Moderate Resolution CavityRingdown Spectroscopy of

Reactive Intermediates Relevantto Hydrocarbon Oxidation

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor ofPhilosophy in the Graduate School of The Ohio State University

By

Neal Douglas Kline, B.A.

Graduate Program in Chemistry

The Ohio State University

2014

Dissertation Committee:

Professor Terry A. Miller, Advisor

Professor Anne McCoy

Professor Walter Lempert

c© Copyright by

Neal Douglas Kline

2014

Abstract

The oxidation of hydrocarbons affects nearly every living organism on the planet. Two of

the most prevalent examples of this process are the combustion of hydrocarbon fuels and

oxidative degradation of hydrocarbons in the atmosphere. The study of combustion and

atmospheric chemistry is necessary as these are two of the most ubiquitous processes known

and they can affect everything ranging from the economy to our health. In order to fully

understand these very complex processes the reactive intermediates that are formed in them

need to be investigated. Isolating and studying reactive intermediates can help provide

better understanding of complicated reaction mechanisms and the elementary reactions

that make up the mechanism. Positive spectral identification of reactive intermediates can

help one gain greater insight into their structural properties, reaction kinetics, and overall

chemical behavior. However, observing and identifying these species is not trivial. Due to

their inherently reactive nature they are very short lived, exist in very small concentrations,

and require spectroscopic techniques with high sensitivity in order to observe them. Once

they are observed, high level quantum mechanical calculations need to be performed in

order analyze and assign their spectra.

Two classes of reactive intermediates, peroxy radicals and Criegee intermediates, were

the focus of this research and were observed in the near infrared (NIR) using moderate res-

olution cavity ringdown spectroscopy (CRDS). The A − X transition of peroxy radicals is

found in the NIR and was used to identify several peroxy radicals: C6-C10 straight chain per-

oxy radicals (hexyl-decyl peroxy), one branched isomer of octyl peroxy (iso-octyl peroxy),

an OH substituted peroxy radical in 2,1-hydroxypropyl peroxy (2,1-HPP), and two singly

halogenated methyl peroxy radicals in chloromethyl peroxy (CH2ClO2) and bromomethyl

ii

peroxy (CH2BrO2). Spectral assignments for all but the C6-C10 peroxy radicals have been

aided by quantum mechanical calculations to determine origin band position and X and A

state vibrational frequencies; assignments for C6-C10 peroxy radicals were determined by

spectral/structural relationships. We also observed what we believe to be a transition from

the 1A′ state to the low lying 3A′ state of the simplest Criegee interemdiate, methylene

peroxy (CH2O2). While theoretical calculations are ongoing for this molecule to aid its

spectral analysis we have developed an argument for the assignment of the carrier of spec-

trum based on a strong analogy between the electronic structure of methylene peroxy and

ozone, comparison with spectra collected for CH2ClO2 and CH2BrO2, and the dependence

of the spectrum on chemical conditions, specifically its reactivity with SO2.

iii

To faith, family, and friends

iv

Acknowledgments

I would like to acknowledge Dr. Miller, Terrance Codd, Phillip Thomas, Dmitri Melnik,

Jinjun Liu, Becky Gregory, Muang Huang, Rabi Chhantyal-Pun, Gabriel Just, Ming Wei

Chen, and any other group former and current group members with which I have had

contact for helping to grow as an individual and a scientist during my time in graduate

school. I would also like to thank Matt George and Tim George from Rock City church for

helping me to grow in my faith and and giving encouragement when I needed it.

v

Vita

April 15, 1986 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Born—Medina, Ohio

2007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B.A., The College of Wooster, Wooster,Ohio

2008-present . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Graduate Associate, The Ohio State Uni-versity, Columbus, Ohio

Publications

Neal D. Kline and Terry A. Miller. Observation of the A − X Electronic Transition ofC6-C10 Peroxy Radicals. Chem. Phys. Lett, 2014, 610, 149.

Neal D. Kline and Terry A. Miller. Detection and Characterization of Reactive ChemicalIntermediates Using Cavity Ringdown Spectroscopy. Chapter 2 in Cavity Enhanced Spec-troscopy and Sensing, G. Gagliardi and P. Loock, eds., Springer-Verlag, Berlin Heidelberg,2014.

Neal D. Kline and Terry A. Miller. Analysis of the A− X Electronic Transition of the 2,1-Hydroxypropyl Peroxy Radical using Cavity Ringdown Spectroscopy. Chem. Phys. Lett,2012, 530, 16.

Phillip S. Thomas, Neal D. Kline, and Terry A. Miller. A− X Absorption of the PropargylPeroxy Radical(H-C≡C-CH2OO·: A Cavity Ring-Down Spectroscopic and ComputationalStudy. J. Phys. Chem. A, 2010 114, 12437.

Rabi Chhantyal-Pun, Neal D. Kline, Phillip S. Thomas, and Terry A. Miller. Observation ofthe A − X Electronic Transition of β-Hydroxyethylperoxy Radical. J. Phys. Chem. Lett.,2010 1, 1846.

Phillip S. Thomas, Rabi Chhantyal-Pun, Neal D. Kline, Terry A. Miller. The A − XAbsorption of Vinoxy Radical Revisited: Normal and Herzberg-Teller Bands Observed viaCavity Ringdown Spectroscopy. J. Chem. Phys., 2010, 132, 114302.

vi

Fields of Study

Major Field: Chemistry

vii

Table of Contents

PageAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivAcknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vVita . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii

Chapters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xii

1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Peroxy Radicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Criegee Intermediates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

2 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.1 Cavity Ringdown Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142.2 Experimental Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182.3 Chemistry Related to the Production of Reactive Species . . . . . . . . . . 20

2.3.1 2,1-Hydroxy Propylperoxy Radical Production . . . . . . . . . . . . 202.3.2 Production of Straight Chain and Iso-octyl Peroxy Radicals . . . . . 242.3.3 Chloromethyl Peroxy and Bromomethyl Peroxy Production Methods 242.3.4 Methylene Peroxy Production Method . . . . . . . . . . . . . . . . . 26

3 Spectroscopic Observation and Identification of the 2,1-HydroxypropylPeroxy Radical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283.2 Computational Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

3.3.1 Computational Results . . . . . . . . . . . . . . . . . . . . . . . . . . 293.3.2 Overview of Experimental Spectra . . . . . . . . . . . . . . . . . . . 343.3.3 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4 Observation of C6-C10 Peroxy Radicals . . . . . . . . . . . . . . . . . . . . 454.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454.2 Spectral/Structural Relationships . . . . . . . . . . . . . . . . . . . . . . . . 454.3 Peroxy Radical Spectra and Assignments . . . . . . . . . . . . . . . . . . . 47

viii

5 Observation of the A−X Electronic Transitions of Chloromethyl Peroxyand Bromomethyl Peroxy . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605.2 Computational Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

5.3.1 Computational Results . . . . . . . . . . . . . . . . . . . . . . . . . . 635.3.2 Overview of Experimental Spectra . . . . . . . . . . . . . . . . . . . 68

5.4 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 705.4.1 G Conformer Origin and ∠XCOO Torsional Region: 6700-7500 cm−1 705.4.2 G Conformer OO Stretching and T Conformer Origin Region: 7500-

8400 cm−1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 765.4.3 G Conformer Combination Band and T Conformer OO Stretching

Region: 8400-9200 cm−1 . . . . . . . . . . . . . . . . . . . . . . . . . 815.4.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

6 Possible Observation of the a3A′ − X1A′ Electronic Transition of theMethylene Peroxy Criegee Intermediate . . . . . . . . . . . . . . . . . . . 866.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 866.2 Overview of Experimental Spectrum . . . . . . . . . . . . . . . . . . . . . . 936.3 Chemical Evidence for Assigning Carrier . . . . . . . . . . . . . . . . . . . . 93

6.3.1 Test 1: Iodine Atom Absorption . . . . . . . . . . . . . . . . . . . . 936.3.2 Test 2: Determination of Self-Reaction Rate . . . . . . . . . . . . . . 966.3.3 Reaction with SO2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99

6.4 Spectroscopic Evidence for Assigning Carrier . . . . . . . . . . . . . . . . . 1046.4.1 Computational Methods for CH2IO2 . . . . . . . . . . . . . . . . . . 1056.4.2 Computational Results for CH2IO2 . . . . . . . . . . . . . . . . . . . 1066.4.3 Spectral Analysis with G Conformer of CH2IO2 . . . . . . . . . . . . 109

6.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110

7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116

Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

A X and A State Frequencies of 2,1-HPP Conformers . . . . . . . . . . . . 126

B Complete X and A State Geometrical Parameters for the Conformersof Bromomethyl Peroxy and Chloromethyl Peroxy . . . . . . . . . . . . . 130

C Comparison of Production Methods for CH2ClO2 and CH2BrO2 . . . . 135

D Computational Details for CH2IO2 . . . . . . . . . . . . . . . . . . . . . . 138

ix

List of Figures

Figure Page

1.1 Mechanism for low temperature hydrocarbon combustion from Reference [1]. 21.2 Mechanism for OH/O2 addition to propene. . . . . . . . . . . . . . . . . . . 41.3 Methyl peroxy potential energy surfaces of the X, A, and B states as a

function of ROO. From Reference [2]. . . . . . . . . . . . . . . . . . . . . . . 71.4 Ozonolysis mechanism predicted by Criegee.3 . . . . . . . . . . . . . . . . . 81.5 Diagram depicting valence orbitals and electrons of methylene peroxy (left)

and ozone (right).4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.1 Schematic diagram of the CRDS experiment. A pulse of light is trappedin a cavity formed between two highly reflective mirrors with leaked lightbeing detected by a photodiode. Exponential decay ringdown curve that isobtained from the photodiode is shown below the cavity. . . . . . . . . . . . 15

2.2 Setup of the room temperature, moderate resolution CRDS experiment. TheNIR radiation is generated via isolation of second Stokes stimulated Ramanshifting of a visible dye laser output. Radicals are produced inside the ring-down cavity via photolysis using either 193 or 248 nm light from an excimerlaser. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2.3 Reaction showing the synthesis of 2-1-iodopropanol and 1-2-iodopropanol. . 212.4 Reaction scheme for producing 2,1-HPP radical. Top panel shows photoly-

sis of two iodohydrin isomers which yield the corresponding hydroxy propylisomers. Middle panel illustrates an alternative route to the hyroxyl propylradicals via Cl atom abstraction of an H atom from isopropanol. Bottompanel shows the (3-body) addition of O2 to the hydroxyl propyl isomers toyield the 1,2- and 2,-1 HPP radicals. . . . . . . . . . . . . . . . . . . . . . . 23

2.5 Diagram showing formation of straight chain (hexyl-decyl) peroxy radicals(top panel) and iso-octyl peroxy radical (bottom panel). In formation of thestraight chain peroxy radicals n=4-8, m=3-7. . . . . . . . . . . . . . . . . . 25

2.6 Diagram showing the photolytic methods to produce chloromethyl peroxy(top panel) and bromomethyl peroxy (bottom panel). . . . . . . . . . . . . 26

2.7 Mechanism showing production of methylene peroxy. . . . . . . . . . . . . . 27

x

3.1 Structure of four lowest energy conformers of the 2,1-HPP radical with impor-tant geometrical parameters of the radical highlighted for X and A states (Astate values are italicized). Dihedral and bond angle values given in degrees(◦); bond lengths given in angstroms (A). . . . . . . . . . . . . . . . . . . . 32

3.2 Experimental spectra from current study obtained using two different pro-duction methods. The black trace was obtained using the chlorine atomreaction with isopropanol and has been offset by 20 ppm for clarity; the redtrace was obtained by photolyzing the iodopropanol precursor at 248 nm andhas been digitally smoothed. Strong correlation in the 7200-7600 cm−1 regionindicates both spectra a common carrier. There are many additional linespresent in the black trace, most notably from 7800-8300 cm−1 and from 8700-8850 cm−1 which are identified as belonging to HO2. There is also a smallamount of methyl peroxy (CH3O2) present as the origin band at 7383 cm−1

and the 1211 band at 7488 cm−1 are observed. There is also an unidentified

carrier present in the black trace at 8529 cm−1. The 2P1/2-2P3/2 transition

at 7602 cm−1 of the iodine atom is present in the red trace. . . . . . . . . . 363.3 The experimental spectra obtained in this study (iodohydrin photolysis, red

trace) and β-HEP (black trace) are compared with different regions of thespectrum labeled to emphasize similarities between the two traces. The β-HEP trace has been shifted +22 ppm for clarity. . . . . . . . . . . . . . . . 37

3.4 The black trace shows the spectrum obtained in this work. Below the trace isa stick plot of the Franck-Condon simulation from the X state vibrationlesslevel that has been weighted by the Boltzmann factors of the conformers andtheir oscillator strengths. The calculated T00 values of the conformers havebeen shifted to match the experimental origin frequencies. Origin bands ofthe different conformers have been labeled with the A, A′, or A′′ notationand have been color coded; subsequent transitions belonging to each of theconformers have the same color and notation as that of the origin band.Black lettering and no prime corresponds to transitions of the G′1G2G3T4T5

conformer, blue lettering and a single prime corresponds to transitions of theG1G2G3G4G5 conformer, and red lettering with a double prime correspondsto transitions of the G1G2G3T4T5 conformer. Table 3.3 lists of assignmentsand measured frequencies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.5 The 7300-7750 cm−1 region of 2,1-HPP. The top (black) trace was obtained byphotolyzing the iodopropanol precursor with 248 nm light under ambient con-ditions. The bottom (red) trace was obtained by running the iodopropanolprecursor through a slit jet expansion and electric discharge to obtain jetcooled conditions as noted previously in Wu et. al.5,6 Conditions for the jetcooled experiment were a rotational temperature of 15-30 K and negligiblevibrational excitation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

4.1 Graph (adapted from data in reference [7]) depicting the typical carbon num-ber distribution of regular unleaded and premium unleaded gasolines basedon volume percent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

xi

4.2 Graph showing the dependence of A− X origin frequency on peroxy radicalspecies and its structure, based on data presented in reference [8] and resultsfrom the present study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.3 A−X spectra of peroxy radicals from hexane-decane (straight chain) precur-sors. Red (top) trace is hexyl peroxy, blue trace is heptyl peroxy, green traceis octyl peroxy, pink trace is nonyl peroxy, and black (bottom) trace is decylperoxy. In the hexyl peroxy trace there is some interference that appears inthe spectrum from 8230-8489 cm−1 caused by incomplete subtraction fromprecursor absorption in that region. The sharp line structure in the n=8likely arises from HO2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

4.4 7300-8000cm−1 region of decyl peroxy A − X spectrum at different timedelays. Black (top) trace is 5 µs time delay, blue trace is 100 µs, pink traceis 500 µs time delay, and red trace is 1 ms delay. Structure near 7383 cm−1

and 7488 cm−1 are the origin and 1211 bands, respectively, of methyl peroxy.9 52

4.5 A − X spectra of iso-octyl peroxy blue (bottom) trace produced from iso-octane precursor (structure in upper right corner of Figure) and n-octyl per-oxy red (top) trace. Primary hydrogen abstraction sites on iso-octane arelabeled with an “a”, secondary sites with a “b”, and tertiary sites with a “c”. 55

4.6 Time delay tests of A − X spectrum of iso-octyl peroxy. Red (top) trace istaken at a 5 microsecond delay of NIR probe from the excimer photolysisbeam, blue trace taken at 500 microsecond delay of NIR probe, and blacktrace taken at 1 millisecond delay of NIR probe. Labeling of bands signifiesassignment of transition to primary or tertiary isomer. A and B bands belongto primary isomer while A′, B′, and C′ bands belong to tertiary isomer. . . 59

5.1 Structure of G and T conformers of CH2BrO2 and CH2ClO2 with geometricalparameters that change significantly upon excitation from the X state to theA state given in the tables (A state values are italicized). Dihedral and bondangle values given in degrees (◦); bond lengths given in angstroms (A). . . . 67

5.2 Experimental A − X spectra of CH2ClO2 (top, blue trace) and CH2BrO2

(bottom, black trace); CH2ClO2 spectrum has been offset by 7 ppm for clar-ity. The CH2ClO2 trace presented was obtained by 248 nm photolysis ofCH2ClI precursor while the CH2BrO2 trace was obtained by 248 nm pho-tolysis of CH2Br2. There are multiple interferences in both spectra. From7000-7180 cm−1 in both spectra there are interferences from the precursorsand HO2. The 2P1/2-

2P3/2 transition of the iodine atom is present at 7603cm−1 in the CH2ClO2 trace, from 7930-8150 cm−1. HO2 is visible in theCH2BrO2 trace and weakly visible in the CH2ClO2 trace. Interferences dueto precursors are present from 8715-8752 cm−1 and 8698-8817 cm−1 in theCH2ClO2 and CH2BrO2 spectra, and there is additional HO2 interference inthe CH2BrO2 trace from 8891-8980 cm−1. . . . . . . . . . . . . . . . . . . . 69

xii

5.3 Experimental spectra of the G conformer origins of CH2ClO2 (top panel,blue trace) and CH2BrO2 (bottom panel, black trace) compared to simu-lations (red traces, both panels). The simulations were obtained throughusing ab initio the rotational constants in Table 5.1. The simulation forthe G conformers of CH2ClO2 and CH2BrO2 use the ratio of componentsof the transition dipole along the inertial axis of a:b:c=0.48:0.33:0.19 anda:b:c=0.44:0.39:0.17, respectively. . . . . . . . . . . . . . . . . . . . . . . . . 72

5.4 Experimental spectra of CH2ClO2 (top panel) and CH2BrO2 (bottom panel)compared to their Franck-Condon simulations. The calculated T00 values ofthe conformers have been shifted to match the experimental origin frequenciesbut the frequencies of the bands relative to the origin are held fixed to theircalculated values. The red lettering in both panels corresponds to transitionsbelonging to the G conformer of each species. There are interferences fromHO2 and precursor absorption from 6890-7240 cm−1 and 6980-7190 cm−1 inthe CH2ClO2 and CH2BrO2 spectra. Tables 5.6 and 5.7 lists the assignmentsand measured experimental frequencies. . . . . . . . . . . . . . . . . . . . . 73

5.5 Experimental spectrum of CH2ClO2 (top, blue trace) compared to a simu-lation of the the T conformer origin of CH2ClO2 (bottom, red trace). Thesimulation was obtained using the ab initio rotational constants in Table 5.1with a pure c-type transition moment with no contributions from the a andb components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

5.6 Experimental spectra of the CH2BrO2 bands at 7654 cm−1 (top panel) and7719 cm−1 (bottom panel) compared to a simulation of the T conformerorigin. (top, blue trace) compared to a simulation of the the T conformerorigin of CH2ClO2 (bottom, red trace). The black traces in both panels arethe experimental traces while the red traces in both panels were obtainedthrough a simulation using the ab initio rotational constants in Table 5.1using a pure c-type transition moments with no contributions from the aand b components. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

5.7 Experimental spectra of CH2ClO2 (top panel) and CH2BrO2 (bottom panel)compared to their Franck-Condon simulations. The calculated T00 values ofthe T conformers have been shifted to match the experimental origin fre-quencies. As in Figure 5.4, the red lettering in both panels corresponds totransitions belonging to the G conformer of each species, while the dark bluelettering corresponds to trasitions belonging to the T conformer. . . . . . . 80

5.8 Experimental spectra of CH2ClO2 (top panel) and CH2BrO2 (bottom panel)compared to their Franck-Condon simulations. Color coding of transitions isthe same is in in Figure 5.7. . . . . . . . . . . . . . . . . . . . . . . . . . . . 82

5.9 A− X spectrum of methyl peroxy. . . . . . . . . . . . . . . . . . . . . . . . 845.10 A− X spectrum of ethyl peroxy. . . . . . . . . . . . . . . . . . . . . . . . . 85

6.1 General diagram showing reaction path for the CH2X(∗) + O2 reaction. All∆H values (kcal/mole) shown are given relative to an entrance channel valueof zero for CH2X + O2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91

xiii

6.2 Experimental spectrum obtained from the 248 nm photolysis of CH2I2. Thereare interferences from precursor absorption between 6940-7000 cm−1 and8700-8750 cm−1, interference from water between 7060-7350 cm−1, and thethe 2P1/2-

2P3/2 transition of the iodine atom is present at 7603 cm−1. . . . 946.3 The spectra of CH2ClO2 (top, blue trace), CH2BrO2 (middle, black trace),

and of the photoproduct obtained from 248 nm photolysis of CH2I2 in thepresence of O2 (bottom, red trace). The CH2ClO2 and CH2BrO2 have beenoff set by 15 and 7 ppm, respectively, for clarity. . . . . . . . . . . . . . . . 95

6.4 Top panel is the proposed mechanism for the formation of methylene peroxyby 248 nm photolysis of CH2I2. Scatterplot showing the relationship betweenthe intensity of the 2P1/2-

2P3/2 transition of the iodine atom is present at 7603cm−1, the intensity of the photoproduct signal, and the pressure of O2 added.Initial total pressure in cell: 85.0 torr: 0.10 torr CH2I2; 84.9 torr N2 mirrorpurge, window purge, backing N2 on CH2I2. O2 pressures represented by datapoints. Black data points represent iodine atom absorption, blue data pointsrepresent absorption of photoproduct. Error bars represent one standarddeviation of measurements made for one data point as measurements weremade in triplicate. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97

6.5 Time decay profile of photoproduct from the CH2I + O2 reaction. Each colortrace represents a different delay time between NIR probe beam and excimerphotolysis beam: black trace is 1 µs delay, blue trace is 10 µs delay, greentrace is 25 µs delay, red trace is 50 µs delay, cyan trace is 75 µs delay, andpink trace is 100 µs delay. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100

6.6 The experimental traces obtained from the 248 nm photolysis of CH2ClI inthe absence of SO2 (top, blue trace) and presence of 7.0 torr SO2 (bottom,green trace). Blue trace was shifted +7 ppm for clairity. . . . . . . . . . . . 101

6.7 The experimental traces obtained from the 248 nm photolysis of CH2Br2 inthe absence of SO2 (top, black trace) and presence of 7.0 torr SO2 (bottom,green trace). Black trace was shifted +7 ppm for clairity. . . . . . . . . . . 102

6.8 The experimental traces obtained from the 248 nm photolysis of CH2I2 inthe absence of SO2 (top, red trace) and presence of 7.0 torr SO2 (bottom,green trace). Green trace was shifted -40 ppm for clarity. . . . . . . . . . . 103

6.9 Spectra obtained from the CH2I + O2 reaction (top, red trace) comparedto a simulation (bottom, black trace) of the G conformer origin band ofCH2IO2. The simulation was obtained through using ab initio the rota-tional constants in Table 6.3. The simulation for the G conformer of CH2IO2

use a ratio of components of the transition dipole along the inertial axis ofa:b:c=0.48:0.40:0.13. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111

6.10 Experimental spectrum of CH2I + O2 reaction compared to the Franck-Condon simulation of the G conformer of CH2IO2. The origin transition ofthe Franck-Condon simulation has been shifted to match what we considerto be the origin band of the experimental spectrum, but the frequencies ofthe bands relative to the origin are held fixed to their calculated values. . . 112

xiv

B.1 Complete X and A geometric parameters for the G conformer of CH2ClO2 (Astate values are italicized). Dihedral and bond angle values given in degrees(◦); bond lengths given in angstroms (A). . . . . . . . . . . . . . . . . . . . 131

B.2 Complete geometric parameters for the T conformer of CH2ClO2 (A statevalues are italicized). Dihedral and bond angle values given in degrees (◦);bond lengths given in angstroms (A). . . . . . . . . . . . . . . . . . . . . . . 132

B.3 Complete geometric parameters for the G conformer of CH2BrO2 (A statevalues are italicized). Dihedral and bond angle values given in degrees (◦);bond lengths given in angstroms (A). . . . . . . . . . . . . . . . . . . . . . . 133

B.4 Complete geometric parameters for the G conformer of CH2BrO2 (A statevalues are italicized). Dihedral and bond angle values given in degrees (◦);bond lengths given in angstroms (A). . . . . . . . . . . . . . . . . . . . . . . 134

C.1 Comparison between the 248 nm photolysis of CH2ClI (top, blue trace) and193 nm photolysis of CH2Cl2 (bottom, pink trace) production methods forCH2ClO2. The top trace was shifted +8 ppm for clarity. . . . . . . . . . . . 136

C.2 Comparison between the 248 nm photolysis of CH2Br2 (top, black trace) and248 nm photolysis of CH2BrI (bottom, pink trace) production methods forCH2BrO2. The top trace was shifted +8 ppm for clarity. . . . . . . . . . . . 137

D.1 X and A structures of G conformer of CH2IO2, A state values are italicized.Dihedral and bond angle values given in degrees (◦); bond lengths given inangstroms (A). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

xv

List of Tables

Table Page

3.1 Ground state properties of 2,1-HPP conformers: calculated X state energies(cm−1), origin frequencies (T00, cm−1), Boltzmann weights (w), oscillatorstrengths (f ), and relative intensities (I=w x 105 f ) for conformers of 2,1-HPP radical. Energies were computed at G2 level of theory and ZPVEcorrected; oscillator strengths were computed at UCIS/6-31G(d) level. Xstate energies reported are relative to the G1G2G3T4T5 conformer. . . . . . 31

3.2 Predicted X and A state vibrations for the G1G2G3T4T5 conformer of 2,1-HPP. Mode numbering follows Herzberg’s notation and is based on the valuesof the A state frequencies. All vibrational frequencies (cm−1) were calculatedvia UB3LYP/aug-cc-pVTZ and scaled (x0.970). . . . . . . . . . . . . . . . . 33

3.3 Assignments (conformation and vibration) of observed transitions in the A-X spectrum of 2,1-HPP. Band “A” corresponds to origin of G′1G2G3T4T5

conformer and subsequent band labels with no prime correspond to transi-tions belonging to this conformer. Band “A′” corresponds to origin of theG1G2G3G4G5 conformer and subsequent bands with a single prime label cor-respond to transitions belonging to this conformer. Band “A′′” correspondsto origin of the G1G2G3T4T5 conformer and subsequent bands with a doubleprime label correspond to transitions belonging to this conformer. . . . . . 40

4.1 Experimental origin frequencies, COO bend vibrations, and OO stretch bandsof hexyl-decyl peroxy radicals. . . . . . . . . . . . . . . . . . . . . . . . . . . 49

4.2 Spectral assignments for iso-octyl peroxy. . . . . . . . . . . . . . . . . . . . 58

xvi

5.1 Results of theoretical calculations for conformers of CH2BrO2 and CH2ClO2:relative X state energies (cm−1), origin frequencies (T00, cm−1), Boltzmannweights (w), oscillator strengths (f ), and relative intensities (I=w x 105 f ) forconformers of CH2BrO2 and CH2ClO2. Energies for conformers of CH2BrO2

were computed at G2 level of theory and ZPE corrected; while CH2ClO2

energies were done at the UB3LYP/aug-cc-pvtz level of theory and zero pointenergy corrected. X state energies for the respective conformers of CH2ClO2

and CH2BrO2 are reported are relative to the G conformers. The rotationalconstants in the ground and excited states are labeled as A′′/B′′/C ′′ andA′/B′/C ′, respectively. Oscillator strengths, Boltzmann wieghts, and I areunitless; all other values are given in cm−1. . . . . . . . . . . . . . . . . . . 64

5.2 Predicted X and A state vibrations for the G conformer of CH2ClO2. Modenumbering follows Herzberg’s notation and is based on the values of the Astate frequencies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

5.3 Predicted X and A state vibrations for the T conformer of CH2ClO2. Modenumbering follows Herzberg’s notation and is based on the values of the Astate frequencies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

5.4 Predicted X and A state vibrations for the G conformer of CH2BrO2. Modenumbering follows Herzberg’s notation and is based on the values of the Astate frequencies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.5 Predicted X and A state vibrations for the T conformer of CH2BrO2. Modenumbering follows Herzberg’s notation and is based on the values of the Astate frequencies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

5.6 Assignments (conformation and vibration) of observed transitions in the A-Xspectrum of CH2BrO2. Band “A” corresponds to origin of the G conformerand subsequent band labels with no prime correspond to transitions belongingto this conformer. Band “A′” corresponds to origin of the T conformerand subsequent bands with a single prime label correspond to transitionsbelonging to this conformer. . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

5.7 Assignments (conformation and vibration) of observed transitions in the A-Xspectrum of CH2ClO2. Band “A” corresponds to origin of the G conformerand subsequent band labels with no prime correspond to transitions belongingto this conformer. Band “A′” corresponds to origin of the T conformerand subsequent bands with a single prime label correspond to transitionsbelonging to this conformer. . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

6.1 Experimentally determined parameters that were used in Equation 6.4 todetermine the amount of internal energy present in the CH2X

∗ fragment. Allvalues are given in kcal/mole. Numbers in parenthesis in Et column representfraction of Et that is deposited into CH2X

∗ fragment. . . . . . . . . . . . . 896.2 Thermochemical data for the reactions of CH2X* + O2 to form CH2XO2 and

CH2O2 + X, all values in kcal/mole. Values of ∆Hex for CH2X* fragmentswere calculated considering what state the Y fragment is produced in. If theCH2X* fragment has more energy, the Y fragment was produced in the 2P3/2

ground state, if it has less energy the Y fragment was produced in the 2P1/2

excited state. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90

xvii

6.3 Results of theoretical calculations for conformers of CH2IO2: relative Xstate energies (cm−1), calibrated origin frequencies (T00, cm−1), Boltzmannweights (w), oscillator strengths (f ), and relative intensities (I=w x 105 f )for conformers of CH2IO2. X state energies for the respective conformersof CH2IO2 are reported are relative to the G conformers. The rotationalconstants in the ground and excited states are labeled as A′′/B′′/C ′′ andA′/B′/C ′, respectively. Oscillator strengths, Boltzmann wieghts, and I areunitless; all other values are given in cm−1. . . . . . . . . . . . . . . . . . . 107

6.4 Predicted X and A state vibrations for the G conformer of CH2IO2. Modenumbering follows Herzberg’s notation and is based on the values of the Astate frequencies. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

A.1 Predicted X and A state vibrations for the G′1G2G3T4T5 conformer of 2,1-HPP. Mode numbering follows Herzberg’s notation and is based on the valuesof the A state frequencies. All vibrational frequencies (cm−1) were calculatedvia UB3LYP/aug-cc-pVTZ. . . . . . . . . . . . . . . . . . . . . . . . . . . . 127

A.2 Predicted X and A state vibrations for the G1G2G3G4G5 conformer of 2,1-HPP. Mode numbering follows Herzberg’s notation and is based on the valuesof the A state frequencies. All vibrational frequencies (cm−1) were calculatedvia UB3LYP/aug-cc-pVTZ. . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

A.3 Predicted X and A state vibrations for the T1G2G3T4T5 conformer of 2,1-HPP. Mode numbering follows Herzberg’s notation and is based on the valuesof the A state frequencies. All vibrational frequencies (cm−1) were calculatedvia UB3LYP/aug-cc-pVTZ. . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

xviii

Chapter 1

Introduction

1.1 Peroxy Radicals

The immense scientific interest in peroxy radicals stems from their importance in combustion

processes and atmospheric oxidation cycles.10–12 In low temperature combustion (<700 K),

the mechanism of peroxy radical formation begins with the abstraction of a hydrogen atom

from the hydrocarbon fuel (RH) by a hydroxyl radical (OH) to form an alkyl radical (R).

The alkyl radical undergoes a 3-body reaction with molecular oxygen to form an alkyl

peroxy radical (RO2).13–16

RH+OH −→ R+H2O (1.1)

R + O2 +M −→ RO2 +M (1.2)

Once the peroxy radical is formed, the reaction can branch in several ways as shown in Fig.

1.1.1 HO2 can react with it yielding an alkyl hydroperoxide which can decompose into an

alkoxy radical and OH, or the peroxy radical can eliminate HO2 directly to give an alkene.

But perhaps the most interesting path the peroxy radical can take is the terminal oxygen

atom of the peroxy moiety internally abstracting an alkyl hydrogen to give the QOOH

species. The rate of QOOH formation is governed by how large the ring structure inter-

mediate is and the type of the hydrogen atoms present (primary, secondary, tertiary). The

formation of the QOOH species can lead to a chain branching step in which two additional

OH radicals are formed along with an oxy radical. This is significant because if enough of

1

the reactions proceed down this pathway it will be able to sustain and accelerate oxida-

tion and, over time, the radical concentrations and reaction rate will build on themselves.

Eventually this will lead to an explosive increase in radical concentration, oxidation rate,

and temperature to give you autoignition, or ”engine knocking”, which is an uncontrolled

burn of the hydrocarbon fuel and the limiting factor in the thermodynamic efficiency of

the engine. In order to decrease engine knock the molecular structure of the fuel must

be considered with regard to preventing the peroxy radical isomerization to QOOH.17–21

When more branched hydrocarbons are used in the fuel, hydrogen atom transfer will have

to proceed through three and four membered rings which have fairly high ring strain and

will prevent formation of QOOH. The branching will also increase the amount of methyl

groups as side chains which has also been shown to reduce engine knock.17,18RH + OH-H2ORROOQOOH+ O2OOQOOHHOO Q-HOOHHOOQ-HO + OHOQ-HQ + 2OH

+ O2alkyl radicalalkyl peroxy radicalinternal H abstractionhydroperoxy alkyl radicalsecond O2 additioninternal H abstractionketohydroperoxideROOH + O2HO2 + alkeneOH + O-heterocycle RO + OHalkyl hydroperoxidechain branchingdirect HO2 eliminationchain propagation

chain branching+ HO2

Figure 1.1: Mechanism for low temperature hydrocarbon combustion from Reference [1].

2

Peroxy radical chemistry is also important in the atmospheric oxidation of volatile or-

ganic compounds (VOC’s), such as alkanes. According to Guenther22 up to 1300 Tg/year

of hydrocarbons are emitted into the atmosphere from biogenic and anthropogenic sources.

The formation of peroxy radicals in the atmosphere proceeds again through the reaction of

alkyl radicals with molecular oxygen. The alkyl radicals are formed by the abstraction of hy-

drogen from a hydrocarbon by ·OH, or Cl· which are also present in the atmosphere.23 The

alkyl radical then reacts rapidly with molecular oxygen and a third body, M-typically N2,

which stabilizes the peroxy radical product by collision. The presence of peroxy radicals in

the atmosphere has several consequences, particularly in the chemistry of the troposphere.

For example, O3 can be produced in the troposphere, where it is considered a pollutant,

via peroxy radicals’ reaction with NO:

NO+ RO2 −→ NO2 +RO (1.3)

NO2≤430nm−→ NO+O (1.4)

O(3P) + O2 +M −→ O3 +M (1.5)

Peroxy radicals’ reaction with NO produces excess amounts of NO2 in the troposphere

where it can be photolyzed by sunlight to generate oxygen atoms. The oxygen atoms then

react with molecular oxygen to produce tropospheric O312,24 which is considered a health

hazard. A second reaction of peroxy radicals with NO in the troposphere can produce excess

amounts of hydroxyl radical:

NO + RO2 −→ NO2 +RO (1.6)

RO +O2 −→ R′C = O+HO2 (1.7)

HO2 +NO −→ NO2 +OH (1.8)

OH +O3 −→ HO2 +O2 (1.9)

The OH radicals can then migrate from the troposphere to the stratosphere, where O3 is

considered a filter against UV radiation, and reduce the amount of O3 present putting our

health at risk.12

3

In addition to the atmospheric oxidation of alkanes by OH radicals, large quantities

of alkenes emitted into the atmosphere are readily oxidized by them as well and undergo

reactions that can pollute the troposhpere.25–28 The double bond of an alkene can be

attacked by OH to give a hydroxy alkyl radical that can react with molecular oxygen in a

three-body reaction to yield hydroxyperoxy radicals.29–35 In the case of propene, the OH

radical can add to the primary carbon of the double bond or the secondary carbon of the

double bond giving two hydroxyalkyl radical isomers which will produce two hydroxypropyl

peroxy radical isomers (Fig. 1.2).

Figure 1.2: Mechanism for OH/O2 addition to propene.

Propene is an alkene with fairly significant emissions into the atmosphere predominantly

originating from vegetation, fossil fuel combustion, and marine life. There have been several

studies to determine the annual emissions of propene from biogenic sources with estimates

ranging from 1 Tg/year to as much as 20 Tg/year.29,36 Due to the large annual emissions

of propene each year there have been several investigations trying to elucidate the details of

4

its reaction with OH.37–39 Loison and co-workers39 studied the gas phase kinetics of the OH

radical with propene while analyzing the product branching ratios (hydrogen abstraction

vs. addition of OH to the C=C double bond) and the site-specificity of OH addition.

They conducted their experiments at 300 K and determined the hydrogen atom abstraction

channel branching ratio to be <2% (and most likely to be <1%) and also determined the

site specificity of OH addition to be 72%(±15) for addition to the terminal carbon atom

and 28%(±15) for addition to the secondary carbon atom. The OH addition site specificity

reported by Loison is consistent with earlier experimental results reported by Cvetanovic40

and Peeters et al.41

The atmospheric oxidation of propene can also serve as a prototype for larger alkenes,

such as isoprene and other unsaturated hydrocarbons, which are emitted in even greater

quantities. After methane (emission rates of ∼600 Tg/yr),22 isoprene is the most abun-

dant hydrocarbon emitted into the atmosphere with global emission rates reaching between

450-500 Tg/yr and an atmospheric concentration between 1-30 ppb.22,42,43 As with other

alkenes isoprene has a very high reactivity towards OH radical which adds across a double

bond followed by reaction of the resulting radical with molecular oxygen. This in combina-

tion with the high concentration of isoprene in the atmosphere makes the OH/O2+isoprene

reaction an extremely important atmospheric process and it has been the subject of many

theoretical and experimental studies.30,31,34,35,44–51

The combustion of alcohols can also result in the formation of hydroxyperoxy radicals.

1-propanol and isopropanol are two of many oxygenated species including ethers, esters,

and other alcohols that are being investigated as fuel additives due to their ability to

prevent the formation of polyaromatic hydrocarbons (PAH) and soot and their tendency to

improve ignition and combustion performance.52–56 Under combustion conditions the main

reaction pathway for the alcohols will be hydrogen abstraction by OH radical at the α− or

β- hydrogen atom with the newly formed hydroxy alkyl adduct reacting with O2 to yield

hydroxy peroxy radicals.57–60

For many years the best way to spectroscopically observe peroxy radicals was to probe

their B − X absorption located in the UV. Many studies have been conducted citing this

5

transition as a way to identify peroxy radicals and to monitor their self reaction kinetics

using time-resolved UV absorption spectroscopy.61–65 Monitoring the B − X transition of

peroxy radicals also proved to be experimentally facile as it has a relatively large absorp-

tion cross section on the order of 10−18 cm2/molecule. However, there is a considerable

disadvantage to using this transition to distinguish among different peroxy radicals. The

B−X transition involves promotion of an electron from the second highest occupied molec-

ular orbital (HOMO-1) to the singly occupied molecular orbital (SOMO). The HOMO-1 is

bonding and the SOMO is antibonding with respect to the O-O bond. Hence, the B − X

transition involves the promotion of an electron from a bonding orbital to an antibonding

orbital so that when a peroxy radical undergoes a transition to the B state, the O-O bond

dissociates.2 This gives rise to a repulsive B state potential energy surface (Fig. 1.3) and

the resulting spectrum for all peroxy radicals is a broad featureless absorption near 240 nm

which is roughly 40 nm wide. This lack of any observable structure makes the UV transition

a poor diagnostic with which to identify different peroxy radicals.

Alternatively, the A − X transition of peroxy radicals is located in the NIR and is a

HOMO to SOMO transition. The HOMO of peroxy radicals is nonbonding along the O-

O bond making the A − X transition a promotion of an electron from a nonbonding to

antibonding orbital.2 Therefore, the A state is a bound state and upon excitation to the A

state there is only a lengthening of the O-O bond instead of complete dissociation. These

qualities indicate that the A − X spectra of peroxy radicals will be well structured and

characteristic of a given RO2 radical. This allows the differentiation among different peroxy

radicals and makes the A − X transition an excellent diagnostic with which to observe

peroxy radicals. However, observing the A − X transition may be experimentally difficult

as it is based on the highly forbidden a1∆g-X3Σg

− transition of the O2 chromophore.66

Consequently the cross section for this transition is ≈104 times smaller than the B − X

transition. Unfortunately traditional absorption experiments do not have the sensitivity

to detect this weak of a transition. However cavity ringdown spectroscopy, with its long

effective path lengths for light to travel through the absorber (∼105 passes through the

sample in a typical ringdown cavity, with mirrors of >99.99%), can be used to overcome the

6

Figure 1.3: Methyl peroxy potential energy surfaces of the X, A, and B states as a functionof ROO. From Reference [2].

obstacle of studying a trace gas species with a weak electronic transition. This dissertation

will focus on observation of the A− X electronic transition for hexyl-decyl peroxy radicals,

iso-octyl peroxy, (2,1)-HPP, chloromethyl peroxy, and bromomethyl peroxy. All of the

aforementioned peroxy radicals were observed in the NIR using cavity ringdown absorption

spectroscopy. The quantum mechanical calculations that were needed to assign the complex

spectra that was acquired will be described as well. However, due to the large size of hexyl-

decyl peroxy radicals and iso-octyl peroxy, it was not feasible to do the kinds of calculations

that were used on 2,1-HPP, chloromehtyl peroxy, and bromomethyl peroxy. Therefore,

spectral/structural relationships that have been derived through the prior work done on

the A− X transition of smaller peroxy radicals will be used to assign the spectra.

7

1.2 Criegee Intermediates

As discussed in the previous section, the reactions of alkenes with OH radical followed by

reaction with O2 produce tropospheric pollutants that can have serious implications for our

atmosphere. There are other oxidation reactions alkenes can undergo in the atmosphere

including reaction with ozone. The mechanism for ozonolysis was first proposed by Criegee3

in 1949 and as part of the mechanism molecules called Crigee intermediates were predicted

to form.29,67–75 Given the ozonolysis of ethene for example, Fig. 1.4, it has been reasonably

well established that the initiation of ozonolysis reactions involve cycloaddition of O3 to the

C=C double bond to form an energy-rich, five membered ring called a primary ozonide.

Due to its instability, the primary ozonide decomposes by cleaving at the C-C bond and

one of the O-O bonds to produce a stable carbonyl molecule (formaldehyde) and an un-

stable carbonyl oxide (methylene peroxy, CH2O2) also known as a Criegee intermediate.

Subsequent product studies of ozonolysis reactions have supported Criegee’s mechanism by

identifying the main products as the carbonyl compounds predicted by Criegee while also

identifying fragmentation products of Criegee interemdiates (CO, CO2).76–82

Figure 1.4: Ozonolysis mechanism predicted by Criegee.3

8

However, despite Criegee’s proposed mechanism and the indirect evidence that Criegee

intermediates are indeed formed during ozonolyis of alkenes, direct observation of these

radicals eluded scientists for many decades. That changed in 2008 when Taatjes et al.83

observed the mass spectrum of a species with a mass/charge ratio (m/z ) of 46, which could

come from methylene peroxy (molecular formula CH2O2), by reacting CH2S(O)CH3 with

O2. There are several molecules which could be formed from this chemistry and be respon-

sible for the m/z=46 signal including dioxirane, formic acid, ethanol, and dimethyl ether.

To prove the m/z=46 signal was indeed methylene peroxy the photoionization spectrum of

the species was acquired and it was shown not to match the known photoionization spectra

of any of the other molecules and they were ruled out. Then in 2012 Welz et al.84 dis-

covered another synthetic method to produce CH2O2 by reacting CH2I radicals with O2.

Again a signal was observed at m/z=46, the photoionization spectrum was obtained and

it matched the earlier one recorded by Taatjes.83 This discovery has led to many more

spectroscopic investigations of Criegee intermediates in the past two years. Su et al.85 used

IR spectroscopy to investigate the ground state of methylene peroxy while Nakajima and

Endo,86 McCarthy et al.,87 and Daly et al.88 have done microwave studies. Beames et

al.89 and Sheps90 have both studied the B(1A′) - X(1A′) transition of methylene peroxy

located in the UV and recently studies have been extended to include work91–93 on an alkyl

substituted Criegee intermediate, CH3CHO2.

Theoreticians have also taken an interest in studying the electronic nature and structure

of Criegee interemdiates with a focus on methylene peroxy. The first theoretical studies

on methylene peroxy appeared in the literature in the 1970s and concentrated on eluci-

dating its ground state structure which can be represented as either a zwitterion or a

1,3-biradical.4,94–96 The main difference between the two representations is the biradical

structure has been predicted to have equivalent C-O and O-O bond lengths of ∼1.30 A

while the zwitterion is predicted to have a shorter C-O bond (1.26-1.28 A) and a longer O-

O bond(1.34-1.36 A).97–101 Studies by Hiberty,94 Goddard,4,95 and Cremer96,100 indicated

that the ground state of methylene peroxy is a singlet state that is a superposition of a 1,3-

biradical and zwitterionic structure where the biradical character prevails. This thought

9

was accepted to be true up until the high level CCSD(T) study by Cremer et al.101 which,

for the first time, predicted that methylene peroxy was better represented as a zwitterion

than a biradical. The latter prediction turned out to be true as experimental IR studies on

the ground state by Su et al.85 obtained vibrational frequencies more consistent with the

zwitterionic structure and a subsequent microwave study determined C-O and O-O bond

lengths of 1.272(3) A and 1.345(3) A,86 respectively, proving that the zwitterionic structure

dominates over the biradical structure in the ground state.

While there have numerous theoretical and experimental investigations into the ground

state and a few select excited state of methylene peroxy, there are other excited electronic

states that have received little to no attention. Lee and co-workers97 performed a very thor-

ough theoretical investigation of the B− X transition, which was subsequently investigated

by Beames et al.89 and Sheps,90 and in the same study performed some initial calculations

on the A − X transition. One excited electronic state that has gotten no theoretical or

experimental consideration is the 3A′ state, which is the lowest lying excited state above

the 1A′ state. The relationship between the 3A′ and 1A′ states of methylene peroxy is

illustrated in Fig. 1.5 where an analogy to an isoelectronic analogue, ozone, is also made.

10

1A’ 3A’ 1A1 3A21A’3A’ 3A21A1 9553.021(78) cm-1

Figure 1.5: Diagram depicting valence orbitals and electrons of methylene peroxy (left) andozone (right).4

11

In Fig. 1.5 the top panel shows the lewis structures of methylene peroxy and ozone,

directly below the lewis structures are orbital diagram depicting only the p-orbitals and the

electrons in those orbitals with orbitals in the plane of the paper and perpendicular to the

paper being represented by∞ and ◦, respectively. In methylene peroxy and ozone there are

singly occupied p-orbitals on the terminal ends of the molecules perpendicular to the plane

of the paper and, depending on how the electron spins are paired, the molecules can either

be in a singlet or triplet state. The resulting singlet and triplet states would be degenerate

if the two singly occupied p-orbitals had no way to interact, however, in both methylene

peroxy and ozone there is a lone pair of electrons that delocalizes over the entire molecule

providing a mechanism for their interaction.4 This breaks the degeneracy of the singlet and

triplet states with the singlet state being lower in energy in methylene peroxy and ozone.

The absorption spectrum of ozone is well known and contains four absorption bands:

the Hartley, Huggins, Chappuis, and Wulf bands.102–105 For a period of time there was a

controversy in the literature as to what the Wulf band transition could be attributed to.

Some theoreticians claimed the Wulf band was a transition from the 1A1 ground state to the

1A2 state,106–108 while others claimed it was a transition to the 3A2 state.109–111 Bouveir

et al.112–114 performed a high-resolution, FTIR study on the Wulf band determining an

excited state geometry and excited state rotational constants most consistent with the 3A2

state confirming the assignment. This result is highly relevant and related to the current

problem of methylene peroxy given that the molecules are isoelectronic and the singlet

ground states and lowest lying triplet states of both molecules are produced through the

same mechanism. The singlet-triplet energy gap is directly related to the size of overlap

between the two singly occupied p-orbitals. Wadt and Goddard4 claim that the overlap,

and consequently the singlet-triplet energy gap, will be smaller in methylene peroxy than

in ozone because the central lone pair of electrons will not delocalize as equally across

methylene peroxy as it will in ozone. And since the origin of the 3A2-1A1 transition in

ozone is 9553.021(78)113 cm−1, the analogous 3A′-1A′ transtion in methylene peroxy should

be at lower energies.

In addition to the work done on peroxy radicals mentioned in the previous section, exper-

12

iments were done to observe the a3A′−X1A′ of the methylene peroxy Criegee intermediate,

CH2OO. This radical was also observed in the NIR using cavity ringdown absorption spec-

troscopy. While we currently do not have ab initio calculations to support our assignment of

our experimental spectrum to CH2OO, we have formed several circumstantial arguments to

support our claim. The spectral carrier’s decay rate, reactivity with SO2, and dependence

on O2 pressure were all investigated as a part of this study and compared to the behavior

of known peroxy radicals. The experimental evidence obtained in this study indicates that

the spectral carrier is indeed CH2O2 and is consistent with what is currently present in the

literature describing the behavior of this molecule.

13

Chapter 2

Experimental

2.1 Cavity Ringdown Theory

The Beer-Lambert Law is the mathematical description for absorption experiments,115

I = I0.exp(−σNL) (2.1)

where I0 and I are the incident and transmitted light measured before and after an absorbing

species, N and σ are the number density and absorption cross section at a given wavelength

of the absorbing species, and L is the path length that the absorbing species interacts

with the incident light. In a conventional, single pass absorption experiment I and I0 are

monitored as a function of wavelength; sensitivity of the experiment is determined by how

well I and I0 can be differentiated from each other. Hence, if one is attempting to detect

a species that absorbs very weakly and is present in small concentrations, conventional

absorption experiments will not have the sensitivity to be successful if the difference in I

and I0 is undetectable.

Cavity ringdown spectroscopy (CRDS) is a modification of a conventional absorption

experiment that utilizes a multipass absorption technique to effectively increase the path

length along a physically small dimension. The basic principles of CRDS are shown in Fig.

2.1. In a typical experiment a pulse of light is injected into an optical cavity formed by

two highly reflective mirrors, typically >99.99%. On each pass a small amount of light is

transmitted through the mirrors (known as cavity loss) and if a detector is placed behind

the exit mirror the transmission of this light can be recorded. As Fig. 2.1 shows, with

14

each successive round trip that the light makes the intensity of the light in the optical

cavity decreases. This multipass set up generally gives effective path lengths on the order

of hundreds of kilometers for a physical dimension measured in centimeters.

Photo-diodeL

TimeIntensity τ0Figure 2.1: Schematic diagram of the CRDS experiment. A pulse of light is trapped ina cavity formed between two highly reflective mirrors with leaked light being detected bya photodiode. Exponential decay ringdown curve that is obtained from the photodiode isshown below the cavity.

The theory behind CRDS can be derived by adapting the Beer-Lambert Law to include

the parameters of a CRDS experiment. To describe the fractional loss of light experienced

in a CRDS experiment the following equation can be used:

It = I0 · exp[−

(number of reflections

round trip

) (loss

reflection

) (number of round trips

)](2.2)

where I0 is the intensity of the initial pulse and It is the intensity of the light after time t.

For a ringdown cavity terminated by two mirrors, there are two reflections per round trip

15

and the number of round trips n in time t will be

n =tc

2L(2.3)

where c is the speed of light and L is path length between the two mirrors. If the loss per

reflection is equal to 1-R, where R is the reflectivity of the mirrors, then Eq. 2.2 can be

rewritten as

It = I0.exp

[− (1− R)

(tc

L

)](2.4)

Eq. 2.4 describes the decrease of the intensity of light in the empty cavity. The pulse of

light will make on the order of 105 passes before decaying to 1/e of its initial intensity. The

amount of time it takes for the light pulse to decay to 1/e of its initial intensity is called an

“empty cavity ringdown” time (τ0) and can be calculated using the following equation:

τ0 =L/c

1−R(2.5)

The CRDS apparatus used for this dissertation consisted of a ringdown cavity with L=54

cm with mirrors of 99.995% reflectivity to give τ0=30 µs. In addition to determining a

ringdown time of an empty cavity, a ringdown time (τ) can also be calculated when there

is an absorbing species present. To derive the mathematical form of τ , absorption loss of

the cavity must be determined:

absorption loss = (loss due to absorber per round trip)

(number of round trips) = (2σNl)tc

2L

(2.6)

where l corresponds to the absorber path length. The total loss of light intensity including

the loss from the absorber and cavity loss can be expressed as

total loss = (1− R)

(tc

L

)+ (σNl)

(tc

L

)= [(1− R) + (σNl)]

tc

L(2.7)

By modifying Eq. 2.5 the ringdown time for a cavity containing an absorbing species can

16

then be written as

τ =L/c

1−R+Nσ(ν)l(2.8)

The quantity Nσ(ν0)l can be defined as the absorbance, A, of the species at frequency ν0,

which can be calculated from the measured ringdown times by combining Eqs. 2.5 and 2.8:

A =L

cτ− L

cτ0(2.9)

The first historical mention of CRDS appears in the early 1980s when Herbelin et al.116

attempted to measure mirror reflectivity by measuring the phase shift of a continuous

wave (CW) laser exiting an optical cavity formed between two mirrors. CRDS was first

demonstrated as an absorption technique by O’Keefe and Deacon117 in 1988 who observed

the doubly forbidden b1Σg-X3Σg transition of O2 in air; this experiment proved to be the

prototype for modern day CRDS experiments.

Since the fateful experiment of O’Keefe and Deacon there have been many efforts to

use cavity ringdown spectroscopy (CRDS) as a method to detect trace species in the gas

phase. Wang and coworkers118 developed a CRDS spectrometer and conducted prelimi-

nary clinical trials to analyze acetone concentrations in breath samples to find a correlation

with traditional diabetes biomarkers, since acetone has been identified as a biomarker for

diabetes. Snels et al.119 implemented a cw-CRDS spectrometer for the trace detection of

explosive materials in the 1560-1680 nm region. The Ravishankara120,121 group and the

Brown122 group have explored the relevance of CRDS to study trace gases in the atmo-

sphere. They demonstrated that CRDS techniques possessed the sensitivity and selectivity

to measure trace gas concentrations below 1 ppb in ambient air gas mixtures by monitor-

ing NOX species. Anderson123–125 has developed a CRDS spectrometer capable of being

flown on NASA’s WB-57 high altitude research aircraft. This spectrometer has been used

to detect water isotopologues and determine methane isotopologue ratios in the ambient

atmosphere. Lehmann126 has also demonstrated the capability of using CRDS to detect

single cell biological agents. These are just a few of the applications for CRDS, recent

reviews illustrate these applications summarized here and more.127–129

17

2.2 Experimental Setup

The apparatus used to observe the various peroxy radicals and methylene peroxy in the

NIR shown in Fig. 2.2. A PrecisionScan Sirah Dye Laser was pumped with the second

harmonic of a Nd:YAG laser at 20 Hz and tuned over the range 575-670 nm using four

different dyes: DCM, Rhodamine B, Rhodamine 101, and Pyrromethene 597. The output

of the dye laser (60-110 mJ/pulse) was focused by a lens (f=50 cm) into a 70 cm single pass

Raman cell purged with 300 psi H2. This will generate the required NIR radiation (1.51-

1.10 µm) by stimulated Raman shifting (SRS). When the radiation exits the Raman cell

the second Stokes component is isolated using 1000 nm long pass filters (Corion-LL-1000)

and the resulting 1-3 mJ/pulse of NIR radiation is coupled into the ringdown cell. The

cell is fabricated of stanless steel, is 54 cm in length and terminated by two 6 m radius-

of-curvature, plano-concave mirrors (R≥99.995%, Los Gatos Research) which are housed

in custom made flanges that allow for accurate alignment using finely threaded screws for

adjustment of the mirrors. To cover the entire spectral region of interest four sets of mirrors

centered at 1.4, 1.3, 1.2, and 1.064 µm were used with each mirror set having sufficient

overlap to ensure complete wavelength coverage. To prevent damage to the mirrors from

the chemistry a constant purge of nitrogen was used to shield them. In addition to the

mirror purge inlets, the ringdown cell was constructed with inlets for precursor gases, a

Baratron pressure gauge, and an exhaust port leading to a mechanical pump. Once the

NIR light exits the cell it is focused onto an amplified InGaAS photodiode (Thorlabs, PDA

400) where the detector output is recorded by a 12 bit 20 MHz digitizing card. During

a typical scan approximately 20-40 consecutive laser shots are averaged at each dye laser

frequency and, using the nonlinear Levenberg-Marquardt algorithm, the digitized signal of

the decay of radiation is fit to a single exponential to acquire the ringdown time, initial

amplitude, and baseline. The calculated ringdown time is converted to cavity absorption

per pass and saved as a data point for a given frequency in the spectrum.

18

Sirah dye laser Nd:YAG Pump LaserRaman cell (H2)

Photolysis:ExcimerLaserSecondStokesPDCRDSComputer

Figure 2.2: Setup of the room temperature, moderate resolution CRDS experiment. TheNIR radiation is generated via isolation of second Stokes stimulated Raman shifting of avisible dye laser output. Radicals are produced inside the ringdown cavity via photolysisusing either 193 or 248 nm light from an excimer laser.

The cell is also fashioned with two rectangular apertures (2 x 16 cm) in the center of

the cell separated by 15.6 cm for UV grade quartz windows which allow for photolysis of

the precursor gases. An excimer laser is used to initiate the chemistry necessary to produce

the radicals studied in this dissertation. The excimer is operated with a gas mixture of ArF

or KrF to generate 193 or 248 nm light depending on the experiment being done. Once

the radiation leaves the excimer the beam is shaped by a cylindrical and spherical lens to

a rectangular shape (13.0 cm along the cavity axis by 0.1 cm height) and sent through

the photolysis windows in the central part of the cell. The excimer laser is fired 1-5 µs

before each pulse of NIR light entered the cell, which allows enough time for the reactive

intermediates to form, but not enough time for them to react, be pumped out, or diffuse

19

to the walls of the cell. At each dye laser frequency, a ringdown time is acquired with the

photolysis excimer laser on and off. This will generate two traces: an on trace containing

photolysis products and an off trace that contains background structures (i.e. precursor

bands, background water lines, etc). These two traces are subtracted (on-off) to produce a

trace dependent only upon the photolysis products.

2.3 Chemistry Related to the Production of Reactive Species

To produce the radicals that we want to observe two main methods were employed: photol-

ysis of halogenated precursor and chlorine atom reaction with hydrocarbon precursor. The

next few sections detail the chemistry and the conditions of these experiments.

2.3.1 2,1-Hydroxy Propylperoxy Radical Production

The 2,1-HPP radical is a reactive intermediate formed in the atmospheric oxidation of

propene and the combustion of propanol. The study of this molecule represents an ex-

tension of our work in studying the hydroxy peroxy radicals as we have already studied

β-hydroxyethyl peroxy (β-HEP) radical.130,131

One of the methods used to produce 2,1-HPP was photolysis of 1-iodo-2-propanol. Un-

fortunately, the iodohydrin precursor necessary for this experiment is not commercially

available and needed to be synthesized with the synthesis being adapted from Ghosh et

al.132 Fig. 2.3 represents the chemical reaction to produce 1-2-iodopropanol and involves

opening propylene oxide with hydriodic acid to yield a mixture of isomers: 1-iodo-2-propanol

and 2-iodo-1-propanol.

To synthesize the iodohydrin precursor, the follwing procedure was used:

1. 160 mL of distilled water was put into 500 mL round bottom flask.

2. 20 mL of propylene oxide (99%, Sigma Aldrich) was added to the round bottom flask.

3. The round bottom flask was suspended over a stirrer with a clamp and a stir bar was

added to the flask. Reaction mixture was clear and colorless.

4. The stirrer was set on medium speed and 40 mL of hydriodic acid (HI, 57% by weight

20

Figure 2.3: Reaction showing the synthesis of 2-1-iodopropanol and 1-2-iodopropanol.

in H2O, 99.99%, Sigma Aldrich) was added drop-wise over a 15 minute period. Upon adding

all of the HI, the reaction mixture had turned a brown color.

5. After drop-wise addition of HI was complete, reaction mixture was allowed to stir for

additional two minutes.

6. Reaction mixture was quenched with saturated solution of sodium pyrosulfate. Upon

addition of sodium pyrosulfate solution reaction mixture became a cloudy brown color.

7. Transferred half of reaction mixture to 500 mL separatory funnel. Performed two

extractions of reaction mixture with 125 mL of ethyl acetate for each extraction. After

extraction procedure, ethyl acetate layer turned a clear, brown color while the reaction

mixture became a lighter brown color and was still cloudy.

8. Ethyl acetate layer was deposited into 1000 mL beaker.

9. Ethyl acetate extraction procedure was repeated for remaining reaction mixture.

Ethyl acetate layers were deposited into same 1000 mL beaker as previous extractions.

10. Half of ethyl acetate extract was taken from 1000 mL beaker and put into new,

clean separatory funnel.

11. Ethyl acetate was washed twice with 100 mL of a 5% by weight aqueous solution of

21

sodium thiosulfate. Upon washing with sodium thiosulfate solution, ethyl acetate turned

from a brown, colorless to a clear, colorless solution.

12. Ethyl acetate layer was put into a new, clean 1000 mL beaker. Washing procedure

was repeated with remaining ethyl acetate solution.

13. After washing, remaining ethyl acetate solution was combined with the rest of the

ethyl acetate in 1000 mL beaker.

14. Dried the combined ethyl acetate layers with anhydrous Na2SO4

15. After drying, ethyl acetate was filtered into a 1000 mL round bottom flask.

16. 1000 mL round bottom flask was put under a rotary evaporator for approximately

45 minutes to remove ethyl acetate. After ethyl acetate was removed, a clear liquid with a

yellowish tint remained in bottom of flask.

17. Clear, yellowish liquid was then put under vacuum for 30 minutes to remove any

remaining ethyl acetate. Obtained product yields ∼95%.

The final product was obtained as a mixture of 2-iodo-1-propanol and 1-iodo-2-propanol.

1H and 13C NMR were used to characterize them with their identities being verified by

comparison to 1H and 13C NMR from a previous study.133 The proportion of the 1-iodo-

2-propanol (87%) to 2-iodo-1-propanol (13%) isomers was determined by comparing the

integrated intensity of the peaks at δ=1.22 ppm and δ=1.80 ppm, respectively, which rep-

resent hydrogen atoms on the terminal methyl group of both molecules.

The synthesized mixture was transferred from the round bottom flask to a sample bomb

for the experiment; the sample bomb containing the precursor was heated to 40◦ C while a

stream of N2 (backing pressure of 1 psi) with 40 torr of this gas mixture being maintained in

the ringdown cell. Typical partial pressures in the cell for this method were [N2]=110 torr,

[O2]=40.0 torr, and [1-iodo-2-propanol]≈0.5 torr. The gas mixture was photolyzed with 248

nm light allowing the iodine atom to photolytically dissociate creating two isomers of the

hydroxy propyl alkyl radical (Fig. 2.4) which which rapidly underwent a 3-body reaction

with O2 and N2 to yield 2,1-hpp radical and 1,2-hpp radical.

The second method used to produce 2,1-HPP was hydrogen atom abstraction from

isopropanol using chlorine atoms (Fig. 2.4). The chlorine atoms were generated in the

22

(COCl)2 193 nm 2Cl + 2COOH + Cl OH + OHOH OHOH O2+O2+ OH OOOO HO2 + O15% 85%

I OH 248 nm OHHO I 248 nm

OOO2+ HO 2,1-HPP1,2-HPP

HO

HOFigure 2.4: Reaction scheme for producing 2,1-HPP radical. Top panel shows photolysisof two iodohydrin isomers which yield the corresponding hydroxy propyl isomers. Middlepanel illustrates an alternative route to the hyroxyl propyl radicals via Cl atom abstractionof an H atom from isopropanol. Bottom panel shows the (3-body) addition of O2 to thehydroxyl propyl isomers to yield the 1,2- and 2,-1 HPP radicals.

CRDS cell by photolyzing oxalyl chloride (COCl)2 with 193 nm light. Oxalyl chloride has

a large absorption cross section at 193 nm (σ=3.8 x 10−18 cm2) and has been shown to

be a clean source of chlorine atoms.134,135 The chlorine atoms can abstract a hydrogen

atom from the isopropanol molecule from either the α- or β- positions, with a branching

ratio of 85% to 15%, respectively.136 The newly formed hydroxy alkyl radicals then add

O2 to give the 2,1-HPP and 2,2-HPP radical. However, under our conditions the 2,2-

HPP radical is very unstable and falls apart very rapidly to form HO2 and acetone. To

carry out the chlorine atom reaction, a stream of N2 was bubbled through isopropanol at

room temperature with ≈ 12.0 psi backing pressure. Approximately 8.0 torr of this gas

23

mixture was maintained in the ringdown cell. The typical experimental conditions for this

experiment were [isopropanol]≈0.4 torr,[(COCl)2]≈1.3 torr, [N2]=79.0 torr, and [O2]=40.0

torr.

2.3.2 Production of Straight Chain and Iso-octyl Peroxy Radicals

The largest peroxy radicals we have studied in our group have contained five carbon atoms.

Studying the C6-C10 peroxy radicals represents an extension of this work to include the

larger peroxy radicals that will be more relevant to combustion chemistry.

The straight chain (C6-C10) peroxy radicals and iso-octyl peroxy radicals were produced

in the CRDS cell via a three-body reaction involving the appropriate alkyl radical, N2, and

O2. The alkyl radicals were generated by hydrogen abstraction of the corresponding alkane

precursor with chlorine atoms produced from photolysis of oxalyl chloride as discussed in

Section 2.3.1. As shown in Fig. 2.5, when producing straight chain peroxy radicals, the

chlorine atoms may abstract a hydrogen atom from one of the terminal carbon atoms, which

will result in a primary alkyl radical and primary peroxy radical; or they can abstract a

hydrogen atom from one of the interior carbon atoms yielding a secondary alkyl radical

which will produce a secondary peroxy radical. The chlorine atom reaction with iso-octane

could involve hydrogen abstraction from a primary, secondary, or tertiary site which would

yield the corresponding primary, secondary, or tertiary peroxy radical. A stream of N2

with a backing pressure ≈4.0 psi was bubbled through the liquid alkane precursors while

they were heated to 40◦ to 50◦ C with 4.0-10.0 torr of this gas mixture maintained in the

ringdown cell. Typical partial pressures in the cell for this experiment were [N2]≈30.0-33.0

torr, [O2]≈20.0 torr, [(COCl)2]≈0.3-0.5 torr, and hydrocarbon precrsors=0.2-0.4 torr.

2.3.3 Chloromethyl Peroxy and Bromomethyl Peroxy Production Meth-

ods

Chloromethyl peroxy and bromomethyl peroxy were both produced using two methods.

Photolysis of CH2ClI (97%, Sigma Aldrich) and CH2Cl2 (≥99.8%,Sigma Aldrich) with

248 and 193 nm light, respectively, were used to produce the chloromethyl peroxy radical

24

193nm, (COCl)2193 nm, (COCl)2Figure 2.5: Diagram showing formation of straight chain (hexyl-decyl) peroxy radicals (toppanel) and iso-octyl peroxy radical (bottom panel). In formation of the straight chainperoxy radicals n=4-8, m=3-7.

(CH2ClO2), while CH2Br2 (99%, Sigma Aldrich) and CH2BrI (97%, Sigma Aldrich) were

both photolyzed with 248 nm light to produce the bromomethyl peroxy radical (CH2BrO2),

Fig. 5.1. The iodine atom will preferentially be photolyzed off of the precursors which

contain iodine, forming a CH2X (X=Br or Cl) radical, while the CH2X2 precursors have

an X photolyzed off to create the aforementioned radicals. Once the CH2X radical are

photolytically created, they add O2 in a 3-body process to produce the corresponding peroxy

radical. Liquid precursor samples were obtained commercially and were kept in glass sample

bombs while performing experiments. The glass bombs containing the liquid precursors were

heated to 40◦-50◦ C while N2 was bubbled through with a backing pressure of 1-5 psi with

≈5.0-7.0 torr of this gas mixture maintained in the ringdown cell. Typical partial pressures

25

in the cell for this experiment were [N2]≈27.0-30.0 torr, [O2]≈30.0 torr, and halogenated

hydrocarbon precrsors=0.2-0.4 torr.

Figure 2.6: Diagram showing the photolytic methods to produce chloromethyl peroxy (toppanel) and bromomethyl peroxy (bottom panel).

2.3.4 Methylene Peroxy Production Method

While Criegee intermediates were proposed as part of the ozonolysis mechanism in 1949,3

scientists lacked an appropriate synthetic method with which to prepare the molecule so

it could be directly observed. This changed recently as Welz et al.84 discovered a way of

preparing the methylene peroxy Criegee intermediate in a very clean manner which involved

photolysis of CH2I2 to produce CH2I radicals and allowing them to react with molecular

oxygen, Fig. 2.7. This path is thought to produce CH2IO2 in a vibrationally excited state,

but instead of the iodomethyl peroxy molecule being collisionally stabilized, it dissociates

26

the other iodine atom to yield methylene peroxy. This has been used by several other

research groups to probe the vibrational structure of the ground state and to investigate

the B state.85,89,90 We will likewise use this synthetic method in our study of the 3A′-1A′

transition of methylene peroxy. Diiodomethane was obtained commercially and was kept

in glass bomb while performing experiments. Diiodomethane was heated to 50◦ C while N2

was bubbled through with a backing pressure of 1 psi with ≈30 torr of this gas mixture was

maintained in the ringdown cell. Typical partial pressures in the cell for this method were

[N2]≈85.0 torr, [O2]≈0.3-70 torr, and [CH2I2]=0.04 torr.

Figure 2.7: Mechanism showing production of methylene peroxy.

27

Chapter 3

Spectroscopic Observation andIdentification of the

2,1-Hydroxypropyl PeroxyRadical

3.1 Introduction

The study of alkyl peroxy radicals has been one of the main focuses of our research group for

over a decade.8 In that time span we have identified more than a dozen alkyl peroxy radicals

including saturated, unsaturated, halogenated, and cyclical peroxy radicals.137–140 Recently

we have extended our studies to include -OH substituted peroxy radicals, or hydroxy peroxy

radicals. Our first study of this class of molecules was done on (β-HEP) which included a

room temperature, moderate resolution study and a jet-cooled, high resolution study.130,131

This chapter represents an extension of those studies to include the room temperature,

moderate resolution investigation of the 2,1-hydroxypropyl peroxy (2,1-HPP) radical.

3.2 Computational Methods

To aid the in analysis of the experimental spectra obtained, quantum chemical calculations

using Gaussian 09W141 were performed to predict relative energies, oscillator strengths,

origin bands, and vibrational frequencies for the different conformers of 2,1-HPP. Relative

energies of the different conformers were determined from their calculated G2 energies for

the X state; these relative energies were subsequently used in a Boltzmann population anal-

28

ysis of the conformers at room temperature. Origin values were obtained by calculating zero

point energy corrected energies of the X and A states at the G2 level of theory and taking

the difference. For calculation of the A state, the appropriate electron configuration was ob-

tained by permuting the HOMO and SOMO orbitals in an initial ROHF step; the converged

A state electronic wave functions were used in subsequent calculations as described in pre-

vious reports on peroxy radicals.8 Past studies of peroxy radicals have shown the G2 level

of theory to be reliable in predicting origin frequencies to within tens of wavenumbers.130

Transition moments of the A − X transition were calculated by UCIS/6-31G(d) method.

Predicted relative intensities of the spectra of different conformers were calculated as prod-

ucts of oscillator strengths and Boltzmann weights. Equilibrium geometries, normal mode

vectors, and harmonic vibrational frequencies were computed at the UB3LYP/aug-cc-pVTZ

level of theory and used in Franck-Condon simulations using the MolFC142 program.

3.3 Results and Discussion

3.3.1 Computational Results

Based on a search of the X state potential energy surface of 2,1-HPP, 25 stable conformers

are expected for this radical. Table 3.1 lists all 25 conformers along with a summary of the

results from the calculations mentioned in the previous paragraph. Based on Boltzmann

populations at room temperature and calculated oscillator strengths the three, or possibly

four, lowest energy conformers are predicted to have significantly stronger A − X spectra

than the others. Figure 3.1 shows the X state equilibrium geometries of the four most sta-

ble conformers: G1G2G3T4T5, G′1G2G3T4T5, G1G2G3G4G5, and T1G2G3T4T5. The dif-

ferent conformers are principally distinguished by five different dihedral angles: ∠OOC1C2,

∠OC1C2O, ∠HOC1C2, ∠OC1C2C3, and ∠HOC2C3 (indicated 1, 2, 3, 4, and 5, respectively

in the conformer labels). The different dihedral angles take on values of ∼±60◦, correspond-

ing to a gauche (G) configuration, or 180◦, corresponding to a trans (T) configuration. The

conformers are labeled specifying as G or T the different values of the dihedral angles.

Two of the most stable conformers, G1G2G3T4T5 and G1G2G3G4G5, assume a 6-

29

membered cyclic geometry which enables an intramolecular hydrogen bonding interaction

between the hydroxyl hydrogen and the terminal oxygen on the peroxy moiety. The other

two have a 5-membered cyclic ring H-bond involving the hydroxyl hydrogen and the penul-

timate O atom. Similar results were seen previously in our study of β-HEP.130,131 Figure

3.1 also contains geometrical parameters of interest in the X and A states of the four most

stable conformers. As with other peroxy radicals, upon excitation to the A state a lengthen-

ing of the O-O bond is predicted along with a moderate change in the ∠OOC1 bond angle.8

There is also a significant change in the ∠HOC2C1 and ∠HOC2C3 dihedral angles, which

is consistent with observations from β-HEP.130 However, the ∠OOC1C2 dihedral angle is

also predicted to have a relatively large change for G1G2G3T4T5 (11.6◦) and G1G2G3G4G5

(8.6◦) and a significantly smaller change for the other two conformers. The large change in

the ∠OOC1C2 dihedral angle is generally not seen in peroxy radicals.

Table 3.2 presents a list of the normal mode vibrational frequencies of the G1G2G3T4T5

conformer calculated at the UB3LYP/aug-cc-pVTZ level of theory with qualitative descrip-

tions of each normal mode for the ground and excited states. Appendix A lists the X and

A state frequencies for the G′1G2G3T4T5, G1G2G3G4G5, and T1G2G3T4T5 conformers.

30

Table 3.1: Ground state properties of 2,1-HPP conformers: calculated X state energies(cm−1), origin frequencies (T00, cm

−1), Boltzmann weights (w), oscillator strengths (f ),and relative intensities (I=w x 105 f ) for conformers of 2,1-HPP radical. Energies werecomputed at G2 level of theory and ZPVE corrected; oscillator strengths were computed atUCIS/6-31G(d) level. X state energies reported are relative to the G1G2G3T4T5 conformer.

Conformer X T00 (G2) w 105 f I

G1G2G3T4T5 0 7371 1.000 10.50 10.50

G′1G2G3T4T5 23 7211 0.894 2.41 2.15

G1G2G3G4G5 242 7261 0.310 7.31 2.26

T1G2G3T4T5 311 6917 0.222 2.78 0.62

G′1G2G3G4G5 478 0.099 0.82 0.08

G1G2T3T4G5 560 0.067 2.17 0.14

G1T2T3G4G5 569 0.064 3.03 0.19

T1G2G3G4G5 649 0.043 2.59 0.11

G1T2G3G4G5 676 0.038 2.10 0.080

G′1T2T3G4G5 745 0.027 1.94 0.052

G′1T2G3G4G5 794 0.022 1.65 0.036

G1G2T3G4G5 826 0.018 0.70 0.013

G1G2G3T4G5 837 0.018 3.02 0.054

G1T2G3G4T5 871 0.015 1.73 0.026

T1G2T3T4G5 880 0.014 2.61 0.037

T1T2T3G4G5 905 0.013 2.21 0.029

T1T2G3G4T5 954 0.010 2.63 0.026

T1T2G3G4G5 981 0.009 2.27 0.02

T′1G2T3G4G5 1024 0.007 2.99 0.02

G1G2G3G4T5 1062 0.006 0.76 0.005

T1G2G3T4G5 1072 0.006 2.78 0.02

G′1G2T3T4G5 1110 0.005 1.33 0.007

T1G2G3G4T5 1140 0.004 3.01 0.01

G′1G2T3G4G5 1216 0.003 0.84 0.003

31

Angle X A∠OOC1C2 61.6 73.2∠OC1C2O -77.4 -69.9∠HOC1C2 50.2 62.4∠HOC2C3 170.8 176.9Angle X A∠OOC1C2 -74.8 -72.1∠OC1C2O -58.8 -54.7∠HOC1C2 59.6 51.0∠HOC2C3 179.4 171.7Angle X A∠OOC1C2 -61.4 -70.0∠OC1C2O 78.9 74.5∠HOC2C1 -52.8 -68.8∠HOC2C3 74.5 57.7Angle X A∠OOC1C2 170.5 169.7∠OC1C2O -62.0 -63.1∠HOC1C2 58.5 51.9∠HOC2C3 179.5 172.9

~ ~~ ~~ ~~ ~

G1G2G3T4T5G1’G2G3T4T5G1G2G3G4G5T1G2G3T4T5

123123

123 123Figure 3.1: Structure of four lowest energy conformers of the 2,1-HPP radical with impor-tant geometrical parameters of the radical highlighted for X and A states (A state valuesare italicized). Dihedral and bond angle values given in degrees (◦); bond lengths given inangstroms (A).

32

Table 3.2: Predicted X and A state vibrations for the G1G2G3T4T5 conformer of 2,1-HPP.Mode numbering follows Herzberg’s notation and is based on the values of the A statefrequencies. All vibrational frequencies (cm−1) were calculated via UB3LYP/aug-cc-pVTZand scaled (x0.970).

Mode X Description Mode A Description

ν33 72 CCOO torsion ν33 89 CCOO torsion

ν32 157 CCOO + CCOH torsion ν32 146 OCCO torsion

ν31 217 CH3 twist ν31 210 CH3 twist

ν30 251 OCCO rock ν30 259 CCOH rock + COO bend

ν29 326 COH bend ν29 315 CCOO rock + COO bend

ν28 389 CCC bend ν28 367 CCC bend

ν27 443 CCOH torsion ν27 443 CCOH torsion

ν26 466 CCOH rock ν26 461 COO bend

ν25 556 COO bend ν25 524 CCOH torsion + CCOO torsion

ν24 814 molecular breathing ν24 819 molecular breathing

ν23 870 C-O(O) stretch ν23 867 C-O(O) stretch

ν22 919 CH2 + CH3 + CH bend/wag ν22 910 CH2 + CH3 + CH bend/wag

ν21 922 CH3/CH2/CH assymetric stretch ν21 919 O-O stretch + CH3 wag

ν20 1035 CH3-CH stretch ν20 966 O-O stretch + C-O(O) stetch

ν19 1096 CCC bend ν19 1025 CCC bend

ν18 1114 CH3-CH stretch ν18 1101 CH3/CH/CH2 rock/bend

ν17 1137 O-O/C-O(H) stretch CH3/CH/CH2 wag ν17 1127 C-O(H) stretch CH3/CH/CH2 wag

ν16 1235 C-H2 asymmetric bending ν16 1223 C-H2 asymmetric bending C-H wag

ν15 1258 O-H/C-H wag ν15 1250 O-H/C-H wag

ν14 1304 C-H2 bend ν14 1296 C-H2/C-H bend

ν13 1340 C-H bend ν13 1352 C-H bend

ν12 1367 C-H3 breathing ν12 1366 C-H3 breathing

ν11 1385 C-C(OH) stretch OH bend ν11 1381 CH3/CH breathing

ν10 1422 C-H2 scissor ν10 1420 C-H2 scissor

ν9 1443 C-H3 scissor ν9 1442 C-H3 scissor

ν8 1458 C-H3 scissor ν8 1457 C-H3 scissor

ν7 2885 C-H stretch ν7 2902 C-H stretch

ν6 2947 C-H3 symmetric stretch ν6 2945 C-H3/C-H2 symmetric stretch

ν5 2970 C-H2 symmetric stretch ν5 2950 C-H3/C-H2 symmetric stretch

ν4 3010 C-H3 asymmetric stretch ν4 3009 C-H3/C-H2 asymmetric stretch

ν3 3021 C-H3 asymmetric stretch ν3 3019 C-H2 asymmetric stretch

ν2 3032 C-H2 asymmetric stretch ν2 3022 C-H3 asymmetric stretch

ν1 3622 O-H stretch ν1 3638 O-H stretch

33

3.3.2 Overview of Experimental Spectra

Figure 3.2 shows the CRDS spectrum of obtained using hydrogen atom abstraction (black

trace) and photolysis of iodohydrin precursor (red trace). There is a strong correlation

between the red and black spectra in the 7200-7600 cm−1 region while in the 7800-8400

cm−1 and 8700-8850 cm−1 regions, much of spectrum obtained using the chlorine atom

reaction is obscured by the presence of the HO2143 radical prohibiting detailed comparison

between the two experimental traces. There are also lines from the A− X transition of the

methyl peroxy radical (CH3O2) in the black trace with the origin band at 7383 cm−1 and

the 1211 band at 7488 cm−1 clearly visible. The presence of methyl peroxy is the reult of a

secondary photolysis process. According to Fig. 2.4, a hydrogen atom can be abstracted

from an α position which will yield α-hydroxypropyl peroxy radical which will rapidly fall

apart to form HO2 and acetone. Acetone has a large absorption cross section at 193 nm

and produces methyl radicals when photolyzed by this specific wavelength. The methyl

radicals react with O2 to form methyl peroxy. There is also one additional transition from

an unknown carrier seen near 8529 cm−1.

Figure 3.3 shows traces from the iodohydrin photolysis method (red trace) and β-HEP

(black trace) for comparison. Looking at the two traces one can identify characteristics that

they have in common with other organic peroxy radicals such as the origin region (7250-

7500 cm−1) and OO stretch regions (8100-8400 cm−1). However, there is another region

(7700-8100 cm−1) which seems uniquely characteristic of β-hydroxy peroxy radicals in that

it pertains to HOCC bending/torsion motions. Observing transitions related to HOCC

torsional motion does add a level of complexity to the spectra of β-hydroxy peroxy radicals

compared to the spectra of non-substituted peroxy radicals and is quite interesting. These

bands are the second most intense bands in the spectrum (after the origin) and indicates

that modes pertaining to HOCC torsional motion are highly active upon excitation to the

A state and signifies that there is a significant change in geometry along that particular

coordinate upon excitation.

Chemical tests were also performed to confirm that the features observed were from

34

peroxy radicals. First, the delay between the excimer photolysis laser pulse and the probe

NIR laser pulse was varied from 5-1000 µs in both production methods. As the delay

between the photolysis and probe laser was increased, the signal in the spectrum decayed

over this time period. This is consistent with observations of other peroxy radicals as

well.138 Experiments were also performed in which O2 was omitted from the reaction and

replaced by an equal pressure of N2. The result was that there were no visible features in

the trace from either production method further confirming that the absorptions observed

are from a peroxy radical.

35

Wavenumber

7500 8000 8500 9000

ppm

/pas

s

20

30

40

50

60

70 CH3O2 HO2 HO2AlternateCarrier

Figure 3.2: Experimental spectra from current study obtained using two different productionmethods. The black trace was obtained using the chlorine atom reaction with isopropanoland has been offset by 20 ppm for clarity; the red trace was obtained by photolyzing theiodopropanol precursor at 248 nm and has been digitally smoothed. Strong correlationin the 7200-7600 cm−1 region indicates both spectra a common carrier. There are manyadditional lines present in the black trace, most notably from 7800-8300 cm−1 and from8700-8850 cm−1 which are identified as belonging to HO2. There is also a small amountof methyl peroxy (CH3O2) present as the origin band at 7383 cm−1 and the 121

1 band at7488 cm−1 are observed. There is also an unidentified carrier present in the black trace at8529 cm−1. The 2P1/2-

2P3/2 transition at 7602 cm−1 of the iodine atom is present in thered trace.

36

Wavenumber

7200 7400 7600 7800 8000 8200 8400 8600 8800 9000

ppm

/pas

s

10

20

30

40

50

60

70

80Origin Region -CCOH Bending/Torsion -OO Stretch High Frequency Combination Bands

Figure 3.3: The experimental spectra obtained in this study (iodohydrin photolysis, redtrace) and β-HEP (black trace) are compared with different regions of the spectrum labeledto emphasize similarities between the two traces. The β-HEP trace has been shifted +22ppm for clarity.

37

3.3.3 Discussion

Comparison between the spectra spectra obtained in this study and β-HEP is helpful for

initial assignments. There are several broad bands in in the 7200-7450 cm−1 likely cor-

responding to origin transitions for different conformers of each species. The 7700-8170

cm−1 region contains strong transitions originating from ∠HOCC torsion fundamentals,

combination bands, and overtones. The signals in the 8200-8400 cm−1 and 8500-8900 cm−1

correspond to, respectively, OO stretching fundamentals and to combination bands. How-

ever, there are low frequency transitions present at 7491 cm−1, 7568 cm−1, and 7655 cm−1 in

the present spectrum that have no corresponding transitions in the β-HEP spectrum. The

origin of these transitions could be from sequence bands of low frequency modes, as has

been observed in previous studies of peroxy radicals,9 or from the fundamental transitions

of low frequency modes.

In Figure 3.4 the A-X spectrum obtained in this study is shown along with the Franck-

Condon simulation. The transitions are also labeled according to the band assignments

that are listed in Table 3.3. The first step in assigning the spectrum is to assign the

origin bands of the different conformers. As stated previously, there are predicted to be 25

minima on the potential energy surface of 2,1-HPP corresponding to different conformers.

However, according to Boltzmann populations at room temperature and calculated oscillator

strengths (Table 3.1) only three or possibly four conformers are likely to be observed with

sufficient signal to noise ratio. Looking at Figure 3.4, there is one quite strong band with

a peak at 7406 cm−1 (band A′′) that also has a very prominent shoulder to it. According

to the calculated intensities from the oscillator strengths and Boltzmann populations in

Table 3.1, the G1G2G3T4T5 conformer origin is predicted to have the strongest intensity

and its G2 calculated T00 value is 7371 cm−1; band A′′ was therefore assigned as the origin

of the G1G2G3T4T5 conformer. The shoulder that is present at 7360 cm−1 (band A′) was

assigned as the origin (calculated T00=7261 cm−1) of the G1G2G3G4G5 conformer and the

band present at 7248 cm−1 (band A) was assigned as the origin (calculated T00=7211 cm−1)

of the G′1G2G3T4T5 conformer. There is no obvious band attributable to the T1G2G3T4T5

38

origin, and as its spectrum is expected to have the least intensity of the four and calculated

to be significantly red-shifted (T00=6917 cm−1), it is henceforth not considered.

Wavenumber (cm-1)

7200 7400 7600 7800 8000 8200 8400 8600 8800 9000

Ab

sorb

ance

(p

pm

/pass

)

0

10

20

30

40

50

000 000000

3301 3201 2701 270127013301 2702 200120013301 2001270120012701200127013301A A’ A’’

B’’ C’’ B B’ D’’ E’’ F’’ G’’ H’’ C C’ I’’ J’’28012701 2001280120012701

Figure 3.4: The black trace shows the spectrum obtained in this work. Below the trace isa stick plot of the Franck-Condon simulation from the X state vibrationless level that hasbeen weighted by the Boltzmann factors of the conformers and their oscillator strengths.The calculated T00 values of the conformers have been shifted to match the experimentalorigin frequencies. Origin bands of the different conformers have been labeled with the A,A′, or A′′ notation and have been color coded; subsequent transitions belonging to each ofthe conformers have the same color and notation as that of the origin band. Black letteringand no prime corresponds to transitions of the G′

1G2G3T4T5 conformer, blue lettering anda single prime corresponds to transitions of the G1G2G3G4G5 conformer, and red letteringwith a double prime corresponds to transitions of the G1G2G3T4T5 conformer. Table 3.3lists of assignments and measured frequencies.

39

Table 3.3: Assignments (conformation and vibration) of observed transitions in the A-X spectrum of 2,1-HPP. Band “A” corresponds to origin of G′

1G2G3T4T5 conformer andsubsequent band labels with no prime correspond to transitions belonging to this conformer.Band “A′” corresponds to origin of the G1G2G3G4G5 conformer and subsequent bandswith a single prime label correspond to transitions belonging to this conformer. Band “A′′”corresponds to origin of the G1G2G3T4T5 conformer and subsequent bands with a doubleprime label correspond to transitions belonging to this conformer.

Band Experimental

Frequency

(cm−1)

Experimental

Shift (cm−1)

Assignment

A 7248 0 000

A′ 7360 0 000

A′′ 7407 0 000

B′′ 7492 85 3310

C′′ 7568 162 3210

B 7656 249 2810(2710)

B′ 7773 412 2710

D′′ 7817 409 2710

E′′ 7894 487 27103310

F′′ 8161 754 2720

G′′ 8288 881 2010

H′′ 8335 929 20103310

C 8576 1169 20102810(20

1027

10)

C′ 8654 1294 20102710

I′′ 8694 1288 20102710

J′′ 8754 1347 2010271033

10

40

Earlier in the discussion it was mentioned that bands at 7491 cm−1 (B′′), 7568 cm−1

(C′′), and 7655 cm−1 (B) in the 2,1-HPP spectrum have no corresponding transitions in the

β-HEP spectrum and could be from sequence band structure or low frequency fundamentals.

According to the Franck-Condon simulation there are low frequency fundamentals that will

have a significant Franck-Condon factor and should be present in the spectrum. Additional

experimental work was done to identify these transitions. Our lab has recently implemented

a new technique for obtaining broad scan range spectral data of jet-cooled molecules. Figure

3.5 shows the 7300-7750 cm−1 region of the 2,1-HPP spectra with data obtained at room

temperature (top, black trace) and data obtained under jet-cooled conditions (bottom, red

trace). The advantage gained from the jet-cooled experiment is that vibrational excitation is

essentially eliminated, which means that no hot band structure will appear in the jet cooled-

spectrum. In the jet-cooled spectrum the origin band of the G1G2G3T4T5 conformer is very

prominent as band A′′ followed by bands B′′ and C′′. As Figure 3.5 shows, band B is likely

absent in the jet-cooled spectrum. (This observation is subject to a small caveat in that

band B is a little bit weaker than bands B′′ and C′′ and due to the fact that a factor of

8.5 in signal/noise is lost in going from the ambient temperature experiment to the jet-

cooled experiment, so its absence is not completely definitive). The clear presence of bands

B′′ and C′′ indicates that they are not hot band structure. These bands will therefore be

assigned as the fundamentals of ω33 and ω32, respectively. The disappearance of band B in

the jet-cooled spectrum can be reasonably explained by assigning it to the G′1G2G3T4T5

conformer.

The band structure present at 7600-8170 cm−1 in the 2,1-HPP spectrum has analo-

gous transitions in the β-HEP spectrum and it is where the HOCC bending/torsion of a

β-hydroxy peroxy radical is expected to appear. Bands B, B′ and D′′ look very similar

to the origin bands for the G′1G2G3T4T5, G1G2G3G4G5 and G1G2G3T4T5 conformers, re-

spectively, and are the second strongest features in the spectrum. Bands B′ and D′′ were

assigned as the fundamental of ν27 for the G1G2G3G4G5 and G1G2G3T4T5 conformers.

Band B was also assigned to the HOCC torsion motion of the molecule in its G′1G2G3T4T5

conformation. However, according to the frequency calculations performed for this con-

41

Wavenumber (cm-1)

7300 7400 7500 7600 7700

Ab

so

rptio

n (

pp

m/p

ass)

10

20

30

40

50

0

1

2

3

4A’’

B’’ C’’

Figure 3.5: The 7300-7750 cm−1 region of 2,1-HPP. The top (black) trace was obtained byphotolyzing the iodopropanol precursor with 248 nm light under ambient conditions. Thebottom (red) trace was obtained by running the iodopropanol precursor through a slit jetexpansion and electric discharge to obtain jet cooled conditions as noted previously in Wuet. al.5,6 Conditions for the jet cooled experiment were a rotational temperature of 15-30K and negligible vibrational excitation.

former, the HOCC torsion motion is spread over two modes, ν27 and ν28; whereas for the

other two conformers, the HOCC torsion motion is mainly localized to one mode, ν27.

42

Therefore band B could have several possible assignments: it could be the fundamental of

ν27, the fundamental of ν28, or fundamentals from both modes could be present and band B

is their convolution. Band E′′ has roughly the same spacing from the fundamental of ν27 of

the G1G2G3T4T5 conformer as ν33 has from the origin band of the G1G2G3T4T5 conformer

and was assigned as a combination band of the fundamentals of ν33 and ν27.

The β-HEP spectrum aided in the assignment of band F′′ in that the first overtone of

the HOCC torsion mode was present in the β-HEP spectrum. Band F′′ was assigned as

the first overtone of ν27 of the G1G2G3T4T5 conformer as ν27 is the HOCC torsional mode

of the 2,1-HPP molecule. At first glance it will appear that band F′′ is too far red shifted

to be assigned as the first overtone of ν27. However, there is expected to be a great deal

of anharmonicity present in the overtone band, which was also observed in the overtone

of the HOCC torsional mode in the β-HEP spectrum, which can account for the red shift

observed.

The bands present at 8200-8400 cm−1 involve the OO stretch vibrations of the molecule.

Bands G′′ and H′′ were assigned as the fundamental of ν20 and a combination band of ν20

and ν33, respectively. Going further to the blue there is a series of bands: C, C′, I′′, and

J′′ corresponding to high frequency combination bands built upon the OO stretch band.

Bands K′′ and L′′ are transitions from the G1G2G3T4T5 conformer and were assigned as

combination bands of ν20ν27 and ν20ν27ν33, respectively. Band C′ was assigned as the

combination of ν20ν27 of the G1G2G3G4G5 conformer. The assignment of band C could

have a few possibilities as the assignment of band B had. Band C could be a combination

of ν20ν27 or ν20ν28, or both combination bands could be present and band C a convolution.

Returning to bands B′′ and C′′ their observation and assignment as low frequency fun-

damentals is significant because low frequency fundamentals are typically not observed in

peroxy spectra. Historically, in peroxy radicals, the lowest frequency modes pertain to

the OOCC bending or torsional modes. In some cases, the transitions were forbidden by

symmetry.8,9 However, in other cases, such as β-HEP which has C1 symmetry, there are

no symmetry restrictions and still no low frequency fundamental bands are observed. Of

course, 2,1-HPP also has C1 symmetry, but we do observe the low frequency fundamental

43

bands in that spectrum. To understand why we see the bands in the 2,1-HPP spectrum and

not in the β-HEP spectrum we can look at the predicted geometries of the molecules in the

X and A states. In Figure 3.1 the geometrical parameters for the different conformers of

2,1-HPP are given; specifically looking at the ∠OOC1C2 dihedral angle of the G1G2G3T4T5

in the X and A states there is a change of 11.6◦. The two conformers that were assigned in

the β-HEP paper, G1G2G3 and G′1G2G3, had changes of 1.0◦ and 3.0◦, respectively in the

∠OOC1C2 angle.130 The large change in the ∠OOC1C2 dihedral angle of the G1G2G3T4T5

conformer is likely the reason that progressions are observable on those low frequency modes

in the 2,1-HPP spectrum, but not the β-HEP spectrum.

There is a question of whether any lines in the iodohydrin-precursor spectrum are at-

tributable to 1,2-HPP, particularly as there are large sections of the chlorine atom reaction

trace which cannot be compared directly to the iodohydrin trace. However, based on the

ratios of 1-iodo-2-propanol (which will give 2,1-HPP) and 2-iodo-1-propanol (which will

give 1,2-HPP) in the precursor mixture, 87% and 13%, respectively, it is estimated that

there will be a factor of 6.7 lost in the intensity of the 1,2-HPP spectrum vs. the 2,1-HPP

spectrum. In addition the oscillator strength of the most stable conformer of the 1,2-HPP

radical is predicted to be 1.5 times smaller than the most stable conformer of 2,1-HPP

radicla, yielding roughly a total factor of 10 in intensity difference between the 1,2-HPP

spectrum and the 2,1-HPP spectrum. Moreover, the agreement between the Franck-Condon

simulations and the overall intensity of the bands of the different conformers of the 2,1-HPP

radical gives us confidence that no 1,2-HPP features contribute to the assigned spectrum.

44

Chapter 4

Observation of C6-C10 PeroxyRadicals

4.1 Introduction

Larger hydrocarbons (≥C6) make up a significant portion of the gasoline mixtures used

in spark ignition engines.7,144,145 Fig. 4.1 shows that the most prevalent hydrocarbons

in regular unleaded fuel contain five carbon atoms with 73% of the blend being composed

of hydrocarbons larger than C5; in the premium blend the most prevalent hydrocarbons

contain eight carbon atoms with hydrocarbons larger than C5 making up 75% of the mix-

ture. Consequently, the majority of peroxy radicals formed from the combustion of these

hydrocarbon fuels will contain more than 5 carbon atoms. This chapter will investigate the

A− X electronic transition of straight chain C6-C10 peroxy radicals along with a branched

C8 peroxy isomer, iso-octyl peroxy.

4.2 Spectral/Structural Relationships

Historically ab initio calculations were done to aid in the assignment of peroxy radical spec-

tra. Our group has successfully used the G2 method of calculation to accurately predict

origin frequencies (within 100 cm−1) and used DFT methods to calculate vibrational fre-

quencies for both the X and A states.130,146 However, due to the large size of the peroxy

radicals in this study it was not feasible to do the kind of calculations that have been done

previously. Therefore, to assign the spectra obtained we will use the spectral/structural

45

Carbon Number

2 4 6 8 10 12

Vo

lum

e P

erc

en

t

0

5

10

15

20

25

30

Regular Unleaded (87 Octane)Premium Unleaded (92 Octane)

Figure 4.1: Graph (adapted from data in reference [7]) depicting the typical carbon numberdistribution of regular unleaded and premium unleaded gasolines based on volume percent.

relationships that have been derived through the prior work done on the A− X transition

of smaller peroxy radicals.8 Fig. 4.2 shows the dependence of the origin frequency on sub-

stitution and specific conformer of the peroxy radical. Origin frequencies for the peroxy

radicals appear in the 7300-7800 cm−1 range and highly depend on the substitution of the

radical and its geometrical conformation. In addition to the origin band, other characteris-

tic transitions of peroxy radicals include an OO stretch band and sometimes a COO bend

band that appear ∼900 cm−1 and 400-500 cm−1 to the blue of the origin, respectively.

These relationships between substitution/conformer of the peroxy radical and origin values

and the typical shifts from the origin of COO bending and OO stretching will be applied

to assign the current spectra.

46

4.3 Peroxy Radical Spectra and Assignments

Fig. 4.3 shows the CRDS spectra resulting from the straight chain hydrocarbon precursors

that were studied (hexane-decane). For each hydrocarbon precursor, oxygen dependent and

time dependent tests were performed to provide chemical evidence that a peroxy radical

was indeed the carrier of the observed spectrum. In each trace there is a very intense

band ∼7590 cm−1 and another almost as intense band about 900 cm−1 to the blue. The

very intense bands ∼7590 cm−1 can be assigned to the origin bands of the corresponding

alkyl peroxy radical. The aforementioned transitions ≈900 cm−1 to the blue of the origin

bands in all the spectra are the OO stretch bands of the peroxy radicals. In the hexyl

peroxy (red) trace there also appears to be another transition ≈450 cm−1 to the blue of

the origin which is most likely the COO bend transition of the peroxy radical. The best

experimentally determined frequencies for the observed transitions are given in Table 4.1

and have an estimated error of ± 10 cm−1 due to there being several isomers/conformers

that may overlap under a single peak and a limited signal-to-noise ratio.

47

Carbon Number

0 1 2 3 4 5 6 7 8 9 10 11

Origin

Fre

quency

7200

7300

7400

7500

7600

7700

7800

7900

Primary Peroxies (T1... Conformers)Primary Peroxies (G1G2...Conformers)Primary Peroxies (G1T2..G1'G2...Conformers)Primary Peroxies (Current Study) Secondary Peroxies (T1...Conformers)Secondary Peroxies (G1...Conformers)Secondary Peroxies (Current Study)Tertiary Peroxy Radicals (t-butyl peroxy)Tertiary Peroxy Radicals (Current Study)

Figure 4.2: Graph showing the dependence of A − X origin frequency on peroxy radicalspecies and its structure, based on data presented in reference [8] and results from thepresent study.

48

Table 4.1: Experimental origin frequencies, COO bend vibrations, and OO stretch bandsof hexyl-decyl peroxy radicals.

Peroxy Radical Origin (cm−1) COO Bend (cm−1) OO Stretch (cm−1)

Hexyl 7584(10) 8053(10) 8511(10)

Heptyl 7587(10) - 8515(10)

Octyl 7591(10) - 8520(10)

Nonyl 7592(10) - 8526(10)

Decyl 7576(10) - 8523(10)

The origin frequencies for the straight chain peroxy radicals show a slight upward trend

for C6-C9 and then there is a decrease for C10. It was somewhat surprising to see the

decrease at C10, however, there may be additional chemical processes occurring as there

is evidence of HO2 (sharp structure from ∼7800-8100 cm−1) which is not present in any

other spectra obtained. Observing the spectrum at longer delay times gives some insight to

what is occurring. Fig. ?? shows the 7300-7800 cm−1 region of the decyl peroxy spectrum

at different delay times. At the 5 microsecond delay time (black trace) there is some

unidentifiable structure ≈7385 cm−1; as we increase the delay time it becomes obvious that

this is the origin of methyl peroxy.9 The appearance of methyl peroxy in this situation

is quite curious as now both it and HO2 appear in the decyl peroxy spectrum while not

appearing in any others. As a check of the chemistry, a photolysis experiment was performed

in the absence of the oxalyl chloride to see if the methyl peroxy or HO2 formation could be

due to photolysis of the decane precursor itself. When oxalyl chloride was not present none

of the above spectra (decyl peroxy, HO2, nor methyl peroxy) are observed. This means that

the initial hydrogen atom abstraction of the alkane precursor is required. We also believe it

suggests the formation of HO2 and methyl peroxy is related and could be part of additional

chemistry that the decyl peroxy radical is undergoing. The exact mechanism to produce

HO2 and methyl peroxy from decyl peroxy is unclear but there are a number of chemical

pathways that are available to decyl peroxy once it is formed including direct elimination of

49

Wavenumber (cm-1)

7200 7400 7600 7800 8000 8200 8400 8600 8800

pp

m/p

ass

-20

0

20

40

60

80

100

CH3-(CH2)n-CH3n=4n=5n=6n=7n=8

Figure 4.3: A−X spectra of peroxy radicals from hexane-decane (straight chain) precursors.Red (top) trace is hexyl peroxy, blue trace is heptyl peroxy, green trace is octyl peroxy, pinktrace is nonyl peroxy, and black (bottom) trace is decyl peroxy. In the hexyl peroxy tracethere is some interference that appears in the spectrum from 8230-8489 cm−1 caused byincomplete subtraction from precursor absorption in that region. The sharp line structurein the n=8 likely arises from HO2.

50

HO2 to give an alkene (as stated in the introduction), and the terminal oxygen atom of the

peroxy moiety internally abstracting an alkyl hydrogen to give the QOOH species which

can also eliminate HO2 to give an alkene. One could easily imagine methyl peroxy, and

possibly larger alkyl peroxy, radicals being formed as part of a reaction pathway including

the QOOH species.

Given that HO2 and methyl peroxy are formed, the effect their presence can have on

the spectrum should be addressed. The HO2 signals are far enough from any decyl peroxy

signal that they probably have very little influence on the decyl peroxy spectrum itself.

However, methyl peroxy absorbs at lower energies than decyl peroxy and its spectrum

partially overlaps with decyl peroxy. Larger alkyl peroxy radicals will still absorb to the red

of decyl, but closer in frequency so that their bands are even more overlapped. This may

be enough to give the appearance that the decyl peroxy band center has been red shifted

when in actuality there are other peroxy radical absorbers present that are influencing the

band. This could explain why the decyl peroxy origin and OO stretch bands do not seem

to follow the upward trend in frequency of the C6-C9 peroxy radicals. However, while the

argument is suggestive it is not definitive.

Comparing with the values presented in Fig. 4.2, the positions of the A−X origin bands

demonstrate that the predominant carriers are the secondary peroxy radical isomers in the

G1... conformation. This result is primarily an outcome of the chemistry used to generate

the peroxy radicals. Each straight chain hydrocarbon used as a precursor has six primary

sites from which a hydrogen atom can be abstracted. However, in each precursor there

are more secondary sites than primary sites with ratios (secondary:primary) of 1.33:1 for

hexane, 1.67:1 for heptane, 2:1 for octane, 2.33:1 for nonane, and 2.67:1 for decane giving

more secondary sites to abstract a hydrogen from. In addition the reported, relative reactiv-

ities147,148 of hydrogen atoms to abstraction, 2:9:18 (primary:secondary:tertiary), indicate

that secondary hydrogens are 4.5 times more reactive to abstraction than primary hydro-

gens. Moreover, the origin frequencies (see Fig. 4.2) of the primary peroxy radicals and

secondary peroxy radicals (T1... conformation) are not consistent with our experimental

data. Origin positions for primary peroxy radicals in the T1... and G1G2... conformation

51

Wavenumber

7400 7600 7800 8000

ppm

/pass

0

2

4

6

8

10

12

14

Figure 4.4: 7300-8000cm−1 region of decyl peroxy A− X spectrum at different time delays.Black (top) trace is 5 µs time delay, blue trace is 100 µs, pink trace is 500 µs time delay,and red trace is 1 ms delay. Structure near 7383 cm−1 and 7488 cm−1 are the origin and1211 bands, respectively, of methyl peroxy.9

52

are noticeably red shifted from our experimental data, while origin values for primary and

secondary peroxy radicals in the G1T2...G1′G2... and T1... conformation, respectively, are

much more scattered and do not correlate well with our data. (One exception is the T1...

conformer of 2-propyl peroxy which has an origin of 7692 cm−1.) Taking into account the

relative abundances and reactivities of primary to secondary hydrogen atoms as well as ori-

gin frequencies support assigning the carriers of the spectra in Fig. 4.3 to the corresponding

straight chain peroxy radicals, secondary isomers in the G1... conformation.

Fig. 4.5 shows the spectrum when iso-octane is the precursor and it is compared to the

n-octyl peroxy spectrum. The same chemical tests were performed on the iso-octane trace

and it again displayed behavior consistent with that of a peroxy radical. The iso-octyl

peroxy spectrum appears more complex than the straight chain peroxy radicals studied

presumably due to the fact that it has primary, secondary, and tertiary sites for hydrogen

atom abstraction. In the origin region the iso-octyl peroxy trace has two intense bands

(≈7500 and ≈7800 cm−1) compared to the one intense band of the straight chain peroxy

radicals. The ≈7800 cm−1 band is significantly blue shifted from the other origin bands

that have been observed in this study. In Fig. 4.2, there is a data point at 7757 cm−1

representing the origin of t-butyl peroxy, a tertiary peroxy radical. We therefore assign the

≈7800 band of the iso-octyl peroxy spectrum as the likely origin of the tertiary isomer.

However, the band near 7500 cm−1 has more than one possible assignment. In Fig. 4.2,

the primary peroxy radicals in the G1G2... conformation have origin frequencies from 7480-

7592 cm−1, with all but one origin falling into the 7480-7551 cm−1 range. One could easily

argue that an origin value of 7503 cm−1 falls well within this range and make the assignment

that this band belongs to a primary isomer of iso-octyl peroxy in the G1G2... conformation.

But if a rough calculation is made to predict band intensities of the different isomers using

the relative abundances of primary:secondary:tertiary (p:s:t) abstraction sites (15:2:1) along

with relative reactivities of Cl atoms to p:s:t H atoms (2:9:18) we obtain relative intensities

for the p:s:t isomer bands of 30:18:18. This predicts that the primary isomer origin will

be 2/3 stronger than either the secondary or tertiary origin bands while the transitions we

observe in our spectrum are the same intensity. This argument would be consistent with the

53

band at 7503 cm−1 belonging to the secondary isomer even though it is red shifted nearly

60 cm−1 from any other secondary isomer origin band. However this assignment would also

leave unanswered the lack of any observation of a band attributable to the primarily isomer

despite predictions that it should be most abundant. Furthermore we searched for spectra

to the red of 7300 cm−1 and could not see any structure we could attribute to any peroxy

radical, whereas assignment of the 7503 cm−1 band to a secondary isomer would lead to

the expectation of a primary isomer origin in this region. Regardless of the assignment it

is surprising that we do not observe origin bands from each of the primary, secondary, and

tertiary conformer.

54

Wavenumber (cm-1)

7200 7400 7600 7800 8000 8200 8400 8600 8800 9000 9200

pp

m/p

ass

20

30

40

50

60

70

aa a b a ac

Figure 4.5: A−X spectra of iso-octyl peroxy blue (bottom) trace produced from iso-octaneprecursor (structure in upper right corner of Figure) and n-octyl peroxy red (top) trace.Primary hydrogen abstraction sites on iso-octane are labeled with an “a”, secondary siteswith a “b”, and tertiary sites with a “c”.

55

It is probably worth noting that the relative reactivities of Cl atoms to p:s:t H atoms was

determined through experimentation on n-butane and iso-butane. Iso-octane is relatively

complex compared to either of those molecules and it is unknown whether the relative

reactivities of Cl atoms to p:s:t H atoms will hold quantitatively for iso-octane in this

situation. We therefore tentatively assign the band near 7500 cm−1 as the primary isomer

based on the spectral/structural relationships that we have derived, but clearly note that

we are puzzled by the absence of a third band for iso-octane.

In addition to the spectral/structural relationships, one can also utilize self-reaction rate

information to determine to which isomer a given band belongs. In our experiments we have

the ability to delay the time between the excimer photolysis pulse and the NIR probe pulse.

Fig. 4.6 shows the results of these time delay experiments. At long delay times there are

three bands, labeled as A′, B′, and C′, that are still clearly visible. Long lived spectral

lines were observed in our previous investigation of tertiary radicals.147 In that study a

much slower self-reaction rate constant was found for the t-butyl peroxy radical than for its

secondary and primary isomer counterparts. These additional observations strongly support

the assignment of bands A′, B′, C′ to the tertiary decyl peroxy isomer.

Additional information about the isomers can be obtained by estimating the initial

radical concentrations. Melnik et al. [149] found their chlorine atom concentration to

be 2 x 1015 molecules/cm−3; since we are using similar conditions that should be a good

approximation for the concentration of chlorine atoms in our current study. Looking at

Fig. 4.5, the two bands near 7500 and 7800 cm−1 have nearly the same intensity. If we

assume equal oscillator strengths we can infer the isomers are formed in a 1:1 ratio. Given

our initial chlorine atom concentration, this will yield a peroxy radical concentration of 1 x

1015 molecules/cm3 for each of the species present in our spectrum.

We can estimate the half-life of the peroxy radicals from the data in Fig. 4.6. The

bands that we have assigned as the tertiary peroxy radicals do not seem to decay at all.

This consistent with a reported self-reaction rate constant of 2 x 10−17 cm3 molecule−1

s−1 [62] for tertiary peroxy radicals, which would make their self reaction not observable

on our millisecond timescale, with comparable observation being obtained in similar ex-

56

periments.62,147 From Fig. 4.6 we can estimate a half-life of ≈500 µs for bands A and B

and obtain self reaction rate constants of ≈2 x 10−12 cm3 molecule−1 s−1. Unfortunately

this rate constant does not allow us to differentiate between isomers, since it is roughly

consistent with either primary or secondary peroxy radicals. Based on a combination of

the self-reaction rates and the spectral/structural relationships our conclusion therefore is

that we can definitively assign bands A′, B′, and C′ to the tertiary isomer. Using the spec-

tral/structural relationships, we feel it is more likely, but not certain, that bands A and B

belong to the primary isomer.

57

Table 4.2: Spectral assignments for iso-octyl peroxy.

Band Frequency Assignment

A 7503 Origin of Primary Isomer

A′ 7804 Origin of Tertiary Isomer

B′ 8230 COO Bend of Tertiary Isomer

B 7597 OO Stretch of Primary Isomer

C′ 8734 OO Stretch of Tertiary Isomer

58

Wavenumber (cm-1)

7200 7400 7600 7800 8000 8200 8400 8600 8800 9000 9200

pp

m/p

ass

10

20

30

40

50

60 A A’ B’ B C’

Figure 4.6: Time delay tests of A − X spectrum of iso-octyl peroxy. Red (top) trace istaken at a 5 microsecond delay of NIR probe from the excimer photolysis beam, blue tracetaken at 500 microsecond delay of NIR probe, and black trace taken at 1 millisecond delayof NIR probe. Labeling of bands signifies assignment of transition to primary or tertiaryisomer. A and B bands belong to primary isomer while A′, B′, and C′ bands belong totertiary isomer.

59

Chapter 5

Observation of the A− XElectronic Transitions ofChloromethyl Peroxy and

Bromomethyl Peroxy

5.1 Introduction

Halogenated methanes (CCl2F2, CH3Br, CH3I, CH2I2, CH2Cl2) have been discovered in the

atmosphere.150 These halogenated species serve as the main sources for the reactive Cl, Br,

and I atoms which have been found to participate in ozone destruction in the troposphere

and lower stratosphere.151,152 Freons (CClY FY ) are used in consumer products such as

refrigerants and aerosols and are emitted into the atmosphere from anthropogenic sources.

Other organic halogens (CHY XY ; X=Cl, Br, I) are mainly emitted from natural sources i.e.

algae, snow pack, and salt pans.150 The main removal mechanism of freons is photolysis of

the C-Cl bond creating a free chlorine atom in the atmosphere that was shown by Molina

and Rowland153 to have devastating effects on ozone concentrations. Since then freon use

has decreased dramatically although many problems from freon will persist for years due to

its long lifetime of 102 years.154 For methyl bromide and methyl chloride, their main source

of removal in the lower atmosphere is thought to be reaction with OH155

60

CH3Br + OH −→ CH2Br + H2O (5.1)

CH2Br + O2 +M −→ CH2BrO2 +M (5.2)

CH3Cl + OH −→ CH2Cl + H2O (5.3)

CH2Cl + O2 +M −→ CH2ClO2 +M (5.4)

For both molecules, OH will abstract a hydrogen atom yielding an alkyl radical which reacts

with oxygen and a third body to give a halogenated peroxy radical. CH2ClO2 and CH2BrO2

have been studied in the UV and have a very broad absorption near 240 nm similar to the

absorption for non-halogenated peroxy radicals.155–157 This chapter will explore the A− X

electronic transition of chloromethyl and bromomethyl peroxy radicals in the near infrared.

While the CH2ClO2 and CH2BrO2 radicals may be formed in the atmosphere by the

mechanism shown above, in the laboratory they will be made by a different method which

includes photolysis of an appropriate precursor (e.g. CH2Cl2, CH2ClI, CH2Br2CH2BrI) by

193 nm or 248 nm light. This initial photolysis will produce an alkyl radical that reacts

with O2 to produce the desired peroxy radical. Details of this reaction mechanism are given

in Chapter 6 where the reactions involving the photolysis of CH2I2 are discussed.

5.2 Computational Methods

In order to facilitate assignment of spectroscopic features, quantum chemical calculations

using the Gaussian 09W141 package were performed to predict relative energies, oscillator

strengths, origin bands, and vibrational frequencies for the different conformers of CH2ClO2

and CH2BrO2. Relative energies of the different conformers were determined from their

calculated G2 energies for the X state; these relative energies were subsequently used in a

Boltzmann population analysis of the conformers at room temperature. For the conformers

of CH2ClO2, origin values were obtained by using the zero point corrected energies of the

X and A state at the UB3LYP level of theory with an aug-cc-pvtz basis set and taking

the difference. A similar procedure was repeated to obtain origin values for the conformers

61

of CH2BrO2 using the G2 level of theory. For calculation of the A state, the appropriate

electron configuration was obtained by permuting the HOMO and SOMO orbitals in an

initial ROHF step; the converged A state electronic wave functions were used in subse-

quent calculations. Oscillator strengths for the electronic transitions were computed at the

UCIS/6-31G* level of theory.

The G2 method of calculation was attempted on the conformers of the CH2ClO2 radical

prior to the UB3LYP calculation, however, there was difficulty in obtaining the A state of

the CH2ClO2 molecule for both conformers. As mentioned earlier, the HOMO and SOMO

orbitals are permuted to get the electronic configuration of the A state and the molecules

stays in this electronic configuration throughout the G2 calculation. During the A state

G2 calculation of the CH2ClO2 molecule, the molecule collapsed back down to the X state.

Several things were tried without success to solve this problem and ultimately the X and

A state energies were calculated at the UB3LYP level of theory for CH2ClO2.

For the purpose of qualitatively predicting band origin profiles of the different conform-

ers of CH2ClO2 and CH2BrO2, room temperature simulations of the conformers’ rotational

contours were produced using our group’s SpecView158 program using purely ab initio rota-

tional constants. The rotational constants for the conformers of CH2ClO2 were obtained at

the UB3LYP/aug-cc-pvtz level of theory for the ground and excited states; the rotational

constants for the conformers of CH2BrO2 were obtained from the G2 calculations after the

MP2 geometry optimization had converged. The simulations also required the values for

the transition dipole moment along the a, b, and c inertial axes which were calculated by

the UCIS/6-31G* method.

We also undertook Franck-Condon simulations in order to help assign the vibrational

structure present in each spectrum. For the purpose of calculating multi-dimensional

Franck-Condon factors we used Gaussian09W.141 Equilibrium geometries, normal mode vec-

tors, and harmonic vibrational frequencies used as input were calculated at the B3LYP/aug-

cc-pvtz level of theory with the simulations being performed in the limit of cold (0 K) absorp-

tion. We obtained stick plots of the vibrational spectrum for each conformer of CH2ClO2

and CH2BrO2 that were weighted by oscillator strengths and Boltzmann populations.

62

5.3 Results

5.3.1 Computational Results

Based on searches of the X state potential energy surface of CH2ClO2 and CH2BrO2,

each molecule is predicted to have two stable conformers. Table 5.1 lists the conformers

for each molecule along with a summary of the results from the calculations mentioned

in the previous paragraph. Based on Boltzmann populations at room temperature and

calculated oscillator strengths both conformers of CH2ClO2 and CH2BrO2 are predicted to

be observed. The T conformers of both halo-peroxy radicals are predicted to be roughly 1/3

the intensity of the G conformers. Figure 5.1 shows the X state equilibrium geometries for

both conformers of CH2ClO2 and CH2BrO2. The different conformers are distinguished by

the ∠XCOO (X=Cl or Br) dihedral angle. If the ∠XCOO dihedral angle takes on values of

∼±60◦the conformer is labeled as gauche (G), if the ∠XCOO angle takes on a value of 180◦

the conformer is labeled as trans (T) configuration. Figure 5.1 also contains geometrical

parameters of the two species’ conformers that change significantly upon excitation from the

X and A. As with other peroxy radicals, upon excitation to the A state a lengthening of the

O-O bond is predicted along with a moderate change in the ∠OOC1 bond angle.8 There is

also a fairly large change in the ∠XCOO angle upon excitation to the A state which was also

observed in the study of the 2,1-HPP radical. The fact that there are several geometrical

changes occurring in the vicinity of the O2 group should come as no surprise as it is the

chromophore of the A− X transition.

Tables 5.4 to 5.3 presents a list of the normal mode vibrational frequencies for the

conformers of CH2ClO2 and CH2BrO2 calculated at the UB3LYP/aug-cc-pVTZ level of

theory with qualitative descriptions of each normal mode.

63

Table 5.1: Results of theoretical calculations for conformers of CH2BrO2 and CH2ClO2:relative X state energies (cm−1), origin frequencies (T00, cm

−1), Boltzmann weights (w),oscillator strengths (f ), and relative intensities (I=w x 105 f ) for conformers of CH2BrO2

and CH2ClO2. Energies for conformers of CH2BrO2 were computed at G2 level of theoryand ZPE corrected; while CH2ClO2 energies were done at the UB3LYP/aug-cc-pvtz levelof theory and zero point energy corrected. X state energies for the respective conformersof CH2ClO2 and CH2BrO2 are reported are relative to the G conformers. The rotationalconstants in the ground and excited states are labeled as A′′/B′′/C ′′ and A′/B′/C ′, re-spectively. Oscillator strengths, Boltzmann wieghts, and I are unitless; all other values aregiven in cm−1.

CH2BrO2 CH2ClO2

G T G T

X 0 268 0 240

T00 6820 7629 6681 7695

w 1.000 0.137 1.000 0.157

105 f 1.48 2.90 1.25 2.56

I 1.48 0.40 1.25 0.40

A′′ 0.527 1.219 0.543 1.290

B′′ 0.078 0.062 0.116 0.092

C ′′ 0.072 0.060 0.103 0.087

A′ 0.485 1 .208 0.604 1.273

B′ 0.081 0.062 0.113 0.092

C ′ 0.073 0.060 0.103 0.087

64

Table 5.2: Predicted X and A state vibrations for the G conformer of CH2ClO2. Modenumbering follows Herzberg’s notation and is based on the values of the A state frequencies.

Mode X Description Mode A Description

ν12 97 OOCCl torsion ν12 147 OOCCl torsion

ν11 338 OCCl bend ν11 333 OCCl bend

ν10 518 COO bend ν10 447 COO bend

ν9 716 C-Cl stretch ν9 679 C-Cl stretch

ν8 895 C-O stretch ν8 904 C-O stretch + CH2 rock

ν7 1011 C-H2 rock ν7 944 O-O stretch + C-H2 rock

ν6 1134 O-O stretch ν6 1037 C-O stretch + O-O stretch

ν5 1268 C-H2 torsion ν5 1256 C-H2 torsion

ν4 1333 C-H2 symmetric sway ν4 1333 C-H2 sway

ν3 1458 HCH bend ν3 1446 HCH bend

ν2 3105 C-H2 symmetric stretch ν2 3093 C-H2 symmetric stretch

ν1 3189 C-H2 asymmetric stretch ν1 3179 C-H2 asymmetric stretch

Table 5.3: Predicted X and A state vibrations for the T conformer of CH2ClO2. Modenumbering follows Herzberg’s notation and is based on the values of the A state frequencies.

Mode X Description Mode A Description

ν12 62 OOCCl torsion ν12 87 OOCCl torsion

ν11 280 OCCl bend ν11 269 OCCl bend

ν10 420 COO bend ν10 371 COO bend

ν9 793 C-Cl stretch ν9 783 C-Cl stretch

ν8 940 C-O stretch ν8 947 O-O stretch + C-O stretch

ν7 960 C-H2 rock ν7 1006 O-O stretch

ν6 1166 O-O stretch ν6 1007 C-H2 rock

ν5 1199 C-H2 torsion ν5 1207 C-H2 torsion

ν4 1339 C-H2 symmetric sway ν4 1352 C-H2 sway

ν3 1474 HCH bend ν3 1501 HCH bend

ν2 3099 C-H2 symmetric stretch ν2 3064 C-H2 symmetric stretch

ν1 3178 C-H2 asymmetric stretch ν1 3130 C-H2 asymmetric stretch

65

Table 5.4: Predicted X and A state vibrations for the G conformer of CH2BrO2. Modenumbering follows Herzberg’s notation and is based on the values of the A state frequencies.

Mode X Description Mode A Description

ν12 91 OOCBr torsion ν12 127 OOCBr torsion

ν11 284 OCBr bend ν11 294 OCBr bend

ν10 508 COO bend ν10 400 COO bend

ν9 603 C-Br stretch ν9 584 C-Br stretch

ν8 866 C-O stretch ν8 801 O-O stretch + C-O stretch + CH2 rock

ν7 905 C-H2 twist ν7 830 O-O stretch + C-H2 rock

ν6 1002 O-O stretch ν6 945 C-O stretch + O-O stretch

ν5 1261 C-H2 torsion ν5 1246 CH2 torsion

ν4 1293 C-H2 symmetric sway ν4 1291 C-H2 sway

ν3 1451 HCH bend ν3 1461 HCH bend

ν2 3113 C-H2 symmetric stretch ν2 3101 C-H2 symmetric stretch

ν1 3203 C-H2 asymmetric stretch ν1 3193 C-H2 asymmetric stretch

Table 5.5: Predicted X and A state vibrations for the T conformer of CH2BrO2. Modenumbering follows Herzberg’s notation and is based on the values of the A state frequencies.

Mode X Description Mode A Description

ν12 50 OOCBr torsion ν12 78 OOCBr torsion

ν11 236 OCBr bend ν11 235 OCBr bend

ν10 380 COO bend ν10 314 COO bend

ν9 710 C-Br stretch ν9 684 C-Br stretch

ν8 898 C-H−2 torsion ν8 944 O-O stretch

ν7 933 C-O stretch ν7 947 C-H2 rock

ν6 1157 O-O stretch ν6 983 C-O stretch

ν5 1189 C-H2 symmetric sway ν5 1201 C-H2 asymmetric sway

ν4 1306 C-H2 asymmetric sway ν4 1316 C-H2 symmetric sway

ν3 1467 HCH bend ν3 1497 HCH bend

ν2 3106 C-H2 symmetric stretch ν2 3069 C-H2 symmetric stretch

ν1 3196 C-H2 asymmetric stretch ν1 3140 C-H2 asymmetric stretch

66

Parameterof Interest X AO-O 1.32 Å 1.39 Å∠OOC 111.8o 110.7o∠OCCl 111.3o 113.5o∠OOCCl -84.9o -73.9oParameterof Interest X AO-O 1.31 Å 1.39 Å∠OOC 108.2o 107.1o∠OCCl 106.8o 106.2o∠OOCCl 180.0o 180.0oParameterof Interest X AO-O 1.32 Å 1.39 Å∠OOC 110.3o 107.3o∠OCBr 111.3o 113.1o∠OOCBr -77.6o -70.3oParameterof Interest X AO-O 1.31 Å 1.40 Å∠OOC 108.0o 104.3o∠OCBr 106.7o 105.4o∠OOCBr 180.0o 180.0o

~ ~~ ~~ ~~ ~

CH2ClO2 GCH2ClO2 TCH2BrO2 GCH2BrO2 T

Figure 5.1: Structure of G and T conformers of CH2BrO2 and CH2ClO2 with geometricalparameters that change significantly upon excitation from the X state to the A state givenin the tables (A state values are italicized). Dihedral and bond angle values given in degrees(◦); bond lengths given in angstroms (A).

67

5.3.2 Overview of Experimental Spectra

Figure 5.2 shows the CRDS spectrum obtained using 248 nm photolysis of CH2ClI (blue

trace) and CH2Br2 (black trace); the experimental conditions were given in Section 2.3.3.

There are interferences from several sources in both traces; however, similarities can be

identified between both of these spectra and the spectra of typical peroxy radicals. Both

experimental traces in Figure 5.2 have a defined origin region near 6800 cm−1 and then

what appears to be an OO stretch region ∼900 cm−1 to the blue of the origin. The ex-

perimental traces produced from the 248 nm photolysis of CH2ClI and CH2Br2 were also

reproduced using the 193 nm photolysis of CH2Cl2 and 248 nm photolysis of CH2BrI (Ap-

pendix C), respectively. In addition to the spectral evidence, the chemical behavior of the

carriers is consistent with what is expected of peroxy radicals. The delay between the ex-

cimer photolysis laser pulse and the probe NIR laser pulse was varied from 5-100 µs during

experiments using all precursors. As the delay between the photolysis and probe laser was

increased, the signal in the spectrum decayed over this time period. This is consistent

with observations of other peroxy radicals as well.138 While the decay times observed were

faster than typical peroxy radicals, Catoire et al.155 and Nielsen et al.157 determined self-

reaction rate constants for halogenated peroxy radicals, CH2ClO2 and CH2BrO2, of 3.7 x

10−12 cm3 molecule−1 s−1 and 1.05 x 10−12 cm3 molecule−1 s−1, respectively. Compared

to a self-reaction rate constant of methyl peroxy159 3.40 x 10−13 cm3 molecule−1 s−1, the

rate constants for the halogenated peroxy radicals are about an order of magnitude larger.

Experiments were also performed in which O2 was omitted from the reaction and replaced

by an equal pressure of N2. The result was that there were no visible features from pho-

toproducts in any traces from any of the production methods further confirming that the

absorptions observed are from a peroxy radical.

68

Wavenumber

7000 7500 8000 8500 9000

pp

m/p

ass

0

5

10

15

20

25

HO2/Precursor

HO2PrecursorHO2

Figure 5.2: Experimental A− X spectra of CH2ClO2 (top, blue trace) and CH2BrO2 (bot-tom, black trace); CH2ClO2 spectrum has been offset by 7 ppm for clarity. The CH2ClO2

trace presented was obtained by 248 nm photolysis of CH2ClI precursor while the CH2BrO2

trace was obtained by 248 nm photolysis of CH2Br2. There are multiple interferences inboth spectra. From 7000-7180 cm−1 in both spectra there are interferences from the pre-cursors and HO2. The

2P1/2-2P3/2 transition of the iodine atom is present at 7603 cm−1 in

the CH2ClO2 trace, from 7930-8150 cm−1. HO2 is visible in the CH2BrO2 trace and weaklyvisible in the CH2ClO2 trace. Interferences due to precursors are present from 8715-8752cm−1 and 8698-8817 cm−1 in the CH2ClO2 and CH2BrO2 spectra, and there is additionalHO2 interference in the CH2BrO2 trace from 8891-8980 cm−1.

69

5.4 Discussion

5.4.1 G Conformer Origin and ∠XCOO Torsional Region: 6700-7500 cm−1

Looking at Figure 5.2, it is clear that in each spectrum there is a closely-spaced doublet

near 6800 cm−1 which is where the origin bands of the G conformers of both CH2ClO2 and

CH2BrO2 are predicted to be. While this is suggestive, simulating the rotational contour

of the two spectra using ground and excited state rotational constants from ab initio cal-

culations of these species would add additional evidence to the argument. Figure 5.3 shows

the experimental traces compared to the simulated spectra of the two G conformers. The

simulations reproduce the experimental spectra very well further confirming the assignment

to the G conformer origin bands of the halomethylperoxy radicals. (Simulations of the T

conformers are shown to be quite different in Section 5.4.3.)

After assignment of the origin bands, other transitions observed in the spectrum need

to be assigned in terms of excited state vibrations of the conformers. This was accom-

plished by performing Franck-Condon simulations and assigning bands based on how well

their predicted intensities and frequencies match with the experimental spectra. In Fig-

ure 5.4, the 6700-7500 cm−1 regions of the CH2ClO2 (top panel) and CH2BrO2 (bottom

panel) spectra are shown with selected transitions labeled and with their assigned Franck-

Condon simulations; in both panels the red color correspond to transitions belonging to

the G conformer of each species. The Franck-Condon simulations for both CH2ClO2 and

CH2BrO2 predict that there will be fundamental bands with enough intensity to be ob-

served corresponding to XCOO tosional motion, XCO bending, and COO bending; which

is expected due to the large geometrical changes these parameters undergo upon excitation

to the A state. Unfortunately, the region where the majority of these bands are predicted

is obscured by precursor absorption and HO2 in the spectra of both species. While many

of these fundamental transitions will be unassigned, assignments will be made more to the

blue which include combination bands built off of these fundamental bands. In addition to

the fundamental transitions that are predicted in this region, we should also expect to see

hot bands (1201, 1101) or sequence bands (1211, 11

11) as ν12 and ν11 should have significant

70

population at room temperature in the G conformers of CH2ClO2 and CH2BrO2. However,

the hot bands are predicted to be roughly 100 cm−1-300 cm−1 to the red of the origin bands

of CH2ClO2 and CH2BrO2 which is where there is interference present in the form of HO2

and makes assignment difficult. The sequence bands are predicted to be 30-50 cm−1 to the

blue of the origin bands of both species, but there does not seem to be anything that can

be clearly attributed to sequence bands in either the CH2ClO2 or CH2BrO2 trace.

In the spectrum of CH2ClO2 the simulation predicts three active modes involving

ClCOO torsional motion along with ClCO and COO bending (ν12, ν11, and ν10, respec-

tively). However, due to interferences only 1010 and 12101110 are assigned as B and C (Table

5.7). Similar to the CH2ClO2 Franck-Condon simulation, the simulation of CH2BrO2 pre-

dicts ν12, ν11, and ν10 to be active in this spectrum as well. While interferences make

assignment somewhat difficult there are still some that can be made. B was assigned as

the 1210 band and C was assigned as the 12101110 combination band but could also have 1010

contributing to it as it is also predicted to be observed in the Franck-Condon simulation.

Bands D, E, and F were then assigned as combination bands and overtones involving ν12,

ν11, and ν10.

71

Wavenumber

6740 6760 6780 6800 6820 6840 6860

ppm/pass

0

5

10

15

20

Wavenumber

6740 6760 6780 6800 6820 6840 6860

ppm/pass

0

5

10

15

20

Figure 5.3: Experimental spectra of the G conformer origins of CH2ClO2 (top panel, bluetrace) and CH2BrO2 (bottom panel, black trace) compared to simulations (red traces, bothpanels). The simulations were obtained through using ab initio the rotational constants inTable 5.1. The simulation for the G conformers of CH2ClO2 and CH2BrO2 use the ratioof components of the transition dipole along the inertial axis of a:b:c=0.48:0.33:0.19 anda:b:c=0.44:0.39:0.17, respectively.

72

Wavenumber (cm-1)

6800 7000 7200 7400

ppm/pass

0

2

4

6

8

10

12

14

16

18 B C000 1201 1202 1101 120111011001

Wavenumber (cm-1)

6800 7000 7200 7400

ppm/pass

0

2

4

6

8

10

12

14

16

18 A B DC E F000 1201 1202 1101 12011101 120211011001 110110011102 G12011102

A

Figure 5.4: Experimental spectra of CH2ClO2 (top panel) and CH2BrO2 (bottom panel)compared to their Franck-Condon simulations. The calculated T00 values of the conformershave been shifted to match the experimental origin frequencies but the frequencies of thebands relative to the origin are held fixed to their calculated values. The red lettering inboth panels corresponds to transitions belonging to the G conformer of each species. Thereare interferences from HO2 and precursor absorption from 6890-7240 cm−1 and 6980-7190cm−1 in the CH2ClO2 and CH2BrO2 spectra. Tables 5.6 and 5.7 lists the assignments andmeasured experimental frequencies.

73

Table 5.6: Assignments (conformation and vibration) of observed transitions in the A-Xspectrum of CH2BrO2. Band “A” corresponds to origin of the G conformer and subsequentband labels with no prime correspond to transitions belonging to this conformer. Band“A′” corresponds to origin of the T conformer and subsequent bands with a single primelabel correspond to transitions belonging to this conformer.

Band Experimental

Frequency

(cm−1)

Experimental

Shift (cm−1)

Assignment

A 6800 0 000

B 6918 118 1210

C 7209 409 12101110 (1010)

D 7332 532 12201110

E 7365 565 1120

F 7453 653 11101010

G 7487 687 12101120

H 7521 721 810

I 7566 766 710

A′ 7654 0 000

J 7719 919 610

K 7811 1011 1210610

L 8226 1426 121011106

10

B′ 8505 851 810

M 8578 1778 620

N 8664 1864 1210620

74

Table 5.7: Assignments (conformation and vibration) of observed transitions in the A-Xspectrum of CH2ClO2. Band “A” corresponds to origin of the G conformer and subsequentband labels with no prime correspond to transitions belonging to this conformer. Band“A′” corresponds to origin of the T conformer and subsequent bands with a single primelabel correspond to transitions belonging to this conformer.

Band Experimental

Frequency

(cm−1)

Experimental

Shift (cm−1)

Assignment

A 6811 0 000

B 7258 447 1010

C 7275 464 12101110

D 7585 774 11101010

A′ 7703 0 000

E 7830 1019 610

F 8120 1309 1110610

G 8270 1459 1010610

H 8325 1514 121011106

10

B′ 8520 817 810

C′ 8628 925 710

I 8795 1984 620

J 8891 2080 1210620

K 8984 2173 1220620

75

5.4.2 G Conformer OO Stretching and T Conformer Origin Region: 7500-

8400 cm−1

In the 7500-8400 cm−1 portion of the spectra we expect to observe bands corresponding to

OO stretching of the G conformer and, according to ab initio calculations, the T conformer

origin bands of both species. Looking at the CH2ClO2 trace in Figure 5.2, there are two

strong bands at 7703 cm−1 and 7830 cm−1. The G2 calculation predicts the T conformer

origin of CH2ClO2 to be at 7695 cm−1 making the band at 7703 cm−1 an excellent candidate

for the origin of the T conformer. Figure 5.5 shows the experimental trace of CH2ClO2

compared the simulated spectrum of the T conformer. The simulation reproduces the

spectrum very well and confirms the assignment of T conformer origin band of CH2ClO2.

The CH2BrO2 spectrum is slightly more complex in this region and has three relatively

strong bands that occur at 7654 cm−1, 7719 cm−1, and 7811 cm−1. The G2 calculation

predicts the T conformer origin of CH2BrO2 to be at 7629 cm−1 making the bands at 7654

cm−1 and 7719 cm−1 candidates for the origin band. The simulation for the T conformer

origin can be compared to the bands at 7654 cm−1 and 7719 cm−1 to make a determination

as to which one is the origin of the T conformer. Figure 5.6 shows the 7654 cm−1 and 7719

cm−1 bands of CH2BrO2 compared with the simulation of the T conformer origin. The

simulation seems to describe the band contour of both bands making it difficult to assign

one band definitively to the T conformer origin. To aid in the assignemnt of the T conformer

origin band of CH2BrO2 and other bands in this region, the Franck-Condon simulation of

the G conformer can be used to make vibrational assignments of the G conformer and

eliminate bands as candidates for the origin of the T conformer of CH2BrO2.

In Figure 5.7, the 7500-8400 cm−1 regions of CH2ClO2 and CH2BrO2 are shown along

with their Franck-Condon simulations. In the CH2BrO2 trace (bottom panel) there are

several normal modes of vibration that are predicted to have intensity for the G conformer.

The 610 band seems to line up in frequency position and intensity with the band at 7719

cm−1 (J) while the 1210610 combination band lines up well with the band at 7811 cm−1 (K)

reinforcing the assignment of the 7719 cm−1 band to 610. That leaves the 7654 cm−1 band

76

(A′) as the only possible assignment for the T conformer origin of CH2BrO2. The intensity

of the OO stretch vibration is spread over several modes in the G conformer of CH2BrO2: ν6,

ν7, and ν8. The fundamentals of ν6 and ν7 ((H and I) along with the 121011106

10 combination

band (L) are all observed in this spectrum.

In CH2ClO2, the intensity of the OO stretch vibration is only spread over two normal

modes of vibration ν7 and ν6 which are predicted to appear in this part of the spectrum. The

ν7 fundamental is expected to occur near the T conformer origin of CH2ClO2 which makes

assignment of distinct features to either the 710 band of the G conformer or the origin of the

T conformer extremely difficult, hence, the transition at 7703 cm−1 is assigned to the origin

of the T conformer (A′) as it will contribute the most to the intensity of the absorption

band. The ν6 (E) fundamental appears clearly in the spectrum and also forms combination

bands with low frequency torsional and bending vibrations which are also assigned (F, G,

and H). The CH2ClO2 spectrum also contains a ν11ν10 (D) combination band in this region

near where the 2P1/2-2P3/2 transition of the iodine atom is observed

77

Wavenumber

7640 7660 7680 7700 7720 7740 7760

ppm

/pa

ss

4

6

8

10

12

Figure 5.5: Experimental spectrum of CH2ClO2 (top, blue trace) compared to a simulationof the the T conformer origin of CH2ClO2 (bottom, red trace). The simulation was obtainedusing the ab initio rotational constants in Table 5.1 with a pure c-type transition momentwith no contributions from the a and b components.

78

Wavenumber

7580 7600 7620 7640 7660 7680 7700 7720

ppm/pass

0

5

10

15

20

Wavenumber

7660 7680 7700 7720 7740 7760 7780 7800

ppm/pass

0

5

10

15

20

Figure 5.6: Experimental spectra of the CH2BrO2 bands at 7654 cm−1 (top panel) and7719 cm−1 (bottom panel) compared to a simulation of the T conformer origin. (top, bluetrace) compared to a simulation of the the T conformer origin of CH2ClO2 (bottom, redtrace). The black traces in both panels are the experimental traces while the red traces inboth panels were obtained through a simulation using the ab initio rotational constants inTable 5.1 using a pure c-type transition moments with no contributions from the a and bcomponents.

79

Wavenumber (cm-1)

7600 7800 8000 8200 8400

ppm/pass

0

2

4

6

8

10

12 D E F G HA’701 601 1101601000 10016011201110160111011001

Wavenumber (cm-1)

7600 7800 8000 8200 8400

ppm/pass

0

2

4

6

8

10

12

14

16

18 H I J K LA’801 701 601 1201601 12011101601000

Iodine Atom HO2

Figure 5.7: Experimental spectra of CH2ClO2 (top panel) and CH2BrO2 (bottom panel)compared to their Franck-Condon simulations. The calculated T00 values of the T con-formers have been shifted to match the experimental origin frequencies. As in Figure 5.4,the red lettering in both panels corresponds to transitions belonging to the G conformerof each species, while the dark blue lettering corresponds to trasitions belonging to the Tconformer.

80

5.4.3 G Conformer Combination Band and T Conformer OO Stretching

Region: 8400-9200 cm−1

Above 8400 cm−1 the spectra of CH2BrO2 and CH2ClO2 are predicted to display combina-

tion bands belonging to the G conformers and overtone bands of the OO stretch, in addition

to the fundamental OO stretching bands of the T conformers. The OO stretching of the T

conformer for both halogenated species is much more localized in one normal mode than in

the G conformer. Hence, in the spectrum of CH2BrO2 we attribute one transition to the

OO stretching of the T conformer (B′); while two transitions pertaining to OO stretching

vibrations of the T conformer are assigned in the CH2ClO2 spectrum (B′, C′). There are

also transitions attributed to the G conformer of CH2BrO2 and CH2ClO2 in this region

of the spectrum. The overtone of the 620 band (I) is observed in the CH2ClO2 spectrum

along with the 1210620 and 12206

20 combination bands (J and K). The 620 band (M) is also

observed in the CH2BrO2 spectrum along with the 1210620 combination band (N). There

are also interferences from precursors and HO2 which are obscuring parts of the CH2ClO2

and CH2BrO2 spectra making assignments difficult. There is also a noticeable shift in the

predicted transitions of the Franck-Condon simulation vs. the experimental spectrum. This

is because the Franck-Condon simulations makes predictions within the harmonic approxi-

mation, while in reality there may be a fair amount of anharmonicity present in the excited

state potentials of the halogenated methyl peroxy radicals especially exciting this high in

the excited state potential.

5.4.4 Conclusions

While the chemical mechanisms that produce peroxy radicals in the lab and in real world

settings are well known, the following chapter will elaborate on a somewhat controversial

mechanism involving the photolysis of CH2I2 and subsequent reaction with O2 and the

photoproduct(s) that result. Because of this it is important to organize the conclusions

from this chapter so they may be contrasted with the information presented in Chapter 6.

Through looking at the spectra in Figure 5.2 one can see that they do share some similarities

81

Wavenumber (cm-1)

8400 8600 8800 9000 9200

ppm/pass

0

2

4

6

8

10

I JB’ C’ 701 602 1201602801 1202602K

Wavenumber (cm-1)

8400 8600 8800 9000 9200

ppm/pass

0

2

4

6

8

10

12

14 M NB’801 602 1201602

Figure 5.8: Experimental spectra of CH2ClO2 (top panel) and CH2BrO2 (bottom panel)compared to their Franck-Condon simulations. Color coding of transitions is the same is inin Figure 5.7.

82

in their appearance. However, through careful analysis and assignment, it becomes very

clear that they are two unique spectra that represent two distinct chemical species. By

comparing the spectra in Figure 5.2 to the two most closely related alkyl peroxy radicals,

CH3O2 and CH3CH2O2, there are observations that can be made as to what the effect of

halogen substitution can have on the A− X spectra of peroxy radicals.

When substituting a halogen in place of a hydrogen atom in CH3O2, you create two

conformers (G and T) as opposed to the one G conformer of CH3O2. Only considering the

G conformers of the halogenated methyl peroxy radicals, their origin transition energies

are red shifted by ∼600 cm−1 from the origin transition energy of CH3O2 (7381 cm−1,

Figure 5.9).9 This is seen in the spectrum and supported by theoretical calculations and

suggests that the presence of the halogen may shift the A state down leading to the red

shift of the origin transition energy. There are also low frequency torsional bands that are

predicted (some have assigned fundamentals, others show up in combination bands) for the

halogenated peroxy radicals that are not present for CH3O2.

Comparing the spectra of the halogenated peroxy radicals to CH3CH2O2 (Figure 5.10)

can give insight to the effect of having a halogen substituted on instead of a methyl group.

In CH3CH2O2 there are two conformers present (G and T) like in the halogenated peroxy

radicals. The origin transitions of the G and T conformers in CH3CH2O2 are 7362 cm−1 and

7592 cm−1, respectively.160 Like in CH3O2, there is a nearly 600 cm−1 red shift between the

origin bands of the G conformers of CH3CH2O2 and the halogenated peroxy radicals, but

there is not much of a difference between the T conformer origin of CH3CH2O2 and the T

conformer origins of the CH2ClO2 and CH2BrO2, 7654 cm−1 and 7703 cm−1, respectively.

This suggests that when the molecule is in the G conformation, having the methyl group

present is almost the equivalent of having a hydrogen present but the halogens have a

major effect on the relative positions of the X and A states. However, when in the T

conformation and the methyl group or halogens are 180◦ relative to the peroxy group, there

is not much of a difference in the relative positions of the X and A states depending on

what the substituent is.

83

Wavenumber (cm-1)

7200 7400 7600 7800 8000 8200 8400 8600

Absorb

an

ce

(p

pm

/pa

ss)

0

20

40

60

80

100

120

Figure 5.9: A− X spectrum of methyl peroxy.

84

Wavenumber (cm-1)

7200 7400 7600 7800 8000 8200 8400 8600

Ab

sorb

ance

(p

pm

/pa

ss)

0

20

40

60

80

100

120

140

Figure 5.10: A− X spectrum of ethyl peroxy.

85

Chapter 6

Possible Observation of the

a3A′ − X1A′ Electronic Transitionof the Methylene Peroxy

Criegee Intermediate

6.1 Introduction

The previous three chapters have demonstrated our capabilities to observe and interpret

the spectra of alkyl peroxy radicals, including halogenated and OH substituted derivatives.

In the last chapter we shift our focus to investigating another class of reactive intermediates

associated with hydrocarbon oxidation, Criegee intermediates. As mentioned in Chapter

1.2, Criegee intermediates were first proposed in 19493 but had not been directly detected

until only a few years ago when methylene peroxy was observed.84 That initial discovery

was followed by several spectroscopic investigations of the ground and excited states of

methylene peroxy. In most studies methylene peroxy was produced using the mechanism

described in Chapter 2.6 (Fig. 2.7), wherein CH2I2 is photolyzed with 248 nm light in

the presence of O2 to generate CH2O2.84,85,89,90 However, there has been some controversy

regarding this mechanism and the production of CH2IO2 vs. CH2O2 which will be discussed

here before describing our investigations into methylene peroxy.

The initial study by Welz et al.84 was the first to establish that methylene peroxy

could indeed be produced from the reaction of CH2I radicals with O2, but it is an inter-

esting mechanism that deserves extra scrutiny as a vibrationally excited CH2IO2 molecule

86

is formed before CH2O2. It is believed that prior to collisonal stabilization of CH2IO2, the

molecule dissociates the iodine atom to yield CH2O2. The proposal of this mechanism has

motivated several investigations into the CH2I + O2 reaction to understand more clearly

what is happening. Huang et al.161 investigated the CH2I + O2 reaction by monitoring the

2P1/2 ←−2 P3/2 of iodine atom at 7603.138 cm−1 as a function of buffer gas (N2, He, O2,

SF6) pressure. They observed a striking decrease in iodine atom signal at high pressures

of O2 which was not seen with the other buffer gases. They concluded that the decrease in

iodine atom signal was because of collisional stabilization of the CH2IO2 molecule by O2.

This result was consistent with the proposed mechanism but required a large cross section

for deactivation of CH2IO2 by O2. The formation of CH2IO2 must be considered for our

study as it is a peroxy radical and will have its A− X transition in the NIR, which is where

we searched for the a3A′ − X1A′ transition of methylene peroxy.

However, since the above mentioned study there have been two other studies published

that seem to directly contradict the conclusion of Huang et al.161 Su et al.85 reported

the ground state IR spectrum of methylene peroxy and carried out their experiment under

conditions (94.0 total torr: 91.0 torr O2, 2.5 torr N2, 0.1 torr CH2I2) that the results of

Huang et al.161 imply there would be significant collisional stabilization of CH2IO2 and low

yields of methylene peroxy. But Su et al.85 claim that they observed no CH2IO2 in their

spectrum and only observed CH2O2. Additionally, Stone et al.162 repeated the experiments

of Huang et al.,161 but did not observe the large decrease in iodine atom signal at high O2

pressures that Huang et al.161 had reported. This led Huang et al.161 to re-examine their

results and publish a correction of their original work in which they stated that they now

found similar observations to what Stone et al.162 had seen.163

In addition to the spectroscopic work done on the CH2I + O2 reaction, the thermody-

namics of the reaction have been examined to map out a reaction diagram. The results of

Lee et al.97 have been used in combination with our work to examine the thermochemistry

of the following three reactions:

87

CH2XY+ hν −→ CH2X∗ +Y(Y∗) (6.1)

CH2X∗ +O2 −→ CH2OO+X (6.2)

CH2X∗ +O2 +M −→ CH2XOO+M (6.3)

where X,Y=Cl, Br, or I. After initial photolysis of the CH2XY precursor in Equation 6.1, the

CH2X fragment will contain some amount of excess internal energy which can be determined

from Equation 6.4

hν = D0(XH2C − Y ) + ESO + Eint(CH2X) + Et (6.4)

where hν=148.1 kcal/mole (193 nm) or 115.3 kcal/mole (248 nm). D0(XH2C-Y) is the C-Y

bond dissociation energy. ESO is the spin orbit splitting of the Y atom (ESO for Y∗ is given

in Table 6.1), ESO=0.0 kcal/mole for Y channel. Eint(CH2X) is the internal energy of the

CH2X fragment. The values for Et represent the total amount of energy that is converted

to translational energy for the CH2X∗ and Y(Y∗) fragments. The numerical values for these

quantities are given in Table 6.1 for the combination of photolysis energies and precursors

experimentally employed .

As mentioned above, Et represent the amount of translational energy in the CH2X∗

and Y(Y∗) fragments. Some Et values were determined experimentally and are given in

Table 6.1 with appropriate references. Studies on the 248 nm photodissociation of CH2ClI

and CH2BrI could not be found, so their Et values were estimated from the Et values

for the 193 nm photodissociation of CH2Cl2 and 248 nm photodissociation of CH2Br2.

In the experimental studies concerning the photodissociation of CH2Cl2164 and CH2Br2

165

approximately 42% and 33%, respectively, of the excess energy after breaking the C-X bond

was converted to translational energy. For the 248 nm photolysis of CH2ClI and CH2BrI, we

determined the percentage of energy converted to translation by assuming it was the same

as for the 193 nm photodissociation of CH2Cl2 and 248 nm photodissociation of CH2Br2.

The value of Et obtained using this assumption is listed Et column of Table 6.1 for the

88

photolysis of CH2ClI and CH2Br2. With these values in hand Equation 6.4 can be solved

for Eint(CH2X).

The Et values listed need are partitioned between the CH2X∗ and Y(Y∗) fragments by

using the conservation of momentum of these fragments. The portion of the translational

energy that belongs to the CH2X∗ fragment is represented in Table 6.1 by the number in

parenthesis in the Et column. The amount of translational energy that is determined to

belong to the CH2X∗ fragment is then added to Eint to obtain ∆Hex, the total excitation en-

ergy of CH2X∗. The values of ∆Hex(CH2X

∗) are given in Table 6.2 and shown schematically

in Figure 6.1.

Table 6.1: Experimentally determined parameters that were used in Equation 6.4 to deter-mine the amount of internal energy present in the CH2X

∗ fragment. All values are given inkcal/mole. Numbers in parenthesis in Et column represent fraction of Et that is depositedinto CH2X

∗ fragment.

D0 ESO Et

C-Cl 80.9166 - -

C-Br 69.8166 - -

C-I 51.8167 - -

Cl∗ - 2.52168 -

Br∗ - 10.5169 -

I∗ - 21.7170 -

CH2Cl∗ + Cl(Cl∗) - - 30.7164 (17.8)

CH2Cl∗ + I(I∗) - - 28.6 (8.0)

CH2Br∗ + Br(Br∗) - - 16.0165 (8.6)

CH2Br∗ + I(I∗) - - 19.1 (8.1)

CH2I∗ + I(I∗) - - 9.4171 (4.9)

To complete Table 6.1 and Figure 6.2 the enthalpies of reaction for 6.2 (∆Hrxn CH2XO2)

and 6.3 (∆Hrxn CH2O2 + X) were calculated for X=Cl, Br, and I using the following

enthalpies of formation (∆Hf,298K) in kcal/mole: CH2O2 25.3;98 CH2I 51.9±0.7; CH2Br

89

39.8±1.0;172 CH2Cl 27.7±2.0;,173 CH2BrO2 (G) 10.7 (T) 11.5; CH2IO2 (G) 22.7 (T) 23.8;97

I 25.52±0.01; Br 26.74±0.03; 28.99±0.01;.174 The ∆Hrxn CH2XO2 for X=Br and I were

taken from [97]; values for X=Cl were approximated from the values for X=Br and I.

The ∆Hrxn CH2XO2 and ∆Hrxn CH2O2 + X are represented in Figure 6.1 relative to the

entrance channel CH2X + O2 being 0.0 kcal/mole.

Table 6.2: Thermochemical data for the reactions of CH2X* + O2 to form CH2XO2 andCH2O2 + X, all values in kcal/mole. Values of ∆Hex for CH2X* fragments were calculatedconsidering what state the Y fragment is produced in. If the CH2X* fragment has moreenergy, the Y fragment was produced in the 2P3/2 ground state, if it has less energy the Yfragment was produced in the 2P1/2 excited state.

X

Cl Br I

∆Hex CH2X∗ (CH2XI + 248 nm) 42.9 or 21.2 52.5 or 30.8 62.9 or 41.2

∆Hex CH2Cl∗ (CH2Cl2 + 193 nm) 55.3 or 52.8 - -

∆Hex CH2Br∗ (CH2Br2 + 248 nm) - 39.7 or 29.3 -

∆Hrxn CH2XO2 ≈-28 (T), ≈-29 (G) -28.32 (T), -29.14 (G) -28.15 (T), -29.19 (G)

∆Hrxn CH2O2 + X 26.0 12.2 -1.1

In addition to calculating the thermochemistry of the CH2X + O2 reactions and mapping

out the reaction path in Figure 6.1, calculations were done to locate transition states both

for the initial addition of O2 to the CH2X radical and for the dissociation of the halogen

to yield methylene peroxy. Calculations for CH2BrO2 and CH2IO2 were performed in

Reference [97], while calculations for CH2ClO2 were performed in this study. No transition

state was found for the addition of O2 to the CH2X radical, which is not surprising as

addition of O2 to an alkyl radical is a well known barrierless, or near-barrierless, reaction.

However, there was also no transition state found for the dissociation of the halogen atom

for any species to yield CH2O2. This would seem to indicate that the pathway to produce

CH2O2 is energetically available for all of the reactions that are described above. This is

somewhat surprising given that the CH2ClO2 and CH2BrO2 radicals were readily produced

90

CH2X* + O2 + Y(Y*)

CH2XO2 (G) CH2O2 + I CH2O2 + Br CH2O2 + Cl ΔHexCH2XO2 (T)ΔHrxn (CH2XO2)

ΔHrxn (CH2O2 + X)CH2X + O2CH2XY + hν

Figure 6.1: General diagram showing reaction path for the CH2X(∗) + O2 reaction. All ∆Hvalues (kcal/mole) shown are given relative to an entrance channel value of zero for CH2X+ O2.

91

using the above reactions and their spectra analyzed in the previous chapter.

However, there is a similar situation with the production of methyl peroxy, CH3O2,

from the 248 nm photolysis of CH3I. Using the previously described procedure and best

available bond dissociation energies and heats of formations,175–177 the internal energy of

the CH3 fragment (∆Hex CH3) is found to be 32.36 kcal/mole or 10.6 kcal/mole depending

on if the I atom is produce in the 2P3/2 or 2P1/2 state, ∆Hrxn CH3O2 is -30.4 kcal/mole

(there is no barrier for the addition of O2 to CH3),178 and ∆Hrxn CH3O + O is 27.7

kcal/mole (there is no barrier for the addition of O2 to CH3)178 and likewise no barrier

for the dissociation of the terminal oxygen atom to form CH3O).178 From these computed

values it seems that the CH3O + O pathway is energetically available to the CH3 + O2

reaction, but the CH3O2 molecule needs to be still collisionally stabilized to be observed

using this production method.9

The interpretation of the spectrum we have obtained is complicated by the fact that

there is controversy surrounding the mechanism and subsequent product(s) produced; the

interpretation becomes even more problematic when we consider that we do not have ac-

curate calculations for the excited states of methylene peroxy. In Section 1.2, an argument

using ozone was presented to show that ozone and methylene peroxy are similar in many

ways: they are isoelectronic, have similar electronic states, and the energy gap between these

states will also be similar. It was proposed that the a3A′ − X1A′ transition in methylene

peroxy is analogous to the 3A2-1A1 transition (Wulf Band) in ozone and that the transition

for methylene peroxy should appear in the region that we have investigated. Therefore,

in order to assign the carrier of the spectrum we have done a series of chemical tests that

should allow differentiation between the CH2IO2 and CH2O2 carriers. Ab inito calculations

have been performed for the CH2IO2 molecule and a spectral analysis similar to that done

for CH2BrO2 and CH2ClO2 will also be presented. The results of the above spectral and

chemical analyses along with a preliminary assignment of the carrier of the spectrum will

presented in this chapter.

92

6.2 Overview of Experimental Spectrum

Figure 6.2 shows the CRDS spectrum obtained using the 248 nm photolysis of CH2I2. There

are interferences from H2O, the CH2I2 precursor, and the iodine atom which is marked in

the trace). While there is a question as to what the carrier of the spectrum is, it does

display distinct characteristics of a peroxy radical. There is a defined origin region near

6800-6900 cm−1 and ∼900 cm−1 to the blue there appears to be an OO stretch region. This

should come as no surprise as both the CH2IO2 and CH2O2 molecules have a peroxy moiety

present which can account for OO stretching in the spectrum. The spectrum also displays

a high level of complexity as there are more than a dozen transitions that are observed.

As with other reactive species studied, the behavior of the spectrum was monitored as a

function of delay time between NIR probe beam and excimer photolysis, and also in the

presence and absence of oxygen. The absorbance of the photoproduct was found to decrease

with increasing delay between NIR probe beam and excimer photolysis and when O2 was

emitted from the cell, no absorption signals were observed confirming the necessity of O2

being present to produce a spectrum.

Figure 6.3 shows a comparison between the spectra of CH2ClO2, CH2BrO2, and the

photoproduct(s) of the CH2I + O2 reaction. The comparison shows that there is a good

deal of similarity among the spectra, but they all are different spectra that represent distinct

chemical carriers; although it appears that the spectrum resulting from CH2I + O2 reaction

does contain some unique features.

6.3 Chemical Evidence for Assigning Carrier

6.3.1 Test 1: Iodine Atom Absorption

There have been studies done to infer the identity of the photoproduct that is being formed

in the CH2I + O2 reaction by monitoring the concentration of iodine atom as a function

of O2 pressure.161,162 These studies have photolyzed CH2I2 in the absence of O2 to obtain

a baseline reading of iodine atom and then add O2 incrementally. It has been found that

93

Wavenumber (cm-1)

7000 7500 8000 8500 9000

Ab

so

rptio

n (

ppm

/pa

ss)

0

5

10

15

20 H2OContaminationPrecursor Precursor

Iodine atom2P1/2�2P3/2

Figure 6.2: Experimental spectrum obtained from the 248 nm photolysis of CH2I2. Thereare interferences from precursor absorption between 6940-7000 cm−1 and 8700-8750 cm−1,interference from water between 7060-7350 cm−1, and the the 2P1/2-

2P3/2 transition of theiodine atom is present at 7603 cm−1.

94

Wavenumber

7000 7500 8000 8500 9000

Abso

rba

nce

(p

pm

/pa

ss)

0

10

20

30

Figure 6.3: The spectra of CH2ClO2 (top, blue trace), CH2BrO2 (middle, black trace),and of the photoproduct obtained from 248 nm photolysis of CH2I2 in the presence of O2

(bottom, red trace). The CH2ClO2 and CH2BrO2 have been off set by 15 and 7 ppm,respectively, for clarity.

95

there is an initial increase in iodine atom concentration, implying that methylene peroxy

is being formed, but there is a slight decrease in iodine atom as O2 is continually added

to the cell. One drawback of these experiments is that they only selectively monitored the

iodine atom signal, not iodine atom and the organic photoproduct signal simultaneously.

Our experiment can overcome this as our organic photoproduct signal occurs in the 6700-

9100 cm−1 region which overlaps with the 2P1/2-2P3/2 transition of iodine atom and allows

simultaneous observation of iodine atom and our photoproduct.

Figure 6.4 shows the proposed mechanism to make methylene peroxy along with a scat-

terplot that monitors iodine atom and photoproduct signal as a function of O2 pressure. As

predicted by the mechanism, with no O2 present in the cell there is a baseline absorption

of iodine atom from photolysis of CH2I2 and no other photoproduct detected. After O2

is added there is an increase in the iodine absorption along with appearance of the photo-

product of the spectrum, which is expected if the photoproduct is methylene peroxy. As

O2 pressure is increased the iodine atom absorption decreases slightly while the photoprod-

uct absorption does not change. The data shown in Figure 6.4 indicates that methylene

peroxy is being produced by reacting CH2I radicals with O2. But the decrease in iodine

atom signal has been interpreted in other studies as indicating CH2IO2 is being colisionally

stabilized. However, the study by Welz et al. shows a mass spectrum obtained during the

CH2I + O2 reaction and shows the presence of IO and IO2. Iodine atom could be reacting

with O2 to produce those species which would contribute to the decrease in iodine atom

absorption. While the exact reason why the iodine atom signal seems to decrease with

additional O2 pressure is yet to be determined, the data presented in Figure 6.4 suggests

that methylene peroxy is being formed in our experiment. However it does not establish

how much methylene peroxy is being formed or whether it is the carrier of the spectrum in

Figure 6.2.

6.3.2 Test 2: Determination of Self-Reaction Rate

To advance our argument further, and to begin establishing what the carrier of the spectrum

is, we made direct comparisons between our experimental conditions and those of Su et al.85

96

Pressure of O2 Added (Torr)

0 10 20 30 40 50 60 70

pp

m/p

ass

0

20

40

60

80

Figure 6.4: Top panel is the proposed mechanism for the formation of methylene peroxyby 248 nm photolysis of CH2I2. Scatterplot showing the relationship between the intensityof the 2P1/2-

2P3/2 transition of the iodine atom is present at 7603 cm−1, the intensity ofthe photoproduct signal, and the pressure of O2 added. Initial total pressure in cell: 85.0torr: 0.10 torr CH2I2; 84.9 torr N2 mirror purge, window purge, backing N2 on CH2I2. O2

pressures represented by data points. Black data points represent iodine atom absorption,blue data points represent absorption of photoproduct. Error bars represent one standarddeviation of measurements made for one data point as measurements were made in triplicate.

97

In the Su study the ground state of the methylene peroxy species was observed in the IR

without any observable features from the CH2IO2 molecule, which their experiment would

be sensitive to. We used the same synthetic procedure they used (248 nm photolysis of CH2I2

precursor in the present of O2) but under slightly different different pressure conditions. The

Su study used pressures of 94.0 total torr: 0.1 torr CH2I2/2.5 torr N2/91.5 torr O2, while

we used pressures of 85.5 total torr: 0.1 torr CH2I2/84.9 torr N2/0.5 torr O2 to produce the

spectrum in Figure 6.2. However, we did change our pressure conditions to those of the Su

study and conducted the experiment again and did not see any changes to our spectrum.

Following the publication of [85], the authors determined a self-reaction rate constant

of (4±2) x 10−10 cm3 molecule−1 s−1.179 With our apparatus we are also able to determine

self-reaction rates of reactive intermediates by varying the delay between the NIR probe

beam and the excimer photolysis beam and obtaining a time decay profile of the reactive

intermediate (Figure 6.5). To obtain the self-reaction rate constant the initial concentration

of the reactive species needs to be estimated as well. This was done by calculating the the

number of photons absorbed along the path of our precursor by measuring the power output

of the excimer behind our ringdown cell in the absence (82.0 mJ/pulse) and presence (80.0

mJ/pulse) of the CH2I2 precursor. 2.0 mJ/pulse of 248 nm light is absorbed by the CH2I2

precursor, which is 2.50 x 1015 photons. CH2I2 has a quantum yield of 1180 yielding 2.50

x 1015 CH2I radicals, and assuming all CH2I radicals produce methylene peroxy (as Su

et al.179 did) we get 2.50 x 1015 methylene peroxy radicals formed. To determine the

concentration of radicals we take the dimensions of our excimer beam (12.0 cm x 0.1 cm)

and multiply by the path length that the beam is propagating through (15.0 cm). Those

dimensions give us a volume of 18.0 cm3 and a radical concentration of 1.39 x 1014 molecules

cm−3. We then estimate the half-life of the reactive species from our time-delay tests (15

µs) and use the second order integrated rate equation to estimate the self-reaction rate

constant

1

At=

1

A0+ kt (6.5)

98

where At is the concentration of the reactive species at time t, A0 is the initial concentration

of the reactive species, and k is the rate constant. Using this procedure we obtain a self-

reaction rate constant of 4.8±3 x 10−10 cm3 molecule−1 s−1 which is consistent with the

value reported by Su et al.179 While the self-reaction rate constant of CH2IO2 is not

known, those rate constants are known for CH2ClO2 and CH2BrO2 and are 3.7 x 10−12

cm3 molecule−1 s−1 and 1.05 x 10−12 cm3 molecule−1 s−1, respectively. Assuming the rate

constant of CH2IO2 is similar to those for the Cl and Br substituted methyl peroxy radicals,

the rate constant we have determined for the species we are producing in our experiment

indicates that the carrier of the spectrum is methylene peroxy.

6.3.3 Reaction with SO2

Another test we can use to determine the carrier of the spectrum in Figure 6.2 is to compare

the chemical behavior of it to the chemical behavior of the CH2ClO2 and CH2BrO2 in the

presence of SO2. The reaction rate of Criegee intermediates with SO2 has been found to be

orders of magnitude faster than the reaction rate of peroxy radicals with SO2. Methylene

peroxy’s84,162 reaction rate with SO2 is (3.9±0.7) x 10−11 cm3 molecule−1 s−1 while the

reaction rate of methyl peroxy181,182 with SO2 is ≤1 x 10−16 cm3 molecule−1 s−1, which is

essentially no observable reaction between methyl peroxy and SO2. Therefore, when SO2

is added to the reaction cell when the species shown in Figure 6.3 are produced, if the

carriers are typical peroxy radicals there should be no reaction and if the carrier is some

kind of Criegee intermediate there should be a reaction. Figures 6.6 to 6.8 show the results

obtained when SO2 is added to the reaction cell when the species in Figure 6.3 are formed.

Figures 6.6 and 6.7 show what happens to the spectra of CH2ClO2 and CH2BrO2 when 7.0

torr of SO2 is added to the reaction cell. As one would expect if CH2ClO2 and CH2BrO2

are the carriers, there is no discernable decrease in the signal when SO2 is added. However,

Figure 6.8 shows the result when 7.0 torr is added to the reaction cell when the photolysis

of CH2I2 is performed, there is clearly a reaction that takes place between the carrier of

that spectrum and SO2. Based on the reaction rate information provided above, the carrier

would appear to be the methylene peroxy Criegee intermediate.

99

Wavenumber (cm-1)

7400 7600 7800 8000

Absorb

ance (

ppm

/pass)

0

5

10

15

20

25

Figure 6.5: Time decay profile of photoproduct from the CH2I + O2 reaction. Each colortrace represents a different delay time between NIR probe beam and excimer photolysisbeam: black trace is 1 µs delay, blue trace is 10 µs delay, green trace is 25 µs delay, redtrace is 50 µs delay, cyan trace is 75 µs delay, and pink trace is 100 µs delay.

100

Wavenumber (cm-1)7000 7500 8000 8500 9000

Abso

rban

ce

(p

pm

/pass)

0

5

10

15

20

25

Figure 6.6: The experimental traces obtained from the 248 nm photolysis of CH2ClI in theabsence of SO2 (top, blue trace) and presence of 7.0 torr SO2 (bottom, green trace). Bluetrace was shifted +7 ppm for clairity.

101

Wavenumber (cm-1)

7000 7500 8000 8500 9000

Ab

so

rba

nce

(p

pm

/pa

ss)

0

5

10

15

20

25

Figure 6.7: The experimental traces obtained from the 248 nm photolysis of CH2Br2 in theabsence of SO2 (top, black trace) and presence of 7.0 torr SO2 (bottom, green trace). Blacktrace was shifted +7 ppm for clairity.

102

Wavenumber (cm-1)7000 7500 8000 8500 9000

Absorb

ance (

pp

m/p

ass)

-5

0

5

10

15

Figure 6.8: The experimental traces obtained from the 248 nm photolysis of CH2I2 in theabsence of SO2 (top, red trace) and presence of 7.0 torr SO2 (bottom, green trace). Greentrace was shifted -40 ppm for clarity.

103

6.4 Spectroscopic Evidence for Assigning Carrier

The previous section presented several pieces of chemical evidence in an attempt to assign

the carrier of the spectrum in Figure 6.2; in this section spectroscopic arguments will be

advanced in an attempt to assign the carrier. As was presented in Chapters 3 and 5, a

powerful way that spectroscopic assignments are made is through theoretical calculations.

However, at the present time no theoretical calculations exist for the 3A′ state of methylene

peroxy and a detailed theoretical analysis is not possible to predict the a3A′− X1A′ CH2O2

spectrum, but in section 6.4.1 we consider spectral predictions for the A2A−X2A spectrum

of CH2IO2. In addition to the theoretical arguments that are made later in the chapter

general observations regarding the spectrum shown in Figure 6.2 and possible assignment

can be made.

Spectral observations that are consistent with the methylene peroxy carrier include

examining how the origin frequency shifts between CH3O2, CH2ClO2, CH2BrO2, and the

spectrum in Figure 6.2. Figure 6.3 shows the spectra of CH2ClO2, CH2BrO2, and the

spectrum from Figure 6.2. If one looks at the CH2ClO2 and CH2BrO2 spectra and even at

the methyl peroxy spectrum (Figure 5.9), a trend can be seen concerning the identity of the

substituent on the carbon atom and position of the origin band of the G conformer. When

the substituents are only hydrogen atoms (methyl peroxy) an origin value of 7381 cm−1 is

observed. Upon halogen substitution with chlorine and bromine atoms the origin band red

shifts to 6811 and 6800 cm−1, respectively. If this trend is extrapolated to include iodine

atom substitution (CH2IO2), the origin band would be expected to red shift even further.

But this is not what is observed experimentally as there is a triplet of bands that occurs

in the CH2I + O2 spectrum that occurs at 6821, 6851, and 6902 cm−1 which are all blue

shifted from the origins of CH2ClO2 and CH2BrO2. This result would seem to go against

the trend that as the substituent gets heavier the origin band will continue to red shift.

The second observation to note is that what seems to be the OO stretch region of the

spectrum is moderately red shifted from what is usually observed with peroxy radicals.

The triplet structure between 6800-6900 cm1 was mentioned in the previous paragraph;

104

this structure is repeated in Figure 6.2 at 7697, 7727, and 7779 cm−1 which is the same

separation these peaks had in the 6800-6900 cm−1 region and is only shifted by 876 cm−1.

Typically OO stretches for peroxy radicals are between 900-1000 cm−1 from the origin

Spectral observations consistent with the CH2IO2 carrier include the direct comparison

made between spectra in Figure 6.3. Qualitatively, all of the spectra look similar: they have

origin bands in the 6800-6900 cm−1 and the step structure in the 7500-7700 cm−1 region

of the CH2BrO2 and CH2I + O2 spectra is also very similar and quite unique. Forming the

opinion that the spectrum resulting from the CH2I + O2 reaction is CH2IO2 based on these

comparisons would be reasonable. Another option available to advance a spectroscopic

argument for the carrier of the spectrum in Figure 6.2 is to perform calculations on the

CH2IO2 molecule. However, the presence of the iodine atom presents a problem as a

pseudopotential is required to describe it theoretically and therefore limits what we have

been able to previously do in terms of computations of peroxy radicals. The theoretical

methods used for CH2IO2 and their results are presented in the following sections.

6.4.1 Computational Methods for CH2IO2

Quantum chemical calculations using the Gaussian 09W141 package were performed to pre-

dict relative energies, oscillator strengths, origin bands, and vibrational frequencies for the

different conformers of CH2IO2. Relative energies of the different conformers were deter-

mined from their calculated UB3LYP aug-cc-pvtz pp energies for the X state; these relative

energies were subsequently used in a Boltzmann population analysis of the conformers at

room temperature. For the conformers of CH2IO2, origin values were obtained by using the

zero point corrected energies of the X and A state at the UB3LYP aug-cc-pvtz pp basis set

and TDUB3LYP aug-cc-pvtz pp, respectively, and taking the difference. From previous

studies we have discovered that the TD method of calculating excited states gives overes-

timated A − X transition energies. To counter this overestimation, the origin transition

of several standard peroxy radicals (methyl peroxy, ethyl peroxy, trifluoromethyl peroxy,

bromomethyl peroxy) were calculated with the G2 method and the TD method described

here. The origin transition energies from each method were calculated and the difference

105

between the G2 and TD origin transition energies were determined to calculate a calibration

that could be applied generally to TD calculations to get a more accurate origin transition

energy. A calibration of 1444 cm−1 was determined and applied to the calculated origin

transition energy of CH2IO2.

As with CH2ClO2 and CH2BrO2, a simulation of the rotational contour and Franck-

Condon simulations were performed for the conformers predicted to be in the spectrum.

The details of which can be found in Appendix D.

6.4.2 Computational Results for CH2IO2

Based on searches of the X state potential energy surface of CH2IO2, it is predicted that

there will be two stable conformers. The different conformers are distinguished by the

∠ICOO dihedral angle. If the ∠ICOO dihedral angle takes on a values of ∼±60◦the con-

former is labeled as gauche (G), if the ∠ICOO angle takes on a value of 180◦ the conformer

is labeled as trans (T) configuration. Table 6.3 lists the conformers along with a summary of

the results from the calculations mentioned in the previous paragraph. Based on Boltzmann

populations at room temperature and calculated oscillator strengths, only the G conformer

of CH2IO2 is predicted to be observed. The X state equilibrium geometry for the G con-

former of CH2IO2 and its X and A state geometrical parameters ar presented in Appendix

D. As with other peroxy radicals, upon excitation to the A state a lengthening of the O-O

bond is predicted along with a moderate change in the ∠OOC bond angle.8 There is also

a fairly large change in the ∠ICOO angle upon excitation to the A state which was also

observed in the study of the 2,1-HPP radical. The fact that there are several geometrical

changes occurring in the vicinity of the O2 group should come as no surprise as it is the

chromophore of the A− X transition.

Table 6.4 presents a list of the normal mode vibrational frequencies for the G conformer

of CH2IO2 calculated at the UB3LYP/aug-cc-pvtz pp level of theory (TD for excited state)

with qualitative descriptions of each normal mode.

106

Table 6.3: Results of theoretical calculations for conformers of CH2IO2: relative X stateenergies (cm−1), calibrated origin frequencies (T00, cm

−1), Boltzmann weights (w), oscilla-tor strengths (f ), and relative intensities (I=w x 105 f ) for conformers of CH2IO2. X stateenergies for the respective conformers of CH2IO2 are reported are relative to the G conform-ers. The rotational constants in the ground and excited states are labeled as A′′/B′′/C ′′

and A′/B′/C ′, respectively. Oscillator strengths, Boltzmann wieghts, and I are unitless; allother values are given in cm−1.

CH2IO2

G T

X 0 317

T00 6724 7545

w 1.000 0.108

105 f 5.81 1.47

I 5.81 0.16

A′′ 0.544 1.19

B′′ 0.056 0.0474

C ′′ 0.053 0.0460

A′ 0.480 1.19

B′ 0.059 0.0476

C ′ 0.054 0.0462

107

Table 6.4: Predicted X and A state vibrations for the G conformer of CH2IO2. Modenumbering follows Herzberg’s notation and is based on the values of the A state frequencies.

Mode X Description Mode A Description

ν12 85 OOCI torsion ν12 100 OOCI torsion

ν11 250 OCI bend ν11 279 OCI bend

ν10 502 COO bend ν10 398 COO bend

ν9 544 C-I stretch ν9 519 C-I stretch

ν8 847 C-O stretch ν8 847 CH2 rick

ν7 930 C-H2 rock ν7 905 O-O stretch + C-O stretch

ν6 1124 O-O stretch ν6 943 O-O stretch + C-O stretch

ν5 1249 C-H2 torsion ν5 1232 C-H2 torsion

ν4 1260 C-H2 symmetric sway ν4 1252 C-H2 sway

ν3 1445 HCH bend ν3 1457 HCH bend

ν2 3114 C-H2 symmetric stretch ν2 3103 C-H2 symmetric stretch

ν1 3207 C-H2 asymmetric stretch ν1 3201 C-H2 asymmetric stretch

108

6.4.3 Spectral Analysis with G Conformer of CH2IO2

In Chapter 5 one of the ways we were able to confirm that CH2ClO2 and CH2BrO2 were

the carriers of the spectra presented was through comparison of simulations of origin band

rotational contours with the origin bands in the experimental spectrum. We do the same

analysis here as the simulation of the G conformer of CH2IO2 is compared to what we

believe is the origin band of the spectrum presented in this chapter, Figure 6.9. One thing

to consider is that the origin band of the G conformer is predicted to be at 6725 cm−1 (Table

6.3) and the band we consider to be the origin is at 6902 cm−1 which is 177 cm−1 blue shifted

from the predicted origin position. The simulation in Figure 6.9 seems to do a relatively good

job of predicting the band contour. A sharp peak is predicted in the simulation and is seen

in the experimental trace and there is also some additional structure predicted to the blue

of the sharp peak in the simulation which could account for the shoulder that is observed

in the experimental trace. However, the splitting between the peaks in the simulation

is 3-4 cm−1 which we would expect to resolve under our experimental conditions. This

means that if we are indeed producing CH2IO2 in the experiment, the origin band should

appear as a multiplet instead of a sharp peak with a shoulder on the blue side. While the

simulation does seem to describe the band shape of what we are considering the origin band

of the spectrum, numerous questions remain trying to reconcile the simulation with what

is observed experimentally.

Another method employed that aids in the assignment of the carrier (and also the way

we vibrationally analyze the spectrum) is by producing a Franck-Condon simulation. In

Figure 6.10 the spectrum obtained from the CH2I + O2 reaction is shown with the Franck-

Condon simulation of the G conformer of CH2IO2. Let us first examine the 6800-6900cm−1

region of the spectrum before looking at the data holistically. In the 6800-6900 cm−1 region

we observe what we believe to be the origin band of the spectrum (6902 cm−1) and have

shifted the Franck-Condon simulation to line up with it. But looking at the spectrum we

also observe two bands to the red at 6821 and 6851 cm−1. These bands are not produced

in the Franck-Condon simulation, but that could be due to these bands being hot bands

109

or sequence bands which would not be produced by our Franck-Condon simulation as the

simulation is done at 0 K. Looking at Table 6.4 could give an idea as to where the sequence

bands would appear. According to the vibrational frequencies listed, sequence bands for the

two lowest frequency modes (1211 and 1111) would be blue shifted relative to the origin, so

they cannot account for the transitions at 6821 and 6851 cm−1. However, hot bands (1201,

1202, 1101) would be red shifted from the origin and could be the source of these transitions.

For example the 1201 transition would occur 85 cm−1 to the red of the origin band, 1101

would occur 250 cm−1 to the rd of the origin, and the 1212 band would be 60 cm−1 to the

red of the origin. While we can make statements about the frequency position of these hot

bands based upon calculated the calculate frequencies, the intensities of the bands is another

matter. To obtain accurate intensities for these hot bands, the Franck-Condon factors and

oscillator strengths need to be calculated which is currently being done by another member

of the group and will hopefully shed light on this situation.

Looking at the rest of the Franck-Condon simulation, there is a strong progression on

modes 12 and 11 involving fundamental, combination, and overtone bands along with a

clearly defined OO stretch region and then combination bands built off of the OO stretch.

There is difficulty in trying to assign the transitions predicted in the simulation with tran-

sitions in the spectrum as there is not a clear correlation between the two (there was a

clear correlation in 2,1-HPP, CH2ClO2, CH2BrO2) and the difficulty is amplified by the

fact that the simulation fails at accurately describing the relative intensities of the bands

in the experimental spectrum. Overall, it seems that the Franck-Condon simulation of the

G conformer of CH2IO2 does a poor job of describing our experimental spectrum.

6.5 Conclusion

This chapter has presented spectral data of the photoproduct from the CH2I + O2 reac-

tion and discussed several arguments that could aid in the determination of the carrier

of the spectrum. These arguments have included chemical evidence, qualitative spectral

evidence regarding general appearance of the spectrum, trends concerning halogen substi-

110

Wavenumber (cm-1)

6800 6820 6840 6860 6880 6900 6920 6940

Ab

so

rba

nce

(p

pm

/pa

ss)

0

2

4

6

8

10

12

14

Figure 6.9: Spectra obtained from the CH2I + O2 reaction (top, red trace) compared to asimulation (bottom, black trace) of the G conformer origin band of CH2IO2. The simulationwas obtained through using ab initio the rotational constants in Table 6.3. The simulationfor the G conformer of CH2IO2 use a ratio of components of the transition dipole along theinertial axis of a:b:c=0.48:0.40:0.13.

111

Wavenumber (cm-1)

7000 7500 8000 8500 9000

Ab

so

rba

nce

(p

pm

/pa

ss)

0

5

10

15

20

25

Figure 6.10: Experimental spectrum of CH2I + O2 reaction compared to the Franck-Condonsimulation of the G conformer of CH2IO2. The origin transition of the Franck-Condon simu-lation has been shifted to match what we consider to be the origin band of the experimentalspectrum, but the frequencies of the bands relative to the origin are held fixed to their cal-culated values.

112

tution of methyl peroxy radical, and finally an argument focused on theoretical analysis of

the spectrum in terms of the G conformer of CH2IO2. The chemical evidence presented for

determining the carrier of the spectrum in Figure 6.2 is all consistent with the carrier of the

spectrum being methylene peroxy, especially the compelling evidence of the self-reaction

rate that was determined and the reaction with SO2. The theoretical analysis resulted

in simulations of the rotational contour and vibrational structure of of the G conformer of

CH2IO2 and provided a chance to compare them to the experimental spectrum to determine

how well they describe the spectrum. While an argument can be made that the rotational

simulation does capture some aspects of the origin band in our experimental spectrum, the

Franck-Condon simulation did not seem to accurately describe our experimental spectrum

at all leading to serious question as the whether the G conformer of CH2IO2 is the carrier

of our spectrum.

113

Chapter 7

Conclusions

Cavity ringdown spectroscopy is a very powerful and versatile technique that can be used

to probe species in very small concentrations under varying conditions. This dissertation

has demonstrated how CRDS can be applied to probe the NIR A − X spectra of long

chain peroxy radicals, OH substituted peroxy radicals, halogenated peroxy radicals, and

(possibly) the a3A′ − X1A′ transition of the methylene peroxy under room temperature

conditions and (in some experiments) jet-cooled conditions. The A − X spectra of the

OH substituted and halogenated peroxy radicals were assigned in terms of the excited

state vibrations of the molecules’ respective conformers by utilizing ab initio calculations to

produce Franck-Condon simulations and simulations of rotational contours for the different

conformers predicted to be in the spectrum. The power of spectral/structural relationships

was demonstrated by applying those relations to assign the spectra of the long chain peroxy

radicals. Typical transitions involving OO stretching were observed for the all of peroxy

radical systems studied while COO bending and low frequency torsional fundamentals were

identified for the halogenated and OH substituted species.

However, as of the writing of this thesis, the carrier of the spectrum presented in Chapter

6 is still not definitively known. There are theoretical calculations in progress that will

hopefully allow the definitive assignment of the carrier and allow the vibrational assignment

of the spectrum. In lieu of theoretical calculations, a series of chemical tests were performed

with the results presented in Chapter 6 along with a presentation of computations done on

the G conformer of CH2IO2. The chemical tests were designed to help determine the carrier

114

of the spectrum from the two possibilities: CH2O2 and CH2IO2. The results of looking at the

iodine atom intensity and photoproduct intensity as a function of O2 pressure indicates that

methylene peroxy is being formed. Comparison between our experimental conditions and

those of Su et al., and ultimately performing the experiment under identical experimental

conditions to Su et al., would seem to indicate that we are producing methylene peroxy.

The correlation between our calculated self-reaction rate constant of (4.8±3) x 10−10 cm3

molecule−1 s−1 and the rate constant in the literature179 of (4±2) x 10−10 cm3 molecule−1

s−1 gives us evidence that not only is methylene peroxy being formed, but that it is the

carrier our spectrum. Ultimately the chemical tests done with SO2 seem be a clear indication

that the carrier of the spectrum is indeed methylene peroxy.

Chapter 6 also included a summary of the theoretical work done on the CH2IO2 molecule

and compared simulations we have obtained with the experimental spectrum. While those

results may stimulate some discussion about CH2IO2 being the carrier of the spectrum, I

do not believe that what is presented in terms of CH2IO2 being the carrier is convincing

at all and those results in conjunction with the results of the chemical tests presented in

Chapter 6 along with the entire body of work done around the world on the CH2I + O2

reaction clearly indicates that CH2O2 is the carrier of our spectrum.

115

Bibliography

[1] J. Zador, C. A. Taatjes, and R. X. Fernandes, PECS 37, 371 (2011).

[2] J. A. Jafri and D. H. Phillips, J. Am. Chem. Soc. 112, 2586 (1990).

[3] R. Criegee and G. Wenner, Chem. Ber. 9, 564 (1949).

[4] W. R. Wadt and W. A. Goddard, J. Am. Chem. Soc. 97, 3004 (1975).

[5] S. Wu, P. Dupre, P. Rupper, and T. A. Miller, J. Chem. Phys. 127, 224305 (2007).

[6] S. Wu, P. Dupre, and T. A. Miller, Phys. Chem. Chem. Phys. 8, 1682 (2006).

[7] L. Gibbs, B. Anderson, K. Barnes, G. Engeler, J. Freel, J. Horn, M. Ingham, D.Kohler, D. Lesini, R. MacArthur, M. Mortier, D. Peyla, B. Taniguchi, A. Tiedman,S. Welstand, D. Bernhardt, K. Collini, A. Farr, J. Jones, J. Lind, C. Tom, and J.Benson, Technical report, Chevron (unpublished).

[8] E. N. Sharp, P. Rupper, and T. A. Miller, Phys. Chem. Chem. Phys. 10, 3955 (2008).

[9] C.-Y. Chung, C.-W. Cheng, Y.-P. Lee, H.-Y. Liao, E. N. Sharp, P. Rupper, and T. A.Miller, J. Chem. Phys. 127, 044311 (2007).

[10] P. D. Lightfoot, R. A. Cox, J. N. Crowley, M. Destriau, G. D. Hayman, M. E. Jenkin,G. K. Moortgat, and F. Zabel, Atmos. Envir. 26A, 1805 (1992).

[11] G. S. Tyndall, R. A. Cox, C. Granier, R. Lesclaux, G. K. Moortgat, M. J. Pilling,A. R. Ravishankara, and T. J. Wallington, J. Geophys. Res. 106, 12157 (2001).

[12] O. J. Nielsen and T. J. Wallington, Peroxyl Radicals (John Wiley and Sons, NewYork, NY, 1997).

[13] I. Glassman, Combustion (Academic Press, Orlando, FL, 1987).

[14] H. J. Curran, P. Gaffuri, W. J. Pitz, and C. K. Westbrook, Combust. Flame 114, 149(1998).

[15] S. Wang, D. L. Miller, N. P. Cernansky, H. J. Curran, W. J. Pitz, and C. K. Westbrook,Combust. Flame 118, 415 (1999).

116

[16] P. Gaffuri, T. Faravelli, E. Ranzi, N. P. Cernansky, D. Miller, A. D’Anna, and A.Ciajolo, Am. Inst. Chem. Eng. 43, 1278 (1997).

[17] A. Perdih and F. Perdih, Acta Chim. Slov. 53, 306 (2006).

[18] A. T. Bablan and I. Motoc, MATCH (Commun. Math. Chem.) 5, 197 (1979).

[19] C. Morley, Combust. Sci. Tech. 55, 115 (1987).

[20] J. Warnatz, U. Mass, and R. W. Dibble, Combustion: Physical and Chemical Fun-damentals, Modeling and Simulation, Experiments, Pollutant Formation, 3rd ed.(Springer, Berlin, 2001).

[21] C. K. Westbrook, Chem. Ind. 15, 562 (1992).

[22] A. Guenther, C. N. Hewitt, D. Erickson, R. Fall, C. Geron, T. Graedel, P. Harley,L. Klinger, M. Lerdau, W. A. Mckay, T. Pierce, B. Scholes, R. Steinbrecher, R.Tallamraju, J. Taylor, and P. Zimmerman, J. Geophys. Res. 100, 8879 (1995).

[23] A. R. Ravishankara, PNAS 106, 13639 (2009).

[24] M. J. Pilling and I. W. M. Smith, in Modern Gas Kinetics: Theory, Experiment, andApplication (Blackwell Scientific Publications, Oxford, 1987), Chap. C1 and C2.

[25] R. A. Rasmussen and M. A. Khalil, J. Geophys. Res. 93, 1417 (1988).

[26] S. Sawada and T. Totsuka, Atmos. Environ. 20, 821 (1986).

[27] W. L. Chameides, R. Lindsay, J. Richardson, and C. S. Kiang, Science 241, 1473(1988).

[28] M. Trainer, E. J. Williams, D. D. Parrish, M. P. Buhr, E. J. Allwine, H. H. Westberg,F. C. Fehsenfeld, and S. C. Liu, Nature 329, 705 (1987).

[29] J. G. Calvert, R. Atkinson, J. A. Kerr, S. Madronich, G. K. Moortgat, T. J. Walling-ton, and G. Yarwood, The Mechanisms of Atmospheric Oxidation of the Alkenes(Oxford University Press, New York, New York, 2000).

[30] J. D. Surrat, A. W. H. Chan, N. C. Eddingsaas, M. Chan, C. L. Loza, A. J. Kwan,S. P. Hersey, T. C. Flagan, P. O. Wenneberg, and J. H. Seinfeld, Proc. Natl. Acad.Sci. U.S.A. 0911114107 (2009).

[31] B. Ghosh, A. Bugarin, B. T. Connell, and S. W. North, J. Phys. Chem. A 114, 2553(2010).

[32] T. S. Dibble, J. Comput. Chem. 26, 836 (2005).

[33] W. Lei, R. Zhang, W. S. McGivern, A. Derecskei-Kovacs, and S. W. North, J. Phys.Chem. A 105, 471 (2001).

[34] R. Zhang and W. Lei, J. Chem. Phys. 113, 8574 (2000).

[35] W. Lei and R. Zhang, J. Phys. Chem. A 105, 3808 (2001).

117

[36] S. Guo, J. Tan, Y. Ma, F. Yang, Y. Yu, and J. Wang, Atmos. Res. 101, 701 (2011).

[37] I. Dıaz-Acosta, J. R. Alvarez-Iabaoy, and A. Vivier-Bunge, Int. J. Chem. Kinet 31,29 (1999).

[38] J. Zador, A. W. Jasper, and J. A. Miller, Phys. Chem. Chem. Phys. 11, 11040 (2009).

[39] J. Loison, J. Daranlot, A. Bergeat, F. Caralp, R. Mereau, and K. M. Hickson, J. Phys.Chem. A 114, 13326 (2010).

[40] R. J. Cvetanovic, 12th International Symposium on Free Radicals (PUBLISHER, AD-DRESS, 1976).

[41] J. Peeters, W. Boullart, V. Pultau, S. Vandenberk, and L. Vereecken, J. Phys. Chem.A 111, 1618 (2007).

[42] R. S. Martin, H. Westberg, E. Allwine, L. Ashman, J. C. Farmer, and B. Lamb, J.Atmos. Chem. 13, 1 (1991).

[43] C. Wiedinmyer, S. Friedfeld, W. Baugh, J. Greenberg, A. Guenther, M. Fraser, andD. Allen, Atmos. Environ. 35, 1001 (2001).

[44] G. D. Silva, C. Graham, and Z. Wang, Environ. Sci. Tech. 44, 250 (2010).

[45] F. Paulot, J. D. Crounse, H. G. Kjaergarrd, A. Kurten, J. M. S. Clair, J. H. Seinfeld,and P. O. Wenneberg, Science 325, 730 (2009).

[46] P. Campuzano-Jost, M. B. Williams, L. D’Ottone, and A. J. Hynes, J. Phys. Chem.A 108, 1537 (2004).

[47] W. Lee, M. Baasandorj, P. S. Stevens, and R. A. Hites, Environ. Sci. Technol. 39,1030 (2005).

[48] J. Peeters, T. L. Nguyen, and L. Vereecken, Phys. Chem. Chem. Phys. 11, 5935(2009).

[49] S. S. Brown, J. A. DeGouw, C. Warneke, T. B. Ryerson, W. P. Dube, E. Atlas,R. J. Weber, R. E. Peltier, J. A. Neuman, J. M. Roberts, A. Swanson, F. Flocke,S. A. McKeen, J. Brioude, R. Sommariva, M. Trainer, F. C. Fehsenfeld, and A. R.Ravishankara, Atmos. Chem. Phys. 9, 3027 (2009).

[50] E. E. Greenwald, B. Ghosh, K. C. Anderson, K. S. Dooley, P. Zou, T. Selby, D. L.Osborn, G. Meloni, C. A. Taatjes, F. Goulay, and S. W. North, J. Phys. Chem. A114, 904 (2010).

[51] M. E. Jenkin, A. A. Boyd, and R. Lesclaux, J. Atmos. Chem. 29, 267 (1998).

[52] M. M. Maricq, R. E. Chase, D. H. Podsiadlik, W. O. Seigl, and E. W. Kaiser, Societyof Automotive Engineers SAE 952518, (1995).

[53] J. Li, A. Kazakov, and F. L. Dryer, Int. J. Chem. Kinet. 33, 859 (2001).

118

[54] N. Miyamoto, H. Ogawa, N. M. Nurun, K. Obata, and T. Arima, Society of Automo-tive Engineers SAE 980506, (1998).

[55] K. H. Song, P. Nag, T. A. Litzinger, and D. C. Haworth, Combust. Flame 135, 341(2003).

[56] C. K. Westbrook, W. J. Pitz, and H. J. Curran, J. Phys. Chem. A. 110, 6912 (2006).

[57] A. Frassoldati, C. A., T. Faravelli, U. Niemann, R. E., R. Seiser, and K. Seshadri,Combust. Flame 157, 2 (2010).

[58] Z. Serinyel, P. O’Toole, E. Larkin, G. O’Malley, and H. J. Curran, Energ. Fuel 23,5886 (2009).

[59] T. Kasper, P. Oswald, U. Struckmeier, K. Kohse-Hoinghaus, C. A. Taatjes, J. Wang,T. A. Cool, M. E. Law, A. Morel, and P. R. Westmoreland, Combust. Flame 156,1181 (2009).

[60] Y. Li, L. Wei, Z. Tian, B. Yang, J. Wang, T. Zhang, and F. Qi, Combust. Flame 152,336 (2008).

[61] T. P. Murrells, M. E. Jenkin, S. J. Shalliker, and G. D. Hayman, J. Chem. Soc.,Faraday Trans. 87, 2351 (1991).

[62] T. J. Wallington, P. Dagaut, and M. J. Kurylo, Chem. Rev. 92, 667 (1992).

[63] R. Atkinson, D. L. Baulch, R. A. Cox, R. F. Hampson, J. A. Kerr, and J. Troe, J.Phys. Chem. Ref. Data 18, 881 (1989).

[64] E. Villenave and R. Lesclauz, J. Phys. Chem. 100, 14372 (1996).

[65] O. J. Nielsen, J. Sehested, S. Langer, E. Ljungstroem, and I. Wangberg, Chem. Phys.Lett. 238, 359 (1995).

[66] S. Newman, I. Lane, A. Orr-Ewing, D. Newnham, and J. Ballard, J. Chem. Phys.110, 10749 (1999).

[67] O. Horie and G. Moortgat, Acc. Chem. Res. 31, 387 (1998).

[68] W. Sander, Agnew. Chem. Int. Ed. Engl. 29, 344 (1990).

[69] W. H. Bunnelle, Chem. Rev. 91, 335 (1991).

[70] N. M. Donahue, G. T. Drozd, S. A. Epstein, A. A. Presto, and J. H. Kroll, Phys.Chem. Chem. Phys. 13, 10848 (2011).

[71] C. A. Taatjes, D. E. Shallcross, and C. J. Percival, Phys. Chem. Chem. Phys. 16,1704 (2014).

[72] G. Copeland, M. V. Ghosh, D. E. Shallcross, C. J. Percival, and J. M. Dyke, Phys.Chem. Chem. Phys. 13, 14839 (2011).

[73] R. Criegee, Agnew. Chem. Int. Ed. 14, 745 (1975).

119

[74] D. Johnson and G. Marston, Chme. Soc. Rev. 37, 699 (2008).

[75] G. Copeland, M. V. Ghosh, D. E. Shallcross, C. J. Percival, and J. M. Dyke, Phys.Chem. Chem. Phys. 13, 17461 (2011).

[76] Y. K. Wei and R. J. Cvetonovic, Can. J. Chem 41, 913 (1963).

[77] J. T. Herron and R. E. Huie, J. Phys. Chem 78, 2085 (1974).

[78] J. T. Herron and R. E. Huie, Int. J. Chem. Kinet. 10, 1019 (1978).

[79] H. Niki, P. D. Maker, C. M. Savage, and L. P. Breitenbach, Chem. Phys. Lett. 46,327 (1977).

[80] F. Su, J. G. Calvert, and J. H. Shaw, J. Phys. Chem. 84, 239 (1980).

[81] S. Hatakeyama, H. Kobayashi, and H. Akimoto, J. Phys. Chem 88, 4736 (1984).

[82] T. Vrbaski and R. J. Cvetonovic, Can. J. Chem 38, 1063 (1960).

[83] C. A. Taatjes, G. Merloni, T. M. Selby, A. J. Trevitt, D. L. Osborn, C. J. Percival,and D. E. Shallcross, J. Am. Chem. Soc. 130, 11883 (2008).

[84] O. Welz, J. D. Savee, D. L. Osborn, S. S. Vasu, C. J. Percival, D. E. Shallcross, andC. A. Taatjes, Science 335, 204 (2012).

[85] Y. T. Su, Y. H. Huang, H. A. Witek, and Y. P. Lee, Science 340, 174 (2013).

[86] M. Nakajima and Y. Endo, J. Chem. Phys. 139, 101103 (2013).

[87] M. C. McCarthy, L. Cheng, K. N. Crabtree, O. Martinez, T. L. Nguyen, C. C. Wom-ack, and J. F. Stanton, J. Chem. Phys. Lett. 4, 4133 (2013).

[88] A. M. Daly, B. J. Drouin, and S. S. Yu, J. Mol. Spec. (2014).

[89] J. M. Beames, F. Liu, L. Lu, and M. I. Lester, J. Am. Chem. Soc. 134, 20045 (2012).

[90] L. Sheps, J. Phys. Chem. Lett. 4, 4201 (2013).

[91] J. M. Beames, F. Liu, L. Lu, and M. I. Lester, J. Chem. Phys. 138, 244307 (2013).

[92] C. A. Taatjes, O. Welz, A. J. Eskola, J. D. Savee, A. M. Scheer, D. E. Shallcross, B.Rotavera, E. P. F. Lee, J. M. Dyke, D. K. Mok, D. L. Osborn, and C. J. Percival,Science 340, 177 (2013).

[93] M. Nakajima and Y. Endo, J. Chem. Phys. 140, 011101 (2014).

[94] P. C. Hiberty, J. Am. Chem. Soc. 98, 6088 (1976).

[95] L. B. Harding and W. A. Goddard, J. Am. Chem. Soc. 100, 7180 (1978).

[96] D. Cremer, J. Am. Chem. Soc. 101, 7199 (1979).

[97] E. P. F. Lee, D. K. W. Mok, D. E. Shallcross, C. J. Percival, D. L. Osborn, C. A.Taatjes, and J. M. Dyke, Chem. Eur. J. 18, 12411 (2012).

120

[98] M. T. Nguyen, T. L. Nguyen, V. T. Ngan, and H. M. T. Nguyen, Chem. Phys. Lett.448, 183 (2007).

[99] R. Crehuet, J. M. Anglada, D. Cremer, and J. M. Bofill, J. Phys. Chem. A 106, 3917(2002).

[100] J. Gauss and D. Cremer, Chem. Phys. Lett. 133, 420 (1987).

[101] D. Cremer, J. Gauss, E. Kraka, J. F. Stanton, and R. J. Bartlett, Chem. Phys. Lett.1993, 547 (209).

[102] H. Han, B. Suo, D. Xie, Y. Lei, Y. Wang, and Z. Wen, Phys. Chem. Chem. Phys. 13,2723 (2011).

[103] Y. Matsumi and M. Kawasaki, Chem. Rev. 103, 4767 (2003).

[104] R. Bacis, A. J. Bouvier, and J. M. Flaud, Spect.rochim. Acta, Part A 54, 17 (1998).

[105] S. M. Anderson and K. Mauersberger, J. Geopjys. Res. 100, 3033 (1995).

[106] S. M. Anderson, J. Morton, and K. Mauersberger, J. Chem. Phys. 93, 3826 (1990).

[107] P. J. Hay and T. D. Dunning, J. Chem. Phys. 67, 2290 (1977).

[108] S. M. Anderson, J. Maeder, and K. Mauersberger, J. Chem. Phys. 94, 6351 (1991).

[109] B. Minaev and H. A. gren, Chem. Phys. Lett. 217, 531 (1994).

[110] M. Braunstein and T. T. Pack, J. Chem. Phys. 96, 6378 (1992).

[111] B. F. Minaev and E. M. Kozlo, J. Struct. Res. 38, 895 (1997).

[112] A. J. Bouvier, R. Bacis, B. Bussery, S. Churassy, D. Inard, M. Nota, J. Brion, J.Malicet, and S. M. Anderson, Chem. Phys. Lett. 255, 263 (1996).

[113] A. J. Bouvier, D. Inard, V. Veyret, B. Bussery, R. Bacis, S. Churassy, J. Brion, J.Malicet, and R. H. Judge, J. Mol. Spec. 190, 189 (1998).

[114] D. Inard, A. J. Bouvier, R. Bacis, S. C. F. Bohr, J. Brion, J. Malicet, and M. Jacon,Chem. Phys. Lett. 287, 515 (1998).

[115] I. N. Levine, in Physical Chemistry (McGraw-Hill, Inc., New York, 1995), Chap. Spec-troscopy and Photochemistry.

[116] J. M. Heberlin, J. A. McKay, M. A. Kwok, R. Ueunten II, D. S. Urevig, D. J. Spencer,and D. J. Bernard, Appl. Opt. 19, 144 (1980).

[117] A. O’Keefe and D. A. G. Deacon, Rev. Sci. Instrum. 59, 12 (1988).

[118] C. Wang, A. Mbi, and M. Shepherd, IEEE Sensors J. 10, 54 (2010).

[119] M. Snels, T. Venezia, and L. Belfiore, Chem. Phys. Lett. 489, 134 (2010).

121

[120] H. D. Osthoff, S. S. Brown, T. B. Ryerson, T. J. Fortin, B. M. Lerner, E. J. Williams,A. Pettersson, T. Baynard, W. P. Dube, S. J. Ciciora, and A. R. Ravishankara, J.Geophys. Res. 111, D12305(1) (2006).

[121] S. S. Brown, H. Stark, and A. R. Ravishankara, Appl. Phys. B 75, 173 (2002).

[122] M. Aldener, S. S. Browns, H. Stark, J. S. Daniel, and A. R. Ravishankara, J. Mol.Spec. 232, 223 (2005).

[123] D. S. Sayres, E. J. Moyer, T. F. Hanisco, J. M. St. Clair, F. N. Keutsch, A. O’ Brien,N. T. Allen, L. Lapson, J. N. Demusz, M. Rivero, T. Martin, M. Greenberg, C. Tuoz-zolo, G. S. Engel, J. H. Kroll, J. B. Paul, and J. G. Anderson, Rev. Sci. Instruments80, 044102(1) (2009).

[124] M. F. Witinski, D. S. Sayres, and J. G. Anderson, Appl. Phys. B 102, 375 (2011).

[125] E. J. Moyer, D. S. Sayres, G. S. Engel, J. M. St. Clair, F. N. Keutsch, N. T. Allen,J. H. Kroll, and J. G. Anderson, Appl. Phys. B 92, 467 (2008).

[126] P. B. Tarsa, A. D. Wist, P. Rabinowitz, and K. K. Lehmann, Appl. Phys. Lett. 85,4523 (2004).

[127] J. J. Scherer, J. B. Paul, A. O’Keefe, and R. J. Saykally, Chem. Rev. 97, 25 (1997).

[128] M. D. Wheeler, S. M. Newman, A. J. Orr-Ewing, and M. N. R. Ashfold, J. Chem.Soc. Faraday Trans. 94, 337 (1998).

[129] G. Berden, R. Peeters, and G. Meijer, Int. Rev. Phys. Chem. 19, 565 (2000).

[130] R. Chhantyal-Pun, N. D. Kline, P. S. Thomas, and T. A. Miller, J. Phys. Chem. Lett.1, 1846 (2010).

[131] M.-W. Chen, G. M. P. Just, T. Codd, and T. A. Miller, J. Chem. Phys. 135, 184304(2011).

[132] B. Ghosh, A. Bugarin, B. T. Connel, and S. W. North, J. Phys. Chem. A 114, 5299(2010).

[133] I. Mathieu-Pelta and S. A. Evans, J. Org. Chem. 57, 3409 (1992).

[134] A. V. Baklanov and L. V. Krasnoperov, J. Phys. Chem. A 105, 97 (2001).

[135] M. Ahmed, D. Blunt, D. Chen, and A. G. Suits, J. Chem. Phys. 106, 7617 (1997).

[136] T. Ohta, H. Bandow, and H. Akimoto, Int. J. Chem. Kinet 14, 173 (1982).

[137] G. M. P. Just, P. Rupper, T. A. Miller, and W. L. Meerts, Phys. Chem. Chem. Phys.112, 4773 (2010).

[138] P. S. Thomas and T. A. Miller, Chem. Phys. Lett. 491, 123 (2010).

[139] P. S. Thomas, R. Chhantyal-Pun, and T. A. Miller, J. Phys. Chem. A 114, 218 (2010).

122

[140] S. J. Zalyubovsky, D. Wang, and T. A. Miller, Chem. Phys. Lett. 335, 298 (2001).

[141] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheese-man, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Cari-cato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg,M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima,Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr., J. E. Peralta, F.Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R.Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar,J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V.Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J.Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G.Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels,O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski, and D. J. Fox, Gaussian 09,Revision A.01, Gaussian, Inc., Wallingford, CT, 2009.

[142] R. Borrelli and A. Peluso, J. Chem. Phys. 119, 8437 (2003).

[143] E. H. Fink and D. A. Ramsay, J. Mol. Spec. 185, 304 (1997).

[144] B. Hamilton and R. J. Falkiner, in Fuels and Lubricants Handbook: Technology, Prop-erties, Performance, and Testing, 37 ed., (Astm. Intl., PA, U.S.A, 2003), Chap. 3:Motor Gasoline, pp. 61–87.

[145] A. Bhasin, Ph.D. thesis, University of Iowa, 2010.

[146] N. D. Kline and T. A. Miller, Chem. Phys. Letts. 530, 16 (2012).

[147] B. G. Glover and T. A. Miller, J. Phys. Chem. A 109, 11191 (2005).

[148] D. Sarzynski and B. Sztuba, Int. J. Chem. Kinet. 34, 651 (2002).

[149] D. Melnik, R. Chhantyal-Pun, and T. A. Miller, J. Phys. Chem. A 114, 11583 (2010).

[150] L. J. Carpenter, P. S. Liss, and S. A. Penkett, J. Geophys. Res. 108(D9), 4256 (2003).

[151] D. L. Albritton and R. T. Watson, Montreal Protocol Assessment Supplement (PUB-LISHER, Nairobi, Kenya, 1993).

[152] S. A. S. A. P. H. Butler, R. J. Cicerone, A. Golombek, M. J. Kurylo, J. S. Levine, M.Molina, R. Prinn, C. E. Reeves, J. M. Rodriguez, H. Singh, D. Toohey, and R. Weiss,Montreal Protocol Assessment Supplement (PUBLISHER, Nairobi, Kenya, 1994).

[153] M. J. Molina and F. S. Rowland, Nature 249, 810 (1974).

[154] L. Wachowski, P. Kirszensztejn, and Z. Foltynowics, Pol. J. Environ. Stud. 10, 415(2001).

[155] V. Catoire, R. Leaslaux, and P. D. P. D. L. Rayez, J. Phys. Chem 98, 2889 (1994).

[156] E. Villenave and R. Lesclaux, Chem. Phys. Lett. 236, 376 (1995).

123

[157] O. J. Nielsen, J. Munk, G. Locke, and T. J. Wallington, J. Phys. Chem 95, 8714(1991).

[158] V. L. Stakhursky and T. A. Miller, 56th International Symposium on Molecular Spec-troscopy (The Ohio State University, Columbus, Ohio, 2001) .

[159] R. Atkinson, D. L. Baulch, R. A. Cox, J. N. Crowley, R. F. Hamspon, R. G. Hynes,M. E. Jenkin, M. J. Rossi, and J. Troe, Atmos. Chem. Phys. 6, 3625 (2006).

[160] P. Rupper, E. N. Sharp, G. Tarczay, and T. A. Miller, J. Phys. Chem. A 111, 832(2007).

[161] H. Huang, A. J. Eskola, and C. S. Taatjes, J. Phys. Chem. Lett. 3, 3399 (2012).

[162] D. Stone, M. Blitz, L. Daubney, T. Ingham, and P. Seakins, Phys. Chem. Chem.Phys. Comm. (2013).

[163] H. Huang, B. Rotavera, A. J. Eskola, and C. A. Taatjes, J. Phys. Chem. Lett. 4, 3824(2013).

[164] X. Yang, P. Felder, and J. R. Huber, Chem. Phys. (1994).

[165] Y. Lee, C. Chen, and S. Lin, J. Chem. Phys. 118, 10494 (2003).

[166] E. Tschuikow-Roux and S. Paddison, Int. J. Chem. Kinet. 19, 15 (1987).

[167] A. F. Lago, J. P. Kerchner, A. Bodi, B. Sztaray, B. Miller, D. Wurzelman, and T.Baer, J. Phys. Chem. A 109, 1802 (2005).

[168] H. Uehara and K. Horiai, J. Opt. Soc. Am. B 4, 1217 (1987).

[169] J. V. Kasper, C. R. Pollock, R. F. Curl, and F. K. Tittle, Chem. Phys. Lett. 77, 211(1981).

[170] Y. He, M. Quack, R. Ranz, and G. Seyfang, Chem. Phys. Lett 215, 228 (1993).

[171] J. H. Lehman, H. Lei, and M. I. Lester, Chem. Phys. Lett. 590, 16 (2103).

[172] P. Marshall, G. N. Srinivas, and M. Schwartz, J. Phys. Chem. A 109, 6371 (2005).

[173] C. F. Rodriquez, D. K. Bohme, and A. C. Hopkins, J. Phys. Chem. A 100, 2942(1996).

[174] J. D. J. D. C. D. Wagman and V. A. Medvedev, CODATA Key Values for Thermo-dynamics (Hemisphere Publishing Company, New York, 1989).

[175] R. Zhu and M. C. Lin, J. Phys. Chem. A. 105, 6243 (2001).

[176] A. Eppink and D. H. Parker, J. Chem. Phys. 46, 4758 (1999).

[177] G. Li, H. J. Hwang, and H. C. Jung, Rev. Sci. Inst 76, 023105 (2005).

[178] S. P. Walch, Chem. Phys. Lett. 215, 81 (1993).

124

[179] Y. Su, H. Lin, R. Putikam, H. Matsui, M. C. Lin, and Y. P. Lee, Nature Chem.(2014).

[180] S. Y. Chen, P. Y. Tsai, H. C. Lin, C. C. Wu, K. S. Lin, B. J. Sun, and A. H. H.Chang, J. Chem. Phys. 134, 034315 (2011).

[181] P. D. L. P. L. A. Cox, J. N. J. N. C. Destriau, G. D. Hayman, M. E. Jenkin, M. J.Rossi, and J. Troe, Atmos. Chem. Phys. 6, 3625 (2006).

[182] C. S. Kan and J. G. J. G. C. H. Shaw, J. Phys. Chem. 85, 1126 (1981).

125

Appendix A

X and A State Frequencies of2,1-HPP Conformers

126

Table A.1: Predicted X and A state vibrations for the G′1G2G3T4T5 conformer of 2,1-HPP.

Mode numbering follows Herzberg’s notation and is based on the values of the A statefrequencies. All vibrational frequencies (cm−1) were calculated via UB3LYP/aug-cc-pVTZ.

Mode X Description Mode A Description

ν33 82 CCOO torsion ν33 83 CCOO torsion

ν32 109 CCOO + CCOH torsion ν32 137 CCOO rock

ν31 215 CH3 twist ν31 216 CH3 twist

ν30 229 CCOH rock ν30 227 OCCO torsion

ν29 338 CCC bend ν29 312 CCC bend

ν28 351 OH wag ν28 376 OH wag

ν27 451 CCOH bend ν27 405 CCOH torsion

ν26 470 H2C-CH3 symmetric bend ν26 481 CCOH/OCCO torsion

ν25 570 COO bend ν25 541 COO bend

ν24 840 CH2/CH/CH3 molecular breathing ν24 842 C-O(H) stretch + HC-CH2 stretch

ν23 876 C-O(O) stretch ν23 880 C-O(O) stretch + C-O(H) stretch

ν22 937 H2C-C-CH3 breathing ν22 934 CH2/CH3/CH bend

ν21 962 C-OO stretch + C-OH stretch ν21 957 O-O stretch + C-O(H) stretch

ν20 1055 O-H wag/H3-C stretch ν20 993 O-O stretch + C-O(O) stetch

ν19 1133 CCC bend ν19 1058 CCC bend

ν18 1150 O-O stretch + C-OH stretch ν18 1141 CH3/CH/CH2 bend

ν17 1168 O-O stretch + OCCO torsion ν17 1167 C-O(H) stretch + CH3 wag

ν16 1257 O-H wag/stretch ν16 1264 C-H2/C-H symmetric bend

ν15 1284 O-H/C-H wag ν15 1282 O-H/C-H wag

ν14 1331 CH2/CH bend ν14 1344 CH3/CH2/CH/OH wag

ν13 1385 CH2/CH bend ν13 1389 C-H/CH2 wag

ν12 1406 CH3/CH/CH2 bend ν12 1411 CH2/CH3 bend

ν11 1427 H3-C(OH) stretch ν11 1427 CH3/CH2/CH/OH breathing

ν10 1463 C-H2 scissor ν10 1466 C-H2 scissor

ν9 1487 C-H3 scissor ν9 1487 C-H3 scissor

ν8 1502 C-H3 scissor ν8 1503 C-H3 scissor

ν7 3002 C-H stretch ν7 2989 C-H stretch

ν6 3038 C-H3 symmetric stretch ν6 3035 C-H3/C-H2 symmetric stretch

ν5 3065 C-H2 symmetric stretch ν5 3041 C-H3/C-H2 symmetric stretch

ν4 3102 C-H3 asymmetric stretch ν4 3099 C-H3/C-H2 asymmetric stretch

ν3 3114 C-H3 asymmetric stretch ν3 3111 C-H2 asymmetric stretch

ν2 3130 C-H2 asymmetric stretch ν2 3113 C-H3 asymmetric stretch

ν1 3804 O-H stretch ν1 3790 O-H stretch

127

Table A.2: Predicted X and A state vibrations for the G1G2G3G4G5 conformer of 2,1-HPP. Mode numbering follows Herzberg’s notation and is based on the values of the Astate frequencies. All vibrational frequencies (cm−1) were calculated via UB3LYP/aug-cc-pVTZ.

Mode X Description Mode A Description

ν33 69 CCOO torsion ν33 89 CCOO torsion

ν32 152 CCOO + CCOH torsion ν32 146 CCOO rock

ν31 238 OCCC torsion ν31 240 CH3 twist

ν30 242 CH3 twist ν30 257 CCOH torsion + OCCO torsion

ν29 369 CCC bend ν29 347 CCC bend

ν28 399 OH wag + CCOH torsion ν28 395 H3C-CH-CH2 scissor

ν27 409 OH wag + CCOH torsion ν27 412 OH wag

ν26 470 COO bend ν26 445 COO bend

ν25 662 CCOO torsion + OCCO torsion ν25 663 CCOH torsion + OCCO torsion

ν24 782 CH2-CH-CH3 symmetric stretch ν24 791 C-O(H) stretch + HC-CH2 stretch

ν23 875 C-O(O)/HC-CH3 symmetric stretch ν23 877 C-O(O) stretch + HC-CH3 stretch

ν22 943 CCCO torsion ν22 939 CH2-CH-CH3 stretch

ν21 960 CCOH torsion + C-OH stretch ν21 958 O-O stretch + C-O(H) stretch

ν20 1043 CCOH torsion ν20 997 O-O stretch + C-O(O) stetch

ν19 1132 O-O stretch + HC-CH3 stretch ν19 1042 CCOH torsion

ν18 1144 CCC bend ν18 1125 CCC bend

ν17 1164 O-O stretch + C-O(H) stretch ν17 1145 H2-C-CH-CH3 symmetric stretch

ν16 1278 CH2 bend ν16 1263 CH2/CH/OH wag

ν15 1287 CCOH bend + OH wag ν15 1296 CH3CH/CH2/OH/ wag

ν14 1356 CH2/CH bend ν14 1349 CH3/CH2/CH wag

ν13 1377 CH2/CH bend ν13 1382 C-H/CH2 wag

ν12 1414 CH3 breathing ν12 1412 CH3 wag

ν11 1425 OH wag + CH wag ν11 1417 CH3/CH2/CH breathing

ν10 1468 CH2 scissor ν10 1467 C-H2 scissor

ν9 1492 C-H3 scissor ν9 1494 C-H3 scissor

ν8 1502 C-H3 scissor ν8 1498 C-H3 scissor

ν7 3023 C-H stretch ν7 3020 C-H stretch

ν6 3034 C-H3 symmetric stretch ν6 3036 C-H3 symmetric stretch

ν5 3067 C-H2 symmetric stretch ν5 3049 C-H/C-H2 symmetric stretch

ν4 3098 C-H3 asymmetric stretch ν4 3098 C-H3/ asymmetric stretch

ν3 3117 C-H3 asymmetric stretch ν3 3112 C-H2 asymmetric stretch

ν2 3126 C-H2 asymmetric stretch ν2 3115 C-H3 asymmetric stretch

ν1 3741 O-H stretch ν1 3751 O-H stretch

128

Table A.3: Predicted X and A state vibrations for the T1G2G3T4T5 conformer of 2,1-HPP.Mode numbering follows Herzberg’s notation and is based on the values of the A statefrequencies. All vibrational frequencies (cm−1) were calculated via UB3LYP/aug-cc-pVTZ.

Mode X Description Mode A Description

ν33 60 CCOO torsion ν33 96 CCOO torsion

ν32 129 CCOO + CCOH torsion ν32 132 CCOO + CCOH torsion

ν31 183 OCCC torsion ν31 184 OCCC torsion

ν30 218 CH3 twist ν30 219 CH3 twist

ν29 340 CCC bend ν29 324 CCC bend

ν28 347 OH wag ν28 369 O-H wag

ν27 419 COO bend ν27 380 OH wag + CCC bend

ν26 480 OCC bend ν26 485 COO bend

ν25 525 COO bend ν25 498 CCO bend

ν24 858 CH2-CH-CH3 symmetric stretch ν24 854 C-O(H) stretch + HC-CH2 stretch

ν23 917 C-O(O)/HC-CH3 symmetric stretch ν23 919 C-O(O) stretch + HC-CH3 stretch

ν22 938 CCCO torsion ν22 944 CH2-CH-CH3 asymmetric stetch

ν21 961 C-O(H) stretch ν21 974 C-O(O) stretch + C-O(H) stretch

ν20 1058 CCOH torsion ν20 1007 O-O stretch

ν19 1124 CCOH torsion + C-O(H) stretch ν19 1062 CCOH torsion

ν18 1170 CCC bend ν18 1145 CCOH torsion

ν17 1191 O-O stretch ν17 1159 H2-C-CH-CH3 symmetric stretch

ν16 1234 CH2 wag ν16 1239 CH2/CH/OH wag

ν15 1278 CH + OH wag ν15 1280 CH3CH/CH2/OH/ wag

ν14 1339 CH2/CH bend ν14 1345 CH3/CH2/CH wag

ν13 1393 Molecular breathing ν13 1388 C-H/CH2 wag

ν12 1412 CH2/CH/CH3/OH wag ν12 1409 CH3 wag

ν11 1426 H3-C(OH) stretch ν11 1426 CH3/CH2/CH breathing

ν10 1476 CH2 scissor ν10 1486 C-H2 scissor

ν9 1487 C-H3 scissor ν9 1500 C-H3 scissor

ν8 1503 C-H3 scissor ν8 1503 C-H3 scissor

ν7 2988 C-H stretch ν7 2989 C-H stretch

ν6 3040 C-H3 symmetric stretch ν6 3027 C-H3 symmetric stretch

ν5 3056 C-H2 symmetric stretch ν5 3041 C-H/C-H2 symmetric stretch

ν4 3103 C-H3 asymmetric stretch ν4 3081 C-H3/ asymmetric stretch

ν3 3116 C-H3 asymmetric stretch ν3 3106 C-H2 asymmetric stretch

ν2 3119 C-H2 asymmetric stretch ν2 3115 C-H3 asymmetric stretch

ν1 3802 O-H stretch ν1 3795 O-H stretch

129

Appendix B

Complete X and A StateGeometrical Parameters for the

Conformers of BromomethylPeroxy and Chloromethyl

Peroxy

130

Parameterof Interest X AC-H1 1.08 Å 1.08 ÅC-H2 1.08 Å 1.09 ÅC-Cl 1.79 Å 1.81 ÅC-O1 1.43 Å 1.40 ÅO-O 1.32 Å 1.39 Å∠H1CH2 114.7o 112.7o∠H1CO1 105.7o 104.7o∠H2CO1 109.3o 111.5o∠H1CCl 108.2o 107.6o∠H2CCl 108.6o 106.7o∠O1CCl 111.3o 113.5o∠O2O1C 111.8o 110.7o∠O2O1CCl -84.9o -73.9o∠H1CO1O2 157.8o 169.0o∠H2CO1O2 35.1o 46.8o~ ~CH2ClO2 G

1 21 2

Figure B.1: Complete X and A geometric parameters for the G conformer of CH2ClO2

(A state values are italicized). Dihedral and bond angle values given in degrees (◦); bondlengths given in angstroms (A).

131

Parameterof Interest X AC-H1 1.08 Å 1.08 ÅC-H2 1.08 Å 1.08 ÅC-Cl 1.77 Å 1.77 ÅC-O1 1.45 Å 1.43 ÅO-O 1.32 Å 1.39 Å∠H1CH2 112.8o 112.3o∠H1CO1 108.6o 110.1o∠H2CO1 108.6o 110.1o∠H1CCl 109.7o 109.0o∠H2CCl 109.7o 109.0o∠O1CCl 107.3o 106.2o∠O2O1C 109.1o 107.1o∠O2O1CCl 180.0o 180.0o∠H1CO1O2 61.5o 62.1o∠H2CO1O2 -61.5o -62.1o~ ~CH2ClO2 T1 2 1 2

Figure B.2: Complete geometric parameters for the T conformer of CH2ClO2 (A state valuesare italicized). Dihedral and bond angle values given in degrees (◦); bond lengths given inangstroms (A).

132

Parameterof Interest X AC-H1 1.09 Å 1.08 ÅC-H2 1.09 Å 1.09 ÅC-Br 1.94 Å 1.96 ÅC-O1 1.43 Å 1.40 ÅO-O 1.32 Å 1.39 Å∠H1CH2 113.7o 113.0o∠H1CO1 105.5o 104.8o∠H2CO1 109.6o 111.9o∠H1CBr 108.5o 107.8o∠H2CBr 108.2o 106.5o∠O1CBr 111.3o 113.1o∠O2O1C 110.3o 107.3o∠O2O1CBr -77.6o -70.3o∠H1CO1O2 42.1o 49.9o∠H2CO1O2 164.9o 172.6o~ ~CH2BrO2 G

1 212

Figure B.3: Complete geometric parameters for the G conformer of CH2BrO2 (A statevalues are italicized). Dihedral and bond angle values given in degrees (◦); bond lengthsgiven in angstroms (A).

133

1 212 Parameterof Interest X AC-H1 1.09 Å 1.08 ÅC-H2 1.09 Å 1.09 ÅC-Br 1.92 Å 1.92 ÅC-O1 1.45 Å 1.43 ÅO-O 1.31 Å 1.40 Å∠H1CH2 112.9o 112.2o∠H1CO1 109.0o 110.9o∠H2CO1 109.0o 110.9o∠H1CBr 109.5o 108.6o∠H2CBr 109.5o 108.6o∠O1CBr 106.7o 105.4o∠O2O1C 108.0o 104.3o∠O2O1CBr 180.0o 180.0o∠H1CO1O2 -61.9o -62.7o∠H2CO1O2 61.9o 62.7o~ ~CH2BrO2 T

Figure B.4: Complete geometric parameters for the G conformer of CH2BrO2 (A statevalues are italicized). Dihedral and bond angle values given in degrees (◦); bond lengthsgiven in angstroms (A).

134

Appendix C

Comparison of ProductionMethods for CH2ClO2 and

CH2BrO2

135

Wavenumber

7000 7500 8000 8500 9000

Ab

sorb

an

ce (

pp

m/p

ass

)

0

5

10

15

20

Figure C.1: Comparison between the 248 nm photolysis of CH2ClI (top, blue trace) and193 nm photolysis of CH2Cl2 (bottom, pink trace) production methods for CH2ClO2. Thetop trace was shifted +8 ppm for clarity.

136

Wavenumber (cm-1)

7000 7500 8000 8500 9000

Abso

rba

nce

(p

pm

/pass)

0

5

10

15

20

25

Figure C.2: Comparison between the 248 nm photolysis of CH2Br2 (top, black trace) and248 nm photolysis of CH2BrI (bottom, pink trace) production methods for CH2BrO2. Thetop trace was shifted +8 ppm for clarity.

137

Appendix D

Computational Details forCH2IO2

For the purpose of qualitatively predicting band origin profiles of the G conformer of

CH2IO2, room temperature simulations of its rotational contour was produced using our

group’s SpecView158 program using purely ab initio rotational constants. The rotational

constants for the G conformer of CH2IO2 were obtained at the UB3LYP/aug-cc-pvtz pp

(TD for excited state) level of theory for the ground and excited states. The simulations

also required the values for the transition dipole moment along the a, b, and c inertial axes

which were calculated by the UCIS/aug-cc-pvtz pp method.

We also undertook Franck-Condon simulations in order to help assign the vibrational

structure present in each spectrum. For the purpose of calculating multi-dimensional

Franck-Condon factors we used Gaussian09W.141 Equilibrium geometries, normal mode vec-

tors, and harmonic vibrational frequencies used as input were calculated at the UB3LYP/aug-

cc-pvtz pp level of theory for the X state and the TDUB3LYP/aug-cc-pvtz pp level of theory

for the Astate with the simulations being performed in the limit of cold (0 K) absorption.

138

Parameterof Interest X AC-H1 1.08 Å 1.08 ÅC-H2 1.08 Å 1.09 ÅC-I 2.18 Å 2.21 ÅC-O1 1.42 Å 1.39 ÅO1-O2 1.32 Å 1.38 Å∠H1CH2 114.2o 113.5o∠H1CO1 110.1o 112.2o∠H2CO1 106.6o 106.6o∠H1CI 107.4o 105.3o∠H2CI 106.7o 106.2o∠O1CI 112.0o 114.0o∠O2O1C 112.4o 111.3o∠O2O1CI -88.3o -72.8o∠H1CO1O2 31.1o 46.8o∠H2CO1O2 155.4o 171.0o~ ~CH2IO2 G1 2 12

Figure D.1: X and A structures of G conformer of CH2IO2, A state values are italicized.Dihedral and bond angle values given in degrees (◦); bond lengths given in angstroms (A).

139