (AV)
where 𝛥 H
rxn
𝑠𝑡𝑒𝑝 1
and 𝛥 H
rxn
𝑠𝑡𝑒𝑝 2
are the heats of reaction for the first and second fragmentation
steps, respectively. When the maximum KER release corresponds to zero internal energy in the
fragments, the equations can be simplified to yield:
ℎ𝜈 − 𝛥 H
rxn
𝑠𝑡𝑒𝑝 1
= 𝐸 C
kin
+ 𝐸 AB
kin
+ 𝐸 AB
int
(AVI)
𝐸 AB
int
− 𝛥 H
rxn
𝑠𝑡𝑒𝑝 2
= 𝐸 A
kin,CM
+ 𝐸 B
kin,CM
(AVII)
During each fragmentation step, linear momentum is also conserved in the frame of reference of
each step of bond breaking into two fragments. Therefore,
𝑚 𝐶 𝑣 𝐶 ⃗⃗⃗⃗ + 𝑚 𝐴𝐵
𝑣 AB
⃗⃗⃗⃗⃗⃗⃗ = 0 (AVIII)
𝑚 𝐴 𝑣 𝐴 𝐶𝑀 ⃗⃗⃗⃗⃗⃗⃗
+ 𝑚 𝐵 𝑣 𝐵 𝐶𝑀 ⃗⃗⃗⃗⃗⃗⃗
= 0 (AIX)
Whereas equation (AIX) gives the velocities of A and B in the c.m. frame of AB, their laboratory
frame velocities are obtained from the vector additions:
𝑣 𝐴 ⃗⃗⃗⃗ = 𝑣 𝐴 𝐶𝑀 ⃗⃗⃗⃗⃗⃗⃗
+ 𝑣 𝐴𝐵
⃗⃗⃗⃗⃗⃗ (AX)
𝑣 𝐵 ⃗⃗⃗⃗ = 𝑣 𝐵 𝐶𝑀 ⃗⃗⃗⃗⃗⃗⃗
+ 𝑣 𝐴𝐵
⃗⃗⃗⃗⃗⃗ (AXI)
In the limiting case where the fragments’ velocity vectors in both dissociation steps are collinear
(θ=0 in Figure A5), the maximum velocities of A and B are:
|𝑣 𝐴 | = |𝑣 𝐴 𝐶𝑀
+ 𝑣 𝐴𝐵
| and |𝑣 𝐵 | = |𝑣 𝐵 𝐶𝑀
+ 𝑣 𝐴𝐵
|
180

Thus, the maximum fragment kinetic energies for a given first step AB with internal energy 𝐸 AB
int

are:
𝐾𝐸
max
C
=
(𝑚 A
+𝑚 𝐵 )(ℎ𝜈 −𝛥 H
rxn
𝑠𝑡𝑒𝑝 1
−𝐸 AB
int
)
(𝑚 A
+𝑚 𝐵 +𝑚 C
)
(AXII)
𝐾𝐸
max
𝐴 =
1
2
𝑚 A
(
√
2𝑚 C
(ℎ𝜈 −𝛥 H
rxn
𝑠𝑡𝑒𝑝 2
−𝐸 AB
int
)
(𝑚 A
+𝑚 𝐵 +𝑚 C
)(𝑚 A
+𝑚 𝐵 )
+
√
𝑚 B
(𝐸 AB
int
−𝛥 H
rxn
𝑠𝑡𝑒𝑝 1
)
𝑚 A
(𝑚 A
+𝑚 𝐵 )
)
2
(AXIII)
𝐾𝐸
max
B
=
1
2
𝑚 B
(
√
2𝑚 C
(ℎ𝜈 −𝛥 H
rxn
𝑠𝑡𝑒𝑝 2
−𝐸 AB
int
)
(𝑚 A
+𝑚 𝐵 +𝑚 C
)(𝑚 A
+𝑚 𝐵 )
+
√
𝑚 A
(𝐸 AB
int
−𝛥 H
rxn
𝑠𝑡𝑒𝑝 1
)
𝑚 B
(𝑚 A
+𝑚 𝐵 )
)
2
(AXIV)
These equations are used to generate the values listed in Table 5.5 in Chapter 5.
References
(1) Eidelsberg, M.; Roncin, J.-Y.; Le Floch, A.; Launay, F.; Letzelter, C.; Rostas, J.
Reinvestigation of the Vacuum Ultraviolet Spectrum of CO and Isotopic Species: The
B
1
Σ
+
↔ X
1
Σ
+
Transition. J. Mol. Spectrosc. 1987, 121 (2), 309–336.
(2) Eidelsberg, M.; Rostas, F. Spectroscopic, Absorption and Photodissociation Data for CO
and Isotopic Species between 91 and 115 nm. Astron. Astrophys. 1990, 235, 472–489.
(3) Baker, J.; Tchang‐Brillet, W.-Ü. L.; Julienne, P. S. First Observation of the v= 3 Level of
the B
1
Σ
+
Rydberg State of CO. J. Chem. Phys. 1995, 102 (10), 3956–3961.
(4) Baker, J. The Diffuse v= 4 and 5 Vibrational Levels of the B
1
Σ
+
Rydberg State of Carbon
Monoxide. Chem. Phys. Lett. 2005, 408 (4), 312–316.
(5) Eidelsberg, M.; Launay, F.; Ito, K.; Matsui, T.; Hinnen, P. C.; Reinhold, E.; Ubachs, W.;
Huber, K. P. Rydberg-Valence Interactions of CO, and Spectroscopic Evidence
Characterizing the C′
1
Σ
+
Valence State. J. Chem. Phys. 2004, 121 (1), 292–308.
181

(6) Mina-Camilde, N.; Manzanares, C.; Caballero, J. F. Molecular Constants of Carbon
Monoxide at v = 0, 1, 2, and 3: A Vibrational Spectroscopy Experiment in Physical
Chemistry. J Chem Educ 1996, 73 (8), 804.
(7) Berden, G.; Jongma, R. T.; Meijer, G. A Reanalysis of the k
3
П State of CO. J. Chem.
Phys. 1997, 107 (20), 8303–8310.
(8) Ruscic, B.; Bross, D. Active Thermochemical Tables (ATcT) Values Based on Ver. 1.122
d of the Thermochemical Network. Available ATcT Anl Gov 2018.
(9) Emel’yanenko, V. N.; Turovtsev, V. V.; Fedina, Y. A. Thermodynamic Properties of
Pyruvic Acid and Its Methyl Ester. Thermochim. Acta 2018, 665, 70–75.
(10) Maul, C.; Gericke, K.-H. Photo Induced Three Body Decay. Int. Rev. Phys. Chem. 1997,
16 (1), 1–79.

Abstract (if available)

Abstract The spectroscopy and UV induced photodissociation of hydroxymethyl(CH₂OH/CH₂OD) radical, hot CO₂ and pyruvic acid in molecular beams are studied using time-of-flight (TOF) mass spectrometry and sliced velocity map imaging (SVMI). Rotational, vibrational and electronic states of formaldehyde and cis-hydroxymethylene products are generated in the photodissociation of CH₂OH following excitation of the radical to its 3pₓ and 3pz Rydberg states. With CH₂OD precursors, formaldehyde and hydroxymethylene products are examined separately by monitoring D and H, respectively. Whereas the main dissociation channels lead to formaldehyde and cis-hydroxymethylene in their ground electronic states, at higher excitation energies the kinetic energy distributions (KEDs) of the H and D photofragments exhibit additional peaks, which are assigned as the triplet states of formaldehyde and hydroxymethylene. The rotational excitation of cis-hydroxymethylene depends on the excited Rydberg state of CH₂OD, and is lower in dissociation via the 3pz state than via the lower lying 3pₓ and 3s states. Results obtained with the deuterated isotopologs of CH₂OH demonstrate that the yield of the triplet state of formaldehyde decreases upon increasing deuteration, suggesting that the conical intersection seams that govern the dynamics depend on the degree of deuteration. Vibrational excitation of cis-HCOD, which spans the entire allowed internal energy range, consists mostly of the CO-stretch and CH₂ in-plane bend modes. When the internal energy of cis-HCOD exceeds the dissociation threshold to D + HCO, slow D and H photofragments deriving from secondary dissociation are observed. The yields of these H and D fragments are comparable, and we propose that they are generated from cis-HCOD via prior isomerization to HDCO. The 205-230 nm photodissociation of vibrationally excited CO₂ at temperatures up to 1800 K was studied using Resonance Enhanced Multiphoton Ionization (REMPI) and SVMI. CO₂ molecules seeded in He were heated in a SiC tube attached to a pulsed valve and supersonically expanded to create a molecular beam of rotationally cooled but vibrationally hot CO₂. Photodissociation was observed from vibrationally excited CO₂ with internal energies up to about 20,000 cm⁻¹, and CO(X¹Σ⁺), O(³P) and O(¹D) products were detected by REMPI. The large enhancement in the absorption cross-section with increasing CO₂ vibrational excitation made this investigation feasible. The internal energies of heated CO₂ molecules that absorbed 230 nm radiation were estimated from the kinetic energy release (KER) distributions of CO(X¹Σ⁺) products in v"" = 0. CO₂ internal energies in excess of 16,000 cm⁻¹ were confirmed by observing O(¹D) products. It is likely that initial absorption from levels with high bending excitation accesses both the A¹B₂ and B¹A₂ states, explaining the nearly isotropic angular distributions of the products. The CO product internal energy distributions change with increasing CO₂ temperature, suggesting that more than one dynamical pathway is involved when the internal energy of CO₂ (and the corresponding available energy) increases. The KER distributions of O(¹D) and O(³P) show broad internal energy distributions in the CO(X¹Σ⁺) co-fragment, extending up to the maximum allowed by energy but peaking at low KER values. Although not all the observations can be explained at this time, with the aid of available theoretical studies of CO₂ VUV photodissociation and O+CO recombination, it is proposed that following UV absorption, the two lowest lying triplet states, a³B₂ and b³A₂, and the ground electronic state are involved in the dynamical pathways that lead to product formation. The UV photodissociation of pyruvic acid (PA) is studied in molecular beams using TOF mass spectroscopy and SVMI following excitation to the first absorption band (S₁←S₀) at 330-380 nm. CH₃CO, HOCO, CO, CH₃, and H are detected as photodissociation products. The photofragment yield (PFY) spectrum of the H product is recorded at 350-380 nm in He and Ar carrier gases. The spectrum shows sharp vibrational features reflecting the significant rotational cooling achieved in the molecular beam. It matches well the broad features observed in the room temperature absorption spectrum, and indicates that the S₁ state lives longer than a picosecond. The origin band of the S₁←S₀ transition is identified at 26,710 cm⁻¹, and progressions in the CH₃ and C-C torsional modes are tentatively assigned. KER and angular distributions of CH₃CO, HOCO, CO, CH₃, and H fragments indicate that additional photon absorption from S₁ to the S₂/S₃ states is facile, and is followed by rapid dissociation to the observed fragments. Based on the energetics of the different dissociation pathways and analyses of the observed KER distributions, three-body fragmentation processes are proposed as major contributors to the formation of the observed products.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Dissociation energy and dynamics of HCl and aromatic-water clusters

PDF

Unraveling photodissociation pathways in pyruvic acid and the role of methylhydroxycarbene

PDF

Development and implementation of methods for sliced velocity map imaging. Studies of overtone-induced dissociation and isomerization dynamics of hydroxymethyl radical (CH₂OH and CD₂OH)

PDF

Dissociation energy and dynamics of water clusters

PDF

Photodissociation dynamics of group V hydrides

PDF

Photoelectron and ion imaging investigations of spectroscopy, photoionization, and photodissociation dynamics of diazomethane and diazirine

PDF

A multifaceted investigation of tris(2-phenylpyridine)iridium

PDF

Surface plasmon resonant enhancement of photocatalysis

PDF

Artificial photosynthesis on titanium oxide passivated III-V semiconductors

PDF

Temperature-dependent photoionization and electron pairing in metal nanoclusters

Asset Metadata

Creator Sutradhar, Subhasish (author)

Core Title Photodissociation dynamics of atmospherically relevant small molecules in molecular beams

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry (Chemical Physics)

Publication Date 10/21/2019

Defense Date 07/31/2019

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

Tag carbon dioxide,gas phase,hydroxymethylene, hydroxymethyl,laser radiation,molecular beam,OAI-PMH Harvest,photodissociation,pyruvic acid,radicals,reaction dynamics,small molecules,supersonic expansion,time of flight,unimolecular photodissociation,velocity map imaging,VMI

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Reisler, Hanna (committee chair), Benderskii, Alex (committee member), Vitaly, Kresin (committee member)

Creator Email subhasish263@gmail.com,sutradha@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-226909

Unique identifier UC11674772

Identifier etd-SutradharS-7868.pdf (filename),usctheses-c89-226909 (legacy record id)

Legacy Identifier etd-SutradharS-7868.pdf

Dmrecord 226909

Document Type Dissertation

Rights Sutradhar, Subhasish

Type texts

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

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA