Citation
Determination of bond dissociation energies using Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometry

Material Information

Title:
Determination of bond dissociation energies using Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometry
Creator:
Yang, Yarjing
Publication Date:
Language:
English
Physical Description:
xiv, 236 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Anions ( jstor )
Ethers ( jstor )
Halides ( jstor )
Ionization ( jstor )
Ions ( jstor )
Lasers ( jstor )
Mass spectroscopy ( jstor )
Photodetachment ( jstor )
Signals ( jstor )
Solvents ( jstor )
Chemical bonds ( lcsh )
Chemistry thesis, Ph. D ( lcsh )
Dissertations, Academic -- Chemistry -- UF ( lcsh )
Dissociation ( lcsh )
Ion cyclotron resonance spectrometry ( lcsh )
Mass spectrometry ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 223-234).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Yarjing Yang.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
028143177 ( ALEPH )
38253662 ( OCLC )

Downloads

This item has the following downloads:


Full Text










DETERMINATION OF BOND DISSOCIATION ENERGIES USING FOURIER
TRANSFORM ION CYCLOTRON RESONANCE
(FT-ICR) MASS SPECTROMETRY















By

YARJING YANG













A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1996








ACKNOWLEDGMENTS



I would like to express my appreciation to my research advisor, Dr. John R. Eyler, for his guidance and assistance, especially for his input of FT-ICR background and theory. I also appreciate the technical knowledge of the instrumentation provided by Dr. Cliff Watson. I would like to thank Drs. Dil Peiris and Nathan Ramanathan for their efforts on the initial IRMPD work of the metal/crown ether complexes. The comments and suggestions on my research projects from Dr. Philip Brucat and Dr. David Richardson are also acknowledged.

My appreciation extends to my parents; without their love and support, I would never be able to finish this work. Thanks to all of the members (Rich Burton, Kevin Goodner, Tom Hayes, David Kage, Lisa Lang, Eric Milgram, and Joe Nawrocki) of the current Eyler group for their assistance and friendship. Thanks also go to Dr. Kitty Williams for her assistance in the preparation of research results for publication.

Last but certainly not least, I would like to thank my dear husband, ChenChan Hsueh, for his support, love and help. He has been so understanding and helpful during my graduate career. Without him, I would not have been able to finish the graduate studies.













TABLE OF CONTENTS



ACKNOW LEDGMENTS ........................................... ii

TABLE OF CONTENTS .......................................... iii

LIST O F TABLES ............................................... vi

LIST O F FIG URES .............................................. vii

A BSTRA C T .................................................. xiii

CHAPTERS

1 INTRO DUCTIO N ..................................... 1

Fourier Transform Ion Cyclotron Resonance Mass
Spectrometry (FTICR-MS) ............................. 1
H isto ry .......................... ... ...... .... 1
Basic Concepts ................................ 2
Ion cyclotron motion ........................ 2
Ion excitation ............................. 6
Ion detection 11
Signal acquisition ......................... 16
A pplications ........................................ 20
Purpose of This W ork ........................... 20
Solvated Anions ............................... 24
Host-guest Complexes .......................... 26
Methodology of Determination of Bond Dissociation Energies 37
Photodetachment .............................. 37
CAD ............ 45
Energy behavior of reaction cross section ...... 45 On-resonance CAD ....................... 47
Doppler broadening ....................... 49
Advantages, Disadvantages, and Comparisons to Other
Techniques ........................................ 50

iii







Photodetachment............................ 50
CAD ..................................... 51

2. INSTRUMENTATION.............................. 53

Introduction ...............................53
Nicolet 2 Tesla ICR Instrument ........................ 56
Bruker 7 Tesla ICR Instrument ........................ 66
Electrospray Ionization (ESI) Source Region .........71 Introduction of ESI Features ..................... 77
pKa or pKb............................ 78
Concentration .......................... 78
Solubility............................. 79
Molecular conformation................... 79
Molecular weight ........................ 80
Solvent properties ....................... 80
Ion Transfer Region ........................... 81
Ion Analyzer region ........................... 81

3. DETERMINATION OF BINDING ENERGIES OF
SINGLY-SOLVATED HALIDE IONS USING
PHOTODETACHMENT ............................. 88

Background..................................... 88
Experimental Procedures and Data Analysis .............. 92
Chemicals and Ion-Molecule Reactions ............. 92
Pulse Sequence............................ 109
Data Analysis .................................. 109
Results and Discussion............................ 120
Conclusion .................................... 141

4. BOND DISSOCIATION ENERGIES OF ALKALI AND
HYDRONIUM CATIONS BOUND TO CROWN ETHERS DETERMINED BY ON-RESONANCE COLLISIONALLYACTIVATED DISSOCIATION (CAD) .................. 144

Background .................................... 144
Experimental ................................... 150
Ion formation.............................. 150
Pulse sequence............................ 154
Data Analysis .................................. 155
Ion Appearance Curve ........................ 155
Internal Energies ............................ 159

iv








Rotational energy ........................ 159
Vibrational energies ...................... 159
Trial Function ................................ 160
Deconvolution ................................ 161
Results .................... 163
Collision Conditions ........................... 163
Fragmentation Pathways ....................... 164
Bond Dissociation Energies ..................... 197
D iscussion ........................................ 210
C onclusion ....................................... 216

5. CO NCLUSION .................................... 218

REFERENCE LIST ............................................ 223

BIOGRAPHICAL SKETCH ...................................... 235






























v













LIST OF TABLES




1. 1. Size of alkali-metal ions and cavity size of crown ethers ....... 36

2.1. Laser dyes used for photodetachment experiments and their
corresponding wavelength range before and after frequency
doubling ...................................... 68

3.1. Electron affinities (EA) and proton affinities (PA) of species
studied in this work ............................... 131

3.2. Photdetachment thresholds (ET) and binding energies (BE's) of
solvated halides ................................. 133

4. 1. Vibrational frequencies of cyclohexane and I ,4-dioxane ...... 152

4.2. Ions observed in mass spectra. The capillary potential was 25 V
for ions labelled with an asterisk, 75 V for all other ions; (s)
strong signal, (w) weak signal....................... 170

4.3. Bond dissociation energies for alkali-metal and hydronium ion
crown ether complexes including present work, previous work,
and theoretically calculated values.................... 209













vi













LIST OF FIGURES


Figure D3gQ
1.1. Standard ICR trapping cells. (a) Cubic cell with two trapping
plates perpendicular to z-axis, two excite plates perpendicular to x-axis, and two detect plates perpendicular to y-axis. The ion is trapped inside the cell with cyclotron motion in the xyplane, moving along the z-axis. (b) Cylindrical cell .......... 5

1.2. Standard pulse sequence for ICR experiment ............... 8

1.3. Ion excitation. (1) Impulse excitation: approximate delta function
vs time, Fourier transform to the frequency domain signal (2) Chirp excitation: an rf frequency sweep vs time, Fourier transform to the frequency domain signal (3) SWIFT: (a) The desired frequency spectrum, (b) Inverse FT to give a timedomain signal, (c) Fourier transform to the frequency domain
sig na l . . . . . . . . . . . . . . . . . . . . . . 10

1.4. (1): (a) Ions' cyclotron motion prior to excitation, (b) Ions are
excited to a larger orbit, (c) Ions achieve coherence after excitation. (2) Phase coherency: (a) Two ions of same m/z 1800 out of phase, (b) The in-phase (initial motion parallel to electric field, E) ion speeds up and out-of-phase ion slows
down until they eventually achieve coherent motion ........ 13

1.5. Ions induce image current alternately on upper and lower
p late s .. . . .. . . .. . . . . . . . . . . . . . . . 15

1.6. Heterodyne mode: The signal mixes with a reference to give a
sum (higher frequency) and difference (lower frequency) signal. The higher frequency (low mass) signal is then filtered.
Fast FT gives the low frequency signal which is the
difference ......................................... 19

1.7. The structures of CH30HX-, X = F, Cl, Br, and I ............. 28


vii








1.8. The structures of polyether macrocycles ................... 30

1.9. The structures of metal/hydronium ion-crown ether complexes... 33

1.10. Photodetachment process: (a) A- + hv A + e, (b) A + hv B
+ C, (c)A + hv- B + C + e. ........................... 40

2.1. Simplified block diagram of FTICR mass spectrometer ........ 55

2.2. 2 tesla ICR instrument (a) 2 T magnet, (b) Inlet source, (c) Three
window flange, (d) One window flange, (e) ICR cell, (f)
Diffusion pump, (g) Ion gauge .......................... 58

2.3. Laser set-up for photodetachment experiment (1) Apple II
microcomputer controls "charge" and "fire" commands of YAG laser, (2) Console sends external TTL pulse to trigger the laser, (3) Power control from console to the cell; (a) Quartz prism, (b) Laser quartz windows, (c) ICR cell, (d) 2 T magnet,
(e) Apple II computer, (f) Console, (g) Photodiode ........... 60

2.4. Pulse sequence for photodetachment experiments ........... 63

2.5. C ylindrical cell ........................................ 65

2.6. 7 T ICR instrument (a) 7 T magnet, (b) InfinityTM cell, (c) 400 Us
cryopumps, (d) 800 Us cryopump, (e) Mechanical pumps, (f) Capillary coated with platinum at entrance and exit ends, (g)
Needle, (h) Syringe pump ............................ 70

2.7. Ion form ation in ESI .. ................................. 73

2.8. ES I design ......................................... 76

2.9. Electrostatic lenses in transfer region ...................... 83

2.10. InfinityTM cell with three "side-kick" plates: (a) PVXl, (b) PVX3,
(c) PVX4 plates, (d) the macor plate which is attached to trapping plates to hold the "side-kick" plates and trapping plates, (e) the rf-shimmed trapping plate, (f) the Infinity TM cell
with four brass plates for excition and detection ............ 85

2.11. Pulse sequence for CAD experiments .................... 87


viii








3.1. Mass spectrum of F-.................................. 94

3.2. Mass spectrum of COCI .............................. 96

3.3. Mass spectrum of (CH3OH)F, m/z= 51 ................... 99

3.4. Mass spectrum of (CH3OH)CI, m/z = 67, 69................ 101

3.5. Mass spectrum of OH, m/z = 17........................ 104

3.6. Mass spectrum of (H20)Br, m/z = 97, 99 .................. 106

3.7. Mass spectrum of (CH3OH)Br, m/z = 111, 113. ............ 108

3.8. Photodetachment spectrum of (C2HsOH)Br: U experimental
data; curve best fit to the threshold region of photodetachment effect vs. (E-ET)"- the arrow points to the
onset threshold.... ................................ 113

3.9. Photodetachment spectrum of (CH3CN)Br: U experimental
data; curve best fit to the threshold region of photodetachment effect vs. (E-ET)'12- the arrow points to the
onset threshold.... ................................ 115

3.10. Photodetachment spectrum of (CH3OH)I: U experimental
data; curve best fit to the threshold region of photodetachment effect vs. (E-ET)1/2- the arrow points to the
onset threshold.... ................................ 117

3.11. Photodetachment spectrum of (CH3OH)F: N experimental
data; curve best fit to the threshold region of photodetachment effect vs. (E-ET)112 the arrow points to the
onset threshold.... ................................ 119

3.12. Photodetachment spectra: N (CH3CN)Br; (CH3OH)Br; V
(C2HsOH)Br; A (i-C3H70H)Br; X (n-C3H7OH)Br......... 122

3.13. Photodetachment spectra of (CH3OH)F obtained using different
thermalization times (t): A t=2.34 sec.; X t=3.20 sec.; U
t=3.77 sec.; X t=4.31 sec.; t=6.05 sec.. ............. 124

3.14. Photodetachment spectra of (CH3OH)CI obtained using
different thermalization times (t): 0 t=2.05 sec.; A

ix








t=3.20 sec.; U t=4.34 sec .. .......................... 126

3.15. Photodetachment spectra of (CH3OH)Br obtained using
different thermalization times (t): U t=2.05 sec.; A t=3.77
sec.; 0 t=4.34 sec ................................ 128

3.16. Potential energy curves for alcohol-solvated halide anions
and corresponding neutral species. The lower curve represents the interaction of the solvent and halide from infinite separation to a stable bound state. a) The upper solid curve represents the hydrogen abstraction reaction of ROH + X to form RO + HX. (X = F and Cl). The arrows indicate the thermochemical cycle for the photodetachment process, including the threshold (Ethreshold), electron affinity (EA) of X, binding energy (BE) of (ROH)X-, and the energy difference (AE) between the reported and actual BE. b) Same as a) with dotted curve to indicate hydrogen abstraction reaction for X=
B r a nd I ...................... .... ......... ... ..... 136

4.1. Potential energy curve for ion crown ether complexes and
corresponding neutral species. The D298 represents the bond dissociation energy at room temperature, the Do represents the
bond dissociation energy at zero Kelvin temperature ........ 149

4.2. Appearance curve of (1 8-C-6)Rb The solid circles are the
experimental data (mean of five measurements with error bars corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program ................ 158

4.3. Appearance curves of (1 8-C-6)Rb at different pressures (50 ms
CAD delay time); V 3.0 x 10.8 torr, 0 5.5 x 108 torr, 0 7.3
x 10.8 torr, A 9.2 x 10-B torr ......................... 166

4.4. Appearance curves of (1 8-C-6)Rb at different CAD delay times
(P = 3.3 x 10-8 torr); 0 0.05 s, A 0.08 s, U 0.1 s, V 0.15
s . . . . . . . . . . . . . . . . . . . . . . . . 16 8
4.5. Mass spectrum of (12-C-4)Rb at AV=20 volts. Peaks
corresponding to (12-C-4)Rb* and (12-C-4)2Rb ions were
observed ........................................ 173

4.6. Mass spectrum of (15-C-5)H30 at AV=20 volts. Peaks
corresponding to (15-C-5)H3O (15-C-5)2H+, and (15-Cx








5)2H30' ions were observed ......................... 175

4.7. Mass spectrum of (12-C-4)H3O0 at AV=20 volts. Peaks
corresponding to (12-C-4)H (12-C-4)2H*, and (12-C-4)2H30 ions were observed. The peak intensity of (12-C-4)2H' is higher
than that of (12-C-4)H .............................. 177

4.8. Mass spectrum of (12-C-4)H30' at AV=70 volts. Peaks
corresponding to (12-C-4)H' and (12-C-4)2H ions were observed. The peak intensity of (12-C-4)2H is lower than that
of(12-C-4)H ..................................... 179

4.9. Mass spectrum of (12-C-4)Rb at AV=70 volts. Peak
corresponding to (12-C-4)Rb ion was isolated before
excitation ........................................ 182

4.10. Mass spectrum of (12-C-4)Rb after CAD at AV=70 volts. The
two isotopes of Rb were the only successor ions observed . 184

4.11. Mass spectrum of (18-C-6)H3O0 at AV=20 volts. Peak
corresponding to (18-C-6)H3O0 ion was isolated before
excitation ........................................ 186

4.12. Mass spectrum of (18-C-6)H3O' after CAD at AV=20 volts. Peak
corresponding to (18-C-6)H* ion was the only successor ion
observed ........................................ 188

4.13. Mass spectrum of (15-C-5)H3O0 at AV=20 volts. Peak
corresponding to (15-C-5)H30O ion was isolated before
excitation ........................................ 190

4.14. Mass spectrum of (15-C-5)H3O' after CAD at low activation
energy and AV=20 volts. Peak corresponding to (15-C-5)H' ion
was the only successor ion observed.................... 192

4.15. Mass spectrum of (15-C-5)H3O0 after CAD at high activation
energy and AV=20 volts. Peaks corresponding to (15-C-5)H*,
(C2H40)nH' (n = 2 and 3) were the succesor ions observed... 194
4.16. Absolute intensity plots versus E, of (15-C-5)H3O, (15-C-5)H ,
and (C2H40)nH (n = 2 and 3) ........................ 196

4.17. Mass spectrum of (12-C-4)H at AV=70 volts. Peak

xi







corresponding to (12-C-4)H+ ion was isolated before
excitation....................................... 199

4.18. Mass spectrum of (12-C-4)H+ after CAD at AV=70 volts. Peaks
corresponding to (C2H4O),,H (n = 2 and 3) were ions
observed....................................... 201

4.19. Appearance curve of (15-C-5)Rb+. The solid circles are the
experimental data (mean of five measurements with error bars corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program ...............203

4.20. Appearance curve of (12-C-4)K+. The solid circles are the
experimental data (mean of five measurements with error bars corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program ...............205


4.21. Appearance curve of (12-C-4)Cs+. The solid circles are the
experimental data (mean of five measurements with error bars corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program................ 207























xii













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

DETERMINATION OF BOND DISSOCIATION ENERGIES USING FOURIER
TRANSFORM ION CYCLOTRON RESONANCE (FT- ICR) MASS SPECTROMETRY

By

Yarjing Yang

1996

Chairperson: John R. Eyler
Major Department: Chemistry

The objective of this work was to use Fourier transform ion cyclotron resonance mass spectrometry (FTICR-MS) to determine bond dissociation energies (BDE's) of important species such as solvated anions and host-guest complexes using two different techniques: photodetachment and on-resonance collisionally activated dissociation (CAD).

Photodetachment has been widely used in measuring the electron affinities of various anions. Since the electron affinities of the halides are well known, this part of the work has utilized photodetachment to measure the halides' binding energies with alcohol solvents in the gas phase. An Nd3*:YAG-pumped dye laser with frequency doubler was used to obtain the uv-visible irradiation needed for this work. The binding energies of methanol, ethanol, iso- and normal-propanol, and xiii








acetonitrile with fluoride, chloride, bromide and iodide ions have been measured. The effect of size and polarizibility of the solvents, and comparisons of the solvation energies determined in this work to some reported in earlier thermal-equilibrium studies are discussed.

Collisionally activated dissociation (CAD) processes in mass spectrometry can provide quantitative bond dissociation energy (BDE) estimates if the interaction energy is precisely controlled during collisions. This work has concentrated on improving the accuracy of measuring the thresholds and providing more precise quantitative measurements of host-guest complex interactions in the gas phase by accounting for the effects of multiple ion-molecule collisions, internal energy of precursor ions, thermal spread of energies of the collision gas and the dissociation lifetime of the intermediates. The use of an electrospray ionization (ESI) source for soft ionization that produces intact, multiple charged and solvated gas-phase ions directly from various compounds in solution has been critical for this work. We have measured the binding energies of hydronium and alkali metal ions to the crown ethers (12-crown-4,15-crown-5, and 18-crown-6). The effects of metal size, cavity size and polarizibility of different crown ethers are discussed.











xiv













CHAPTER 1

INTRODUCTION


Fourier Transform Ion Cyclotron Resonance Mass Spectrometry (FTICR-MS)




In 1950, the first practical ion cyclotron mass spectrometer (Omegatron) was used for analytical purposes by Sommer, Thomas and Hippie. 1,2 During the 1960's, a number of research groups exploited the ion cyclotron resonance (ICR) technique, normally defined as "Pre-FTICR" with conventional features such as solenoidal magnets, electromagnets, drift cells, ion trapping, double resonance detection, analog FT and simultaneous excitation and detection. As practiced in the 1960's the technique had slow scan speed, low resolution and a limited mass range.3 In 1974, Comisarow and Marshall"' first applied the FT technique to ICR-MS with frequency chirp excitation, mass analysis in the same region and simultaneous excitation and detection. This has evolved into the versatile and powerful technique often referred to as Fourier transform mass spectrometry (FTMS).

In 1980, superconducting magnets began to be used, resulting in improved performance when compared to electromagnets: ultra-high mass resolution, wide mass range, high mass measurement accuracy, simultaneous detection of all ions,


1








2

ion storage, and many probes of ion structure.12 During the 1980's, a wide variety of applications"-" were reported, including MS/MS experiments, laser desorption, fast atom bombardment (FAB) and high pressure external sources with quadrupole ion guides, gas chromatography mass spectrometry (GC/MS), external sources utilizing electrostatic focussing which lower ultimate pressure at the analyzer, stored waveform inverse Fourier transform (SWIFT) excitation and electrospray ionization (ESI) sources. Both ES120 and matrix assisted laser desorption ionization (MALDI)21 sources began to be used in the early 1990s to increase the range of masses which can be studied with FTICR mass spectrometers. In 1992, Amster and Marshall22 demonstrated quadrupolar axialization to improve resolution and increase sensitivity. In 1993, Smith's group23 developed Time Resolved Ion Correlation (TRIC) to detect highly-charged individual ions with masses higher than 5,000,000 daltons.

Basic Concepts

Ion cyclotron motion

The ions in a magnetic field undergo circular orbital motion, at a cyclotron frequency given by, 2427


w = qBIm = vlr (1.1)


where q is the ion charge, B is the magnetic field, m is the mass of the ion, v is the ion velocity and r is the orbit of the cyclotron motion. In a uniform magnetic field (no electric field), the mass-to-charge ratio, m/q is linearly related to the observed








3

cyclotron frequency, w. The Lorentz force, F, is given by,




9 = q ( v X A')) (1.2)


For BO along the z axis, this force is constrained to the xy-plane (there is no force along the z-axis), so ions can move freely along that axis unless an electric field is applied, 21,21



d 17
M dt2 = q(f V x6o) (1.3)



where E is the electric field.

However, in order to detect the ions, they must be confined within a finite volume bounded by conductive electrodes (in a trapping cell). Figure 1.1 is a schematic drawing of two different ICR trapping cells which are commonly used: cubic and cylindrical. The cell is centered in a homogeneous magnetic field generated by a superconducting magnet and contained in a high vacuum chamber. There are two trapping plates which are perpendicular to the magnetic field (Z-axis), two excite plates (perpendicular to the x-axis) and two detection plates (perpendicular to the y-axis).

The ions may be formed in the ICR cell (using an internal ionization source) or outside the magnetic field (with an external ionization source) by a number of methods, including electron impact (EI), chemical ionization (CI), laser desorption















U) c Q 0




X

0

C.)
CD

N (1)

c
cc

CL
c CL
a) cc
2L 4
(1) V CL c U) 0
(1)
(D .C

C U)
CL (D
CL 0

0
0
CLG Lc :3 G
>1
c


0 a)
0.
U) cc
(D m
a a)
=


-(D 0 W cu c 0

CL
-0 >
c 0
cu E O.C6 6
R c m M

o
C X
4- a)
75
c
c c a) 0
=3 0. CD L- 0 F a) L a E

































. ...................
................






























................... .








6

(LID), fast atom bombardment (FAB), electrospray ionization (ESI), and matrix assisted laser desorption ionization (MALDI). The basic FTICR-MS experiment is conducted using a series of computer-controlled pulses, as depicted in Figure 1.2. Initially a quench pulse is applied to the trapping plate(s) to remove all the ions from the cell prior to starting the next pulse sequence. After ion formation, ions are trapped in the cell by constraining their z-axis motion (along the magnetic field) by applying low voltages (0.5 to 5 V typically) to the trapping plates. Either positive or negative ions can be trapped in the cell by changing the polarity of the voltage applied to the cell plates.

Ion excitation

Three types of ion excitation are generally used for ion detection, as shown in Figure 1.3. (1) In impulse excitation," an approximation to an ideal delta function pulse (infinite amplitude, zero width) is applied. This results in a flat excitation spectrum and should excite all ions equally. The angular cyclotron frequency is of the order of wcmax = 1/At, where At is the pulse width. (2) With chirp excitation,10 a radio frequency (r.f.) pulse is applied whose frequency sweeps rapidly over a range from the lowest to highest frequency desired. The disadvantage of chirp excitation is the nonuniform excitation of the ions, which becomes pronounced for ions near the edge of the swept frequency range. The final cyclotron radius is limited to less than the full dimensions of the cell since ions absorb different amounts of power during the chirp. This introduces peak height distortion and the frequency chirp excitation also contributes to z-axis ejection. (3) For stored waveform inverse





















































c
0
E

CL
x







c






:3
CL
-2





04









CL

0
CL

E .2 .0
0 ig ts
U.
. x
0










IE
E















0

CD

cr
cc E
0 4

0
C.) =3 E c o L- UL
0 =3 I.C5




d)
:3 so
0 c
LL
6 0 E E 0 0

CD
> E

0 cu
:3 4)
0 U

-CU (D 0 (D E 'a
(n LL
4- >
m
E 1D
X
2
CL M a >cu

0 :3
P rr
(D
(D
x
4) c (D M

75 E (D CL 0
Cr
cu



c
0
.I.,
A-0
X cq .0

0
M (D
cn U- =) c ; Cr
(D
%6.
E L
CD o








3 10
L
0 3 :>
0

T

























co




cm








11

Fourier transform (SWIFT) excitation,","" the excitation waveform is generated by taking the desired frequency spectrum (a) (power as a function of frequency) and performing an inverse FT to give a time-domain signal (b). This waveform is then generated by a digital frequency synthesizer. The excitation spectrum (c), which indicates the actual amount of power applied at each frequency, is produced by Fourier transformation of (b). This excitation spectrum is much more uniform over the desired frequency band than either impulse or chirp excitation.

Any of these three excitation modes will cause the ions to absorb energy and be accelerated into larger orbits when the r.f. frequency becomes equal to the ion cyclotron frequency. It is also necessary to excite a packet of ions of a given massto-charge ratio not only to a larger ICR orbital radius, but to coherent motion in order to detect them. Figure 1.4 shows how the ions achieve coherence prior to detection.

Ion detection

The orbiting packet of ions induces a small alternating image current on the receive plates (perpendicular to the y-axis). The image current is given by,"

N q v (1.4)
d



where N is number of ions, v is the ion velocity, q is the charge of the ion, and d is the electrode spacing, i.e. the dimension of the cell as depicted in Figure 1.5. The ion induces image current alternately on upper and lower plates. To obtain a


















o

cu~

CL





2)

%I.

-0


00



N0 C cc E o


E))U



..O 0
C: c4! 00
-00 x C
a)
0
0 =3C: CL

0
Eo0
EU) 0~

0





0


ooU


00



-C C uO o Eo








13 CM








co








































vi
W
4
m
CL


0


m


CL =3
c
0 7a"
CO

CD















cn
c
0

U?




F







4-0 15
C:

=3


CZ

-.E










(D CD cz
E






CD

C.)
(D W cr CL








16

theoretical estimate of the magnitude of the signal produced on the plates by an ion, the Reciprocity Theorem is applied .35 The cyclotron radius, Rc, is given by, E t
C = (1.5)
B



where E is the excitation electric field, t is the duration of the resonant excitation pulse, and B is the magnetic field. The Reciprocity Theorem states that the electrostatic potential created at the position of an ion by 1 volt on the detection plate is directly proportional to the image current induced on the plate by the ion. Thus the predicted FTICR signal strengths can be calculated by a numerical solution of Laplace's equation. The amplitude of the signal per ion only depends on the ratio between the radius of cyclotron orbit and the distance between the receiver plates.

All ions in the cell are excited and detected simultaneously as a time-domain signal. The signal is converted into voltage, amplified, digitized and stored in computer. The time-domain spectrum is a composed of superimposed sine waves: in order to recover the frequency domain spectrum, the complex time-domain spectrum is subjected to a fast Fourier transform (FFT) algorithm. Signal acquisition

For one thousand ions, the r.f. signal at a preamplifier input is on the order of a few hundred microvolts. A low noise electronic design for amplification is required. There are two types of data acquisition used in our ICR instruments. (a)








17

The first produces broadband spectra (all masses). In this mode the signal from a preamplifier is sampled directly by the analog to digital (ND) converter at a rate which must be faster than twice the highest frequency for the lowest mass to be observed (Nyquist criterion). Spectra are signal-averaged by summing the timedomain transients from each repetition in the computer's memory before performing the FT; (b) The second type of data acquisition (heterodyne mode) leads to narrowband spectra.43

As illustrated in Figure 1.6, the heterodyne technique reduces the digitization rate requirements for ion signals and is used to obtain higher mass resolution over a narrow mass range. Given a time-domain spectrum consisting of N points, where N is between 16K and 128K, it is common to zero-fill at least once (add N zeros to the end of the data set), apodize, and obtain the magnitude-mode FT, where the magnitude-mode is defined as [(real)2 + (imaginary)1"2, with a FFT algorithm. Zerofilling has been shown to increase the S/N ratio and increase the displayed resolution of the spectrum.

Since frequency is a physical quantity which can be precisely measured, the exact mass of an ion can be determined very accurately in the FT-ICR technique, since it is directly related to the measured frequency. Typically, low parts-per-million accuracy can be achieved with an internal mass calibration. A high degree of mass accuracy can be maintained due to the stable magnetic field. The main features of FTICR include ultra-high mass resolution, accurate and stable mass calibration, multiple stages of collisionally activated dissociation (MS/MS), extremely long ion















(D
0

C.)
C
T
CD






0
:3 (1) cr = 94



7 U

E cn Um -6
a)


0
C:


cn


ca
cn

cn E
w .x
E 0
m %
c


0 rr
4
CD 0 P E 2 (D 0



CD c 0






-r -C








19


cz C;) c


CO cd co a) CL
E =3 CO





















cz C: 0)








20

storage times (> minute), the ability to study ion/molecule reactions, compatibility with pulsed or continuous ionization techniques, and ease of carrying out photodissociation and photoactivation studies. The disadvantages of FTICR are its limited dynamic range (space charge limits the number of ions which can be trapped to ca. 106), the fact that data acquisition can take milliseconds to seconds, so transient or metastable ions can not be studied directly, and the high cost of the instrument ($300K to $500K).



Applications

Purpose of this Work

Different experimental methods are used in the studies of ion-neutral complexes. The case of interactions which involve charged species requires specific techniques for three main reasons: (1) charged species must be selectively separated from ions of the opposite charge, (ii) they must be stored long enough to reach a state of thermodynamic equilibrium with their surroundings while avoiding collisions with the walls of the reaction chamber and (iii) they must be monitored by highly sensitive methods because of their very low concentrations.

Several reviews have been briefly summarized by Keesee and Castleman et aL 36,37 in a major review of thermodynamic data for ion-molecule reactions in the gas phase. A number of experimental devices have been applied to the study of ion-neutral complexes. Among the devices used to generate and study the gas phase ion complexes similar to those in these experiments are molecular beams,








21
drift tubes, and ion traps.384 Each device offers its own advantages for different types of ions. For the study of ion-neutral reactions, ion cyclotron resonance (ICR) mass spectrometry has been the most extensively applied method. The advantages offered by ICR mass spectrometry include the ability to generate a large variety of molecular ions and high sensitivity. The technique of ICR-MS allows one to generate gas phase ions and trap them for periods of millisecond to several seconds, sometimes even for hours.

The goal of this work was to apply different FTICR-MS techniques, including photodetachment and collisionally activated dissociation (CAD), to obtain bond dissociation energies for solvated anions and host-guest complexes. A secondary goal was obtaining quantitative measurements of bond dissociation energies (BDE's) by improving the threshold determination to account for a variety of systematic effects that would otherwise distort the thermochemistry obtained. Several methods have been used to measure the bond dissociation energies for various ion-neutral complexes, including equilibrium measurements in both high pressure413 and ICR45 mass spectrometers, photodissociation measurements in a time-of-flight (TOF) mass spectrometer,46-50 photodetachment measurements in photoelectron spectroscopy,51-13 collisionally activated dissociation (CAD) in both quadrupole or octopole-guided ion beam54-57 and ICR8 mass spectrometers. The measurement of BDE's by FTICR mass spectrometry has not been widely applied and there is still some uncertainty as to the quality of the measurements. In this section, some recent experimental studies of BDE measurements using different types of techniques are briefly discussed.








22
A few years ago the development of high pressure mass spectrometry reached a stage where it could be applied successfully to chemical problems. The technique and its applications have been summarized by Kebarle.59 An impressive number of clustering equilibria of the type in eq. 1.6 have been studied,



X-Sn X-Sn. + S, In Kn.n.n= AS/R AH/(RT) (1.6)



where X represents a central ion, S a solvent molecule, and XSn and X-S, are two cluster ions of different stoichiometry. Based on the accessibility of suitable ion sources, single ion solvation can be studied by mass spectrometry provided equilibrium is established in such a system.

Photodissociation thresholds can be interpreted as reflecting at least an upper limit to the thermochemical dissociation threshold. Moseley et al. measured the vibrationally resolved photodissociation spectrum of C03, which was analyzed to give a reliable limit of 1.85 eV for D(CO2-O).

Energy-resolved collision-induced dissociation of metal carbonyl anions M(CO)n (M=V, Cr, Mn, and Co) were investigated to determine sequential metalcarbonyl bond energies by Squire's group.57b They used the model described in the Methodology Section of this Chapter to obtain fairly accurate BDE's of 30.8 3.5, 40.6 3.5, 40.6 3.9, and 39.7 3.7 kcal/mol for (CO)sV-CO, (CO)4Cr-CO,

(CO)4Mn-CO, and (CO)3Co-CO, respectively.

Electron photodetachment spectroscopy is one of the most useful methods








23

available for determining the spectroscopic properties of negative ions. This technique may be used to measure the dependence of the cross section on photon energy and therefore to determine the photodetachment threshold. Bowen's group"1 has demonstrated the photodetachment of NH4 using photoelectron spectroscopy (PES). The PES of NH4 resembles that of free hydride, H-, except for being shifted to lower electron kinetic energy due to the stabilization effect of ammonia solvation. They confirmed that this ion cluster is best described as a Hion solvated by NH3, and were able to obtain the solvation energy.

Since the electron affinities of the halides are well known, the threshold for photodetachment of alcohol solvated halides is directly related to the halides' solvation energies in the gas phase. Photodetachment processes in the ICR mass spectrometer provide more accurate quantitative measurement compared to thermal equilibrium methods if the ions are well thermalized and the ion intensity is well controlled. A frequency-doubled dye laser pumped by a Nd3 :YAG laser was used to obtain the UV-near visible output range for this work. Such a laser can cover a very wide wavelength range with narrow bandwidths.

Collisionally activated dissociation (CAD) processes in mass spectrometry can provide quantitative bond dissociation energy (BDE) estimates if the interaction energy is precisely controlled during collisions. One goal of this work was to improve the accuracy of determining the threshold and provide more precise quantitative measurement of the BDE of host-guest complexes in gas phase. The use of an electrospray ionization (ESI) source for soft ionization that produces








24

intact, multiple charged and solvated gas-phase ions directly from various compounds in solution has been critical for this work. In addition, the ions formed using ESI carry little internal energy compared to other ionization techniques such as electron impact (EI), laser desorption (LD) and matrix-assisted laser desorption ionization (MALDI).

Solvated Anions

Living organisms and nonliving matter are affected by interactions with organic chemicals, which are highly dependent upon how molecular structures produce site-specific interactions, such as hydrogen bonds. In general, in the life sciences and in chemistry, the hydrogen-bond (HB) is perhaps the most important kind of specific molecular interaction between molecular species. Examples include solubility, bodily transport and docking to active sites. Hydrogen bonding was first recognized in the physical properties of liquids, and was thought to involve only fluorine, oxygen or nitrogen atoms bridged by an intervening hydrogen." Since then structural studies have indicated that atoms such as chlorine, bromine and sulfur can serve as hydrogen bond acceptors," .12 and that hydrogens bonded to carbon and sulfur can serve as hydrogen bond donors." Another significant type of site-specific noncovalent interaction has been identified as "halogen bonding"." Therefore, study of alcohol solvated halogen ions through hydrogen bonding has been an important issue.

Ion salvation has been studied extensively for many years by both theoretical








25

methods and experimental techniques. Currently, approaches have reached a high degree of accuracy and sophistication. Despite this enormous amount of information, however, no comprehensive and satisfactory theory of ion salvation can be given, due mainly to the lack of a detailed knowledge of the structure of nonsimple liquids. The problems occurring in model calculations on ion salvation are two-fold. Firstly, accurate energy surfaces for ion-molecule complexes are difficult to obtain. They provide information about the most stable geometries and energies of interaction. The static energy of a complex, unfortunately, is only one side of the picture. The shape of potential surfaces for clusters consisting of an ion and many molecules, in general, indicates high flexibility, since several flat energy minima separated by low barriers are usually found. Therefore, the dynamics of the complex is the second important problem for most macroscopic properties at room temperature.

An alternate and more rigorous treatment of ion salvation is based on the study of ion-molecule reactions in the gas phase. Relative enthalpies and free enthalpies of complex formation for solvated ions X-(H20)" with n = 1-6 have been determined .12, The concept of hydrogen bonding (HB) emerged from the study of self-associated liquids (notably water)63 and its importance in complex biological systems is paramount.64 The study of hydrogen-bond, HB, interactions in the gas phase is relevant because it provides direct information on the structure and stability of these complexes in the absence of the perturbations induced by the solvent.

41-45,19
Most of the solvated ions are easily generated and heavily studied.








26
These studies have led to an understanding of the stability of hydrogen bonding and structural variation in these compounds. Such work also provides a better understanding of the characteristic properties of bulk solution such as ions solvated with solvent clusters.129 Among solvated ions the methanol solvated anions are difficult to generate and understanding of their properties is still subject to argument. Also there is still disagreement about the binding energies (BE) of solvated anions using different measurements.4145 Therefore, alternate measurements of the BE using threshold determinations from photodetachment experiments can give more accurate results. Also, since the ICR ion trap can easily form solvated ions using either ion-molecule reactions or electrospray ionization (ESI), it becomes possible to measure BE using photodetachment techniques in the ICR instrument.

The BE measurements of alcohol solvated halides, whose structures are illustrated in Fig. 1.7, using the photodetachment technique with a FTICR mass spectrometer are described in Chapter 3. Host-guest complexes

Polyether macrocycles have been known for decades.6 7 The discovery that certain macrocycles can bind alkali metal and alkaline earth cations gave impetus to the field now known as host-guest chemistry.68 Today, a distinction is made between the classical ring oligoethers (crown ethers) and monocyclic coronands, oligocyclic spherical cryptands and the acyclic podands with respect to topological aspects.69 Multidentate monocyclic ligands with any type of donor atoms are called coronands (crown compounds), while the term crown ether should be reserved for















































m
L.:



LL: 11
x

x
r
0








28









0 < o<
C) CO



o< 0 o<
00 0
(3) 0
CY) CD


o< 0<',
co C"J
CN (DO
C

0 <
C)
C)


0
o<

CY) o<

< CY) C)



o< 0
o< 0 00

C) 0 C)
00 C14
(C)
o<' o< TCO
C

LL co


























































0


E




0
CL 4
0
cn


t5

cn



co
lr-: a)







30


z



c CO
co L4-0 0

0
C,







as
0
L
0




cl


r" a 'a C 0 (1p In c
0 co 0
IWA=.A "a
0
co 0i








31
cyclic oligoethers exclusively containing oxygen as donor atom (see Figure 1.8).

The ability to form host-guest type complexes with metal ions is one of the most important properties of crown ethers and related compounds. These macrocyclic polyethers show a remarkable range of specificity for a wide variety of cations that depends, in part, on the size of the ether, the type of donor atoms (e.g. oxygen, nitrogen, sulfur), and the polarity of the solvent.2 Since their discovery by Pedersen 11,4 in 1967, crown ethers have generated a tremendous amount of interest because of their ability to selectively bind guests such as alkali metal ions. This crown ether/ion specificity can also serve as the basis for many potential practical applications. For example, 90Sr and '37Cs are the two major generators of radiation in nuclear waste, and thus complicate disposal efforts. Horowitz, Dietz, and Fisher"5 have used di-tert-butylcyclohexano-1 8-crown-6 for recovering 11Sr2' from acidic solutions. A more thorough understanding of cation/crown ether chemistry can provide the basis for rational design of new ligands useful in separation of these and other radionuclides from complex waste streams and hazardous waste storage facilities.

The 3-dimensional structure of a crown ether complex is typified by the symmetrical coplanar array of the oxygen atoms. These atoms contact the cation located in the center at an arithmetically calculated interatomic distance as shown in Figure 1.9. Cations which fit less well spatially are tolerated within certain limits, e.g. by deformation of the ligand skeleton or by movement of the cation out of the ring plane. If the cation diameter is much larger, sandwich complexes with 2:1-











































x ci)

E
0




0





0
_0


E

0



V)


Cu


4

.)
0)







W) 33 W)
+
CN





r 0
4-1




















co








34
stoichiometry may form (Fig. 1-9). The degree of stability of crown ether complexes is represented by the stability constant, K, the equilibrium constant of the complexation reaction, which is based on a specific solvent.7 Earlier investigations of cation-binding strengths are based on stability constants for the polyethers and cations in solution.7 However, it has shown that solvents play an important role in binding effects. In solution, the order of complexation selectivity, obtained from log K values, is generally K+ > Rb + > Cs + > Na + > Li + in solvents such as water or methanol.7 In contrast, Na + binds stronger than K + in the gas phase. 78This contradiction is due to the fact that the solvation energy of Na+ is much greater than that of K+. Cation-binding strength has been of great interest to the chemical community. Enhancing cation-binding selectivity is another matter, and the selectivity is also influenced by the ring size. The sizes of cations and crown ethers are listed in Table 1.

Studies of host-guest complexes in the gas phase have permitted better understanding of the fundamental details of molecular recognition in the absence of a solvent .71-11 Gas-phase ion-molecule reactions have also been carried out in order to study the mechanism of ion formation under fast atom bombardment conditions"2 or to determine the relative stabilities of crown ether complexes with alkali metal cations.85

Earlier investigations of crown ether complexes with alkali metal cations have shown the intrinsic nature of binding interactions and the size selectivity of complexation in the gas phase.98 For predicting future applications,
















































(D


-


0


N

U)



L)



0








4-
0
(D
N
C,)





cu














36















r- it co co
(D CN 00 Lo
T-: 1: ci



CN CY) co
V) fl- Lo
T- ci C5


co cj co
C5 ci

w C%4 co CN U)
0 rl- Lo ce)
C; 6 ci
2

0
2

CD 46

Lo C%4 N Z
C cli
04 FI- CO Le V-: T-:


Lzi



co
T- U.) CD

3: 3: 3:
0 2 LO- LOE
N 04 LO 00 +








37
measurements of BDE's would provide better understanding for making use of the ion discrimination properties of crown compounds based on ion-selective membrane electrodes,86 for the study of physiological ion-transport processes in biological membranes, for the investigation of receptor and enzyme interactions as well as for the salt balance and for metabolic processes of the living organism.87

Investigations using CAD to study crown ether complexes with alkali metal cations have been previously reported.8889 However, quantitative information about the BDE's of crown ether complexes has only been reported by the Eyler group.9 In the previous study, approximate binding energies of metal/crown ethers formed using the laser desorption technique were reported. In these studies, we have carefully measured quantitative BDE's of crown ethers with either hydronium or metal ions attached, formed from an external electrospray ionization (ESI) source, using a different approach to determine the dissociation threshold. The details of the measurements are described in Chapter 4.



Methodology of Determination of Bond Dissociation Energies



Photodetachment

Electron photodetachment from gas phase molecular anions has proved to be a remarkable tool in the study of both anions and their neutral photoproducts. A variety of thermochemical and spectroscopic constants can be measured for systems that would be difficult to study by other methods. Among the techniques








38
that have been developed for the study of electron photodetachment are ion beams, trapped ion cells, and drift tubes.'40 The advent of tunable laser systems has increased the precision of measurements made in photodetachment experiments, allowing for the detection of vibrational and even rotational transitions in some systems."2 In addition, the wide wavelength range and narrow bandwidth of the tunable output provide extensive new study areas.

Upon irradiation with light, a molecular anion may either detach an electron (1.7), dissociate (1.8), or undergo photodetachment followed by dissociation (1.9):



A- + hv A + e- (1.7)



A-+ hv B- +C (1.8)



A- +hv B +C +e- (1.9)



Processes (1.7), (1.8), and (1.9) are illustrated in Figure 1.10 (a), (b), and (c), respectively. In photodetachment spectroscopy, these processes are monitored as function of the wavelength of the incident light. This can be accomplished by detecting the decrease in the intensity of precursor ion, A- (processes 1.7, 1.8, and 1.9) or an increase of successor B3- for process 1.8 or the detached electrons for all three processes. The related technique of negative ion photoelectron spectroscopy analyzes the energy of detached electrons produced by the interaction of a fixed






















(D





















































ci


a)


U-










4 On .C








41

frequency light and an ion beam. Photodetachment spectroscopy has been used to determine a number of thermochemical and spectroscopic parameters for a variety of atoms and molecules, including the electron affinity of the neutral photoproduct, the energies and transition moments of electronically excited states of anions, spin-orbit splitting in radicals, and vibrational and rotational spacings in both neutrals and anions. Most anions of organic molecules exhibit only photodetachment; a number of inorganic anions can also photodissociate upon irradiation with visible light. Photodetachment experiments are useful in probing excited states of anions, as well as providing information important in determining binding strengths.

The use of electron affinities derived from either photodetachment experiments or literature data to determine bond dissociation energies is applied in this work. By measuring the depletion of the parent ion as a function of photon energy, the photodetachment threshold, related directly to the binding strength of solvated ions, can be determined.

The raw data in a photodetachment experiment are the measurements of the precursor ion signal with and without laser irradiation. At long wavelengths, the photon energy is not high enough to detach an electron from the precursor anion so the intensity of the precursor ion remains unchanged, i.e. the fraction of photodetachment is zero. When the photon energy is high enough, the intensity of the precursor ion decreases and the fraction of photodetachment is nonzero, increasing linearly near threshold. In order to convert these data into a








42

photodetachment spectrum, the observed fractional signal decrease must be related to the photodetachment cross section. The probability of photodetaching an electron from a negative ion is described by,9o


N
P(A) = 1- 1- exp(-kp x t) (1.10)
No



Where N is the ion signal with irradiation, No is the ion signal without irradiation, t is the duration of the laser beam pulse, and kP is the rate constant for photodetachment given by

kp(A) = ff-o(A)-p(A)dA (1.11)


f is a unitless factor related to geometric overlap between the ion cloud and the photon beam, o(A) is the cross section for photodetachment in cm2, and p is the photon flux (a function of wavelength) in photons cm-2 s1-'. Assuming o(A) is constant over the narrow laser bandwidth and t is kept constant, substituting eq 1.11 into eq 1.10, rearranging, and taking logarithms yields an expression for the cross section in the terms of the quantities measured during a photodetachment experiment:91


N = No exp(- o(A) tff p(A) dA) (1.12)


In No
OA _N (1.13)
tff-p(A)dA








43



Photodetachment spectra consist of the wavelength dependence of this cross section. The geometric factor may be evaluated from the beam diameter, photon flux, and the size of the ion cloud, if absolute cross sections are needed.2 However, for a given photodetachment spectrum, all parameters in the denominator of eq 1.13 are constant, if the laser flux is kept constant. Thus, the relative photodetachment cross-section spectrum, which is sufficient for determination of the photodetachment threshold, may be obtained by plotting ln(NdN) versus the laser energy, E (A).

In a well known paper published in 1948, Eugene Wigner derived the form of the photodetachment cross section for energies near threshold.3 Valid for final states with two particles whose longest range interaction is the centrifugal barrier, his result gives the variation of the cross section with energy, but not its absolute magnitude. The law is given by,



o(E) (AE) 112 or o(E) k1 (1.14)



where a is the cross section, AE is the energy of the departing electron ( the energy above the minimum required for reaction to occur), and k = (2mAE)'/2 I and I are the linear and orbital angular momenta of the outgoing electron. The essence of the Wigner derivation is that, at threshold, the dynamics are solely governed by the








44

longest range potential interaction between the electron and the neutral atom.

In eq 1. 14, the interaction is the centrifugal potential, proportional to 1(1+ 1)/r2. In contrast to photoionization, where the long-range potential falls off as 1/r due to the Columbic force between electron and cation, the strongest long-range electronic interaction in atomic photodetachment is a centrifugal force contributed by the 1(1+ 1)/r2 term contained in the Schrodinger radial wave equation. Departures from the law can occur above the threshold, where shorter-range terms in the potential become important or when forces with the same range as the centrifugal force are active. Examples of such forces are electron-q uad rupole and electron-dipole interactions, respectively. For atoms, the usual dipole selection rule Al = 1 applies, and I in eq 1.14 can be related to the angular momentum of the atomic orbital from which the electron was detached. The extension of the Vvigner result to include polyatomic anions was accomplished by Reed et al.9 In general, the behavior of the cross section near threshold for a polyatomic anions is given by,



a(E) = Y-A1. Ek'*+%[1+O(Ek)], (1.15)



where I* is the smallest integer value I for which A,. is not identically zero for symmetry reasons. To perform the algorithm, one finds the irreducible representation of the anion's HOMO for the point group of the anion. One then assigns a value I'based on that irreducible representation. For example, I' = 0 if the irreducible representation is totally symmetric; I' = 1 if it transforms as x, y, or z; and








45

I' = 2 if it is labeled by xy, xz, yz, x2-y2, or z2. The allowed I* values for photodetachment are then these I' 1. Since the A" representation transforms as Z in the C, point group for alcohol solvated halides, the HOMO has I = 1 and the selection rule Al = 1 gives r = 0. Therefore, from eq 1.15 we obtain,



o(E) Eklr2 or In (NoIN) = a(E-ET)", (1.16)



The threshold, ET, can be obtained using an iterative procedure in which the parameters, a (a scaling factor) and ET are optimized, to obtain the best fit to the photodetachment spectra.

CAD

Energy behavior of reaction cross section

The maximum reaction cross section for an exothermic ion-molecule reaction has an energy dependence that is determined by the long-range interaction potential. For polarizable molecules without dipole moments, this is given by the Langevin-Gioumousis-Stevenson (LGS) expression,9"



OLGs(E) = ne (2a/E)', (1.17)



where e is the electron charge, a is the polarizability of the neutral reactant, and E is the relative kinetic energy of the reactants. Higher order terms in the long range interaction potential can also be included.96








46

For endothermic reactions, microscopic reversibility can be applied to eq 1.17 to yield,97



o(E) = oo(E-ET) /E, (1.18)



where ET is the reaction endothermicity. The scaling parameter is defined as oo = ne (2a)Y2 (psg, / p pg p), where p is the reduced mass and g is the electronic degeneracy of the precursor (p) or successor (s). Although it would be nice if eq 1.18 accurately predicted the energy dependence of all endothermic ion-molecule reactions, this is not observed experimentally.5s57s.a Indeed, reactions that conform to eq 1.18 are exception rather than rule, as described later.

Alternate procedures for determining the energy dependence of reactions include trajectory calculations, statistical theories, and empirical theories. Trajectory calculations are not generally useful because they require a potential energy surface, which is rarely available. Statistical theories such as phase space theory (PST) in which energy and angular momentum are explicitly observed,98 are much more useful.99-02 Empirical models are the most general means of describing the energy dependence of reactions, and in this work, we use,



o(E) = o0(E-ET)n/Em, (1.19)



where the scaling parameter o0 (with units of area.energym-n) need not be given by








47
the expression above. This model includes the model described by eq 1. 18 as well as most other theoretical predictions for endothermic cross sections.103 Unfortunately, even for atom-diatom reactions, there are many theoretical predicted values of n and m, with no single form that describes the energy dependence of all reactions.

Although eq 1. 19 can adequately represent the kinetic energy dependence of the reaction cross sections, it does not explicitly include the internal energies of the ion or neutral reactants. This can be accomplished by modifying eq 1. 19 to,



o(E) = Eog, (E+E,+Ei-ET )n/Em, (1.20)



where the subscripts r and i refer to a rotational and vibrational states; Er or E1 is the energy of that state; g, is the statistical population of the ith state; and the sum is over all states. ET is thus the reaction endothermicity corresponding to 0 K. Eq

1.20 is the model which is used to determine the CAD threshold in Chapter 4. On-resonance CAD

On-resonance CAD techniques have been demonstrated to measure BDE's in FTICR mass spectrometry in several grup.9-10.05 A short radio frequency (r.f.) pulse is applied to the excite plates on an ICR cell to increase the translational energy of a mass-selected precursor ion. The center-of-mass (c.m.) translational energy imparted to an ion during the excitation in FTICR is given in eq. 1.21








48



Ecm Elab* M (eV)
Mm (1.21)
E,t2e M

8m M~m

where Ecm and Elab are the energies in the center-of-mass and laboratory frame, respectively, M is the collision target molecular mass, m is the precursor ionic mass, t is the r.f. pulse length, Ed is the electric field (which is giving by VP, /d for an ideal infinitely long parallel plate cell, where V is the applied r.f. voltage (peak-to-peak)) and d is the plate spacing diameter of the ICR cell. Unfortunately, the measured Ed for cells of finite dimensions is only a fraction of the ideal value, so it is calculated using aVP /d, where a is a geometry factor for the ICR cell (a=0.90 for the Infinity' cell used in this work).106 The activated ion is then allowed to interact with the collision gas (Ar or Xe) for a short period of time (CAD delay time). The translational energy is converted into internal energies such as rotational and vibrational energies via ion/molecule collisions, allowing dissociation to occur during the CAD delay time.


M- L T -, (M-4) T M" L T (1.22)


where M+-L is the ion-ligand complex, T is the collision target gas (Ar), M+ is the ion, and L is the neutral ligand. After dissociation, the intensities of the fragment ions are measured. The ratio of the intensity of the product ion to the total ion intensity is determined. From plots of the intensity ratio (lM+/(IM++ IML+)) VS. the center-of-mass








49

interaction energy, Ecm (eq 1.21), the threshold for bond dissociation can be determined using a deconvolution process and eq 1.20.

The CAD process is initiated by ion-molecule interactions which are governed by long-range attractive potentials. These potentials are sufficiently strong to overcome small activation barriers. Thus, endothermic ion molecule reactions should generally proceed if the available energy exceeds the thermodynamic threshold for product formation.

Doppler broadening

The kinetic energy in the center-of-mass frame defined earlier does not include the thermal motion of the reactant neutral. These molecules have a Maxwell-Boltzmann distribution of velocities at the temperature of the ICR cell, usually 300 K in this work. The effects of this motion are to obscure sharp features in the true cross sections, especially the threshold for reaction. This so-called Doppler broadening was first described in detail by Chantry.07 As will be discussed in chapter 4, this energy distribution and that of the ion are explicitly considered by using them to convolute a trial function for the energy dependence cross sections (Eq 1.20 in our studies). Only after this convolution is accomplished is the model compared to the experimental data. To achieve the best reproduction of the data, the various parameters in the model (E0 or ET, 00, and n) are allowed to vary and are optimized by using nonlinear least-squares methods. In this work, the value of m was set to one, because this model always gave a threshold energy very close to the average value of all other fits considered in Armentrout's work. This assignment








50

is further justified on the basis of a theoretical model for translationally driven reactions" and empirically by the ability of this model to accurately measure thermochemical quantities in this system discussed later in Chapter 4.



Advantages. Disadvantages and Comparisons to Other Techniques Photodetachment

The BE measurements using thermal-equilibrium methods are relative rather than absolute values which can be obtained using the photodetachment technique. All the equilibria measurements are based on the temperature dependence of gas phase equilibria (n,n-1) observed in a high pressure mass spectrometer ion source. The equilibrium constants, Knfl1,f, are determined. Then AG0 is calculated from Van't Hoff plots of equilibrium constants. The BDE's determined from the linear plots have a certain error due to the uncertainty of the slope and intercept. The accuracy of the temperature measurement is also unreliable. The photodetachment technique provides very accurate and straightforward threshold determination due to the wide wavelength range, high resolution (very narrow bandwidth) and powerful output of the laser. However, in FTICR experiments, the depletion of the precursor ion is the only signal which can be monitored directly since the ICR instrument can not detect neutral products and electron detection is not straightforward. An alternate way to detect electrons in the ICR cell is use of an electron scavenger for electron capture, such as CCI4, SF6. Unfortunately, electron scavengers can not be used with solvated halides. The ion formation involving ion-molecule reactions with








51

halide compounds which is described detail in Chapter 3 is excessively complicated by CCI4and SF6. On the other hand, photoelectron spectroscopy (PES) can obtain the same information by measuring the electron kinetic energy at a fixed laser wavelength.

Photodetach ment threshold spectroscopy and photoelectron spectroscopy have been analyzed to obtain electron affinities and molecular structural information, and each has its particular strength and weaknesses. Photodetachment can be utilized to determine electron affinities with an accuracy limited essentially by the wavelength resolution of the light source, while similar determinations using photoelectron spectroscopy studies require only that the wavelength of the intense light source be shorter than the wavelength corresponding to the photodetachment threshold. As compared to CAD and thermal-equilibrium techniques, detachment thresholds obtained using photodetachment are much sharper and can be measured with greater accuracy. CAD

The use of CAD techniques in FTICR-MS to measure thermodynamic quantities for a variety of ionic species provides reaction cross section (or rate constants) over a wide range of kinetic energies. This technique does not require a laser apparatus and associated laser alignment compared to photodissociation and PIES techniques. CAD is an alternate approach to determining BDE when the ions of interest do not absorb or absorb little laser energy at near threshold ranges. However, the accuracy of such information is dependent on the models used to








52

analyze the cross sections and on accounting for effects such as the translational energy distributions of the precursor ion, the internal energies of the precursor ion, the possibility of multiple collisions and the radiative lifetimes of the intermediates.

One of the powerful aspects of using ion-molecule reactions to study chemistry is the ease with which the kinetic energy of the ion can be varied over a very wide range. Although this capability does not see extensive use, it is a very useful tool for probing the potential energy surfaces of reaction systems and, in particular, for measuring the heights of barriers to reaction. One application of this technique to measure thermochemistry of both ionic and neutral species is the determination of bond dissociation energies.













CHAPTER 2

INSTRUMENTATION

Introduction


The complex electronics instrumentation used in FTICR mass spectrometry is presented in Figure 2.1 as a simplified block diagram. Essential equipment for a Fourier transform ion cyclotron resonance mass spectrometer consists of a superconducting magnet, a high vacuum system, a trapping cell, and a console to detect ions, store signal and process acquired signal. Ionization sources may be categorized as internal or external. The internal ionization source is usually located inside the magnetic field (at the entrance of a trapping cell), produces better signalto-noise (S/N) ratio, and does not require a differential pumping set-up to maintain low base pressure. The external ionization source is located outside the magnetic field and operates at higher pressures (ca. 10' torr for El and MALDI, 760 torr for ESI and glow discharge (GD)). It requires several stages of differential pumping before ions enter the trapping cell region and also electrostatic or rf potential lenses to transfer the ions into the cell. The development of external sources has allowed the analyzer cell region to remain clean and at low pressure; therefore, the resolution is improved. Also, since the external source region can be isolated from the cell region, it is more flexible to switch between different ionization techniques 53















































E
0




U)
m
E



U
%6
0
E





0
Z
_0
CD
L-
CL
E



C14



U-







55







c)-E




CL LM
4) AS
(D
0 W

Li CL 0
ul



F CL
E
FN 0
CL ZE
4-0 C:
cn








56

in half an hour or less. This chapter will describe two ICR instruments with internal and external ionization sources. The cell and ESI design used for this work will also be described. The experimental configurations will be briefly introduced and then discussed in detail in the Experimental sections of Chapter 3 and Chapter 4 for both p hotodetachment and CAD techniques, respectively.



Nicolet 2 Tesla ICR Instrument



The FT-ICR mass spectrometer used for photodetachment studies (Chapter 3) was equipped with a 2 tesla superconducting magnet, a vacuum system (Figure 2.2) and a Bruker Aspect 3000 console with 512K memory. The vacuum chamber had 8" flanges on either end, one with three optical windows for transmission of laser radiation, the other with one window which allowed laser access to the filament side of the cell. This arrangement permitted easy alignment of the laser beam.

Figure 2.3 shows the laser set-up for this work. The laser beam was reflected into the center window of the three window flange by a quartz prism. The laser was externally triggered from the console. The console sent a TTL trigger pulse just before the electron beam pulse to an Apple 11 computer with a John Bell Co. interface card. Unless triggered, the computer and interface card sent both "charge" and "fire" commands to the Nd3":YAG laser so that the flashlamps fired at a rate of ca. 9 Hz. The computer shifted to an interrupt routine which sent not only

















c co


0 "0
C


c
0




c
M


0
-a
C

(D







M
0



C




c 0) CU CM
E =)
cu 0)
04 C
0 cu

C (D
E p E CA CL
c 0
co

cu



C"i C.






58

























0 0















-6
0

0

E
a)
L: (D
-C

0 E
0


0 0

c) <

CL
E
0 m 4) 0- r0 4) 0)

4- E
E
cq





CL
c _j

E
-Fu cli
C

X
(D
E
C
cu

'a 0 0 0 0 0
C: M 0 -j


'40
CL cn :3 C
CL


CD :3 U) a *
M 4-
_j 0
0

C, a 0
m 0
E (D
E ,c- L O)o 0 0)
iz 0 4-








60





"O













tn
c








E
co) 00
-.,,JCV ..
0


+ l l........... ....... ........ 1



z (D
70 r-......... -.. ...O.. .......c
O








61
"charge" and "fire" but also a "Q-switch" pulse which caused the laser to fire when interrupted by an external trigger pulse. The time from the rising edge of the TTL trigger pulse to the laser firing, which was monitored by an oscilloscope using a photodiode, was adjusted to match the FTICR pulse sequence so that the laser was fired right before ion excitation. The wavelength calibration of the dye laser was performed using the output of a He:Ne laser by a Continuum field service engineer immediately before this project was begun. The laser has been used solely for this research since the calibration was done. The wavelength was checked using a monochromator ( 0.2 nm) each time when a different dye was used. The pulse sequence used for this work is depicted in Figure 2.4.

The modified vacuum system, which was pumped by a 700 Us diffusion pump and a mechanical pump as a forepump, maintained the background pressure below 10-8 torr. The partial pressures of the three reagents used for the preparation of the solvated anions were controlled by a pulsed valve and three precision leak valves on the inlet system; this set up allowed very reliable control of the partial pressures which is critical for these experiments.

A cylindrical cell (2.45 inches diameter x 3.0 inches length) was used for the photodetachment experiments (as depicted in Figure 2.5). The fine mesh at the filament side of trapping plate was replaced with a coarse mesh to increase transmission of the laser beam and to allow the desired ions to be trapped without ionization of the metal by the laser beam. Ions were formed by internal electron impact (El), with electrons produced by a filament located in front of the trapping

















































c a)
E

CL
x

c
(D
E
-C
0 ca (D
0
-I.
0





c cu =3
cr

a)

CL
4 cli







(D
.o 63
CLC J U) 0


.0 co (
t5 C ca)
4 ) &Z
Cu

cri 0
0. E


a C


0


























Figure 2.5. Cylindrical cell.








65
cylindrical cell



3 i





li


30 1/21"




25 3/4"








3/4"



flange laser window
0.122" thickness
0.978" radius








66

plate. Because the volume of the cylindrical cell is much larger than that of conventional cubic cells used in our laboratories and elsewhere, space charge effects are reduced. The trapping potential was kept as low as possible (1.2-1.4 V) to prevent high velocity ions from being trapped.

Radiation in the 260-350 nm range after frequency doubling was provided by a tunable dye laser (Quantel TDL 50) with a frequency-doubling accessory. This dye laser was pumped by 532 nm (second harmonic of YAG fundamental) output from a Nd3*:YAG laser (Quantel YG 581C), which was synchronized with the mass spectrometer using laboratory-built interface electronics."9 The wavelength scan was controlled by commercial software provided with the dye laser. The wavelength accuracy of this laser system is 0.5 A, and the resetting stability is 0.02 A; the laser linewidth is 0.08 A. Five different dyes (Fluorescein, Rhodamine 590, 610, and 640, and DCM) were used and the provided wavelength range for each dye is listed in Table 2.1. The laser output power (ca. 8-15 mJ/pulse) was kept constant for each photodetachment spectrum.



Bruker 7 Tesla ICR Instrument



All the CAD experiments were performed on a Bruker FTICR mass spectrometer (Bruker Analytical Systems, Billerica MA) equipped with a Spectrospin (Fallanden, Switzerland), 7.0 T superconducting magnet, an RF-shimmed infinityTM































cc 3: CF)
C
c
0
CL
(1)
2
0
C)

(D
-C

c cc
cn
4
c
0
E


x

c 0)
E




0
*6
-C
CL

.0


:3 c

5, :3
0

> c c

rr

L= Q C(S











68









E E
c C:
co co
0) 14, co CY) C)
04 Tco CY)

C)
E
E
C%4
0 CY)
U
>% 0 cyi
0 co co
c


E
U- CD c
E U?
c m
E
LO
0 "L 6
co 0)
LO

a)
m C)
0)

E E
c c c c
m
w E C) LO
-C m m m
"0 LO N
0 L6
c rLO N


E
c E c
c U')
LO r-a r- 00
c Lo
0 0 1
a :3 C) C)
N
LO
0
-C
CD
-C c

c CD 0) 0-0 .0
m =$ CD =3
m c o i= o
0 m
(D <
3



















E cL = E CL -1 0 CL

L) 0)
C

U) C) co


cli
CL E z
CL
0


C (D C) < 0 a)
C






ca) C:





a) -a


E 3:
CD
c
0




E





CL
E :3
CL


C
a)

M CD








70





CD

(D (D 0
........ .......






0 ....



(D

cc$ ....





cts







....... ...








71

cell with a 2.45 inches diameter, a vacuum system, and an Aspect 3000 console. The vacuum system consisted of an external ion source, ion transfer optics, and cell regions which were pumped by 800 I/s (N2), 400 I/s (N2), and 400 I/s (N2) cryopumps, respectively (Figure 2.6). Ions were formed using an Analytica (Branford,CT) high-pressure ESI source. The ESI source region was pumped by two mechanical pumps and a 800 I/s (N2) cryopump to give pressures in the low1 0" mbar region. The ions were transferred to the analysis region via a series of electrostatic lenses, pumped by a 400 I/s (N2) cryopump. The analyzer region was maintained by another 400 I/s (N2) cryopump in the low 1010 mbar range, as indicated by a cold cathode gauge.

Electrospray Ionization (ESI) Source Region

Electrospray ionization (ESI) has became as the premiere ionization method for LC-MS analysis.10 The main reason for its success is the proven ability of ESI to simultaneously desolvate and ionize fragile polar analytes suspended in a liquid matrix at atmospheric pressure. This process can then be followed by mass analysis in the gas phase carried out under vacuum conditions.

The concept of ESI is simple. As an electrical potential is applied to a conductive needle through which liquid is flowing, a "spray" forms which disperses the liquid through the formation of a Taylor cone. This dispersion creates tiny liquid droplets (- 3 pm in diameter) which are electrically charged. These charged droplets then quickly evaporate to dryness.111114 Figure 2.7 shows the ion formation process in ESI. The work of Iribarne and Thompson is the most accepted











73

C)
2:1 > 0
cli d LO
LO r*--CL co

C) I I :0

0
.2 0.
co
>
(1)
0 0 0


+ C:
+ + o E 'w o 0 75 00 X ++ CD
+ + c ++ ++ .0

(D CFJ > 0 0 CL + co

+ +

+ ++

co




CD
z








74

description of the ESI mechanism."' According to their research, the ion suspended in solution evaporates into the gas phase when the shrinking droplet reaches the "Rayleigh limit" where the energy from repulsion of the charge on the droplet surface exceeds the energy of the surface tension which holds the droplet together. The resulting "coulombic explosion" tears the droplet apart and produces even smaller droplets with an higher charge-to-mass ratio. It is from these smaller droplets that ions actually "evaporate" from the liquid phase into the gas phase.

After the charged droplets enter the capillary, the evaporation process is controlled through the application of electrical fields and a heated "counter-current" drying gas, which helps to evaporate the droplets. The ESI chamber is surrounded by a metal screen so that the spray itself can be seen during the ionization process (Figure 2.8). This region is always maintained at atmospheric pressure. Inside the spray chamber there are two electrodes, excluding the needle itself which is normally maintained at ground potential to protect the user from the possibility of electric shock. The largest of these electrodes is the transparent cylindrical electrode (Vcyl), a wire screen which surrounds the bulk of the spray chamber. The final electrode in the atmospheric region of the source is the capillary (0.5 mm i.d.) entrance or V, which is held at ca. 4kV, in order to form charged droplets between the needle and the capillary. This electrode is fabricated by coating the face and side of the quartz dielectric capillary with metal (gold or platinum). The exit end is also coated with metal so that the capillary entrance and exit can be used independently as electrodes to help guide the ions to and out of the capillary. On







































































R) CO)
(D
_0
Fn
LLJ
06
cl








76




-a









31



LM

0
cn

U)
w


LM

x
LU co
E
E
co
LO
-j
0








77

the vacuum or exit side of the capillary, the ions are transferred through several skimmers (0.8 and 1.0 mm diameters) and lenses with pumping stages in the ESI source prior to entering the transfer region. The manipulation of all these voltages is important as they shape the electric field which causes the electrospray to form and guide the ions toward the capillary so they may be transported into the vacuum region of the source.

At the back end of the capillary where the ions emerge, a supersonic expansion takes place between the capillary exit and the skimmer. The capillary exit itself acts as a lens (1-1) and so has a variable potential. The region between the capillary exit and first skimmer (1-2) is differentially pumped to allow for the passage of as many ions as possible while reducing the gas pressure as well. After this skimmer, there is another differential pumping region, where another lens element, L3, and a second skimmer, L4, are present. The final lens elements were L5 and L6. This pressure region is "open" and therefore pumped by the pump in the source region of the mass spectrometer vacuum system. For this reason, L5 and L6 are visible from the outside of the source. L6 is the last controllable lens before the mass analyzer optics. Once the ion has been formed, it is transferred from atmospheric pressure, where the ionization process occurs, to the mass analyzer which is under vacuum through differential pumping stages. Introduction of ESI Features

The efficiency of ionization in ESI is greatly affected by the chemistry of the analyze with its supporting solution. The reason for this is the characteristic relation








78
between the analyte in the solution phase and the extracted analyte in the gas phase when electrospray ionization is used. Electrospray ionization is a very "soft" ionization technique, so minimal distortion of ionic structures, in both chemical and physical terms, occurs during the process of analyte ionization and desolvation.

There are a number of important chemical properties of the analyte which may affect the ion signal observed in ESI process. These include but are not limited to (1) pKa or pKb116, (2) concentration, (3) solubility, (4) molecular conformation, and

(5) molecular weight. Properties of the supporting solution itself can also have effects on the ion signal such as electric conductivity, surface tension, presence of salts, acids and bases, solvent polarity and volatility, and interference from other ionizable analytes.

pK 3orpK

If a protonated base desorbs from the droplet surface during ionization, the analytes which have higher PKa'S will have higher concentrations of the protonated forms of the bases, and therefore higher signals in ESI/MS. Concentration

ESI functions as a concentration sensitive detector when operated within a reasonable range of sample concentration. The range of concentration over which a linear response from the MS can be obtained, or linear dynamic range (LDR), is generally several orders of magnitude.'16 At both high and low ends of possible concentration levels, some deviation from linearity is expected to occur. The formation of the cluster species is frequently observed in ESI-MS when small








79

compounds having finite solubilities are used. The analyze is actually precipitating out in solution at higher concentrations, thus forming clusters on the original species. The concentration of sample may actually lead to decreased ion signal in the MS for the analyze of interest.

Solubilily

As solubility decreases, the likelihood of precipitation, and thus the formation of clusters, in solution increases. As shown above, this may lead to reduced ion signal. For this reason, it is necessary to consider the solubility of the compound of interest when choosing solvents and sample concentrations. Generally, ESI-MS spectra cannot be obtained for very large proteins (especially membrane-bound types) due to solubility issues. In these and other cases, usually a better solvent can be used to increase signal.

Molecular conformation

Most evidence concerning ESI leads to the conclusion that the efficiency of generating protonated ion is related to the pK,, as shown above. Also important is the accessibility of basic sites which attract the charges. This feature is important in larger macromolecules, such as proteins, which have tertiary and even quaternary structural components. In such species, the available basic sites may be "hidden" from the protons they seek.' 17 For this reason, it is sometimes necessary to change the solvent system, by adding dilute acid for example, in order to change the structure of the protein. In severe cases, the protein may be subjected to enzymatic or chemical degradation in order to increase ion signal.








80

Molecular weight

The possible effect of molecular weight on ionization is the mass discrimination"'8 which may occur in both the mass analyzer and the ion optics. Usually by adjusting the voltage parameters of the ion optics one can observe the effect of charge states for multiple charged ions. Solvent properties

High conductivity solutions will not electrospray well under "normal" conditions since the formation of a Taylor cone is prevented in this case.19 Solutions which are high in conductivity usually have acids, bases or salts added to them for various reasons. When using these solvents for electrospray, it is necessary to remove these additives to the sample, as the ion signal and stability are strongly dependent on the quality of the spray generated and the formation of the Taylor cone.

In a similar manner, solutions with a high surface tension are also difficult to use with ESI. It is for this reason that pure water does not electrospray well. This problem is generally overcome through the addition of an organic modifier, such as methanol, to the sample solution. The supporting solution is at least 20% in organic makeup to spray well. The accepted standard for optimum performance is 50%. Finally, the solvent volatility may also slightly affect the ion signal observed in ESI for a particular analyte. After a droplet has been formed by ESI, it must evaporate substantially before the electric field on the droplet surface is high enough to induce a coulombic explosion and subsequent ion evaporation of the analyte. This rate of








81

evaporation will depend on the solvent composition of the droplet. For this reason, solutions with higher organic content may give higher signal since the droplets formed will evaporate faster, leaving more time for analyte ionization to occur. Ion Transfer Region

After formation ions are then accelerated toward the analyzer by a series of electrostatic lenses (3 kV maximum potential between PL5 and PL 7) as shown in Figure 2.9. After acceleration the ions pass x,y-deflection, flight tubes (FLi, FL2), and PL9 at decreasing potential to relative ground prior to entering the cell region. A set of deflection lenses is used to gate the continuous ion beam, so that ions enter the analyzer region only during the ion accumulation period. Finally, the ions are transported into the mass analyzer where their mass is determined. The gate valve is used to isolate the analyzer region from the ionization source region for more rapid changes of the ionization source to El, MALDI, or GD. Ion Analyzer Region

The trapping cell used for CAD work is a rf-shimmed infinityTM cell as shown in Figure 2.10. At the front of the ICR cell there were two "side-kick" plates with pulsed low voltages (1-5 V) during the ion formation pulse. Ions entered the cell with a kinetic energy identical to the first "side-kick" plate potential. After the ions entered the cell, the voltages on the side-kick plates were dropped to the trapping potential (0.5 -1 V). To help trap the ions, the second set of side-kick plates (with two halves) is designed to convert the z-axis translational energy into the xy-plane cyclotron motion. Once the ions were trapped inside the cell, the CAD experiments were performed using a pulse sequence illustrated in Figure 2.11.
























































C
0
7


..D

cc





(D



V)
0





C
CD L
D 0)













83







0i
e,







> fl.

co
x



L cn
-j
0

4)


>













Lr)









Cj -i
OQ CL





cn
-J C"j


C"j X
(A ool/
















-C

(D cu
-4-- am 0)

0
CL

E

-C
E E


CD

-C

x a)

CL
Q
Co
C'i CL
0) C
CL 'a
CL .M cu
C: 0 x c > m CL co

C




x (D
cn

-C
cn (D
_0
0
-C cn o in



76 0 L2
CD C



C
Om

(D








85





4B "a CC
0
E E >
:2 -7m .6"
C:
2
4"





co



--- iwcm






cm cm



























Figure 2.11. Pulse sequence for CAD experiments.




Full Text
71
cell with a 2.45 inches diameter, a vacuum system, and an Aspect 3000 console.
The vacuum system consisted of an external ion source, ion transfer optics, and cell
regions which were pumped by 800 I/s (N2), 400 I/s (N2), and 400 I/s (N2)
cryopumps, respectively (Figure 2.6). Ions were formed using an Analytica
(Branford,CT) high-pressure ESI source. The ESI source region was pumped by
two mechanical pumps and a 800 I/s (N2) cryopump to give pressures in the Iow10'5
mbar region. The ions were transferred to the analysis region via a series of
electrostatic lenses, pumped by a 400 I/s (N2) cryopump. The analyzer region was
maintained by another 400 I/s (N2) cryopump in the low 10'10 mbar range, as
indicated by a cold cathode gauge.
Electrospray Ionization (ESH Source Region
Electrospray ionization (ESI) has became as the premiere ionization method
for LC-MS analysis.110 The main reason for its success is the proven ability of ESI
to simultaneously desolvate and ionize fragile polar analytes suspended in a liquid
matrix at atmospheric pressure. This process can then be followed by mass
analysis in the gas phase carried out under vacuum conditions.
The concept of ESI is simple. As an electrical potential is applied to a
conductive needle through which liquid is flowing, a "spray" forms which disperses
the liquid through the formation of a Taylor cone. This dispersion creates tiny liquid
droplets (~ 3 pm in diameter) which are electrically charged. These charged
droplets then quickly evaporate to dryness.111'114 Figure 2.7 shows the ion formation
process in ESI. The work of Iribarne and Thompson is the most accepted


ACKNOWLEDGMENTS
I would like to express my appreciation to my research advisor, Dr. John R.
Eyler, for his guidance and assistance, especially for his input of FT-ICR
background and theory. I also appreciate the technical knowledge of the
instrumentation provided by Dr. Cliff Watson. I would like to thank Drs. Dil Peiris
and Nathan Ramanathan for their efforts on the initial IRMPD work of the
metal/crown ether complexes. The comments and suggestions on my research
projects from Dr. Philip Brucat and Dr. David Richardson are also acknowledged.
My appreciation extends to my parents; without their love and support, I
would never be able to finish this work. Thanks to all of the members (Rich Burton,
Kevin Goodner, Tom Hayes, David Kage, Lisa Lang, Eric Milgram, and Joe
Nawrocki) of the current Eyler group for their assistance and friendship. Thanks
also go to Dr. Kitty Williams for her assistance in the preparation of research results
for publication.
Last but certainly not least, I would like to thank my dear husband, Chen-
Chan Hsueh, for his support, love and help. He has been so understanding and
helpful during my graduate career. Without him, I would not have been able to
finish the graduate studies.
ii


180
M+L-> M+ + L,
M= Na, K, Rb and Cs
(4.10)
as shown in Figures 4.9 and 4.10 for (12-C-4)Rb+. The (18-C-6)H30 + complex
underwent desolvation and formed (18-C-6)H+ (Figs. 4.11 and 4.12).
(4.11)
(18-C-6)H30+ (18-C-6)H+ + H20
as observed in previous IRMPD studies at low laser fluence.160
The (15-C-5)FI30+ underwent desolvation, followed by fragmentation of the
crown ether. At low rf excitation energies, it desolvated into (15-C-5)H+ (Figs. 4.13
and 4.14), but at high energies, desolvation was followed sequentially by
dissociation into (C2FI40)3FI+ and (C2FI40)2H+ fragments (Fig. 4.15). The absolute
intensities plot versus center-of-mass translational energy, Ecm, for (15-C-5)FI30+
complex including precursor, successor, and all the fragment ions is shown in
Figure4.16. The intensity of precursor ion, (15-C-5)FI30+, decays exponentially with
Ecm, the intensity of successor ion, (15-C-5)FT, decreases with Ecm at low
translational energy. At high translational energy, the signal of (15-C-5)FT
decreases as the signal of the fragments increases.
(15-C-5)H30+ -+ (15-C-5)H+ (C2H40)3H+
(4.12)
(C2H40)2H+


100
80
60
40
20
0
AV=70 volts
177.1
(12-crown-4)H+
(12-crown-4)2H+
1
:
353.2
u
7.0
113.6
210.2
306.8
403.4
500
Mass-to-charge ratio
179


Table 3.1. Electron affinities (EA) and proton affinities (PA) of species studied in this
work.


120
Results and Discussion
The relative photodetachment spectra for bromide ion solvated with CH3OH,
C2H5OH, /-C3H7OH, n-C 3H 7OH, and CH 3CN are presented in Figure 3.12. As
explained above, a delay period was required to assure that the gas-phase
exchange reactions for the preparation of the Br and I' adducts (eq 3.9) reached
equilibrium and that (CH3OH)F and (CH 3OH)CI' (eqs 3.5 & 3.6) were properly
thermalized. Figures 3.13, 3.14, and 3.15 show photodetachment spectra for
(CH3OH)F', (CH 3OH)CI, and (CH 3OH)Br', respectively, obtained with increasing
delay time. Studies of the bromide systems showed results similar to (CH3OH)CI\
The reaction of fluoride ion with methyl formate (eq 3.5) is very exothermic, and a
delay of about 4 seconds was required for collisional cooling of the (CH3OH)F ions.
The overall shift in the threshold (cf., Figure 3.13) from 272.0 nm to 264.4 nm
corresponds to a cooling of 87.6 K for the (CH3OH)F' complex. Because the newly
formed (CH3OH)CI' ions were less energetic, they reached thermal equilibrium in
about 2.5 seconds. The formation of (CH3OH)Br reqiures high pressure and
achievement of thermal-equilibrium, therefore, the 2.5-second delay was also
sufficient for the exchange reactions (eq 3.9) to reach equilibrium.
By a thermochemical cycle for the photodetachment process (eq 3.2) the
threshold is equal to the sum of the binding energy of the solvent anion complex
and the electron affinity of the halogen, with a small correction as described below.
Thus, the BE's may be calculated from the thresholds and the known EA's154 given


52
analyze the cross sections and on accounting for effects such as the translational
energy distributions of the precursor ion, the internal energies of the precursor ion,
the possibility of multiple collisions and the radiative lifetimes of the intermediates.
One of the powerful aspects of using ion-molecule reactions to study
chemistry is the ease with which the kinetic energy of the ion can be varied over a
very wide range. Although this capability does not see extensive use, it is a very
useful tool for probing the potential energy surfaces of reaction systems and, in
particular, for measuring the heights of barriers to reaction. One application of this
technique to measure thermochemistry of both ionic and neutral species is the
determination of bond dissociation energies.


233
169. Li, H.; Jiang, T; Butler, I. S. J. Raman Sped. 1989, 20, 569.
170. Takeuchi, H.; Arai, T; Harada, I. J. Mol. Strudure 1986, 146, 197.
171. Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chem. Rev. 1991,
91, 1721.
172. Izatt, R. M.; Bradshaw, J. S.; Nielsen, S. A.; Lamb, J. D.; Christinsen, J.;
Sen, D. Chem. Rev. 1985, 85, 271.
173. Glendening, E. D.; Feller, D.; Thompson, M. A. J. Am. Chem. Soc. 1994,
116, 10657.
174. Yeh, C. S.; Willey, K. F.; Robbins, D. L.; Pilgrim, J. S.; Duncan, M. A. Chem.
Phys. Lett. 1992, 196, 233.
175. Pozniak, B.; Dunbar, R. C. Int. J. Mass Sped.; Ion Processes, 1994, 133,
97.
176. Dawson, P. H. Int. J. Mass Sped., Ion Processes. 1982, 43, 195.
177.Gentry, W. R. Gas Phase Ion Chemistry, Bowers, M. T. Ed. Academic
Press: New York, 1979, Vol. 2, p. 221.
178. Ervin, K. M.; Armentrout, P. B. J. Chem. Phys. 1985, 83, 166.
179. (a) Meyer, F.; Chen, Y.; Armentrout, P. B. J. Am. Chem. Soc. 1995, 117,
4071. (b) Hales, D. A.; Su, C.; Lian, L.; Armentrout, P. B. J. Chem. Phys. 1994,
100, 1049.
180. Dalleska, N. F.; Tjelta, B. L.; Armentrout, P. B. J. Phys. Chem. 1994, 98,
4191.
181. Marinelli, P. J.; Squires, R. R. J. Am. Chem. Soc. 1989, 111, 4101.
182. Wenthold, P. G.; Squires, R. R. J. Am. Chem. Soc. 1994, 116, 6401.
183. Magnera, T. F.; David, D. E.; Michl, J. J. Am. Chem. Soc. 1989, 111, 4100.
184. Nonose, S.; Tanaka, H.; Nagata, T.; Kondow, T. J. Phys. Chem. 1994, 98,
8866.
185. Shimanouchi, T.; Tables of Molecular Vibrational Frequencies, Part I.
National Standard Reference Data Series-NBS 6: Washington D.C. 1967.


145
such studies,120174'175 but it is not feasible, if, as in this investigation, there is no
absorption of the laser radiation near the threshold range of the ions. An alternative
method is collisionally-activated (formerly collisionally-induced) dissociation (CAD
or CID),88104'105'176'184 which can be applied over a wide energy range, and which
does not require laser apparatus and the associated alignment problems.
Several groups have used CAD and Fourier transform ion cyclotron
resonance mass spectrometry (FTICR-MS) for the measurement of bond
dissociation energies (BDE's).88'104'105,176 After formation of the gas-phase ions,
unwanted ionic species are ejected to isolate the precursor ion (alkali-metal or
hydronium/crown-ether complex in this study) with the neutral collision gas (Ar) in
the ICR cell. A short radiofrequency (rf) pulse of voltage Vpp and duration t is then
applied to the excite plates of the cell. This increases the center-of-mass
translational energy, E^,, of the ions according to the relationship given in equation
1.21.
The energetic ion is then allowed to interact with a neutral collision gas (Ar
or Xe) for a delay time, during which an ion/molecule collision converts translational
energy to internal vibrational and rotational modes. If the energy imparted (eq 1.21)
was sufficient, the complex dissociates to the successor ion and the neutral crown
ether, L:
M+-L - M+ + L (4.1)
The precursor and successor ion intensities are then measured by the usual ICR


Rotational energy 159
Vibrational energies 159
Trial Function 160
Deconvolution 161
Results 163
Collision Conditions 163
Fragmentation Pathways 164
Bond Dissociation Energies 197
Discussion 210
Conclusion 216
5. CONCLUSION 218
REFERENCE LIST 223
BIOGRAPHICAL SKETCH 235
v


Figure 2.4. Pulse sequence for photodetachment experiments.


Figure 3.12. Photodetachment spectra: (CH3CN)Br; (CH3OH)Br; T (C2H5OH)Br; (i-C3H7OH)Br; X (n-
C3H7OH)Br.


Figure 1.8 The structures of polyether macrocycles.


Figure 2.11. Pulse sequence for CAD experiments.


Figure 4.18. Mass spectrum of (12-C-4)H+ after CAD at AV=70 volts. Peaks corresponding to (C2H40)nH+ (n = 2 and
3) were ions observed.


50
is further justified on the basis of a theoretical model for translationally driven
reactions108 and empirically by the ability of this model to accurately measure
thermochemical quantities in this system discussed later in Chapter 4.
Advantages. Disadvantages and Comparisons to Other Techniques
Photodetachment
The BE measurements using thermal-equilibrium methods are relative rather
than absolute values which can be obtained using the photodetachment technique.
All the equilibria measurements are based on the temperature dependence of gas
phase equilibria (n,n-1) observed in a high pressure mass spectrometer ion source.
The equilibrium constants, Kn_1>n, are determined. Then aG0 is calculated from Van't
Hoff plots of equilibrium constants. The BDE's determined from the linear plots
have a certain error due to the uncertainty of the slope and intercept. The accuracy
of the temperature measurement is also unreliable. The photodetachment
technique provides very accurate and straightforward threshold determination due
to the wide wavelength range, high resolution (very narrow bandwidth) and powerful
output of the laser. However, in FTICR experiments, the depletion of the precursor
ion is the only signal which can be monitored directly since the ICR instrument can
not detect neutral products and electron detection is not straightforward. An
alternate way to detect electrons in the ICR cell is use of an electron scavenger for
electron capture, such as CCI4, SF6. Unfortunately, electron scavengers can not be
used with solvated halides. The ion formation involving ion-molecule reactions with


In (No/N)
274.0 277.0 280.0 283.0
Wavelength/nm
126


11
Fourier transform (SWIFT) excitation,24,30'32 the excitation waveform is generated
by taking the desired frequency spectrum (a) (power as a function of frequency) and
performing an inverse FT to give a time-domain signal (b). This waveform is then
generated by a digital frequency synthesizer. The excitation spectrum (c), which
indicates the actual amount of power applied at each frequency, is produced by
Fourier transformation of (b). This excitation spectrum is much more uniform over
the desired frequency band than either impulse or chirp excitation.
Any of these three excitation modes will cause the ions to absorb energy and
be accelerated into larger orbits when the r.f. frequency becomes equal to the ion
cyclotron frequency. It is also necessary to excite a packet of ions of a given mass-
to-charge ratio not only to a larger ICR orbital radius, but to coherent motion in order
to detect them. Figure 1.4 shows how the ions achieve coherence prior to
detection.
Ion detection
The orbiting packet of ions induces a small alternating image current on the
receive plates (perpendicular to the y-axis). The image current is given by,33
where N is number of ions, v is the ion velocity, q is the charge of the ion, and d is
the electrode spacing, i.e. the dimension of the cell as depicted in Figure 1.5. The
ion induces image current alternately on upper and lower plates. To obtain a


Figure 3.1. Mass spectrum of F\


Figure 3.3. Mass spectrum of (CH3OH)F\ m/z = 51.


97
NF3 + e'-> NF2 + F'
(3.3)
COCI2 + e' COCI' + Cl
(3.4)
The anions subsequently reacted with a second reagent gas to form the desired
adduct (Figures 3.3 and 3.4 for (CH3OH)CI' and (CH3OH)F):147
F + HCOOCH3 (CH3OH)P + CO
(3.5)
COCI' + CH3OH (CH3OH)C|- + CO
(3.6)
The neutral carbon monoxide removed some excess energy resulting from
these exoergic reactions, but a relaxation delay of 2.5 to 4 seconds was still
required to allow the (CH3OH)F'or (CH3OH)CI' to approach thermal equilibrium prior
to the photodetachment measurements. For this work the thermalization time was
determined by measuring the photodetachment threshold with increasing delay
times (2-6 seconds), until no significant change was observed. Results of earlier
IR dissociation experiments involving (CH3OH)F' suggested that ca. 100 collisions
between the ions and the bath gas (methyl formate) were required for
thermalization.149'150
The preparation of (CH3OH)l' and the adducts of Br with CH3OH, C2H5OH,
/-C3FI7OH, n-C3H7OH, and CH3CN required an alternate method involving the gas-
phase exchange of (FI20)X' with the desired solvent. Initially, hydroxide ions from


100
80
60
40
20
0
AV=70 volts
261.1
(12-crown-4)Rb+
263.0
J
.0 140.0 230.0 320.0 410.0 500.0
Mass-to-charge ratio
182


Figure 4.14. Mass spectrum of (15-C-5)H30+ after CAD at low activation energy and AV=20 volts. Peak corresponding
to (15-C-5)H+ ion was the only successor ion observed.


25
methods and experimental techniques. Currently, approaches have reached a high
degree of accuracy and sophistication. Despite this enormous amount of
information, however, no comprehensive and satisfactory theory of ion solvation can
be given, due mainly to the lack of a detailed knowledge of the structure of non
simple liquids. The problems occurring in model calculations on ion solvation are
two-fold. Firstly, accurate energy surfaces for ion-molecule complexes are difficult
to obtain. They provide information about the most stable geometries and energies
of interaction. The static energy of a complex, unfortunately, is only one side of the
picture. The shape of potential surfaces for clusters consisting of an ion and many
molecules, in general, indicates high flexibility, since several flat energy minima
separated by low barriers are usually found. Therefore, the dynamics of the
complex is the second important problem for most macroscopic properties at room
temperature.
An alternate and more rigorous treatment of ion solvation is based on the
study of ion-molecule reactions in the gas phase. Relative enthalpies and free
enthalpies of complex formation for solvated ions X'(H20)n with n = 1-6 have been
determined.42 59 The concept of hydrogen bonding (HB) emerged from the study of
self-associated liquids (notably water)63 and its importance in complex biological
systems is paramount.64 The study of hydrogen-bond, HB, interactions in the gas
phase is relevant because it provides direct information on the structure and stability
of these complexes in the absence of the perturbations induced by the solvent.
Most of the solvated ions are easily generated and heavily studied.41-45 59


Figure 4.11. Mass spectrum of (18-C-6)H30+ at AV=20 volts. Peak corresponding to (18-C-6)H30+ ion was isolated
before excitation.


Figure 4.8. Mass spectrum of (12-C-4)H30+ at AV=70 volts. Peaks corresponding to (12-C-4)H+ and (12-C-4)2H+ ions
were observed. The peak intensity of (12-C-4)2H+ is lower than that of (12-C-4)H+.


225
34. Comisarow, M. B. Adv. Mass Spec. 1978, 7, 1042.
35. Comisarow, M. B. Adv. Mass Spec. 1980, 9, 1698.
36. Keesee, R. G.; Castleman, A. W., Jr. J. Phys. Chem. Ref. Data, 1986, 15,
1011.
37. Jennings, K. R. Ed. Fundamentals of Gas Phase Ion Chemistry, NATO ASI
Series, Kluwer: Dordrecht, 1991, Vol. 347.
38. Miller, T. M. Adv. Electron. Electron Phys. 1981, 55, 119.
39. Janousek, B. K.; Brauman, J. I.; Simons, J. J. Chem. Phys. 1979, 71, 2057.
40. Sharma, R. B.; Blades, A. T.; Kebarle, P. J. Am. Chem. Soc, 1984, 106, 510.
41. Meot-Ner (Mautner), M. J. Am. Chem. Soc, 1983, 105, 4906.
42. Blades, A. T.; Jayaweera, P.; Ikonomou, M. G.; Kebarle, P. Int. J. of Mass
Spec. Ion Proc. 1990, 101, 325.
43. Meot-Ner (Mautner), M. J. Am. Chem. Soc, 1983, 105, 4912.
44. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc, 1985, 107, 2612.
45. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc, 1988, 110, 1087.
46. Jarrold, M. F.; lilies, A. J.; Bowers, M. T. J. Am. Chem. Soc, 1985, 107, 7339.
47. Freiser, B. S.; Hettich, R. L.; J. Am. Chem. Soc, 1987, 109, 3537.
48. Brucat, J. P.; Zheng, L.-S.; Pettiette, C. L.; Yang, S. and Smalley, R. E. J.
Chem. Phys. 1986, 84, 3078.
49. Geusic, M. E.; Jarrold, M. F.; Mcllrath, J. J.; Freeman, R. R.; Brown, W. L.; J.
Chem. Phys. 1987, 86, 3862.
50. Moseley, J. T.; Cosby, P. C.; Peterson, J. R. J. Chem. Phys. 1976, 65, 2512.
51. Bowen, K. H. J. Chem. Phys. 1985, 83, 3169.
52. Stevens, A. E.; Feigerle, C. S.; Lineberger, W. C. J. Chem. Phys. 1983, 78,
5420.


LIST OF FIGURES
Figure page
1.1.Standard ICR trapping cells, (a) Cubic cell with two trapping
plates perpendicular to z-axis, two excite plates perpendicular
to x-axis, and two detect plates perpendicular to y-axis. The
ion is trapped inside the cell with cyclotron motion in the xy-
plane, moving along the z-axis. (b) Cylindrical cell 5
1.2.Standard pulse sequence for ICR experiment 8
1.3. Ion excitation. (1) Impulse excitation: approximate delta function
vs time, Fourier transform to the frequency domain signal (2)
Chirp excitation: an rf frequency sweep vs time, Fourier
transform to the frequency domain signal (3) SWIFT: (a) The
desired frequency spectrum, (b) Inverse FT to give a time-
domain signal, (c) Fourier transform to the frequency domain
signal 10
1.4. (1): (a) Ions' cyclotron motion prior to excitation, (b) Ions are
excited to a larger orbit, (c) Ions achieve coherence after
excitation. (2) Phase coherency: (a) Two ions of same m/z
180 out of phase, (b) The in-phase (initial motion parallel to
electric field, E) ion speeds up and out-of-phase ion slows
down until they eventually achieve coherent motion 13
1.5.Ions induce image current alternately on upper and lower
plates 15
1.6.Heterodyne mode: The signal mixes with a reference to give a
sum (higher frequency) and difference (lower frequency)
signal. The higher frequency (low mass) signal is then filtered.
Fast FT gives the low frequency signal which is the
difference 19
1.7. The structures of CH3OHX, X = F, Cl, Br, and 1 28
vii


91
intermediate loosely bound neutral complex may be necessary. The threshold for
photodetachment is obtained by measuring the depletion of SX as a function of
photon energy. This value can be used to calculate the BE of SX', if the electron
affinity (EA) of X is known (either from established values or the photodetachment
threshold of unsolvated X ). Brauman et a/.142"143 have used this method to evaluate
the EA's of several systems including organophosphines and aliphatic oxyl radicals.
Similar measurements of photodissociation reactions (similar process to eq 3.2, but
with the charge remaining on one of the products, presumably X) provide data on
the BE's of solvated cations. For example, Brucat et at.'22 used photodissociation
in a TOF mass spectrometer to obtain the upper limit of the ground-state adiabatic
dissociation energy of (H20)V+. The binding energies of Mg(FI 20) n + (n1) were
determined in Fukes120 and Duncan's121 groups by the same technique. Johnson
et at.'27 also used time-of flight (TOF) mass analysis to demonstrate the competition
between photodissociation and photodetachment in hydrated electron clusters,
(Fl20)n\ They observed that photodetachment is the only process at high excitation
energies. Flowever as the photon energy decreases, photodissociation begins to
compete with photodetachment.
Ion cyclotron resonance mass spectrometry is especially desirable for
photodetachment studies, because ions may be trapped for extended periods and
subjected to laser irradiation. Thus, it is feasible to form solvated anions by suitable
ion-molecule reactions, and to measure their BE's by the photodetachment process.
This report presents results on a number of alcohol-solvated halide ions, (ROFI)X',


3
cyclotron frequency, w. The Lorentz force, F, is given by
F q ( V x So)
(1.2)
For B0 along the z axis, this force is constrained to the xy-plane (there is no force
along the z-axis), so ions can move freely along that axis unless an electric field is
applied,28,29
(1.3)
where E is the electric field.
Flowever, in order to detect the ions, they must be confined within a finite
volume bounded by conductive electrodes (in a trapping cell). Figure 1.1 is a
schematic drawing of two different ICR trapping cells which are commonly used:
cubic and cylindrical. The cell is centered in a homogeneous magnetic field
generated by a superconducting magnet and contained in a high vacuum chamber.
There are two trapping plates which are perpendicular to the magnetic field (z-axis),
two excite plates (perpendicular to the x-axis) and two detection plates
(perpendicular to the y-axis).
The ions may be formed in the ICR cell (using an internal ionization source)
or outside the magnetic field (with an external ionization source) by a number of
methods, including electron impact (El), chemical ionization (Cl), laser desorption


Figure 1.1 Standard ICR trapping cells, (a) Cubic cell with two trapping plates perpendicular to z-axis, two excite plates
perpendicular to x-axis, and two detect plates perpendicular to y-axis. The ion is trapped inside the cell with cyclotron
motion in the xy-plane, moving along the z-axis. (b) Cylindrical cell.


t=3.20 sec.; B-t=4.34 sec 126
3.15. Photodetachment spectra of (CH3OH)Br obtained using
different thermalization times (t): -1=2.05 sec.; -1=3.77
sec.; -1=4.34 sec 128
3.16. Potential energy curves for alcohol-solvated halide anions
and corresponding neutral species. The lower curve
represents the interaction of the solvent and halide from infinite
separation to a stable bound state, a) The upper solid curve
represents the hydrogen abstraction reaction of ROH + X to
form RO + HX. (X = F and Cl). The arrows indicate the
thermochemical cycle for the photodetachment process,
including the threshold (Ethresh0,d), electron affinity (EA) of X,
binding energy (BE) of (ROH)X', and the energy difference
(AE) between the reported and actual BE. b) Same as a) with
dotted curve to indicate hydrogen abstraction reaction for X=
Br and 1 136
4.1. Potential energy curve for ion crown ether complexes and
corresponding neutral species. The D298 represents the bond
dissociation energy at room temperature, the D0 represents the
bond dissociation energy at zero Kelvin temperature 149
4.2. Appearance curve of (18-C-6)Rb+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 158
4.3. Appearance curves of (18-C-6)Rb+ at different pressures (50 ms
CAD delay time); 3.0 10'8 torr, 5.5 x 108 torr, 7.3
x 10'8 torr, A 9.2 x 108 torr 166
4.4. Appearance curves of (18-C-6)Rb+ at different CAD delay times
(P = 3.3 x 10 8 torr); 0.05 s, 0.08 s, 0.1 s, 0.15
s 168
4.5. Mass spectrum of (12-C-4)Rb+ at AV=20 volts. Peaks
corresponding to (12-C-4)Rb+ and (12-C-4)2Rb+ ions were
observed 173
4.6. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peaks
corresponding to (15-C-5)H30+, (15-C-5)2H+, and (15-C-
x


49
interaction energy, Ecm (eq 1.21), the threshold for bond dissociation can be
determined using a deconvolution process and eq 1.20.
The CAD process is initiated by ion-molecule interactions which are governed
by long-range attractive potentials. These potentials are sufficiently strong to
overcome small activation barriers. Thus, endothermic ion molecule reactions
should generally proceed if the available energy exceeds the thermodynamic
threshold for product formation.
Doppler broadening
The kinetic energy in the center-of-mass frame defined earlier does not
include the thermal motion of the reactant neutral. These molecules have a
Maxwell-Boltzmann distribution of velocities at the temperature of the ICR cell,
usually 300 K in this work. The effects of this motion are to obscure sharp features
in the true cross sections, especially the threshold for reaction. This so-called
Doppler broadening was first described in detail by Chantry.107 As will be discussed
in chapter 4, this energy distribution and that of the ion are explicitly considered by
using them to convolute a trial function for the energy dependence cross sections
(Eq 1.20 in our studies). Only after this convolution is accomplished is the model
compared to the experimental data. To achieve the best reproduction of the data,
the various parameters in the model (E0 or ET, o0, and n) are allowed to vary and are
optimized by using nonlinear least-squares methods. In this work, the value of m
was set to one, because this model always gave a threshold energy very close to
the average value of all other fits considered in Armentrout's work. This assignment


This dissertation was submitted to the Graduate Faculty of the Department
of Chemistry in the College of Liberal Art and Sciences and to the Graduate School
and was accepted as partial fulfillment of the requirements for the degree of Doctor
of Philosophy.
May, 1996
Dean, Graduate School


100
80
60
40
20
0
19.001
P
Jm
17.0
21.1
27.7 40.4
Mass-to-charge ratio
74.8
i
mMMMM
499.7
CD


147
ionization methods, such as El, LD, and MALDI, ESI imparts much less internal
energy, which is an important factor in BDE measurements.
The accuracy of BDE's obtained by the CAD technique depends on the
model used to analyze the cross section of the ion/molecule reaction and the
corrections for various systematic effects that may otherwise distort the
thermodynamic data. Both McMahon104 and Eyler88105 have used linear-
extrapolation plots to obtain the observed thresholds, assuming that the cross-
section for dissociation increases linearly with energy above the threshold.
Assuming that the precursor ions were populated at room temperature, and that all
available energy was used for dissociation, the observed thresholds were
interpreted to equal the BDE's with the successor ions produced at zero Kelvin. To
obtain the true thresholds, these values were then deconvoluted with a trial function,
which includes Chantry's107 Doppler-broadening term to account for the thermal
spread of the neutral collision gas. The values were then converted to BDE's at
room temperature by estimating the internal energy of the successor ions and
adding to the 0 K result, (i.e., D298, as shown in Figure 4.1). There are assumptions
and uncertainties in both this type of fitting procedure and the estimation of the
internal energies.
In the present work, a much more careful data analysis was used. The
threshold energies were evaluated using the CRUNCH deconvolution program
developed by Armentrout and coworkers.178'179 This model corrects for the internal
energies of the precursor ions based on calculated frequencies of the normal


Figure 4.16. Absolute intensity plots versus Ecm of (15-C-5)H30+, (15-C-5)H+, and (C2H40)nH+ (n = 2 and 3).


Figure 4.3. Appearance curves of (18-C-6)Rb+ at different pressures (50 ms CAD delay time); 3.0 10'8 torr, -
5.5 x 10'8 torr, 7.3 x 10'8 torr, A 9.2 x 10'8 torr.


CHAPTER 5
CONCLUSION
Photodetachment was first used to measure the binding energies of
(CH3OH)X- (X = F, Cl, Br, I), (CH 3CN)Br, and (ROH)Br' (R= CH3, C2H5, /-C3H 7, n-
C3H 7), in a Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer.
The methanol complexes of F and Cl were prepared by sequential ion-molecule
reactions, while equilibrium solvent exchange was used for preparation of the Br
and I' complexes. The photodetachment process was observed over the
wavelength range of 260-350 nm using a tunable dye laser with a frequency
doubling accessory. The laser can be tuned over a wide wavelength range, and
allows an accurate threshold to be determined due to its very narrow bandwidth.
Assuming that the photodetachment process leads to complete dissociation
to the neutral solvent plus halogen atom, the binding energies were calculated by
subtracting the known electron affinities of halogens from the measured
photodetachment thresholds. The binding energy values obtained for (CH3OH)F',
(CH3OH)C|-, (CH3OH)Br, (CH3OH)l, (C2H5OH)Br, (/-C3H7OH)Br\ (n-C3H7OH)Br\
and (CH3CN)Br are 30.2, 18.9, 15.2, 14.6, 15.7, 16.6, 17.0, and 13.5 kcal/mol,
respectively.
218


100
80
60
40
20
0
17.0
21.1
(CH3OH)CI
66.991
68.992
27.7 40.4
Mass-to-charge ratio
74.8
500.5


236
presentation at the Florida Section ACS Meeting in May and was awarded a Russell
Dissertation Fellowship in December, 1995. In 1996, Mrs. Yang completed her
Ph.D. in physical chemistry under the guidance of Professor John R. Eyler. Her
current research involves determination of bond dissociation energies using
photodetachment, on-resonance collision activated dissociation (CAD) and
photodissociation techniques in FT-ICR mass spectrometry.


Delay Time
i*
Time
Quench pulse
Ion Formation pulse
Excitation pulse
Detection
CO


Figure 1.4 (1): (a) Ions' cyclotron motion prior to excitation, (b) Ions are excited to a larger orbit, (c) Ions achieve
coherence after excitation. (2) Phase coherency: (a) Two ions of same m/z 180 out of phase, (b) The in-phase (initial
motion parallel to electric field, E) ion speeds up and out-of-phase ion slows down until they eventually achieve
coherent motion.


164
where a = nDol represents the average number of target molecules. The ratio of
double to single collisions, qj/q,, is approximately 0.001 with an Ar pressure of 5
x10 8 mbar (3.8 xio8torr) and a 30 msec collision time. In other words, fewer than
0.1 % of the precursor ions should experience multiple collisions with this pressure
and delay time.
To ensure that the single-collision condition was obeyed, the dissociation of
(18-crown-6) Rb+ was studied using different Ar pressures (3.0, 5.5, 7.3, and 9.2 x
10'8 mbar) with a constant CAD delay time of 50 ms and, in a separate series of
experiments, using different CAD delay times (0.05, 0.08, 0.1, and 0.15 s) with the
Ar pressure held at 3.3x1 O'8 mbar. The resulting appearance curves are shown in
Figures 4.3 and 4.4. The curves obtained with pressures of 3.0 and 5.5x 10'8 mbar
and CAD delay times of 0.05 and 0.08 sec are almost identical, and the thresholds
are about the same. With higher pressures and CAD delay times, the thresholds
shifted to lower energy due to multiple collisions. All subsequent appearance
curves were obtained with the pressure and delay times in the acceptable ranges.
Fragmentation Pathways
Table 4.2 presents a compilation of all ions observed in the mass spectra,
both prior to isolation of the precursor ions and subsequent to the CAD process.
Referring to the columns showing the ions present prior to isolation, only the
molecular ions (LM+) were formed by 18-C-6 and 15-C-5 with M+ (M=Na, K, Rb, Cs,


43
Photodetachment spectra consist of the wavelength dependence of this cross
section. The geometric factor may be evaluated from the beam diameter, photon
flux, and the size of the ion cloud, if absolute cross sections are needed.92
However, for a given photodetachment spectrum, all parameters in the denominator
of eq 1.13 are constant, if the laser flux is kept constant. Thus, the relative
photodetachment cross-section spectrum, which is sufficient for determination of the
photodetachment threshold, may be obtained by plotting ln(No/N) versus the laser
energy, E (A).
In a well known paper published in 1948, Eugene Wigner derived the form
of the photodetachment cross section for energies near threshold.93 Valid for final
states with two particles whose longest range interaction is the centrifugal barrier,
his result gives the variation of the cross section with energy, but not its absolute
magnitude. The law is given by,
o(E) (AE)I+1/2 or o(E)~kl+1/2, (1.14)
where o is the cross section, AE is the energy of the departing electron (the energy
above the minimum required for reaction to occur), and k = (2mAE)/j / h and I are the
linear and orbital angular momenta of the outgoing electron. The essence of the
Wigner derivation is that, at threshold, the dynamics are solely governed by the


100
80
60
40
20
0
Upon CAD Low activation energy
239.2
(15-crown-5)H30+
(15-crown-5)H+
22
1. 1
17.1
113.7
210.3 306.9
Mass-to-charge ratio
403.4
500.0
192


215
indicated that the proton coordinates to only two oxygens in the crown ether. The
protonation of polyethers, glymes and crown ethers in the gas phase has been
studied by Kebarle,40who observed that the protonated base, BH+, reacts with water
to form (BH+)H20. He interpreted the behavior of H+ with the crown ether as
analogous to H7(H20)n complexation, in which a cluster of H20's surrounds the
proton. According to Kebarle, the proton coordinates to two oxygens in the crown
ether (two strong hydrogen bonds), and an H20 molecule contributes to the
stabilization of the complex via its permanent dipole and more weakly through its
polarizability (one partial hydrogen bond).
Based on the calculations of Kollman192 et al., the complexation enthalpies
including polarizability of H30+ with crown ethers are about 20 kcal/mol higher than
the values obtained without including polarizability. They found one strong and two
weaker hydrogen bonds in this complex with C4 symmetry. Calculated results on
the H30+ complex do not agree with experimental values.43,202 They suspected the
experimental values corresponded to the complex H20 H+-ether and not
H30+-ether. Further calculations203 using single point 6-31G basis sets based on
the H20 H+-etherstructure were found to be in better agreement with experimental
values. Hence, it is most likely that the H30+ complex is essentially H20 H+-ether
and not H30+ ether.


80
60
40
20
0
AV=20 volts
261.1
(12-crown-4)Rb+
263.0
437.3
(12-crown-4)2Rb+
439.2
ll
.j i
, J
v Li I
l
30.0
140.0
230.0
320.0 410.0
500
Mass-to-charge ratio
173


Figure 3.6. Mass spectrum of (H20)Br, m/z = 97, 99.


146
technique. The procedure is repeated with varying rf pulse lengths to determine the
minimum energy for appearance of the successor ions. Unlike chemical processes
in liquids, gas-phase ion/molecule reactions usually do not have activation barriers.
Thus, the threshold for the dissociation process may be attributed entirely to the
BDE, provided that there is proper control of a number of parameters, as described
below.
The CAD method has been used to observe several alkali-metal/crown-ether
complexes,88-89 and in Eyler's group88 approximate bond dissociation energies were
obtained for the 18-crown-6 complexes of K\ Cs+ and Ba+ using laser desorption
(LD) to form the ions. Subsequent ab initio calculations by Thompson et a/.173 using
the RHF/3-21G, 6-31G, and MP2/6-31G basis sets produced results as much as 30
kcal/mol (1.3 eV) higher than the measured values. The discrepancy can be
attributed in part to the LD process, because shot-to-shot instability of the laser
power results in poor control of precursor ion intensities. Another problem was the
pressure of the argon collision gas (1-2 10-6 torr), which must be kept low (10 8
torr range) to prevent collisions during excitation, as well as multiple collisions
during the CAD delay time.
In this study the BDE's of complexes of 18-C-6, 15-C-5, and 12-C-4 with Na+,
K+, Rb+, Cs+ and H30+ (Figure 1.9) have been measured by CAD in FTICR-MS,
using electrospray ionization (ESI) to produce the gas-phase ions. This method
affords a very reproducible and direct way to form ion/neutral complexes or/and
solvent attached ion complexes from solution. In addition, compared to other


2
ion storage, and many probes of ion structure.12 During the 1980's, a wide variety
of applications13'19 were reported, including MS/MS experiments, laser desorption,
fast atom bombardment (FAB) and high pressure external sources with quadrupole
ion guides, gas chromatography mass spectrometry (GC/MS), external sources
utilizing electrostatic focussing which lower ultimate pressure at the analyzer,
stored waveform inverse Fourier transform (SWIFT) excitation and electrospray
ionization (ESI) sources. Both ESI20 and matrix assisted laser desorption ionization
(MALDI)21 sources began to be used in the early 1990s to increase the range of
masses which can be studied with FTICR mass spectrometers. In 1992, Amster
and Marshall22 demonstrated quadrupolar axialization to improve resolution and
increase sensitivity. In 1993, Smith's group23 developed Time Resolved Ion
Correlation (TRIC) to detect highly-charged individual ions with masses higher than
5,000,000 daltons.
Basic Concepts
Ion cyclotron motion
The ions in a magnetic field undergo circular orbital motion, at a cyclotron
frequency given by,24 27
) = qB/m = M/r (1.1)
where q is the ion charge, B is the magnetic field, m is the mass of the ion, v is the
ion velocity and r is the orbit of the cyclotron motion. In a uniform magnetic field (no
electric field), the mass-to-charge ratio, m/q is linearly related to the observed


Figure 4.10. Mass spectrum of (12-C-4)Rb+ after CAD at AV=70 volts. The two isotopes of Rb+ were the only
successor ions observed.


138
I + CH3OH HI + CH30 AH = 32.6 kcal/mol (3.15)
The reaction barrier must be at least as high as the endothermicity of these
reactions. It is probable that the trend of the barrier heights may be similar to the
AH trend, i.e., much larger for Br and I than for F or Cl. It is generally true that, for
very endothermic reactions, the structure of the transition state is similar to that of
the products. However, the relative bond lengths (discussed below) indicate that
the neutral complex is very similar to the reactants. Thus, the complex must be
located very low on the rising surface, and the corresponding AE must be much less
than the overall barrier height. Given this argument, the value of AEs for Br and I
complexes should be greater than those for Cl complexes, but only by 1-2 kcal/mol.
As explained in the Introduction, there is some uncertainty in the BE values
obtained by equilibrium methods, although error bars are not given by the cited
authors. Examination of their data indicates a random uncertainty of about 1-2
kcal/mol in the reported BE's. Furthermore, binding energies determined by
equilibrium methods are subject to systematic error in the statistical mechanical
estimation of AS, which is based on the vibrational frequency of the H-X bond in
ROH-X'. As the halide becomes heavier, the vibrational frequency of the ROH-X'
bond decreases. It is possible that the vibrational frequency has not been
estimated correctly by previous authors. This would produce a systematic error in
the entropy change, which would subsequently affect the reported BE.
Results of ab initio calculations support the relative bond length predictions,


161
This was the situation in the threshold region of the experimental data. If the
energies are very low (kbT < E0< 1 eV), the distribution function becomes broad
relative to the center-of-mass ion energy. This corresponds to a breakdown of the
stationary target approximation at energies where the velocities of the Ar gas
molecules are comparable to or larger than the ion velocity. At the lowest ion
energies, there is a thermal distribution of the translational energies, and the
interaction energies are determined primarily by the thermal velocities of the target
molecules.
Deconvolution
The form of the convolution for Doppler broadening (eq. 4.6) is such that the
Langevin-Gioumousis-Stevenson (LGS) cross section187, oLGS(E) is directly
proportional to E'1/2 and is unchanged by the broadening effects. The simplest form
of the model threshold law97 is given by equation 1.19. The threshold energy
determined in this manner corresponds to dissociation of room-temperature (298
K) precursor ions to room-temperature successors, i.e. BDE298 (cf., Fig.4.2), and
does not account for the internal energies of precursor ion and probability of
dissociation after the CAD delay time. To correct for these effects the CRUNCH
program uses the relation178
ks -
o0 Igi P(E,E,.,7)
(E*Ei*Erof-ET-)n
E
(4.8)


532 nm
3+
1/
Nd : YAG Laser
Console
f
Doubler
Dye Laser
a
4
g
n.
TV
(3)


Figure 3.5. Mass spectrum of OH', m/z =17.


TABLE OF CONTENTS
ACKNOWLEDGMENTS ii
TABLE OF CONTENTS iii
LIST OF TABLES vi
LIST OF FIGURES vii
ABSTRACT xiii
CHAPTERS
1. INTRODUCTION 1
Fourier Transform Ion Cyclotron Resonance Mass
Spectrometry (FTICR-MS) 1
History 1
Basic Concepts 2
Ion cyclotron motion 2
Ion excitation 6
Ion detection 11
Signal acquisition 16
Applications 20
Purpose of This Work 20
Solvated Anions 24
Host-guest Complexes 26
Methodology of Determination of Bond Dissociation Energies 37
Photodetachment 37
CAD 45
Energy behavior of reaction cross section 45
On-resonance CAD 47
Doppler broadening 49
Advantages, Disadvantages, and Comparisons to Other
Techniques 50


37
measurements of BDE's would provide better understanding for making use of the
ion discrimination properties of crown compounds based on ion-selective membrane
electrodes,86 for the study of physiological ion-transport processes in biological
membranes, for the investigation of receptor and enzyme interactions as well as for
the salt balance and for metabolic processes of the living organism.87
Investigations using CAD to study crown ether complexes with alkali metal
cations have been previously reported.8889 However, quantitative information about
the BDE's of crown ether complexes has only been reported by the Eyler group.89
In the previous study, approximate binding energies of metal/crown ethers formed
using the laser desorption technique were reported. In these studies, we have
carefully measured quantitative BDE's of crown ethers with either hydronium or
metal ions attached, formed from an external electrospray ionization (ESI) source,
using a different approach to determine the dissociation threshold. The details of
the measurements are described in Chapter 4.
Methodology of Determination of Bond Dissociation Energies
Photodetachment
Electron photodetachment from gas phase molecular anions has proved to
be a remarkable tool in the study of both anions and their neutral photoproducts.
A variety of thermochemical and spectroscopic constants can be measured for
systems that would be difficult to study by other methods. Among the techniques


42
photodetachment spectrum, the observed fractional signal decrease must be related
to the photodetachment cross section. The probability of photodetaching an
electron from a negative ion is described by,90
P(A) = 1- = 1- exp( A x t)
(1.10)
Where N is the ion signal with irradiation, N0 is the ion signal without irradiation, t is
the duration of the laser beam pulse, and kp is the rate constant for
photodetachment given by
kp(A) = Jfo(A)-p(A )d\
(1.11)
f is a unitless factor related to geometric overlap between the ion cloud and the
photon beam, o(\) is the cross section for photodetachment in cm2, and p is the
photon flux (a function of wavelength) in photons cnr2 s'1. Assuming a(\) is constant
over the narrow laser bandwidth and t is kept constant, substituting eq 1.11 into eq
1.10, rearranging, and taking logarithms yields an expression for the cross section
in the terms of the quantities measured during a photodetachment experiment:91
N = N0 exp(- o(A) t\f p(A) d\)
(1.12)
tJf-p(A)dA
(1.13)


Figure 4.9. Mass spectrum of (12-C-4)Rb+ at AV=70 volts. Peak corresponding to (12-C-4)Rb+ ion was isolated before
excitation.


16
theoretical estimate of the magnitude of the signal produced on the plates by an ion,
the Reciprocity Theorem is applied.35 The cyclotron radius, Rc, is given by,
where E is the excitation electric field, t is the duration of the resonant excitation
pulse, and B is the magnetic field. The Reciprocity Theorem states that the
electrostatic potential created at the position of an ion by 1 volt on the detection
plate is directly proportional to the image current induced on the plate by the ion.
Thus the predicted FTICR signal strengths can be calculated by a numerical
solution of Laplace's equation. The amplitude of the signal per ion only depends on
the ratio between the radius of cyclotron orbit and the distance between the receiver
plates.
All ions in the cell are excited and detected simultaneously as a time-domain
signal. The signal is converted into voltage, amplified, digitized and stored in
computer. The time-domain spectrum is a composed of superimposed sine waves:
in order to recover the frequency domain spectrum, the complex time-domain
spectrum is subjected to a fast Fourier transform (FFT) algorithm.
Signal acquisition
For one thousand ions, the r.f. signal at a preamplifier input is on the order
of a few hundred microvolts. A low noise electronic design for amplification is
required. There are two types of data acquisition used in our ICR instruments, (a)


224
18. Marshall, A. G.; Wang, T.-C. L; Ricca, T. L. J. Am. Chem. Soc. 1985, 107,
7893.
19. Hunt, D. F.; Shabanowitz, J.; Yates, J. R., Ill; Zhu, N.-Z.; Russell, D. H.;
Castro, M. E. Proc. Natl. Acad. Sci. USA. 1987, 84, 620.
20. Henry, K. D.; Williams, E. R.; Wang, B. H.; McLafferty, F. W.; Shabnowitz, J.;
Hunt, D. F. Proc. Natl. Acad. Sci. USA. 1989, 86, 9075.
21. Castoro, J. A.; Koster, C.; Wilkins, C. L. Rapid Commun. in Mass Spec.
1992, 6, 239.
22. Schweikhard, L; Guan, Z.; Marshall, A. G. J. Mass Spec. Ion Proc. 1992,
120, 71.
23. Bruce, J. E.; Cheng, X.; Bakhiar, R.; J. Am. Chem. Soc. 1994, 116, 7839.
24. Buchanan, M. V. Fourier Transform Mass Spectrometry: Evolution,
innovation, and Applications. Buchanan, M. V. Ed.; ACS Symposium series, 359,
Washington, D.C. 1987
25. Dunbar, R. C. Analitical Applications of Fourier Transform Ion Cyclotron
Resonance Mass Spectrometry, Asamoto, B. Ed.; VCH publishers, Inc.: New
York, 1991.
26. Baykut, G.; Eyler, J. R., Trends in Anal. Chem., 1986, 5, 44.
27. Marshall, A. G.; Grosshans, P. B., Anal. Chem., 1991, 63, 215A.
28. Wang, M.; Marshall, A. G. Int. J. Mass Spec. Ion Proc., 1990, 100, 323.
29. Grosshans, P. B.; Marshall, A. G. Int. J. Mass Spec. Ion Proc., 1990, 100,
346.
30. Marshall, A. G.; Wang, T.-C. L; Ricca, T. L. Chem. Phys. Lett. 1984, 105,
233.
31. Tomlinson, B. L; Hill, H. D. W. J. Chem. Phys. 1973, 59, 1775.
32. Comisarow, M. B.; Melka, J. D. Anal. Chem. 1979, 51, 2198.
33. Comisarow, M. B. J. Chem. Phys. 1978, 69, 4097.


Transfer region
ESI source
"4
o


D [D In D
Podand
o
o o
D D
w
Coronand
pio^
bvJdW
Cryptand
where D = O or S, B = N
-o o
0 I Crown Ether (D = O)


L
ICR
Cell
Sweep
Generator
Disk
Plotter
Display


AV=20 volts
Mass-to-charge ratio
175


In (N0/N)
E/kcalmol"1


216
Conclusion
For 12-crown-4, 15-crown-5, and 18-crown-6 the trend of the alkali-metal
BDE's is Na+ > K+ > Rb+ > Cs +. The BDE's for Na\ K +, Rb+ and Cs+ complexes
follow the order 12-C-4 > 15-C-5 > 18-C-6. The charge density on the metal ion is
a major factor in the observed trends. For the crown ether complexes with H30+,
loss of H20 and ion fragmentation were observed instead of dissociation of H30+.
Although the BDE's for LFT complexes could not be measured, the observations
indicate that H+ binds very strongly. The solvation energies for (15-C-5)H+ and (18-
C-6)H+ are much smaller than the metal ion BDE's, indicating that H+ is bound
strongly to the crown ether cavity, with weak solvation of the entire complex.
This investigation has shown that CAD with FTICR-MS can be used
effectively for measurements of bond dissociation energies. Quantitative results
can be obtained if experimental conditions are properly controlled, and the data
analysis accounts for systematic errors due to the internal energy and kinetic energy
distribution of precursor ion, thermal spread of the collision argon gas and the
lifetime of dissociation.
Although the CAD technique has been widely applied to measure BDEs in
several groups using different types of instruments, such as guided-ion beam and
quadrupole mass spectrometers, this study represents the first time it has been
applied in an ICR mass spectrometer. Different models have been applied to obtain
quantitative BDEs using different instruments; we have successfully applied the


100
80
60
40
20
0
96.913
(H20)Br
98.901
.3.5
16.8 22.2 32.5 61.1 500.5
Mass-to-charge ratio
106


CHAPTER 4
BOND DISSOCIATION ENERGIES OF ALKALI AND HYDRONIUM CATIONS
BOUND TO CROWN ETHERS DETERMINED BY ON-RESONANCE
COLLISIONALLY-ACTIVATED DISSOCIATION TECHNIQUES
Background
Crown ethers constitute an important group of compounds and are especially
well known for their ability to form complexes with alkali-metal cations, which do not
bind favorably to other common ligands. Previous investigations of alkali-
metal/crown-ether complexes have provided information about the intrinsic nature
of the binding interactions and the size-selectivity of complexation.78'89,160'173 In
solution the order of the complexation equilibrium constants is generally K+ >
Rb+ > Cs+ > Na + > Li+ in solvents such as water or methanol. 166'171 However,
competitive equilibrium studies in the gas phase have indicated that Na+ binds to
crown ethers more strongly than k+.78'89160'161 The stability difference is attributed
to the greater solvation enthalpy of Na+ compared to K+. This behavior
demonstrates the importance of gas-phase studies to determine the fundamental
nature of the ion-molecule interaction in the absence of solvent.
Mass spectrometry (MS) can be used to obtain data on the binding of gas-
phase complexes, if there is a means to add a controlled amount of energy to
dissociate the complex. Photodissociation has been used successfully in several
144


230
121. Yeh, C. S.¡ Willey, K. F.; Robbins, D. L; Pilgrim, J. S.; Duncan, M. A. Chem.
Phys. Lett. 1992, 196, 233.
122. Lessen, D. E.; Asher, R. L.; Brucat, P. J. J. Chem. Phys. 1990, 93, 6102.
123. Bradforth, S. E.; Arnold, D. W.; Metz, R. B.; Weaver, A.; Neumark, D. E. J.
Phys. Chem. 1991, 95, 8066.
124. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1985, 107, 766.
125. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1988, 110, 7604.
126. Caldwell, G.; Rozeboom, M. D.; Kiplinger J. P.; Bartmess, J. E. J. Am.
Chem. Soc. 1984, 106, 4660.
127. Posey, L. A.; Campagnola, P. J.; Johnson, M. A. J. Chem. Phys. 1989, 91,
6536.
128. Dallestka, N. F.; Honma, K.; Sunderlin, L. S.; Armentrout, P. B. J. Am.
Chem. Soc. 1994, 116, 3519.
129. Yamabe, S.; Hirao K. Chem. Phys. Lett. 1981, 84, 598.
130. Hiraoka, K.; Yamabe, S. Int. J. Mass Spec. Ion Proc. 1991, 109, 133.
131. (a) Yamdagni, R.; Payzant, J. D.; Kebarle, P. Can. J. Chem. 1973, 51,
2507; (b) Arshadi, M.; Yamdagni, R.; Kebarle, P. J. Phys. Chem. 1970, 74, 1475;
132. Yamdagni, R.; Kebarle, P. J. Am. Chem. Soc. 1972, 94,2940.
133. (a) Caldwell, G.; Kebarle, P. J. Am. Chem. Soc. 1984, 106, 967; (b)
Caldwell, G.; Kebarle, P. J. Am. Chem. Soc. 1985, 107, 2612.
134. (a) Meot-Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 3858; (b) Meot-
Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 3854.
135. Meot-Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 7604.
136. (a) Meot-Ner (Mautner), M. J. Am. Chem. Soc. 1986, 108, 7525; (b) Meot-
Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 3071.
137.(a) Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1983, 105, 2944.; (b)
Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1984, 106, 517.; (c) Larson, J.


E/kcalmol'
In (N/N)
o

o
o
o
o
b
Lx
k>
Ol
cn
cn
cn
III
_IHO HO


Table 2.1. Laser dyes used for photodetachment experiments and their corresponding wavelength range before and
after frequency doubling.


CHAPTER 3
DETERMINATION OF BINDING ENERGIES OF SINGLY-SOLVATED HALIDE
IONS USING THE PHOTODETACHMENT TECHNIQUE
Background
Fundamental studies of the thermodynamics of solvation are important for
several reasons. First there is the chemistry of the solvation process itself. A free
molecule or ion becomes surrounded by solvent molecules in the bulk medium.
This results in changes in the interactions between solvent molecules, as well as the
formation of new solute-solvent interactions, including hydrogen bonds in protic
systems. Solvation is often the key factor in relative solubility effects, which are
essential in many types of chemical separations. Also of practical significance is the
role of solvation in chemical reactions. Since many reactions take place in solution,
the solvent often has a pronounced effect on the reaction rate and position of
equilibrium. Although the optimum solvent for a particular application may be
identified by conducting a series of reactions, it would be more expedient to predict
such effects from knowledge of the solvation of the reacting species. However,
because of the accompanying changes in solvent structure and the solvation of
counterions, it is often difficult to obtain meaningful data from measurements of the
88


acetonitrile with fluoride, chloride, bromide and iodide ions have been measured.
The effect of size and polarizibility of the solvents, and comparisons of the solvation
energies determined in this work to some reported in earlier thermal-equilibrium
studies are discussed.
Collisionally activated dissociation (CAD) processes in mass spectrometry
can provide quantitative bond dissociation energy (BDE) estimates if the interaction
energy is precisely controlled during collisions. This work has concentrated on
improving the accuracy of measuring the thresholds and providing more precise
quantitative measurements of host-guest complex interactions in the gas phase by
accounting for the effects of multiple ion-molecule collisions, internal energy of
precursor ions, thermal spread of energies of the collision gas and the dissociation
lifetime of the intermediates. The use of an electrospray ionization (ESI) source for
soft ionization that produces intact, multiple charged and solvated gas-phase ions
directly from various compounds in solution has been critical for this work. We have
measured the binding energies of hydronium and alkali metal ions to the crown
ethers (12-crown-4,15-crown-5, and 18-crown-6). The effects of metal size, cavity
size and polarizibility of different crown ethers are discussed.
XIV


100
80
60 -
40 -
20 -
0
30.0
AV=20 volts
177.1
(12-crown-4)H+
209.1
a JL
353.2
(12-crown-4)2H+
371.2
(12-crown-4)2H3G
124.0
218.0 312.0
Mass-to-charge ratio
406.0
500.0
177


17
The first produces broadband spectra (all masses). In this mode the signal from a
preamplifier is sampled directly by the analog to digital (A/D) converter at a rate
which must be faster than twice the highest frequency for the lowest mass to be
observed (Nyquist criterion). Spectra are signal-averaged by summing the time-
domain transients from each repetition in the computer's memory before performing
the FT; (b) The second type of data acquisition (heterodyne mode) leads to
narrowband spectra.34'35
As illustrated in Figure 1.6, the heterodyne technique reduces the digitization
rate requirements for ion signals and is used to obtain higher mass resolution over
a narrow mass range. Given a time-domain spectrum consisting of N points, where
N is between 16K and 128K, it is common to zero-fill at least once (add N zeros to
the end of the data set), apodize, and obtain the magnitude-mode FT, where the
magnitude-mode is defined as [(real)2 + (imaginary)2]1'2, with a FFT algorithm. Zero
filling has been shown to increase the S/N ratio and increase the displayed
resolution of the spectrum.
Since frequency is a physical quantity which can be precisely measured, the
exact mass of an ion can be determined very accurately in the FT-ICR technique,
since it is directly related to the measured frequency. Typically, low parts-per-million
accuracy can be achieved with an internal mass calibration. A high degree of mass
accuracy can be maintained due to the stable magnetic field. The main features of
FTICR include ultra-high mass resolution, accurate and stable mass calibration,
multiple stages of collisionally activated dissociation (MS/MS), extremely long ion


26
These studies have led to an understanding of the stability of hydrogen bonding and
structural variation in these compounds. Such work also provides a better
understanding of the characteristic properties of bulk solution such as ions solvated
with solvent clusters.42,59 Among solvated ions the methanol solvated anions are
difficult to generate and understanding of their properties is still subject to argument.
Also there is still disagreement about the binding energies (BE) of solvated anions
using different measurements.41-45 Therefore, alternate measurements of the BE
using threshold determinations from photodetachment experiments can give more
accurate results. Also, since the ICR ion trap can easily form solvated ions using
either ion-molecule reactions or electrospray ionization (ESI), it becomes possible
to measure BE using photodetachment techniques in the ICR instrument.
The BE measurements of alcohol solvated halides, whose structures are
illustrated in Fig. 1.7, using the photodetachment technique with a FTICR mass
spectrometer are described in Chapter 3.
Flost-guest complexes
Polyether macrocycles have been known for decades.65-67 The discovery that
certain macrocycles can bind alkali metal and alkaline earth cations gave impetus
to the field now known as host-guest chemistry.68 Today, a distinction is made
between the classical ring oligoethers (crown ethers) and monocyclic commands,
oligocyclic spherical cryptands and the acyclic podands with respect to topological
aspects.69 Multidentate monocyclic ligands with any type of donor atoms are called
coronands (crown compounds), while the term crown ether should be reserved for


234
186. Khan, F. A.; Clemmer, D. E.; Schultz, R. H.; Armentrout, P. B. J. Phys.
Chem. 1993, 97, 7978.
187. Gioumousis, G.; Stevenson, D. P. J. Chem. Phys. 1958, 29, 294.
188. Faulk, J. D.; Dunbar, R. C.; Lifshitz, C. J. Am. Chem. Soc. 1990, 112, 7893.
189. Fluang, F.-S.; Dunbar, R. C. J. Am. Chem. Soc. 1990, 112, 8167.
190. So, H. Y.; Dunbar, R. C. J. Am. Chem. Soc. 1988, 110, 3080.
191. Kim, M. S. Int. J. Mass Spect. and Ion Physics, 1983, 50, 189.
192. Parzen, E. Modern Probability Theory and It's Applications, Wiley: New
York. 1960, p.251.
193. Wipff, G.; Weiner, P.; Kollman, P. J. Am. Chem. Soc. 1982, 104, 3249.
194. Dang, L. X.; Kollman, P. A. J. Am. Chem. Soc. 1990, 112, 5716.
195. Thompson, M. A. J. Phys. Chem. 1995, 98, 4794.
196. Thompson, M. A.; Glendening, E. D.; Feller, D. J. Phys. Chem. 1994, 98,
10465.
197. Howard, A. E.; Singh, U. C.; Billeter, M.; Kollman, P. A. J. Am. Chem. Soc.
1988, 110, 6984.
198. Dailey, N. K. Synthetic Multidentate Macrocyclic Compounds, Izatt, R.M.;
Christensen, J.J., Eds. Academic Press: New York 1978. p 207.
199. Shannon, R.D. Acta Crystallogr., Sect. A: Found Crystallogr. 1976, 32, 751.
200. Christinsen, J.; Eatough, D. J.; Izatt, R. M. Chem. Rev. 1974, 74, 251.
201. (a) Schori, E.; Jagur-Grodzinski, J. J. Am. Chem. Soc. 1972, 94, 7957. (b)
Nae, N.; Jagur-Grodzinski, J. J. Am. Chem. Soc. 1972, 94, 7957.
202. (a) Lau, Y. K.; Ikuta, S.; Kebarle, P. J. Am. Chem. Soc. 1982, 104, 1462. (b)
Lau, Y. K.; Saluja, P. P. S.; Kebarle, P. J. Am. Chem. Soc. 1980, 102, 7429.
203. Singh, U. C.; Kollman, P. A. QCEP 1982, 2, 17.


Figure 4.15. Mass spectrum of (15-C-5)H30+ after CAD at high activation energy and AV=20 volts,
corresponding to (15-C-5)H\ (C2H40)nH+ (n = 2 and 3) were the succesor ions observed.
Peaks


140
GTO basis set.129 The electronic distribution indicates that CH3CN is described by
H36+-C6'-Cs+=N, rather than CH 3C +=N, with the methyl hydrogens having more
positive character than the center carbon. According to their calculations, the most
stable structure (CH3CN)F has Cs symmetry with the F' bound to one of the three
hydrogens in the methyl group via a linear F-H-C bond with F-H and H-C bond
distances of 1.70 and 1.14 , respectively. The bond lengths in (CH3CN)Cr are
2.70-3.00 and 1.09 for the Cl-H and H-C bonds, respectively. Recently, similar
results were obtained by Riveros et a/.,141 who optimized the geometries for solvated
anions at the MP4(SDTQ) level. In (CH3CN)Br the Br is bonded collectively to the
methyl hydrogens, with Br-H and H-C bond distances of 3.27 and 1.1 ,
respectively. Thus, all the calculated results for (CH3CN)X" complexes indicate a
loose interaction of X' with the solvent. The photodetachment process should,
therefore, be similar to the situation depicted in Figure 3.16. The endothermicity of
the hydrogen abstraction reaction of Br with CH3CN:
Br + CH3CN HBr + CH2CN AH = 4.8 kcal/mol (3.16)
is much less than that for Br + CH3OH (eq 3.14). Considering the very loose
interaction in the neutral (CH3CN)Br complex, the AE is expected to be very small
for this system. This is supported by the close agreement of the BE with previous
results.
Although the threshold data in Table 3.2 were obtained at room temperature,


induced
image current
induced
image current
cn


Figure 4.6. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peaks corresponding to (15-C-5)H30+, (15-C-5)2H+, and
(15-C-5)2H30+ ions were observed.


ee
va
va
7J
O
sa
o
CO
II
00
O


AA
cn
O
I
cn
ro
O
ii
ro
O
va
Ol
O

Ol
ro
IIIIIIMIIMIIIII
mi min ii ii it I


74
description of the ESI mechanism.115 According to their research, the ion
suspended in solution evaporates into the gas phase when the shrinking droplet
reaches the "Rayleigh limit" where the energy from repulsion of the charge on the
droplet surface exceeds the energy of the surface tension which holds the droplet
together. The resulting "coulombic explosion" tears the droplet apart and produces
even smaller droplets with an ehigher charge-to-mass ratio. It is from these smaller
droplets that ions actually "evaporate" from the liquid phase into the gas phase.
After the charged droplets enter the capillary, the evaporation process is
controlled through the application of electrical fields and a heated "counter-current"
drying gas, which helps to evaporate the droplets. The ESI chamber is surrounded
by a metal screen so that the spray itself can be seen during the ionization process
(Figure 2.8). This region is always maintained at atmospheric pressure. Inside the
spray chamber there are two electrodes, excluding the needle itself which is
normally maintained at ground potential to protect the user from the possibility of
electric shock. The largest of these electrodes is the transparent cylindrical
electrode (V^,), a wire screen which surrounds the bulk of the spray chamber. The
final electrode in the atmospheric region of the source is the capillary (0.5 mm i.d.)
entrance or V^, which is held at ca. 4kV, in order to form charged droplets between
the needle and the capillary. This electrode is fabricated by coating the face and
side of the quartz dielectric capillary with metal (gold or platinum). The exit end is
also coated with metal so that the capillary entrance and exit can be used
independently as electrodes to help guide the ions to and out of the capillary. On


150
vibrational modes, the kinetic energy distribution of the precursor ions, the Doppler
broadening mentioned above, and the lifetime of the dissociation process. The
vibrational frequencies for use in the CRUNCH program were obtained using the
HyperChem program for ab initio and ZNDO empirical calculations. All the
calculated vibrational frequencies were scaled to reasonable values by comparing
the calculated and experimental185 results on both cyclohexane and 1,4-dioxane
compounds. The best scaling factor found in this case is {0.89+0.745xsin[(n/2)(f¡-
fm¡ny(fmax-fmin)]}\ where f¡, fmin, and fmax are the frequencies of the ith mode, minimum
and maximum vibrational frequencies obtained from the HyperChem calculation.
The results are shown in Table 4.1. In addition to ground state vibrational
frequencies which were used to estimate the effect of vibrational energy in
CRUNCH program, the vibrational frequencies of the activated complexes were also
obtained using a single-point calculation, and the normal mode which corresponds
to the metal-crown ether stretch mode was removed from the transition state to
account for the dissociation lifetime, again using the CRUNCH program.
Experimental
Ion Formation
All crown ethers and alkali metal salts were obtained from Fisher Scientific,
Inc., and were used without further purification. Solutions were prepared by


Figure 2.10. Infinity cell with three "side-kick" plates: (a) PVX1, (b) PVX3, (c) PVX4 plates, (d) the macor plate which
is attached to trapping plates to hold the "side-kick" plates and trapping plates, (e) the rf-shimmed trapping plate, (f)
the Infinity cell with four brass plates for excition and detection.


On Resonance CAD of (15C-5)H30+
i 1 1 1 1 1 1
0.00 0.80 1.60 2.40 3.20
E,m/eV
196


Figure 2.3. Laser set-up for photodetachment experiment (1) Apple II microcomputer controls "charge" and "fire"
commands of YAG laser, (2) Console sends external TTL pulse to trigger the laser, (3) Power control from console
to the cell; (a) Quartz prism, (b) Laser quartz windows, (c) ICR cell, (d) 2 T magnet, (e) Apple II computer, (f) Console,
(g) Photodiode.


corresponding to (12-C-4)H+ ion was isolated before
excitation
199
4.18. Mass spectrum of (12-C-4)H+ after CAD at AV=70 volts. Peaks
corresponding to (C2H40)nH+ (n = 2 and 3) were ions
observed 201
4.19. Appearance curve of (15-C-5)Rb+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 203
4.20. Appearance curve of (12-C-4)K+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 205
4.21. Appearance curve of (12-C-4)Cs+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 207
XII


CHAPTER 2
INSTRUMENTATION
Introduction
The complex electronics instrumentation used in FTICR mass spectrometry
is presented in Figure 2.1 as a simplified block diagram. Essential equipment for
a Fourier transform ion cyclotron resonance mass spectrometer consists of a
superconducting magnet, a high vacuum system, a trapping cell, and a console to
detect ions, store signal and process acquired signal. Ionization sources may be
categorized as internal or external. The internal ionization source is usually located
inside the magnetic field (at the entrance of a trapping cell), produces better signal-
to-noise (S/N) ratio, and does not require a differential pumping set-up to maintain
low base pressure. The external ionization source is located outside the magnetic
field and operates at higher pressures (ca. 10~5 torr for El and MALDI, 760 torr for
ESI and glow discharge (GD)). It requires several stages of differential pumping
before ions enter the trapping cell region and also electrostatic or rf potential lenses
to transfer the ions into the cell. The development of external sources has allowed
the analyzer cell region to remain clean and at low pressure; therefore, the
resolution is improved. Also, since the external source region can be isolated from
the cell region, it is more flexible to switch between different ionization techniques
53


129
in Table 3.1. Table 3.2 presents the thresholds (including a and ET) and binding
energies for all complexes studied in this work. Results of previous investigations
are also included.
This interpretation of the photodetachment threshold assumes that the
vertical transition proceeds to a repulsive state and not to a neutral bound complex.
Excitation to a bound state would result in the reaction,
SX'-V SX + e (3.12)
instead of eq 3.2. For these solvated anions, it is expected that dissociation to S
plus X (eq 3.2) occurs, as supported by several arguments. Based on Brauman's
experimental and calculated results,91153 the preferred reaction is expected to
depend on the Franck-Condon overlap of the ground state vibrational wavefunction
of the solvated anion with that of the neutral complex at the same geometry.
Considering S to be an alcohol, ROH, interacting with neutral X via the OFI group
(RO-H-X), the position of H between the O and X nuclei depends on the proton
affinities (PA) of RO' and X. If X' has a lower PA than RO', then the FI is closer to
the oxygen (RO'-FT-X), so that the equilibrium internuclear distance for the
interacting neutrals would be similar to that in the solvated ion. Thus, complete
dissociation of ROFI from X is feasible, and a vertical transition may be expected to
follow eq 3.2. On the other hand, if RO' has a lower PA than X', the FI is closer to
X (RO'-FT-X'). The Franck-Condon factor for eq 3.2 would be poor, and the


13


Figure 1.6 Heterodyne mode: The signal mixes with a reference to give a sum (higher frequency) and difference (lower
frequency) signal. The higher frequency (low mass) signal is then filtered. Fast FT gives the low frequency signal which
is the difference.


.X
en


155
to the successor ions and neutral crown ether molecules if the activation energy
exceeded the threshold energy. The precursor and successor ions were
subsequently excited and detected simultaneously using broad-band chirp
excitation with a mass-detection range from 21 to 2000 m/z. Fifty 32K-point time-
domain transients were signal-averaged for all ions. For each rf pulse width, the
intensity measurements for both the precursor and successor ions were repeated
five times, and the respective means and standard deviations were calculated.
Data Analysis
Ion Appearance Curve178
The experimental total rate coefficient for fragmentation, ktot, is related to the
measured intensities, lp and ls, of the precursor and successor ions by the
equation,178
'p VP + z/s) exP( ktot n 0 (4.2)
where n is the number density of the argon collision gas, and t is the effective
reaction time. The rate coefficient for each successor ion, K. is subsequently given
by,


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
DETERMINATION OF BOND DISSOCIATION ENERGIES USING FOURIER
TRANSFORM ION CYCLOTRON RESONANCE (FT- ICR)
MASS SPECTROMETRY
By
Yarjing Yang
1996
Chairperson: John R. Eyler
Major Department: Chemistry
The objective of this work was to use Fourier transform ion cyclotron
resonance mass spectrometry (FTICR-MS) to determine bond dissociation energies
(BDE's) of important species such as solvated anions and host-guest complexes
using two different techniques: photodetachment and on-resonance collisionally
activated dissociation (CAD).
Photodetachment has been widely used in measuring the electron affinities
of various anions. Since the electron affinities of the halides are well known, this
part of the work has utilized photodetachment to measure the halides' binding
energies with alcohol solvents in the gas phase. An Nd3+:YAG-pumped dye laser
with frequency doubler was used to obtain the uv-visible irradiation needed for this
work. The binding energies of methanol, ethanol, iso- and normal-propanol, and
xiii


5)2H30+ ions were observed
175
4.7. Mass spectrum of (12-C-4)H30+ at AV=20 volts. Peaks
corresponding to (12-C-4)H+, (12-C-4)2H+, and (12-C-4)2H30 +
ions were observed. The peak intensity of (12-C-4)2H+ is higher
than that of (12-C-4)H+ 177
4.8. Mass spectrum of (12-C-4)H30+ at AV=70 volts. Peaks
corresponding to (12-C-4)H+ and (12-C-4)2H+ ions were
observed. The peak intensity of (12-C-4)2H+ is lower than that
of (12-C-4)H+ 179
4.9. Mass spectrum of (12-C-4)Rb+ at AV=70 volts. Peak
corresponding to (12-C-4)Rb+ ion was isolated before
excitation 182
4.10. Mass spectrum of (12-C-4)Rb+ after CAD at AV=70 volts. The
two isotopes of Rb+ were the only successor ions observed . 184
4.11. Mass spectrum of (18-C-6)H30+ at AV=20 volts. Peak
corresponding to (18-C-6)H30+ ion was isolated before
excitation 186
4.12. Mass spectrum of (18-C-6)H30+ after CAD atAV=20 volts. Peak
corresponding to (18-C-6)H+ ion was the only successor ion
observed 188
4.13. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peak
corresponding to (15-C-5)H30+ ion was isolated before
excitation 190
4.14. Mass spectrum of (15-C-5)H30+ after CAD at low activation
energy and AV=20 volts. Peak corresponding to (15-C-5)H+ ion
was the only successor ion observed 192
4.15. Mass spectrum of (15-C-5)H30+ after CAD at high activation
energy and AV=20 volts. Peaks corresponding to (15-C-5)H+,
(C2H40)nH+ (n = 2 and 3) were the succesor ions observed. . 194
4.16. Absolute intensity plots versus of (15-C-5)H30+, (15-C-5)H+,
and (C2H40)nH+(n = 2 and 3) 196
4.17. Mass spectrum of (12-C-4)H+ at AV=70 volts. Peak
XI


Figure 3.4. Mass spectrum of (CH3OH)Cr, m/z = 67, 69.


197
The latter two species were the only fragments observed in the IRMPD studies of
(15-C-5)H30+ at high laser fluences. In the case of 12-crown-4, only (12-C-4)H+,
rather than (12-C-4)H30+, was formed during ESI at either high or low applied
capillary voltages. After rf excitation the (12-C-4)H+ underwent fragmentation
completely to (C2H40)3H+ and (C2H40)2H+ (Figure 4.17 and 4.18).
Bond Dissociation Energies
Plots of the relative cross section (ls/ltot) vs. center-of-mass translational
energies and the best fits using eq 4.8 are shown in Figures 4.2, 4.19, 4.20 and
4.21, for (18-C-6)Rb+, (15-C-5)Rb+, (12-C-4)K+, and (12-C-4)Cs+, respectively. The
BDE's determined in this work, as well as values obtained in previous experimental
and theoretical investigations, are presented in Table 4.3.
For all three crown ethers the BDE trend is Na > K > Rb > Cs. The same
stability order was observed in the previous IRMPD work. The results of
calculations on 18-C-6 complexes (with Li+, Na+, K+, Rb+, and Cs+) by Thompson
and coworkers173 showed the same trend, but their values were systematically
higher. For (18-C-6)K+ Wipff et a/.193 reported a calculated BDE of 61.8 kcal/mol for
a D3d structure using an earlier version of the AMBER force field. Later, Dang and
Kollman194 reported a BDE of 57 kcal/mol, also using AMBER. Thompson195 used
the hybrid quantum mechanical/molecular mechanical (QM/MM) molecular
dynamics method to study (18-C-6)K+ and obtained 70.0 kcal/mol vs. 72.0 kcal/mol


34
stoichiometry may form (Fig. 1-9). The degree of stability of crown ether complexes
is represented by the stability constant, K, the equilibrium constant of the
complexation reaction, which is based on a specific solvent.76 Earlier investigations
of cation-binding strengths are based on stability constants for the polyethers and
cations in solution.77 However, it has shown that solvents play an important role in
binding effects. In solution, the order of complexation selectivity, obtained from log
K values, is generally K+ > Rb+ > Cs + > Na + > Li+ in solvents such as water or
methanol.77 In contrast, Na + binds stronger than K + in the gas phase. 78 This
contradiction is due to the fact that the solvation energy of Na+ is much greater than
that of K\ Cation-binding strength has been of great interest to the chemical
community. Enhancing cation-binding selectivity is another matter, and the
selectivity is also influenced by the ring size. The sizes of cations and crown ethers
are listed in Table 1.
Studies of host-guest complexes in the gas phase have permitted better
understanding of the fundamental details of molecular recognition in the absence
of a solvent.78'81 Gas-phase ion-molecule reactions have also been carried out in
order to study the mechanism of ion formation under fast atom bombardment
conditions82 or to determine the relative stabilities of crown ether complexes with
alkali metal cations.83'85
Earlier investigations of crown ether complexes with alkali metal cations
have shown the intrinsic nature of binding interactions and the size selectivity of
complexation in the gas phase.79'81 For predicting future applications,


Table 4.2. Ions observed in mass spectra. The capillary potential was 25 V for ions
labelled with an asterisk, 75 V for all other ions; (s) strong signal, (w) weak signal.


Relative rate coefficient
+
0.4 Delay Time Dependence on (18-C-6)Rb
0123456789
Energy in center-of-mass frame/eV


61
"charge" and "fire" but also a "Q-switch" pulse which caused the laser to fire when
interrupted by an external trigger pulse. The time from the rising edge of the TTL
trigger pulse to the laser firing, which was monitored by an oscilloscope using a
photodiode, was adjusted to match the FTICR pulse sequence so that the laser
was fired right before ion excitation. The wavelength calibration of the dye laser
was performed using the output of a He:Ne laser by a Continuum field service
engineer immediately before this project was begun. The laser has been used
solely for this research since the calibration was done. The wavelength was
checked using a monochromator ( 0.2 nm) each time when a different dye was
used. The pulse sequence used for this work is depicted in Figure 2.4.
The modified vacuum system, which was pumped by a 700 Us diffusion
pump and a mechanical pump as a forepump, maintained the background pressure
below 10'8 torr. The partial pressures of the three reagents used for the preparation
of the solvated anions were controlled by a pulsed valve and three precision leak
valves on the inlet system; this set up allowed very reliable control of the partial
pressures which is critical for these experiments.
A cylindrical cell (2.45 inches diameter x 3.0 inches length) was used for the
photodetachment experiments (as depicted in Figure 2.5). The fine mesh at the
filament side of trapping plate was replaced with a coarse mesh to increase
transmission of the laser beam and to allow the desired ions to be trapped without
ionization of the metal by the laser beam. Ions were formed by internal electron
impact (El), with electrons produced by a filament located in front of the trapping


7T Optics
SI S2 L3
CO
PL1
JZD
PL2
X,Y Deflection
PL7,8 11
Gatevalve
PL9
PVX3
PV2
oo
oo


DETERMINATION OF BOND DISSOCIATION ENERGIES USING FOURIER
TRANSFORM ION CYCLOTRON RESONANCE
(FT-ICR) MASS SPECTROMETRY
By
YARJING YANG
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1996

ACKNOWLEDGMENTS
I would like to express my appreciation to my research advisor, Dr. John R.
Eyler, for his guidance and assistance, especially for his input of FT-ICR
background and theory. I also appreciate the technical knowledge of the
instrumentation provided by Dr. Cliff Watson. I would like to thank Drs. Dil Peiris
and Nathan Ramanathan for their efforts on the initial IRMPD work of the
metal/crown ether complexes. The comments and suggestions on my research
projects from Dr. Philip Brucat and Dr. David Richardson are also acknowledged.
My appreciation extends to my parents; without their love and support, I
would never be able to finish this work. Thanks to all of the members (Rich Burton,
Kevin Goodner, Tom Hayes, David Kage, Lisa Lang, Eric Milgram, and Joe
Nawrocki) of the current Eyler group for their assistance and friendship. Thanks
also go to Dr. Kitty Williams for her assistance in the preparation of research results
for publication.
Last but certainly not least, I would like to thank my dear husband, Chen-
Chan Hsueh, for his support, love and help. He has been so understanding and
helpful during my graduate career. Without him, I would not have been able to
finish the graduate studies.
ii

TABLE OF CONTENTS
ACKNOWLEDGMENTS ii
TABLE OF CONTENTS iii
LIST OF TABLES vi
LIST OF FIGURES vii
ABSTRACT xiii
CHAPTERS
1. INTRODUCTION 1
Fourier Transform Ion Cyclotron Resonance Mass
Spectrometry (FTICR-MS) 1
History 1
Basic Concepts 2
Ion cyclotron motion 2
Ion excitation 6
Ion detection 11
Signal acquisition 16
Applications 20
Purpose of This Work 20
Solvated Anions 24
Host-guest Complexes 26
Methodology of Determination of Bond Dissociation Energies 37
Photodetachment 37
CAD 45
Energy behavior of reaction cross section 45
On-resonance CAD 47
Doppler broadening 49
Advantages, Disadvantages, and Comparisons to Other
Techniques 50

Photodetachment 50
CAD 51
2. INSTRUMENTATION 53
Introduction 53
Nicolet 2 Tesla ICR Instrument 56
Bruker 7 Tesla ICR Instrument 66
Electrospray Ionization (ESI) Source Region 71
Introduction of ESI Features 77
pKa or pKb 78
Concentration 78
Solubility 79
Molecular conformation 79
Molecular weight 80
Solvent properties 80
Ion Transfer Region 81
Ion Analyzer region 81
3. DETERMINATION OF BINDING ENERGIES OF
SINGLY-SOLVATED HALIDE IONS USING
PHOTODETACHMENT 88
Background 88
Experimental Procedures and Data Analysis 92
Chemicals and Ion-Molecule Reactions 92
Pulse Sequence 109
Data Analysis 109
Results and Discussion 120
Conclusion 141
4.BOND DISSOCIATION ENERGIES OF ALKALI AND
HYDRONIUM CATIONS BOUND TO CROWN ETHERS
DETERMINED BY ON-RESONANCE COLLISIONALLY-
ACTIVATED DISSOCIATION (CAD) 144
Background 144
Experimental 150
Ion formation 150
Pulse sequence 154
Data Analysis 155
Ion Appearance Curve 155
Internal Energies 159
IV

Rotational energy 159
Vibrational energies 159
Trial Function 160
Deconvolution 161
Results 163
Collision Conditions 163
Fragmentation Pathways 164
Bond Dissociation Energies 197
Discussion 210
Conclusion 216
5. CONCLUSION 218
REFERENCE LIST 223
BIOGRAPHICAL SKETCH 235
v

LIST OF TABLES
Table page
1.1. Size of alkali-metal ions and cavity size of crown ethers 36
2.1. Laser dyes used for photodetachment experiments and their
corresponding wavelength range before and after frequency
doubling 68
3.1. Electron affinities (EA) and proton affinities (PA) of species
studied in this work 131
3.2. Photdetachment thresholds (ET) and binding energies (BE's) of
solvated halides 133
4.1. Vibrational frequencies of cyclohexane and 1,4-dioxane 152
4.2. Ions observed in mass spectra. The capillary potential was 25 V
for ions labelled with an asterisk, 75 V for all other ions; (s) -
strong signal, (w) weak signal 170
4.3. Bond dissociation energies for alkali-metal and hydronium ion
crown ether complexes including present work, previous work,
and theoretically calculated values 209
VI

LIST OF FIGURES
Figure page
1.1.Standard ICR trapping cells, (a) Cubic cell with two trapping
plates perpendicular to z-axis, two excite plates perpendicular
to x-axis, and two detect plates perpendicular to y-axis. The
ion is trapped inside the cell with cyclotron motion in the xy-
plane, moving along the z-axis. (b) Cylindrical cell 5
1.2.Standard pulse sequence for ICR experiment 8
1.3. Ion excitation. (1) Impulse excitation: approximate delta function
vs time, Fourier transform to the frequency domain signal (2)
Chirp excitation: an rf frequency sweep vs time, Fourier
transform to the frequency domain signal (3) SWIFT: (a) The
desired frequency spectrum, (b) Inverse FT to give a time-
domain signal, (c) Fourier transform to the frequency domain
signal 10
1.4. (1): (a) Ions' cyclotron motion prior to excitation, (b) Ions are
excited to a larger orbit, (c) Ions achieve coherence after
excitation. (2) Phase coherency: (a) Two ions of same m/z
180 out of phase, (b) The in-phase (initial motion parallel to
electric field, E) ion speeds up and out-of-phase ion slows
down until they eventually achieve coherent motion 13
1.5.Ions induce image current alternately on upper and lower
plates 15
1.6.Heterodyne mode: The signal mixes with a reference to give a
sum (higher frequency) and difference (lower frequency)
signal. The higher frequency (low mass) signal is then filtered.
Fast FT gives the low frequency signal which is the
difference 19
1.7. The structures of CH3OHX, X = F, Cl, Br, and 1 28
vii

1.8.The structures of polyether macrocycles.
30
1.9.The structures of metal/hydronium ion-crown ether complexes. . 33
1.10.Photodetachment process: (a) A' + hv A + e', (b) A + hv ET
+ C, (c) A' + hv B + C + e 40
2.1. Simplified block diagram of FTICR mass spectrometer 55
2.2. 2 tesla ICR instrument (a) 2 T magnet, (b) Inlet source, (c) Three
window flange, (d) One window flange, (e) ICR cell, (f)
Diffusion pump, (g) Ion gauge 58
2.3.Laser set-up for photodetachment experiment (1) Apple II
microcomputer controls "charge" and "fire" commands of YAG
laser, (2) Console sends external TTL pulse to trigger the
laser, (3) Power control from console to the cell; (a) Quartz
prism, (b) Laser quartz windows, (c) ICR cell, (d) 2 T magnet,
(e) Apple II computer, (f) Console, (g) Photodiode 60
2.4. Pulse sequence for photodetachment experiments 63
2.5. Cylindrical cell 65
2.6.7 T ICR instrument (a) 7 T magnet, (b) Infinity cell, (c) 400 L/s
cryopumps, (d) 800 L/s cryopump, (e) Mechanical pumps, (f)
Capillary coated with platinum at entrance and exit ends, (g)
Needle, (h) Syringe pump 70
2.7. Ion formation in ESI 73
2.8. ESI design 76
2.9. Electrostatic lenses in transfer region 83
2.10.Infinity cell with three "side-kick" plates: (a) PVX1, (b) PVX3,
(c) PVX4 plates, (d) the macor plate which is attached to
trapping plates to hold the "side-kick" plates and trapping
plates, (e) the rf-shimmed trapping plate, (f) the Infinity cell
with four brass plates for excition and detection 85
2.11. Pulse sequence for CAD experiments 87
viii

3.1. Mass spectrum of F" 94
3.2. Mass spectrum of COCI' 96
3.3. Mass spectrum of (CH3OH)F, m/z = 51 99
3.4. Mass spectrum of (CH3OH)Cr, m/z = 67, 69 101
3.5. Mass spectrum of OH', m/z =17 104
3.6. Mass spectrum of (H20)Br\ m/z = 97, 99 106
3.7. Mass spectrum of (CH3OH)Br, m/z =111,113 108
3.8. Photodetachment spectrum of (C2H5OH)Br: experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2- the arrow points to the
onset threshold 113
3.9. Photodetachment spectrum of (CH3CN)Br: experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2 the arrow points to the
onset threshold 115
3.10. Photodetachment spectrum of (CH3OH)l': experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2 the arrow points to the
onset threshold 117
3.11. Photodetachment spectrum of (CH3OH)F': experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2- the arrow points to the
onset threshold 119
3.12. Photodetachment spectra: (CH3CN)Br; (CH3OH)Br; -
(C2H5OH)Br; (i-C3H7OH)Br; X (n-C3H7OH)Br 122
3.13. Photodetachment spectra of (CH3OH)F'obtained using different
thermalization times (t): -1=2.34 sec.; X -1=3.20 sec.; -
t=3.77 sec.; X -1=4.31 sec.; -1=6.05 sec 124
3.14. Photodetachment spectra of (CH3OH)CI' obtained using
different thermalization times (t): -1=2.05 sec.; -
IX

t=3.20 sec.; B-t=4.34 sec 126
3.15. Photodetachment spectra of (CH3OH)Br obtained using
different thermalization times (t): -1=2.05 sec.; -1=3.77
sec.; -1=4.34 sec 128
3.16. Potential energy curves for alcohol-solvated halide anions
and corresponding neutral species. The lower curve
represents the interaction of the solvent and halide from infinite
separation to a stable bound state, a) The upper solid curve
represents the hydrogen abstraction reaction of ROH + X to
form RO + HX. (X = F and Cl). The arrows indicate the
thermochemical cycle for the photodetachment process,
including the threshold (Ethresh0,d), electron affinity (EA) of X,
binding energy (BE) of (ROH)X', and the energy difference
(AE) between the reported and actual BE. b) Same as a) with
dotted curve to indicate hydrogen abstraction reaction for X=
Br and 1 136
4.1. Potential energy curve for ion crown ether complexes and
corresponding neutral species. The D298 represents the bond
dissociation energy at room temperature, the D0 represents the
bond dissociation energy at zero Kelvin temperature 149
4.2. Appearance curve of (18-C-6)Rb+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 158
4.3. Appearance curves of (18-C-6)Rb+ at different pressures (50 ms
CAD delay time); 3.0 10'8 torr, 5.5 x 108 torr, 7.3
x 10'8 torr, A 9.2 x 108 torr 166
4.4. Appearance curves of (18-C-6)Rb+ at different CAD delay times
(P = 3.3 x 10 8 torr); 0.05 s, 0.08 s, 0.1 s, 0.15
s 168
4.5. Mass spectrum of (12-C-4)Rb+ at AV=20 volts. Peaks
corresponding to (12-C-4)Rb+ and (12-C-4)2Rb+ ions were
observed 173
4.6. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peaks
corresponding to (15-C-5)H30+, (15-C-5)2H+, and (15-C-
x

5)2H30+ ions were observed
175
4.7. Mass spectrum of (12-C-4)H30+ at AV=20 volts. Peaks
corresponding to (12-C-4)H+, (12-C-4)2H+, and (12-C-4)2H30 +
ions were observed. The peak intensity of (12-C-4)2H+ is higher
than that of (12-C-4)H+ 177
4.8. Mass spectrum of (12-C-4)H30+ at AV=70 volts. Peaks
corresponding to (12-C-4)H+ and (12-C-4)2H+ ions were
observed. The peak intensity of (12-C-4)2H+ is lower than that
of (12-C-4)H+ 179
4.9. Mass spectrum of (12-C-4)Rb+ at AV=70 volts. Peak
corresponding to (12-C-4)Rb+ ion was isolated before
excitation 182
4.10. Mass spectrum of (12-C-4)Rb+ after CAD at AV=70 volts. The
two isotopes of Rb+ were the only successor ions observed . 184
4.11. Mass spectrum of (18-C-6)H30+ at AV=20 volts. Peak
corresponding to (18-C-6)H30+ ion was isolated before
excitation 186
4.12. Mass spectrum of (18-C-6)H30+ after CAD atAV=20 volts. Peak
corresponding to (18-C-6)H+ ion was the only successor ion
observed 188
4.13. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peak
corresponding to (15-C-5)H30+ ion was isolated before
excitation 190
4.14. Mass spectrum of (15-C-5)H30+ after CAD at low activation
energy and AV=20 volts. Peak corresponding to (15-C-5)H+ ion
was the only successor ion observed 192
4.15. Mass spectrum of (15-C-5)H30+ after CAD at high activation
energy and AV=20 volts. Peaks corresponding to (15-C-5)H+,
(C2H40)nH+ (n = 2 and 3) were the succesor ions observed. . 194
4.16. Absolute intensity plots versus of (15-C-5)H30+, (15-C-5)H+,
and (C2H40)nH+(n = 2 and 3) 196
4.17. Mass spectrum of (12-C-4)H+ at AV=70 volts. Peak
XI

corresponding to (12-C-4)H+ ion was isolated before
excitation
199
4.18. Mass spectrum of (12-C-4)H+ after CAD at AV=70 volts. Peaks
corresponding to (C2H40)nH+ (n = 2 and 3) were ions
observed 201
4.19. Appearance curve of (15-C-5)Rb+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 203
4.20. Appearance curve of (12-C-4)K+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 205
4.21. Appearance curve of (12-C-4)Cs+. The solid circles are the
experimental data (mean of five measurements with error bars
corresponding to 1 standard deviation); the curve is the best
fit to the data from the CRUNCH program 207
XII

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
DETERMINATION OF BOND DISSOCIATION ENERGIES USING FOURIER
TRANSFORM ION CYCLOTRON RESONANCE (FT- ICR)
MASS SPECTROMETRY
By
Yarjing Yang
1996
Chairperson: John R. Eyler
Major Department: Chemistry
The objective of this work was to use Fourier transform ion cyclotron
resonance mass spectrometry (FTICR-MS) to determine bond dissociation energies
(BDE's) of important species such as solvated anions and host-guest complexes
using two different techniques: photodetachment and on-resonance collisionally
activated dissociation (CAD).
Photodetachment has been widely used in measuring the electron affinities
of various anions. Since the electron affinities of the halides are well known, this
part of the work has utilized photodetachment to measure the halides' binding
energies with alcohol solvents in the gas phase. An Nd3+:YAG-pumped dye laser
with frequency doubler was used to obtain the uv-visible irradiation needed for this
work. The binding energies of methanol, ethanol, iso- and normal-propanol, and
xiii

acetonitrile with fluoride, chloride, bromide and iodide ions have been measured.
The effect of size and polarizibility of the solvents, and comparisons of the solvation
energies determined in this work to some reported in earlier thermal-equilibrium
studies are discussed.
Collisionally activated dissociation (CAD) processes in mass spectrometry
can provide quantitative bond dissociation energy (BDE) estimates if the interaction
energy is precisely controlled during collisions. This work has concentrated on
improving the accuracy of measuring the thresholds and providing more precise
quantitative measurements of host-guest complex interactions in the gas phase by
accounting for the effects of multiple ion-molecule collisions, internal energy of
precursor ions, thermal spread of energies of the collision gas and the dissociation
lifetime of the intermediates. The use of an electrospray ionization (ESI) source for
soft ionization that produces intact, multiple charged and solvated gas-phase ions
directly from various compounds in solution has been critical for this work. We have
measured the binding energies of hydronium and alkali metal ions to the crown
ethers (12-crown-4,15-crown-5, and 18-crown-6). The effects of metal size, cavity
size and polarizibility of different crown ethers are discussed.
XIV

CHAPTER 1
INTRODUCTION
Fourier Transform Ion Cyclotron Resonance Mass Spectrometry FTICR-MS1
History
In 1950, the first practical ion cyclotron mass spectrometer (Omegatron) was
used for analytical purposes by Sommer, Thomas and Hippie.12 During the 1960's,
a number of research groups exploited the ion cyclotron resonance (ICR) technique,
normally defined as "Pre-FTICR" with conventional features such as solenoidal
magnets, electromagnets, drift cells, ion trapping, double resonance detection,
analog FT and simultaneous excitation and detection. As practiced in the 1960's
the technique had slow scan speed, low resolution and a limited mass range.3 8 In
1974, Comisarow and Marshall9'11 first applied the FT technique to ICR-MS with
frequency chirp excitation, mass analysis in the same region and simultaneous
excitation and detection. This has evolved into the versatile and powerful technique
often referred to as Fourier transform mass spectrometry (FTMS).
In 1980, superconducting magnets began to be used, resulting in improved
performance when compared to electromagnets: ultra-high mass resolution, wide
mass range, high mass measurement accuracy, simultaneous detection of all ions,
1

2
ion storage, and many probes of ion structure.12 During the 1980's, a wide variety
of applications13'19 were reported, including MS/MS experiments, laser desorption,
fast atom bombardment (FAB) and high pressure external sources with quadrupole
ion guides, gas chromatography mass spectrometry (GC/MS), external sources
utilizing electrostatic focussing which lower ultimate pressure at the analyzer,
stored waveform inverse Fourier transform (SWIFT) excitation and electrospray
ionization (ESI) sources. Both ESI20 and matrix assisted laser desorption ionization
(MALDI)21 sources began to be used in the early 1990s to increase the range of
masses which can be studied with FTICR mass spectrometers. In 1992, Amster
and Marshall22 demonstrated quadrupolar axialization to improve resolution and
increase sensitivity. In 1993, Smith's group23 developed Time Resolved Ion
Correlation (TRIC) to detect highly-charged individual ions with masses higher than
5,000,000 daltons.
Basic Concepts
Ion cyclotron motion
The ions in a magnetic field undergo circular orbital motion, at a cyclotron
frequency given by,24 27
) = qB/m = M/r (1.1)
where q is the ion charge, B is the magnetic field, m is the mass of the ion, v is the
ion velocity and r is the orbit of the cyclotron motion. In a uniform magnetic field (no
electric field), the mass-to-charge ratio, m/q is linearly related to the observed

3
cyclotron frequency, w. The Lorentz force, F, is given by
F q ( V x So)
(1.2)
For B0 along the z axis, this force is constrained to the xy-plane (there is no force
along the z-axis), so ions can move freely along that axis unless an electric field is
applied,28,29
(1.3)
where E is the electric field.
Flowever, in order to detect the ions, they must be confined within a finite
volume bounded by conductive electrodes (in a trapping cell). Figure 1.1 is a
schematic drawing of two different ICR trapping cells which are commonly used:
cubic and cylindrical. The cell is centered in a homogeneous magnetic field
generated by a superconducting magnet and contained in a high vacuum chamber.
There are two trapping plates which are perpendicular to the magnetic field (z-axis),
two excite plates (perpendicular to the x-axis) and two detection plates
(perpendicular to the y-axis).
The ions may be formed in the ICR cell (using an internal ionization source)
or outside the magnetic field (with an external ionization source) by a number of
methods, including electron impact (El), chemical ionization (Cl), laser desorption

Figure 1.1 Standard ICR trapping cells, (a) Cubic cell with two trapping plates perpendicular to z-axis, two excite plates
perpendicular to x-axis, and two detect plates perpendicular to y-axis. The ion is trapped inside the cell with cyclotron
motion in the xy-plane, moving along the z-axis. (b) Cylindrical cell.

.X
en

6
(LD), fast atom bombardment (FAB), electrospray ionization (ESI), and matrix
assisted laser desorption ionization (MALDI). The basic FTICR-MS experiment is
conducted using a series of computer-controlled pulses, as depicted in Figure 1.2.
Initially a quench pulse is applied to the trapping plate(s) to remove all the ions from
the cell prior to starting the next pulse sequence. After ion formation, ions are
trapped in the cell by constraining their z-axis motion (along the magnetic field) by
applying low voltages (0.5 to 5 V typically) to the trapping plates. Either positive or
negative ions can be trapped in the cell by changing the polarity of the voltage
applied to the cell plates.
Ion excitation
Three types of ion excitation are generally used for ion detection, as shown
in Figure 1.3. (1) In impulse excitation,18 an approximation to an ideal delta function
pulse (infinite amplitude, zero width) is applied. This results in a flat excitation
spectrum and should excite all ions equally. The angular cyclotron frequency is of
the order of wcmax = 1/At, where At is the pulse width. (2) With chirp excitation,10 a
radio frequency (r.f.) pulse is applied whose frequency sweeps rapidly over a range
from the lowest to highest frequency desired. The disadvantage of chirp excitation
is the nonuniform excitation of the ions, which becomes pronounced for ions near
the edge of the swept frequency range. The final cyclotron radius is limited to less
than the full dimensions of the cell since ions absorb different amounts of power
during the chirp. This introduces peak height distortion and the frequency chirp
excitation also contributes to z-axis ejection. (3) For stored waveform inverse

Figure 1.2 Standard pulse sequence for ICR experiment.

Delay Time
i*
Time
Quench pulse
Ion Formation pulse
Excitation pulse
Detection
CO

Figure 1.3 Ion excitation. (1) Impulse excitation: approximate delta function vs time, Fourier transform to the frequency
domain signal (2) Chirp excitation: an rf frequency sweep vs time, Fourier transform to the frequency domain signal
(3) SWIFT: (a) The desired frequency spectrum, (b) Inverse FT to give a time-domain signal, (c) Fourier transform to
the frequency domain signal.

(2)
inverse
7
vor w
(a)
(b)
^ vor co
vor
(0
ft1
(c)
^ vor
(0
o

11
Fourier transform (SWIFT) excitation,24,30'32 the excitation waveform is generated
by taking the desired frequency spectrum (a) (power as a function of frequency) and
performing an inverse FT to give a time-domain signal (b). This waveform is then
generated by a digital frequency synthesizer. The excitation spectrum (c), which
indicates the actual amount of power applied at each frequency, is produced by
Fourier transformation of (b). This excitation spectrum is much more uniform over
the desired frequency band than either impulse or chirp excitation.
Any of these three excitation modes will cause the ions to absorb energy and
be accelerated into larger orbits when the r.f. frequency becomes equal to the ion
cyclotron frequency. It is also necessary to excite a packet of ions of a given mass-
to-charge ratio not only to a larger ICR orbital radius, but to coherent motion in order
to detect them. Figure 1.4 shows how the ions achieve coherence prior to
detection.
Ion detection
The orbiting packet of ions induces a small alternating image current on the
receive plates (perpendicular to the y-axis). The image current is given by,33
where N is number of ions, v is the ion velocity, q is the charge of the ion, and d is
the electrode spacing, i.e. the dimension of the cell as depicted in Figure 1.5. The
ion induces image current alternately on upper and lower plates. To obtain a

Figure 1.4 (1): (a) Ions' cyclotron motion prior to excitation, (b) Ions are excited to a larger orbit, (c) Ions achieve
coherence after excitation. (2) Phase coherency: (a) Two ions of same m/z 180 out of phase, (b) The in-phase (initial
motion parallel to electric field, E) ion speeds up and out-of-phase ion slows down until they eventually achieve
coherent motion.

13

Figure 1.5 Ions induce image current alternately on upper and lower plates.

induced
image current
induced
image current
cn

16
theoretical estimate of the magnitude of the signal produced on the plates by an ion,
the Reciprocity Theorem is applied.35 The cyclotron radius, Rc, is given by,
where E is the excitation electric field, t is the duration of the resonant excitation
pulse, and B is the magnetic field. The Reciprocity Theorem states that the
electrostatic potential created at the position of an ion by 1 volt on the detection
plate is directly proportional to the image current induced on the plate by the ion.
Thus the predicted FTICR signal strengths can be calculated by a numerical
solution of Laplace's equation. The amplitude of the signal per ion only depends on
the ratio between the radius of cyclotron orbit and the distance between the receiver
plates.
All ions in the cell are excited and detected simultaneously as a time-domain
signal. The signal is converted into voltage, amplified, digitized and stored in
computer. The time-domain spectrum is a composed of superimposed sine waves:
in order to recover the frequency domain spectrum, the complex time-domain
spectrum is subjected to a fast Fourier transform (FFT) algorithm.
Signal acquisition
For one thousand ions, the r.f. signal at a preamplifier input is on the order
of a few hundred microvolts. A low noise electronic design for amplification is
required. There are two types of data acquisition used in our ICR instruments, (a)

17
The first produces broadband spectra (all masses). In this mode the signal from a
preamplifier is sampled directly by the analog to digital (A/D) converter at a rate
which must be faster than twice the highest frequency for the lowest mass to be
observed (Nyquist criterion). Spectra are signal-averaged by summing the time-
domain transients from each repetition in the computer's memory before performing
the FT; (b) The second type of data acquisition (heterodyne mode) leads to
narrowband spectra.34'35
As illustrated in Figure 1.6, the heterodyne technique reduces the digitization
rate requirements for ion signals and is used to obtain higher mass resolution over
a narrow mass range. Given a time-domain spectrum consisting of N points, where
N is between 16K and 128K, it is common to zero-fill at least once (add N zeros to
the end of the data set), apodize, and obtain the magnitude-mode FT, where the
magnitude-mode is defined as [(real)2 + (imaginary)2]1'2, with a FFT algorithm. Zero
filling has been shown to increase the S/N ratio and increase the displayed
resolution of the spectrum.
Since frequency is a physical quantity which can be precisely measured, the
exact mass of an ion can be determined very accurately in the FT-ICR technique,
since it is directly related to the measured frequency. Typically, low parts-per-million
accuracy can be achieved with an internal mass calibration. A high degree of mass
accuracy can be maintained due to the stable magnetic field. The main features of
FTICR include ultra-high mass resolution, accurate and stable mass calibration,
multiple stages of collisionally activated dissociation (MS/MS), extremely long ion

Figure 1.6 Heterodyne mode: The signal mixes with a reference to give a sum (higher frequency) and difference (lower
frequency) signal. The higher frequency (low mass) signal is then filtered. Fast FT gives the low frequency signal which
is the difference.

Reference
Mixer
Sum and
Difference Signal
Low Pass Filter
~1,
Difference signal

20
storage times (> minute), the ability to study ion/molecule reactions, compatibility
with pulsed or continuous ionization techniques, and ease of carrying out
photodissociation and photoactivation studies. The disadvantages of FTICR are its
limited dynamic range (space charge limits the number of ions which can be trapped
to ca. 106), the fact that data acquisition can take milliseconds to seconds, so
transient or metastable ions can not be studied directly, and the high cost of the
instrument ($300K to $500K).
Applications
Purpose of this Work
Different experimental methods are used in the studies of ion-neutral
complexes. The case of interactions which involve charged species requires
specific techniques for three main reasons: (I) charged species must be selectively
separated from ions of the opposite charge, (ii) they must be stored long enough to
reach a state of thermodynamic equilibrium with their surroundings while avoiding
collisions with the walls of the reaction chamber and (iii) they must be monitored by
highly sensitive methods because of their very low concentrations.
Several reviews have been briefly summarized by Keesee and Castleman
et at.36 37 in a major review of thermodynamic data for ion-molecule reactions in the
gas phase. A number of experimental devices have been applied to the study of
ion-neutral complexes. Among the devices used to generate and study the gas
phase ion complexes similar to those in these experiments are molecular beams,

21
drift tubes, and ion traps.38-40 Each device offers its own advantages for different
types of ions. For the study of ion-neutral reactions, ion cyclotron resonance (ICR)
mass spectrometry has been the most extensively applied method. The
advantages offered by ICR mass spectrometry include the ability to generate a large
variety of molecular ions and high sensitivity. The technique of ICR-MS allows one
to generate gas phase ions and trap them for periods of millisecond to several
seconds, sometimes even for hours.
The goal of this work was to apply different FTICR-MS techniques including
photodetachment and collisionally activated dissociation (CAD), to obtain bond
dissociation energies for solvated anions and host-guest complexes. A secondary
goal was obtaining quantitative measurements of bond dissociation energies
(BDE's) by improving the threshold determination to account for a variety of
systematic effects that would otherwise distort the thermochemistry obtained.
Several methods have been used to measure the bond dissociation energies for
various ion-neutral complexes, including equilibrium measurements in both high
pressure41^13 and ICR44-45 mass spectrometers, photodissociation measurements in
a time-of-flight (TOF) mass spectrometer,46-50 photodetachment measurements in
photoelectron spectroscopy,51-53 collisionally activated dissociation (CAD) in both
quadrupole or octopole-guided ion beam54-57 and ICR58 mass spectrometers. The
measurement of BDE's by FTICR mass spectrometry has not been widely applied
and there is still some uncertainty as to the quality of the measurements. In this
section, some recent experimental studies of BDE measurements using different
types of techniques are briefly discussed.

22
A few years ago the development of high pressure mass spectrometry
reached a stage where it could be applied successfully to chemical problems. The
technique and its applications have been summarized by Kebarle.59 An impressive
number of clustering equilibria of the type in eq. 1.6 have been studied,
X Sn X S^ + S, In Kntrv1= AS/R AH/(RT) (1.6)
where X represents a central ion, S a solvent molecule, and X Sn and X S^ are two
cluster ions of different stoichiometry. Based on the accessibility of suitable ion
sources, single ion solvation can be studied by mass spectrometry provided
equilibrium is established in such a system.
Photodissociation thresholds can be interpreted as reflecting at least an
upper limit to the thermochemical dissociation threshold. Moseley et al.50 measured
the vibrationally resolved photodissociation spectrum of C03', which was analyzed
to give a reliable limit of 1.85 eV for D(C02-0).
Energy-resolved collision-induced dissociation of metal carbonyl anions
M(CO)n' (M=V, Cr, Mn, and Co) were investigated to determine sequential metal-
carbonyl bond energies by Squire's group.57b They used the model described in the
Methodology Section of this Chapter to obtain fairly accurate BDE's of 30.8 3.5,
40.6 3.5, 40.6 3.9, and 39.7 3.7 kcal/mol for (CO)5V-CO, (CO)4Cr-CO,
(CO)4Mn'-CO, and (CO)3Co -CO, respectively.
Electron photodetachment spectroscopy is one of the most useful methods

23
available for determining the spectroscopic properties of negative ions. This
technique may be used to measure the dependence of the cross section on photon
energy and therefore to determine the photodetachment threshold. Bowen's
group51 has demonstrated the photodetachment of NH4' using photoelectron
spectroscopy (PES). The PES of NH4' resembles that of free hydride, H', except for
being shifted to lower electron kinetic energy due to the stabilization effect of
ammonia solvation. They confirmed that this ion cluster is best described as a H*
ion solvated by NH3, and were able to obtain the solvation energy.
Since the electron affinities of the halides are well known, the threshold for
photodetachment of alcohol solvated halides is directly related to the halides'
solvation energies in the gas phase. Photodetachment processes in the ICR mass
spectrometer provide more accurate quantitative measurement compared to
thermal equilibrium methods if the ions are well thermalized and the ion intensity is
well controlled. A frequency-doubled dye laser pumped by a Nd3+:YAG laser was
used to obtain the UV-near visible output range for this work. Such a laser can
cover a very wide wavelength range with narrow bandwidths.
Collisionally activated dissociation (CAD) processes in mass spectrometry
can provide quantitative bond dissociation energy (BDE) estimates if the interaction
energy is precisely controlled during collisions. One goal of this work was to
improve the accuracy of determining the threshold and provide more precise
quantitative measurement of the BDE of host-guest complexes in gas phase. The
use of an electrospray ionization (ESI) source for soft ionization that produces

24
intact, multiple charged and solvated gas-phase ions directly from various
compounds in solution has been critical for this work. In addition, the ions formed
using ESI carry little internal energy compared to other ionization techniques such
as electron impact (El), laser desorption (LD) and matrix-assisted laser desorption
ionization (MALDI).
Solvated Anions
Living organisms and nonliving matter are affected by interactions with
organic chemicals, which are highly dependent upon how molecular structures
produce site-specific interactions, such as hydrogen bonds. In general, in the life
sciences and in chemistry, the hydrogen-bond (HB) is perhaps the most important
kind of specific molecular interaction between molecular species. Examples include
solubility, bodily transport and docking to active sites. Hydrogen bonding was first
recognized in the physical properties of liquids, and was thought to involve only
fluorine, oxygen or nitrogen atoms bridged by an intervening hydrogen.60 Since
then structural studies have indicated that atoms such as chlorine, bromine and
sulfur can serve as hydrogen bond acceptors,6162 and that hydrogens bonded to
carbon and sulfur can serve as hydrogen bond donors.63 Another significant type
of site-specific noncovalent interaction has been identified as "halogen bonding".63
Therefore, study of alcohol solvated halogen ions through hydrogen bonding has
been an important issue.
Ion solvation has been studied extensively for many years by both theoretical

25
methods and experimental techniques. Currently, approaches have reached a high
degree of accuracy and sophistication. Despite this enormous amount of
information, however, no comprehensive and satisfactory theory of ion solvation can
be given, due mainly to the lack of a detailed knowledge of the structure of non
simple liquids. The problems occurring in model calculations on ion solvation are
two-fold. Firstly, accurate energy surfaces for ion-molecule complexes are difficult
to obtain. They provide information about the most stable geometries and energies
of interaction. The static energy of a complex, unfortunately, is only one side of the
picture. The shape of potential surfaces for clusters consisting of an ion and many
molecules, in general, indicates high flexibility, since several flat energy minima
separated by low barriers are usually found. Therefore, the dynamics of the
complex is the second important problem for most macroscopic properties at room
temperature.
An alternate and more rigorous treatment of ion solvation is based on the
study of ion-molecule reactions in the gas phase. Relative enthalpies and free
enthalpies of complex formation for solvated ions X'(H20)n with n = 1-6 have been
determined.42 59 The concept of hydrogen bonding (HB) emerged from the study of
self-associated liquids (notably water)63 and its importance in complex biological
systems is paramount.64 The study of hydrogen-bond, HB, interactions in the gas
phase is relevant because it provides direct information on the structure and stability
of these complexes in the absence of the perturbations induced by the solvent.
Most of the solvated ions are easily generated and heavily studied.41-45 59

26
These studies have led to an understanding of the stability of hydrogen bonding and
structural variation in these compounds. Such work also provides a better
understanding of the characteristic properties of bulk solution such as ions solvated
with solvent clusters.42,59 Among solvated ions the methanol solvated anions are
difficult to generate and understanding of their properties is still subject to argument.
Also there is still disagreement about the binding energies (BE) of solvated anions
using different measurements.41-45 Therefore, alternate measurements of the BE
using threshold determinations from photodetachment experiments can give more
accurate results. Also, since the ICR ion trap can easily form solvated ions using
either ion-molecule reactions or electrospray ionization (ESI), it becomes possible
to measure BE using photodetachment techniques in the ICR instrument.
The BE measurements of alcohol solvated halides, whose structures are
illustrated in Fig. 1.7, using the photodetachment technique with a FTICR mass
spectrometer are described in Chapter 3.
Flost-guest complexes
Polyether macrocycles have been known for decades.65-67 The discovery that
certain macrocycles can bind alkali metal and alkaline earth cations gave impetus
to the field now known as host-guest chemistry.68 Today, a distinction is made
between the classical ring oligoethers (crown ethers) and monocyclic commands,
oligocyclic spherical cryptands and the acyclic podands with respect to topological
aspects.69 Multidentate monocyclic ligands with any type of donor atoms are called
coronands (crown compounds), while the term crown ether should be reserved for

Figure 1.7 The structures of CH3OHX, X = F, Cl, Br, and I.

Structures of CH3OHX-
0.97A
H ,1.44A
1.47A 1.02
F0.
174.8
O
v1.42A
C
H
H
o
H
o
H
C
H
2.39A *98A o
0 1.43A
Br-""
162.4
H
H
C
H
Cl
H
0 0.98A
2-26A_ H v O.
v1.43A
163.2 'C H
yj
H
o 0.97A
2..82A 0.143X
160.4 XC H
H
/j
H
ro
co

Figure 1.8 The structures of polyether macrocycles.

D [D In D
Podand
o
o o
D D
w
Coronand
pio^
bvJdW
Cryptand
where D = O or S, B = N
-o o
0 I Crown Ether (D = O)

31
cyclic oligoethers exclusively containing oxygen as donor atom (see Figure 1.8).
The ability to form host-guest type complexes with metal ions is one of the
most important properties of crown ethers and related compounds. These
macrocyclic polyethers show a remarkable range of specificity for a wide variety of
cations that depends, in part, on the size of the ether, the type of donor atoms (e.g.
oxygen, nitrogen, sulfur), and the polarity of the solvent.70'72 Since their discovery
by Pedersen73 74 in 1967, crown ethers have generated a tremendous amount of
interest because of their ability to selectively bind guests such as alkali metal ions.
This crown ether/ion specificity can also serve as the basis for many potential
practical applications. For example, 90Sr and 137Cs are the two major generators of
radiation in nuclear waste, and thus complicate disposal efforts. Horowitz, Dietz,
and Fisher75 have used di-tert-butylcyclohexano-18-crown-6 for recovering 90Sr2+
from acidic solutions. A more thorough understanding of cation/crown ether
chemistry can provide the basis for rational design of new ligands useful in
separation of these and other radionuclides from complex waste streams and
hazardous waste storage facilities.
The 3-dimensional structure of a crown ether complex is typified by the
symmetrical coplanar array of the oxygen atoms. These atoms contact the cation
located in the center at an arithmetically calculated interatomic distance as shown
in Figure 1.9. Cations which fit less well spatially are tolerated within certain limits,
e.g. by deformation of the ligand skeleton or by movement of the cation out of the
ring plane. If the cation diameter is much larger, sandwich complexes with 2:1-

Figure 1.9 The structures of metal/hydronium ion-crown ether complexes.

ee
va
va
7J
O
sa
o
CO
II
00
O


AA
cn
O
I
cn
ro
O
ii
ro
O
va
Ol
O

Ol
ro
IIIIIIMIIMIIIII
mi min ii ii it I

34
stoichiometry may form (Fig. 1-9). The degree of stability of crown ether complexes
is represented by the stability constant, K, the equilibrium constant of the
complexation reaction, which is based on a specific solvent.76 Earlier investigations
of cation-binding strengths are based on stability constants for the polyethers and
cations in solution.77 However, it has shown that solvents play an important role in
binding effects. In solution, the order of complexation selectivity, obtained from log
K values, is generally K+ > Rb+ > Cs + > Na + > Li+ in solvents such as water or
methanol.77 In contrast, Na + binds stronger than K + in the gas phase. 78 This
contradiction is due to the fact that the solvation energy of Na+ is much greater than
that of K\ Cation-binding strength has been of great interest to the chemical
community. Enhancing cation-binding selectivity is another matter, and the
selectivity is also influenced by the ring size. The sizes of cations and crown ethers
are listed in Table 1.
Studies of host-guest complexes in the gas phase have permitted better
understanding of the fundamental details of molecular recognition in the absence
of a solvent.78'81 Gas-phase ion-molecule reactions have also been carried out in
order to study the mechanism of ion formation under fast atom bombardment
conditions82 or to determine the relative stabilities of crown ether complexes with
alkali metal cations.83'85
Earlier investigations of crown ether complexes with alkali metal cations
have shown the intrinsic nature of binding interactions and the size selectivity of
complexation in the gas phase.79'81 For predicting future applications,

Table 1.1. Size of alkali-metal ions and cavity size of crown ethers.

Size of Alkali-Metal Ions and Cavity Size of Crown Ethers
rL(A) [ref. 199]
Na
K
Rb
Cs
rm*(A) [ref 198]
1.02
1.38
1.52
1.67
C/rL
12-crown-4
1.2-1.5
0.76
1.02
1.13
1.24
15-crown-5
1.7-2.2
0.52
0.71
0.78
0.86
18-crown-6
2.6-3.2
0.35
0.48
0.52
0.58
rM+- metal ion radius; rL cavity size of crown ethers
CO
CD

37
measurements of BDE's would provide better understanding for making use of the
ion discrimination properties of crown compounds based on ion-selective membrane
electrodes,86 for the study of physiological ion-transport processes in biological
membranes, for the investigation of receptor and enzyme interactions as well as for
the salt balance and for metabolic processes of the living organism.87
Investigations using CAD to study crown ether complexes with alkali metal
cations have been previously reported.8889 However, quantitative information about
the BDE's of crown ether complexes has only been reported by the Eyler group.89
In the previous study, approximate binding energies of metal/crown ethers formed
using the laser desorption technique were reported. In these studies, we have
carefully measured quantitative BDE's of crown ethers with either hydronium or
metal ions attached, formed from an external electrospray ionization (ESI) source,
using a different approach to determine the dissociation threshold. The details of
the measurements are described in Chapter 4.
Methodology of Determination of Bond Dissociation Energies
Photodetachment
Electron photodetachment from gas phase molecular anions has proved to
be a remarkable tool in the study of both anions and their neutral photoproducts.
A variety of thermochemical and spectroscopic constants can be measured for
systems that would be difficult to study by other methods. Among the techniques

38
that have been developed for the study of electron photodetachment are ion beams,
trapped ion cells, and drift tubes.38'40 The advent of tunable laser systems has
increased the precision of measurements made in photodetachment experiments,
allowing for the detection of vibrational and even rotational transitions in some
systems.52 In addition, the wide wavelength range and narrow bandwidth of the
tunable output provide extensive new study areas.
Upon irradiation with light, a molecular anion may either detach an electron
(1.7), dissociate (1.8), or undergo photodetachment followed by dissociation (1.9):
A' + hv-A + e' (1.7)
A' + hv-B' + C (1.8)
A' + hv-B + C + e' (1.9)
Processes (1.7), (1.8), and (1.9) are illustrated in Figure 1.10 (a), (b), and (c),
respectively. In photodetachment spectroscopy, these processes are monitored as
function of the wavelength of the incident light. This can be accomplished by
detecting the decrease in the intensity of precursor ion, A' (processes 1.7, 1.8, and
1.9) or an increase of successor B' for process 1.8 or the detached electrons for all
three processes. The related technique of negative ion photoelectron spectroscopy
analyzes the energy of detached electrons produced by the interaction of a fixed

Figure 1.10. Photodetachment process: (a) A' + hv A + e', (b) A' + hv B' + C, (c) A" + hv B + C + e\

-c-
o

41
frequency light and an ion beam. Photodetachment spectroscopy has been used
to determine a number of thermochemical and spectroscopic parameters for a
variety of atoms and molecules, including the electron affinity of the neutral
photoproduct, the energies and transition moments of electronically excited states
of anions, spin-orbit splitting in radicals, and vibrational and rotational spacings in
both neutrals and anions. Most anions of organic molecules exhibit only
photodetachment; a number of inorganic anions can also photodissociate upon
irradiation with visible light. Photodetachment experiments are useful in probing
excited states of anions, as well as providing information important in determining
binding strengths.
The use of electron affinities derived from either photodetachment
experiments or literature data to determine bond dissociation energies is applied in
this work. By measuring the depletion of the parent ion as a function of photon
energy, the photodetachment threshold, related directly to the binding strength of
solvated ions, can be determined.
The raw data in a photodetachment experiment are the measurements of the
precursor ion signal with and without laser irradiation. At long wavelengths, the
photon energy is not high enough to detach an electron from the precursor anion
so the intensity of the precursor ion remains unchanged, i.e. the fraction of
photodetachment is zero. When the photon energy is high enough, the intensity of
the precursor ion decreases and the fraction of photodetachment is nonzero,
increasing linearly near threshold. In order to convert these data into a

42
photodetachment spectrum, the observed fractional signal decrease must be related
to the photodetachment cross section. The probability of photodetaching an
electron from a negative ion is described by,90
P(A) = 1- = 1- exp( A x t)
(1.10)
Where N is the ion signal with irradiation, N0 is the ion signal without irradiation, t is
the duration of the laser beam pulse, and kp is the rate constant for
photodetachment given by
kp(A) = Jfo(A)-p(A )d\
(1.11)
f is a unitless factor related to geometric overlap between the ion cloud and the
photon beam, o(\) is the cross section for photodetachment in cm2, and p is the
photon flux (a function of wavelength) in photons cnr2 s'1. Assuming a(\) is constant
over the narrow laser bandwidth and t is kept constant, substituting eq 1.11 into eq
1.10, rearranging, and taking logarithms yields an expression for the cross section
in the terms of the quantities measured during a photodetachment experiment:91
N = N0 exp(- o(A) t\f p(A) d\)
(1.12)
tJf-p(A)dA
(1.13)

43
Photodetachment spectra consist of the wavelength dependence of this cross
section. The geometric factor may be evaluated from the beam diameter, photon
flux, and the size of the ion cloud, if absolute cross sections are needed.92
However, for a given photodetachment spectrum, all parameters in the denominator
of eq 1.13 are constant, if the laser flux is kept constant. Thus, the relative
photodetachment cross-section spectrum, which is sufficient for determination of the
photodetachment threshold, may be obtained by plotting ln(No/N) versus the laser
energy, E (A).
In a well known paper published in 1948, Eugene Wigner derived the form
of the photodetachment cross section for energies near threshold.93 Valid for final
states with two particles whose longest range interaction is the centrifugal barrier,
his result gives the variation of the cross section with energy, but not its absolute
magnitude. The law is given by,
o(E) (AE)I+1/2 or o(E)~kl+1/2, (1.14)
where o is the cross section, AE is the energy of the departing electron (the energy
above the minimum required for reaction to occur), and k = (2mAE)/j / h and I are the
linear and orbital angular momenta of the outgoing electron. The essence of the
Wigner derivation is that, at threshold, the dynamics are solely governed by the

44
longest range potential interaction between the electron and the neutral atom.
In eq 1.14, the interaction is the centrifugal potential, proportional to 1(1+1 J/r2.
In contrast to photoionization, where the long-range potential falls off as 1/r due to
the Columbic force between electron and cation, the strongest long-range electronic
interaction in atomic photodetachment is a centrifugal force contributed by the
1(1+1 Vr2 term contained in the Schrodinger radial wave equation. Departures from
the law can occur above the threshold, where shorter-range terms in the potential
become important or when forces with the same range as the centrifugal force are
active. Examples of such forces are electron-quadrupole and electron-dipole
interactions, respectively. For atoms, the usual dipole selection rule Al = 1 applies,
and I in eq 1.14 can be related to the angular momentum of the atomic orbital from
which the electron was detached. The extension of the Wigner result to include
polyatomic anions was accomplished by Reed et al.94 In general, the behavior of
the cross section near threshold for a polyatomic anions is given by,
o(E) = IArEk'*+*[1+0(Ek)], (1.15)
where f is the smallest integer value I for which Ar is not identically zero for
symmetry reasons. To perform the algorithm, one finds the irreducible
representation of the anion's HOMO for the point group of the anion. One then
assigns a value I' based on that irreducible representation. For example, I' = 0 if the
irreducible representation is totally symmetric; I' = 1 if it transforms as x, y, or z; and

45
I' = 2 if it is labeled by xy, xz, yz, x2-y2, or z2. The allowed I* values for
photodetachment are then these I' 1. Since the A" representation transforms as
Z in the Cs point group for alcohol solvated halides, the HOMO has I = 1 and the
selection rule Al = 1 gives I* = 0. Therefore, from eq 1.15 we obtain,
(E) Ek1/2 or In (N0/N) = a(E-ET)y>, (1.16)
The threshold, ET, can be obtained using an iterative procedure in which the
parameters, a (a scaling factor) and ET are optimized, to obtain the best fit to the
photodetachment spectra.
CAD
Energy behavior of reaction cross section
The maximum reaction cross section for an exothermic ion-molecule reaction
has an energy dependence that is determined by the long-range interaction
potential. For polarizable molecules without dipole moments, this is given by the
Langevin-Gioumousis-Stevenson (LGS) expression,95
Olgs(e) = TTe (2a/E),/j, (1.17)
where e is the electron charge, a is the polarizability of the neutral reactant, and E
is the relative kinetic energy of the reactants. Higher order terms in the long range
interaction potential can also be included.96

46
For endothermic reactions, microscopic reversibility can be applied to eq 1.17
to yield,97
o(E) = o0(E-Et)*/E, (1.18)
where ET is the reaction endothermicity. The scaling parameter is defined as o0 =
ne (2a)/j (psgs / p pg p), where p is the reduced mass and g is the electronic
degeneracy of the precursor (p) or successor (s). Although it would be nice if eq
1.18 accurately predicted the energy dependence of all endothermic ion-molecule
reactions, this is not observed experimentally.54'55 57a Indeed, reactions that conform
to eq 1.18 are exception rather than rule, as described later.
Alternate procedures for determining the energy dependence of reactions
include trajectory calculations, statistical theories, and empirical theories. Trajectory
calculations are not generally useful because they require a potential energy
surface, which is rarely available. Statistical theories such as phase space theory
(PST) in which energy and angular momentum are explicitly observed,98 are much
more useful.99'102 Empirical models are the most general means of describing the
energy dependence of reactions, and in this work, we use,
o(E) = o0(E-ET)7Em, (1.19)
where the scaling parameter o0 (with units of area-energy"1") need not be given by

47
the expression above. This model includes the model described by eq 1.18 as well
as most other theoretical predictions for endothermic cross sections.103
Unfortunately, even for atom-diatom reactions, there are many theoretical predicted
values of n and m, with no single form that describes the energy dependence of all
reactions.
Although eq 1.19 can adequately represent the kinetic energy dependence
of the reaction cross sections, it does not explicitly include the internal energies of
the ion or neutral reactants. This can be accomplished by modifying eq 1.19 to,
(E) = Eo0g¡ (E+Er+ErET)n/Em, (1.20)
where the subscripts r and i refer to a rotational and vibrational states; Er or E¡ is the
energy of that state; g, is the statistical population of the ith state; and the sum is
over all states. ET is thus the reaction endothermicity corresponding to 0 K. Eq
1.20 is the model which is used to determine the CAD threshold in Chapter 4.
On-resonance CAD
On-resonance CAD techniques have been demonstrated to measure BDE's
in FTICR mass spectrometry in several groups.89104 105 A short radio frequency (r.f.)
pulse is applied to the excite plates on an ICR cell to increase the translational
energy of a mass-selected precursor ion. The center-of-mass (c.m.) translational
energy imparted to an ion during the excitation in FTICR is given in eq. 1.21

48
M
(eV)
(1.21)
8m M+m
where Ecm and E,ab are the energies in the center-of-mass and laboratory frame,
respectively, M is the collision target molecular mass, m is the precursor ionic mass,
t is the r.f. pulse length, Erf is the electric field (which is giving by Vpp Id for an ideal
infinitely long parallel plate cell, where Vpp is the applied r.f. voltage (peak-to-peak))
and d is the plate spacing diameter of the ICR cell. Unfortunately, the measured Erf
for cells of finite dimensions is only a fraction of the ideal value, so it is calculated
using aVpp Id, where a is a geometry factor for the ICR cell (a=0.90 for the Infinity
cell used in this work).106 The activated ion is then allowed to interact with the
collision gas (Ar or Xe) for a short period of time (CAD delay time). The translational
energy is converted into internal energies such as rotational and vibrational
energies via ion/molecule collisions, allowing dissociation to occur during the CAD
delay time.
- L + T (ML)" + T->M' + L+ T
(1.22)
where M+-L is the ion-ligand complex, T is the collision target gas (Ar), M+ is the ion,
and L is the neutral ligand. After dissociation, the intensities of the fragment ions
are measured. The ratio of the intensity of the product ion to the total ion intensity
is determined. From plots of the intensity ratio (lM7(lM++ lML+)) vs. the center-of-mass

49
interaction energy, Ecm (eq 1.21), the threshold for bond dissociation can be
determined using a deconvolution process and eq 1.20.
The CAD process is initiated by ion-molecule interactions which are governed
by long-range attractive potentials. These potentials are sufficiently strong to
overcome small activation barriers. Thus, endothermic ion molecule reactions
should generally proceed if the available energy exceeds the thermodynamic
threshold for product formation.
Doppler broadening
The kinetic energy in the center-of-mass frame defined earlier does not
include the thermal motion of the reactant neutral. These molecules have a
Maxwell-Boltzmann distribution of velocities at the temperature of the ICR cell,
usually 300 K in this work. The effects of this motion are to obscure sharp features
in the true cross sections, especially the threshold for reaction. This so-called
Doppler broadening was first described in detail by Chantry.107 As will be discussed
in chapter 4, this energy distribution and that of the ion are explicitly considered by
using them to convolute a trial function for the energy dependence cross sections
(Eq 1.20 in our studies). Only after this convolution is accomplished is the model
compared to the experimental data. To achieve the best reproduction of the data,
the various parameters in the model (E0 or ET, o0, and n) are allowed to vary and are
optimized by using nonlinear least-squares methods. In this work, the value of m
was set to one, because this model always gave a threshold energy very close to
the average value of all other fits considered in Armentrout's work. This assignment

50
is further justified on the basis of a theoretical model for translationally driven
reactions108 and empirically by the ability of this model to accurately measure
thermochemical quantities in this system discussed later in Chapter 4.
Advantages. Disadvantages and Comparisons to Other Techniques
Photodetachment
The BE measurements using thermal-equilibrium methods are relative rather
than absolute values which can be obtained using the photodetachment technique.
All the equilibria measurements are based on the temperature dependence of gas
phase equilibria (n,n-1) observed in a high pressure mass spectrometer ion source.
The equilibrium constants, Kn_1>n, are determined. Then aG0 is calculated from Van't
Hoff plots of equilibrium constants. The BDE's determined from the linear plots
have a certain error due to the uncertainty of the slope and intercept. The accuracy
of the temperature measurement is also unreliable. The photodetachment
technique provides very accurate and straightforward threshold determination due
to the wide wavelength range, high resolution (very narrow bandwidth) and powerful
output of the laser. However, in FTICR experiments, the depletion of the precursor
ion is the only signal which can be monitored directly since the ICR instrument can
not detect neutral products and electron detection is not straightforward. An
alternate way to detect electrons in the ICR cell is use of an electron scavenger for
electron capture, such as CCI4, SF6. Unfortunately, electron scavengers can not be
used with solvated halides. The ion formation involving ion-molecule reactions with

51
halide compounds which is described detail in Chapter 3 is excessively complicated
by CCI4 and SF6. On the other hand, photoelectron spectroscopy (PES) can obtain
the same information by measuring the electron kinetic energy at a fixed laser
wavelength.
Photodetachment threshold spectroscopy and photoelectron spectroscopy
have been analyzed to obtain electron affinities and molecular structural
information, and each has its particular strength and weaknesses.
Photodetachment can be utilized to determine electron affinities with an accuracy
limited essentially by the wavelength resolution of the light source, while similar
determinations using photoelectron spectroscopy studies require only that the
wavelength of the intense light source be shorter than the wavelength
corresponding to the photodetachment threshold. As compared to CAD and
thermal-equilibrium techniques, detachment thresholds obtained using
photodetachment are much sharper and can be measured with greater accuracy.
CAD
The use of CAD techniques in FTICR-MS to measure thermodynamic
quantities for a variety of ionic species provides reaction cross section (or rate
constants) over a wide range of kinetic energies. This technique does not require
a laser apparatus and associated laser alignment compared to photodissociation
and PES techniques. CAD is an alternate approach to determining BDE when the
ions of interest do not absorb or absorb little laser energy at near threshold ranges.
However, the accuracy of such information is dependent on the models used to

52
analyze the cross sections and on accounting for effects such as the translational
energy distributions of the precursor ion, the internal energies of the precursor ion,
the possibility of multiple collisions and the radiative lifetimes of the intermediates.
One of the powerful aspects of using ion-molecule reactions to study
chemistry is the ease with which the kinetic energy of the ion can be varied over a
very wide range. Although this capability does not see extensive use, it is a very
useful tool for probing the potential energy surfaces of reaction systems and, in
particular, for measuring the heights of barriers to reaction. One application of this
technique to measure thermochemistry of both ionic and neutral species is the
determination of bond dissociation energies.

CHAPTER 2
INSTRUMENTATION
Introduction
The complex electronics instrumentation used in FTICR mass spectrometry
is presented in Figure 2.1 as a simplified block diagram. Essential equipment for
a Fourier transform ion cyclotron resonance mass spectrometer consists of a
superconducting magnet, a high vacuum system, a trapping cell, and a console to
detect ions, store signal and process acquired signal. Ionization sources may be
categorized as internal or external. The internal ionization source is usually located
inside the magnetic field (at the entrance of a trapping cell), produces better signal-
to-noise (S/N) ratio, and does not require a differential pumping set-up to maintain
low base pressure. The external ionization source is located outside the magnetic
field and operates at higher pressures (ca. 10~5 torr for El and MALDI, 760 torr for
ESI and glow discharge (GD)). It requires several stages of differential pumping
before ions enter the trapping cell region and also electrostatic or rf potential lenses
to transfer the ions into the cell. The development of external sources has allowed
the analyzer cell region to remain clean and at low pressure; therefore, the
resolution is improved. Also, since the external source region can be isolated from
the cell region, it is more flexible to switch between different ionization techniques
53

Figure 2.1. Simplified block diagram of FTICR mass spectrometer.

L
ICR
Cell
Sweep
Generator
Disk
Plotter
Display

56
in half an hour or less. This chapter will describe two ICR instruments with internal
and external ionization sources. The cell and ESI design used for this work will also
be described. The experimental configurations will be briefly introduced and then
discussed in detail in the Experimental sections of Chapter 3 and Chapter 4 for both
photodetachment and CAD techniques, respectively.
Nicolet 2 Tesla ICR Instrument
The FT-ICR mass spectrometer used for photodetachment studies (Chapter
3) was equipped with a 2 tesla superconducting magnet, a vacuum system (Figure
2.2) and a Bruker Aspect 3000 console with 512K memory. The vacuum chamber
had 8" flanges on either end, one with three optical windows for transmission of
laser radiation, the other with one window which allowed laser access to the
filament side of the cell. This arrangement permitted easy alignment of the laser
beam.
Figure 2.3 shows the laser set-up for this work. The laser beam was
reflected into the center window of the three window flange by a quartz prism. The
laser was externally triggered from the console. The console sent a TTL trigger
pulse just before the electron beam pulse to an Apple II computer with a John Bell
Co. interface card. Unless triggered, the computer and interface card sent both
"charge" and "fire" commands to the Nd3+:YAG laser so that the flashlamps fired at
a rate of ca. 9 Hz. The computer shifted to an interrupt routine which sent not only

Figure 2.2. 2 tesla ICR instrument (a) 2 T magnet, (b) Inlet source, (c) Three window flange, (d) One window flange,
(e) ICR cell, (f) Diffusion pump, (g) Ion gauge.

58

Figure 2.3. Laser set-up for photodetachment experiment (1) Apple II microcomputer controls "charge" and "fire"
commands of YAG laser, (2) Console sends external TTL pulse to trigger the laser, (3) Power control from console
to the cell; (a) Quartz prism, (b) Laser quartz windows, (c) ICR cell, (d) 2 T magnet, (e) Apple II computer, (f) Console,
(g) Photodiode.

532 nm
3+
1/
Nd : YAG Laser
Console
f
Doubler
Dye Laser
a
4
g
n.
TV
(3)

61
"charge" and "fire" but also a "Q-switch" pulse which caused the laser to fire when
interrupted by an external trigger pulse. The time from the rising edge of the TTL
trigger pulse to the laser firing, which was monitored by an oscilloscope using a
photodiode, was adjusted to match the FTICR pulse sequence so that the laser
was fired right before ion excitation. The wavelength calibration of the dye laser
was performed using the output of a He:Ne laser by a Continuum field service
engineer immediately before this project was begun. The laser has been used
solely for this research since the calibration was done. The wavelength was
checked using a monochromator ( 0.2 nm) each time when a different dye was
used. The pulse sequence used for this work is depicted in Figure 2.4.
The modified vacuum system, which was pumped by a 700 Us diffusion
pump and a mechanical pump as a forepump, maintained the background pressure
below 10'8 torr. The partial pressures of the three reagents used for the preparation
of the solvated anions were controlled by a pulsed valve and three precision leak
valves on the inlet system; this set up allowed very reliable control of the partial
pressures which is critical for these experiments.
A cylindrical cell (2.45 inches diameter x 3.0 inches length) was used for the
photodetachment experiments (as depicted in Figure 2.5). The fine mesh at the
filament side of trapping plate was replaced with a coarse mesh to increase
transmission of the laser beam and to allow the desired ions to be trapped without
ionization of the metal by the laser beam. Ions were formed by internal electron
impact (El), with electrons produced by a filament located in front of the trapping

Figure 2.4. Pulse sequence for photodetachment experiments.

Quench pulse
Electron impact pulse
Ion-molecule reactions
n
Ejections
Laser fire
Excitation pulse
Detection
Time

Figure 2.5. Cylindrical cell.

65

66
plate. Because the volume of the cylindrical cell is much larger than that of
conventional cubic cells used in our laboratories and elsewhere, space charge
effects are reduced. The trapping potential was kept as low as possible (1.2-1.4 V)
to prevent high velocity ions from being trapped.
Radiation in the 260-350 nm range after frequency doubling was provided by
a tunable dye laser (Quantel TDL 50) with a frequency-doubling accessory. This
dye laser was pumped by 532 nm (second harmonic of YAG fundamental) output
from a Nd3+:YAG laser (Quantel YG 581C), which was synchronized with the mass
spectrometer using laboratory-built interface electronics.109 The wavelength scan
was controlled by commercial software provided with the dye laser. The wavelength
accuracy of this laser system is 0.5 A, and the resetting stability is 0.02 A; the
laser linewidth is 0.08 A. Five different dyes (Fluorescein, Rhodamine 590, 610,
and 640, and DCM) were used and the provided wavelength range for each dye is
listed in Table 2.1. The laser output power (ca. 8-15 mJ/pulse) was kept constant
for each photodetachment spectrum.
Bruker 7 Tesla ICR Instrument
All the CAD experiments were performed on a Bruker FTICR mass
spectrometer (Bruker Analytical Systems, Billerica MA) equipped with a Spectrospin
(Fllanden, Switzerland), 7.0 T superconducting magnet, an RF-shimmed infinity

Table 2.1. Laser dyes used for photodetachment experiments and their corresponding wavelength range before and
after frequency doubling.

Dyes and the Corresponding Wavelength Range Before and After Frequency Doubling
Wavelength
range
Fluorescein
Rhodamine 590
Rhodamine 610
Rhodamine 640
DCM
Before
doubling
520-575 nm
550-590 nm
581-607 nm
567-640 nm
620-696 nm
After
doubling
260-287.5 nm
275-295 nm
290.5-303.5 nm
283.5-320 nm
310-348 nm

Figure 2.6. 7 T ICR instrument (a) 7 T magnet, (b) Infinity cell, (c) 400 L/s cryopumps, (d) 800 L/s cryopump, (e)
Mechanical pumps, (f) Capillary coated with platinum at entrance and exit ends, (g) Needle, (h) Syringe pump.

Transfer region
ESI source
"4
o

71
cell with a 2.45 inches diameter, a vacuum system, and an Aspect 3000 console.
The vacuum system consisted of an external ion source, ion transfer optics, and cell
regions which were pumped by 800 I/s (N2), 400 I/s (N2), and 400 I/s (N2)
cryopumps, respectively (Figure 2.6). Ions were formed using an Analytica
(Branford,CT) high-pressure ESI source. The ESI source region was pumped by
two mechanical pumps and a 800 I/s (N2) cryopump to give pressures in the Iow10'5
mbar region. The ions were transferred to the analysis region via a series of
electrostatic lenses, pumped by a 400 I/s (N2) cryopump. The analyzer region was
maintained by another 400 I/s (N2) cryopump in the low 10'10 mbar range, as
indicated by a cold cathode gauge.
Electrospray Ionization (ESH Source Region
Electrospray ionization (ESI) has became as the premiere ionization method
for LC-MS analysis.110 The main reason for its success is the proven ability of ESI
to simultaneously desolvate and ionize fragile polar analytes suspended in a liquid
matrix at atmospheric pressure. This process can then be followed by mass
analysis in the gas phase carried out under vacuum conditions.
The concept of ESI is simple. As an electrical potential is applied to a
conductive needle through which liquid is flowing, a "spray" forms which disperses
the liquid through the formation of a Taylor cone. This dispersion creates tiny liquid
droplets (~ 3 pm in diameter) which are electrically charged. These charged
droplets then quickly evaporate to dryness.111'114 Figure 2.7 shows the ion formation
process in ESI. The work of Iribarne and Thompson is the most accepted

Figure 2.7. Ion Formation in ESI.

Charge Analyte
Needle
evaporation explosion
Capillary
o@-
+
Ion 5 KV
evaporation 175 C
-j
CO

74
description of the ESI mechanism.115 According to their research, the ion
suspended in solution evaporates into the gas phase when the shrinking droplet
reaches the "Rayleigh limit" where the energy from repulsion of the charge on the
droplet surface exceeds the energy of the surface tension which holds the droplet
together. The resulting "coulombic explosion" tears the droplet apart and produces
even smaller droplets with an ehigher charge-to-mass ratio. It is from these smaller
droplets that ions actually "evaporate" from the liquid phase into the gas phase.
After the charged droplets enter the capillary, the evaporation process is
controlled through the application of electrical fields and a heated "counter-current"
drying gas, which helps to evaporate the droplets. The ESI chamber is surrounded
by a metal screen so that the spray itself can be seen during the ionization process
(Figure 2.8). This region is always maintained at atmospheric pressure. Inside the
spray chamber there are two electrodes, excluding the needle itself which is
normally maintained at ground potential to protect the user from the possibility of
electric shock. The largest of these electrodes is the transparent cylindrical
electrode (V^,), a wire screen which surrounds the bulk of the spray chamber. The
final electrode in the atmospheric region of the source is the capillary (0.5 mm i.d.)
entrance or V^, which is held at ca. 4kV, in order to form charged droplets between
the needle and the capillary. This electrode is fabricated by coating the face and
side of the quartz dielectric capillary with metal (gold or platinum). The exit end is
also coated with metal so that the capillary entrance and exit can be used
independently as electrodes to help guide the ions to and out of the capillary. On

Figure 2.8. ESI design.

External ESI Source
M. P.
~vl
O)

77
the vacuum or exit side of the capillary, the ions are transferred through several
skimmers (0.8 and 1.0 mm diameters) and lenses with pumping stages in the ESI
source prior to entering the transfer region. The manipulation of all these voltages
is important as they shape the electric field which causes the electrospray to form
and guide the ions toward the capillary so they may be transported into the vacuum
region of the source.
At the back end of the capillary where the ions emerge, a supersonic
expansion takes place between the capillary exit and the skimmer. The capillary
exit itself acts as a lens (L1) and so has a variable potential. The region between
the capillary exit and first skimmer (L2) is differentially pumped to allow for the
passage of as many ions as possible while reducing the gas pressure as well. After
this skimmer, there is another differential pumping region, where another lens
element, L3, and a second skimmer, L4, are present. The final lens elements were
L5 and L6. This pressure region is "open" and therefore pumped by the pump in the
source region of the mass spectrometer vacuum system. For this reason, L5 and
L6 are visible from the outside of the source. L6 is the last controllable lens before
the mass analyzer optics. Once the ion has been formed, it is transferred from
atmospheric pressure, where the ionization process occurs, to the mass analyzer
which is under vacuum through differential pumping stages.
Introduction of ESI Features
The efficiency of ionization in ESI is greatly affected by the chemistry of the
analyte with its supporting solution. The reason for this is the characteristic relation

78
between the analyte in the solution phase and the extracted analyte in the gas
phase when electrospray ionization is used. Electrospray ionization is a very "soft"
ionization technique, so minimal distortion of ionic structures, in both chemical and
physical terms, occurs during the process of analyte ionization and desolvation.
There are a number of important chemical properties of the analyte which
may affect the ion signal observed in ESI process. These include but are not limited
to (1) pKa or pKb116, (2) concentration, (3) solubility, (4) molecular conformation, and
(5) molecular weight. Properties of the supporting solution itself can also have
effects on the ion signal such as electric conductivity, surface tension, presence of
salts, acids and bases, solvent polarity and volatility and interference from other
ionizable analytes.
gKa or pKb
If a protonated base desorbs from the droplet surface during ionization, the
analytes which have higher pKa's will have higher concentrations of the protonated
forms of the bases, and therefore higher signals in ESI/MS.
Concentration
ESI functions as a concentration sensitive detector when operated within a
reasonable range of sample concentration. The range of concentration over which
a linear response from the MS can be obtained, or linear dynamic range (LDR), is
generally several orders of magnitude.116 At both high and low ends of possible
concentration levels, some deviation from linearity is expected to occur. The
formation of the cluster species is frequently observed in ESI-MS when small

79
compounds having finite solubilities are used. The analyte is actually precipitating
out in solution at higher concentrations, thus forming clusters on the original
species. The concentration of sample may actually lead to decreased ion signal in
the MS for the analyte of interest.
Solubility
As solubility decreases, the likelihood of precipitation, and thus the formation
of clusters, in solution increases. As shown above, this may lead to reduced ion
signal. For this reason, it is necessary to consider the solubility of the compound
of interest when choosing solvents and sample concentrations. Generally, ESI-MS
spectra cannot be obtained for very large proteins (especially membrane-bound
types) due to solubility issues. In these and other cases, usually a better solvent
can be used to increase signal.
Molecular conformation
Most evidence concerning ESI leads to the conclusion that the efficiency of
generating protonated ion is related to the pKa, as shown above. Also important is
the accessibility of basic sites which attract the charges. This feature is important
in larger macromolecules, such as proteins, which have tertiary and even
quaternary structural components. In such species, the available basic sites may
be "hidden" from the protons they seek.117 For this reason, it is sometimes
necessary to change the solvent system, by adding dilute acid for example, in order
to change the structure of the protein. In severe cases, the protein may be
subjected to enzymatic or chemical degradation in order to increase ion signal.

80
Molecular weight
The possible effect of molecular weight on ionization is the mass
discrimination118 which may occur in both the mass analyzer and the ion optics.
Usually by adjusting the voltage parameters of the ion optics one can observe the
effect of charge states for multiple charged ions.
Solvent properties
High conductivity solutions will not electrospray well under "normal"
conditions since the formation of a Taylor cone is prevented in this case.119
Solutions which are high in conductivity usually have acids, bases or salts added
to them for various reasons. When using these solvents for electrospray, it is
necessary to remove these additives to the sample, as the ion signal and stability
are strongly dependent on the quality of the spray generated and the formation of
the Taylor cone.
In a similar manner, solutions with a high surface tension are also difficult to
use with ESI. It is for this reason that pure water does not electrospray well. This
problem is generally overcome through the addition of an organic modifier, such as
methanol, to the sample solution. The supporting solution is at least 20% in organic
makeup to spray well. The accepted standard for optimum performance is 50%.
Finally, the solvent volatility may also slightly affect the ion signal observed in ESI
for a particular analyte. After a droplet has been formed by ESI, it must evaporate
substantially before the electric field on the droplet surface is high enough to induce
a coulombic explosion and subsequent ion evaporation of the analyte. This rate of

81
evaporation will depend on the solvent composition of the droplet. For this reason,
solutions with higher organic content may give higher signal since the droplets
formed will evaporate faster, leaving more time for analyte ionization to occur.
Ion Transfer Region
After formation ions are then accelerated toward the analyzer by a series of
electrostatic lenses (3 kV maximum potential between PL5 and PL 7) as shown in
Figure 2.9. After acceleration the ions pass x,y-deflection, flight tubes (FL1, FL2),
and PL9 at decreasing potential to relative ground prior to entering the cell region.
A set of deflection lenses is used to gate the continuous ion beam, so that ions
enter the analyzer region only during the ion accumulation period. Finally, the ions
are transported into the mass analyzer where their mass is determined. The gate
valve is used to isolate the analyzer region from the ionization source region for
more rapid changes of the ionization source to El, MALDI, or GD.
Ion Analyzer Region
The trapping cell used for CAD work is a rf-shimmed infinity cell as shown
in Figure 2.10. At the front of the ICR cell there were two "side-kick" plates with
pulsed low voltages (1-5 V) during the ion formation pulse. Ions entered the cell
with a kinetic energy identical to the first "side-kick" plate potential. After the ions
entered the cell, the voltages on the side-kick plates were dropped to the trapping
potential (0.5 -1 V). To help trap the ions, the second set of side-kick plates (with
two halves) is designed to convert the z-axis translational energy into the xy-plane
cyclotron motion. Once the ions were trapped inside the cell, the CAD experiments
were performed using a pulse sequence illustrated in Figure 2.11.

Figure 2.9. Electrostatic lenses in transfer region.

7T Optics
SI S2 L3
CO
PL1
JZD
PL2
X,Y Deflection
PL7,8 11
Gatevalve
PL9
PVX3
PV2
oo
oo

Figure 2.10. Infinity cell with three "side-kick" plates: (a) PVX1, (b) PVX3, (c) PVX4 plates, (d) the macor plate which
is attached to trapping plates to hold the "side-kick" plates and trapping plates, (e) the rf-shimmed trapping plate, (f)
the Infinity cell with four brass plates for excition and detection.

6 mm >1 k-^(a)
(c) nEnJaB
(b)
T
5"
.i;" .-k
i HHimanM
2.25"
I
' y .;-v-
r: ' s',* '.^-iSShfr-
0
0
0
0_ ._c> 0 0
2.625"
-N
Brass plates
(e) rf shimmed
trapping plate
(front view)
00
cn

Figure 2.11. Pulse sequence for CAD experiments.

87
Quench 5 msec
Injection of Ions Formed by ESI -100 msec
Thermalizafjon 2 sec
nn
Ejections
Activation 0-50 psec
CAD delay -10-30 msec
Excite
Detect
Time

CHAPTER 3
DETERMINATION OF BINDING ENERGIES OF SINGLY-SOLVATED HALIDE
IONS USING THE PHOTODETACHMENT TECHNIQUE
Background
Fundamental studies of the thermodynamics of solvation are important for
several reasons. First there is the chemistry of the solvation process itself. A free
molecule or ion becomes surrounded by solvent molecules in the bulk medium.
This results in changes in the interactions between solvent molecules, as well as the
formation of new solute-solvent interactions, including hydrogen bonds in protic
systems. Solvation is often the key factor in relative solubility effects, which are
essential in many types of chemical separations. Also of practical significance is the
role of solvation in chemical reactions. Since many reactions take place in solution,
the solvent often has a pronounced effect on the reaction rate and position of
equilibrium. Although the optimum solvent for a particular application may be
identified by conducting a series of reactions, it would be more expedient to predict
such effects from knowledge of the solvation of the reacting species. However,
because of the accompanying changes in solvent structure and the solvation of
counterions, it is often difficult to obtain meaningful data from measurements of the
88

89
bulk materials. It is, therefore, desirable to study the thermodynamic aspects of
solvation, in particular the binding energies of ion-solvent adducts, in the gas phase,
where there is no interference from other species.
The binding of cations and anions to neutral molecules has been investigated
by a variety of mass spectrometric techniques,120'141 and solvated ions such as the
water and alcohol-solvated halides have also been heavily studied.125'141 These
investigations have provided information about the stability of the hydrogen bond
and the effects of structural variations in these compounds. Work involving ions
surrounded by solvent clusters128'133 has also led to a better understanding of the
characteristic properties of bulk solutions.
Ion solvation energies have been measured by several methods as
described in the Introduction section.41'58 Although a large volume of
thermodynamic data is available, there is still considerable disagreement in the
literature about the binding energies (BE) of solvated anions evaluated by different
methods such as equilibrium measurements in HPMS129'134 and ICR-MS.137'141 This
may be attributed to the use of equilibrium methods, which are often highly
uncertain and may not give direct measurements of the binding energies. Kebarle
and coworkers132 have made extensive use of ion-solvent cluster reactions,
represented by eq 1.6. The equilibrium is established in the ion source of a high
pressure mass spectrometer, and the equilibrium constant, K^.,, is determined from
intensity measurements of the two anions. The enthalpy change can be determined
by repeating the measurements at a series of temperatures.129'131138 Another

90
equilibrium method44'124'126'137141 is a comparative procedure utilizing data at a single
temperature for a series of bimolecular exchange reactions of general form
S.X-+ S2 S2X + S1( AG = RT In K (3.1)
This gives a relative value of the free energy of solvation within the series of
investigated solvent molecules. The enthalpy change is then determined from the
free energy after estimation of the entropy change.
These equilibrium approaches are all prone to large uncertainties. For
example, Hiraoka and Yamabe130 used both a gas-phase solvation reaction (eq 1.6)
and an ab initio MO calculation with a 3-21G basis set to determine the BE of
(CH3OH)F', and obtained results of 23.3 (10-2.0) and 26.2 kcal/mol, respectively.
Using the bimolecular exchange method (eq 3.1), Larson and McMahon137(a)
obtained 29.6 (0.5-10) kcal/mol for the BE of the same anion. There is a
difference of at least six kcal/mol in the two experimental results, and neither agrees
with the ab initio calculation.
An alternative method for determining binding energies makes use of a
photodetachment reaction. To a first approximation, the photodetachment process
may be represented by:
+ X + e\
(3.2)
However, as described further under Results and Discussion, consideration of an

91
intermediate loosely bound neutral complex may be necessary. The threshold for
photodetachment is obtained by measuring the depletion of SX as a function of
photon energy. This value can be used to calculate the BE of SX', if the electron
affinity (EA) of X is known (either from established values or the photodetachment
threshold of unsolvated X ). Brauman et a/.142"143 have used this method to evaluate
the EA's of several systems including organophosphines and aliphatic oxyl radicals.
Similar measurements of photodissociation reactions (similar process to eq 3.2, but
with the charge remaining on one of the products, presumably X) provide data on
the BE's of solvated cations. For example, Brucat et at.'22 used photodissociation
in a TOF mass spectrometer to obtain the upper limit of the ground-state adiabatic
dissociation energy of (H20)V+. The binding energies of Mg(FI 20) n + (n1) were
determined in Fukes120 and Duncan's121 groups by the same technique. Johnson
et at.'27 also used time-of flight (TOF) mass analysis to demonstrate the competition
between photodissociation and photodetachment in hydrated electron clusters,
(Fl20)n\ They observed that photodetachment is the only process at high excitation
energies. Flowever as the photon energy decreases, photodissociation begins to
compete with photodetachment.
Ion cyclotron resonance mass spectrometry is especially desirable for
photodetachment studies, because ions may be trapped for extended periods and
subjected to laser irradiation. Thus, it is feasible to form solvated anions by suitable
ion-molecule reactions, and to measure their BE's by the photodetachment process.
This report presents results on a number of alcohol-solvated halide ions, (ROFI)X',

92
as well as (CH3CN)Br. Riveros et a/.144'147 have developed methods for producing
these adducts (as shown for specific cases in the Procedure section). These
preparation procedures have been used successfully in our laboratory148 to obtain
IR spectra of solvated halides by IR multiphoton dissociation (IRMPD)
measurements. In this work the binding energies of (ROH)X' and (CH3CN)Br have
been determined by the photodetachment method in the 260-350 nm wavelength
range.
Experimental Procedures and Data Analysis
Chemicals and Ion-Molecule Reactions
All chemicals were commercial products of spectral-grade purity, and were
used without further purification, except for several freeze-pump-thaw cycles. NF3
was purchased from Air Products and Chemicals, Inc., phosgene was obtained from
Matheson, optima-grade methanol was purchased from Fisher Scientific, and the
rest of the samples were obtained from Aldrich.
The solvated halide ions were formed by the methods of Riveros.144147 A
two-step procedure was used to produce the methanol-solvated fluoride and
chloride ions.144146 In the first step the pertinent reagent gas (NF3 for the F' adduct;
phosgene for the Cl adduct) underwent dissociative electron capture of 6-12 volt
electrons (Figures 3.1 and 3.2 for F' and COCI ):

Figure 3.1. Mass spectrum of F\

100
80
60
40
20
0
19.001
P
Jm
17.0
21.1
27.7 40.4
Mass-to-charge ratio
74.8
i
mMMMM
499.7
CD

Figure 3.2. Mass spectrum of COCI'.

100
80
60
40
20
COCI
62.959
I
Lili 111, IIlJiiI
O
. i ... i.i
i 1
,¡
ill
I w F *' T *T P m
1 TI T 11 un1' p rlT^ 'H
15.0
18.6
24.5
35.9
67.0
500.5
Mass-to-charge ratio
CD
O)

97
NF3 + e'-> NF2 + F'
(3.3)
COCI2 + e' COCI' + Cl
(3.4)
The anions subsequently reacted with a second reagent gas to form the desired
adduct (Figures 3.3 and 3.4 for (CH3OH)CI' and (CH3OH)F):147
F + HCOOCH3 (CH3OH)P + CO
(3.5)
COCI' + CH3OH (CH3OH)C|- + CO
(3.6)
The neutral carbon monoxide removed some excess energy resulting from
these exoergic reactions, but a relaxation delay of 2.5 to 4 seconds was still
required to allow the (CH3OH)F'or (CH3OH)CI' to approach thermal equilibrium prior
to the photodetachment measurements. For this work the thermalization time was
determined by measuring the photodetachment threshold with increasing delay
times (2-6 seconds), until no significant change was observed. Results of earlier
IR dissociation experiments involving (CH3OH)F' suggested that ca. 100 collisions
between the ions and the bath gas (methyl formate) were required for
thermalization.149'150
The preparation of (CH3OH)l' and the adducts of Br with CH3OH, C2H5OH,
/-C3FI7OH, n-C3H7OH, and CH3CN required an alternate method involving the gas-
phase exchange of (FI20)X' with the desired solvent. Initially, hydroxide ions from

Figure 3.3. Mass spectrum of (CH3OH)F\ m/z = 51.

100
80
60
40
20
0
(CH3OH)F-
51.029
Mfa
HMMu
17.0
t
21.1
27.7
40.4
74.8
499.7
Mass-to-charge ratio
CD
CD

Figure 3.4. Mass spectrum of (CH3OH)Cr, m/z = 67, 69.

100
80
60
40
20
0
17.0
21.1
(CH3OH)CI
66.991
68.992
27.7 40.4
Mass-to-charge ratio
74.8
500.5

102
dissociative electron capture by water reacted with bromo- or iodobenzene to
produce the water adducts (Figures 3.5 and 3.6 for OH" and (H20)Br):
H20 + e" OH" + H (3.7)
OH" + C6H5X (H20)X- + C6H4 (3.8)
Subsequently, H2OX" was allowed to exchange with the solvent (Figure 3.7 for
(CH3OH)Br):
S + H2OX SX" + H20 (3.9)
A period of 2 seconds was needed for the final reaction to reach equilibrium.
Because the SX" ions were already thermalized, no further relaxation delay was
required.
For the electron capture reactions of NF3, COCI2l and H20, 12.0, 6.0, and
10.0 volt electrons were used, respectively, and the gas pressures were 4.0-5.0 x
10'8 torr. The total pressure was 8.0 x 10"8 torr for the production of (CH3OH)F" and
(CH3OH)CI\ and 1.5 x 10"7 torr for the iodide and bromide adducts. The higher
pressure in the latter experiments also helped to assure that the I" and Br adducts
were properly thermalized.

Figure 3.5. Mass spectrum of OH', m/z =17.

100
80
60
40
20
0
17.005
OH'
16.4
28.1
43.9
100.0
Mass-to-charge ratio
104

Figure 3.6. Mass spectrum of (H20)Br, m/z = 97, 99.

100
80
60
40
20
0
96.913
(H20)Br
98.901
.3.5
16.8 22.2 32.5 61.1 500.5
Mass-to-charge ratio
106

Figure 3.7. Mass spectrum of (CH3OH)Br, m/z = 111, 113.

100
80
60
40
20
0
(CH3OH)Br
110.923
112.957
.5.0
18.6
24.5 35.9
Mass-to-charge ratio
67.0
500.5
o
00

109
Pulse Sequence
As shown in Figure 2.4, after a quench pulse removed all the ions from the
cell, the TTL pulse was sent to externally trigger the laser. A 100-150 ms pulse of
6-12 volt electrons was generated for dissociative capture by the first reagent gas
(eqs 3.3, 3.4, and 3.7). This was followed by a 2.5-4 second delay for the ion-
molecule reactions (eqs 3.5, 3.6, and 3.9), as well as thermal relaxation of the
product solvated ions, as described above. After the delay, a series of single
frequency ejections was generated to remove undesired ions. The laser was then
fired, followed by detection of the solvated ions using a standard chirp excitation
sweep. Data were acquired over a broadband frequency range corresponding to
m/z=15 to m/z=1000. One-hundred 32K point time-domain transients were signal-
averaged for all ions except (CH3OH)F' and (CF^OHJCI" (50 transients each). The
absolute intensities of the solvated ions were recorded under the same conditions
with and without blocking the laser beam.
Data Analysis
In a photodetachment experiment the ion signals, both with and without laser
irradiation, are recorded as a function of the laser wavelength. Because the
solvated anion is converted to neutral products, the ion concentration decreases
following the laser pulse. The observed signal ratio, N/N0, where N0 corresponds

110
to the unirradiated sample, must be related to the photodetachment cross section,
o(X), which is the ordinate of an absolute photodetachment spectrum.
The probability, P(A), of removing an electron from a negative ion by laser
irradiation is given by equation 1.10.90 The equations, 1.10 and 1.11, yield an
expression for the cross section in terms of the measured quantities:91
In
V
o(A)
N
/
(1.13)
tff-p(A)dA
For a given photodetachment spectrum, all parameters in the denominator of eq
1.13 are constant, if the laser flux is kept constant. Thus, the relative
photodetachment cross-section spectrum, which is sufficient for determination of the
photodetachment threshold, may be obtained by plotting ln(No/N) versus the laser
wavelength. At each wavelength setting the N and N0 measurements were
repeated 4 to 6 times, and the respective mean and standard deviation were
calculated. The standard deviations of the ln(N0/N) values were propagated from
the standard deviations of the data points, and these were used as error bars for the
photodetachment spectra.
At the threshold the cross section for optically detaching an electron depends
only on the long-range interaction between the two product particles. For
polyatomic anions151 with Cs symmetry, the cross-section is directly proportional to
the square root of the energy above the threshold, (E-ET),/2:

111
o(E) (E-Et)* (3.10)
or In (N0/N) = a(E-ET)% (3.11)
in which a is a constant. To evaluate ET an iterative procedure was used. First, a
plot of ln(N0/N) versus laser energy, E, was prepared, as shown in Figures 3.8,
3.9, 3.10, and 3.11 for (C2H5OH)Br, (CH 3CN)Br \ (CH 3OH)l\ and (CH 3OH)F \
respectively. An approximate ET (ET1), was chosen from the sloping part of the plot,
and the experimental data points (E, ln(N0/N)) were substituted into eq 3.11 to
obtain a series of initial values of a. The mean (a,) and standard deviations of the
points about the mean for the initial values of a were calculated. This procedure
was repeated for decreasing ETs (at intervals of 0.01 kcal/mol) until the minimum
standard deviation about an was obtained. The reported ET was taken as the ETn
producing this minimum standard deviation.
The curves shown in Figures 3.8 to 3.11 are the fit to the data using the ET
obtained in this manner. For each of these plots a least-squares fit was also made
for the points in the near threshold region, and the threshold was taken as the
energy corresponding to the x-intercept. The results from the least-squares fit were
very close to those of obtained using the iterative procedure described above. The
uncertainty in ET was evaluated from the standard deviations of the least-squares
slope and y-intercept.152

Figure 3.8. Photodetachment spectrum of (C2H5OH)Br: experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.

0 35 CoH.OHBr"

Figure 3.9. Photodetachment spectrum of (CH3CN)Br: experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.

In (N0/N)
0.175 CH3CNBr
E/kcaUmol'1
cn

Figure 3.10. Photodetachment spectrum of (CH3OH)l': experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.

E/kcalmol'
In (N/N)
o

o
o
o
o
b
Lx
k>
Ol
cn
cn
cn
III
_IHO HO

Figure 3.11. Photodetachment spectrum of (CH3OH)F': experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.

In (N0/N)
E/kcalmol"1

120
Results and Discussion
The relative photodetachment spectra for bromide ion solvated with CH3OH,
C2H5OH, /-C3H7OH, n-C 3H 7OH, and CH 3CN are presented in Figure 3.12. As
explained above, a delay period was required to assure that the gas-phase
exchange reactions for the preparation of the Br and I' adducts (eq 3.9) reached
equilibrium and that (CH3OH)F and (CH 3OH)CI' (eqs 3.5 & 3.6) were properly
thermalized. Figures 3.13, 3.14, and 3.15 show photodetachment spectra for
(CH3OH)F', (CH 3OH)CI, and (CH 3OH)Br', respectively, obtained with increasing
delay time. Studies of the bromide systems showed results similar to (CH3OH)CI\
The reaction of fluoride ion with methyl formate (eq 3.5) is very exothermic, and a
delay of about 4 seconds was required for collisional cooling of the (CH3OH)F ions.
The overall shift in the threshold (cf., Figure 3.13) from 272.0 nm to 264.4 nm
corresponds to a cooling of 87.6 K for the (CH3OH)F' complex. Because the newly
formed (CH3OH)CI' ions were less energetic, they reached thermal equilibrium in
about 2.5 seconds. The formation of (CH3OH)Br reqiures high pressure and
achievement of thermal-equilibrium, therefore, the 2.5-second delay was also
sufficient for the exchange reactions (eq 3.9) to reach equilibrium.
By a thermochemical cycle for the photodetachment process (eq 3.2) the
threshold is equal to the sum of the binding energy of the solvent anion complex
and the electron affinity of the halogen, with a small correction as described below.
Thus, the BE's may be calculated from the thresholds and the known EA's154 given

Figure 3.12. Photodetachment spectra: (CH3CN)Br; (CH3OH)Br; T (C2H5OH)Br; (i-C3H7OH)Br; X (n-
C3H7OH)Br.

In (N0/N)
298.0 303.0 308.0 313.0 318.0
Wavelength/nm
122

Figure 3.13. Photodetachment spectra of (CH3OH)F" obtained using different thermalization times (t): -1=2.34 sec.
X -1=3.20 sec.; -1=3.77 sec.; X -1=4.31 sec.; -1=6.05 sec.

In (N0/N)
260.0 264.0 268.0 272.0 276.0
Wavelength/nm
124

Figure 3.14. Photodetachment spectra of (CH3OH)Cr obtained using different thermalization times (t): -1=2.05 sec.
A -1=3.20 sec.; -1=4.34 sec.

In (No/N)
274.0 277.0 280.0 283.0
Wavelength/nm
126

Figure 3.15. Photodetachment spectra of (CH3OH)Br obtained using different thermalization times (t): -1=2.05 sec.
A -1=3.77 sec.; -1=4.34 sec.

In (N/N)
302.0 304.0 306.0 308.0 310.0 312.0
Wavelength/nm
128

129
in Table 3.1. Table 3.2 presents the thresholds (including a and ET) and binding
energies for all complexes studied in this work. Results of previous investigations
are also included.
This interpretation of the photodetachment threshold assumes that the
vertical transition proceeds to a repulsive state and not to a neutral bound complex.
Excitation to a bound state would result in the reaction,
SX'-V SX + e (3.12)
instead of eq 3.2. For these solvated anions, it is expected that dissociation to S
plus X (eq 3.2) occurs, as supported by several arguments. Based on Brauman's
experimental and calculated results,91153 the preferred reaction is expected to
depend on the Franck-Condon overlap of the ground state vibrational wavefunction
of the solvated anion with that of the neutral complex at the same geometry.
Considering S to be an alcohol, ROH, interacting with neutral X via the OFI group
(RO-H-X), the position of H between the O and X nuclei depends on the proton
affinities (PA) of RO' and X. If X' has a lower PA than RO', then the FI is closer to
the oxygen (RO'-FT-X), so that the equilibrium internuclear distance for the
interacting neutrals would be similar to that in the solvated ion. Thus, complete
dissociation of ROFI from X is feasible, and a vertical transition may be expected to
follow eq 3.2. On the other hand, if RO' has a lower PA than X', the FI is closer to
X (RO'-FT-X'). The Franck-Condon factor for eq 3.2 would be poor, and the

Table 3.1. Electron affinities (EA) and proton affinities (PA) of species studied in this
work.

131
Electron Affinities3 (EA) and Proton Affinities6 (PA) of Species Studied in this Work
Halogen
EA/kcal-mol'1
()=/kJ-mol1
Anion
PA/kcal-mol'1
()=/kJmol'1
F
78.37 (327.8)
OH
390.8 (1635)
Cl
83.44 (349.0)
ch3o-
381.4 (1595)
Br
77.45 (323.9)
c2h5o-
377.8 (1580)
I
70.53 (295.0)
i-C3H70"
376.0 (1573)
n-C3H70"
376.4 (1574)
t- C3H70
374.9 (1564)
ch2cn-
372.2 (1557)
f-
371.3 (1553)
ci-
333.3 (1394)
Br
323.6 (1353)
I
314.3(1315)
a-Ref. 154
b- Ref. 155

Table 3.2. Photdetachment thresholds (ET) and binding energies (BE's) of solvated
halides.

133
Photodetachment Thresholds (ET) and Binding Energies (BE) of Solvated Halides
This work
Other experimental results
Calculated results
ET/nm BE/kcal-mol'1
() ()=/kJ-mol'1
BE/kcal-mol-1
0=/kJ-mol'1
Method3
[Ref.]
BE/kcal-mol'1 Method [Ref.]
0=/kJ-mol'1 (basis set)
(CH3OH) F-
265.11.2
(1.71)
29.910.5
(125.112.1)
29.6 (123.8)
30.6 (128.0)
23.3 (97.5)
EX [137]
EQ [140]
EQ [130]
26.2 (109.6) 3-21 G
[130]
(CH3OH) Cl-
280.51.4
(2.18)
18.810.5
(78.712.1)
16.8 (70.2)
16.3 (68.2)
14.2 (59.4)
17.4 (72.8)
EX [137]
EQ [134]
EQ [131-133]
EQ [130,138]
16.2(67.8) 3-21 G
[130]
(CH3OH)Br
309.01.3
(1.82)
15.210.4
(63.611.7)
13.9 (58.1)
EQ [130]
16.0(66.9) 3-21 G [130]
12.9(54.0) MP4/SDTQ[141]
(CH3OH)|-
337.21.6
(1.31)
14.410.4
(60.311.7)
11.3 (47.3)
11.2 (46.8)
EQ [131-133]
EQ [130]
10.8 (45.2) 3-21 G
[130]
(C2H5OH)Br
308.811.9
(2.31)
15.410.6
(64.412.5)
14.4 (60.2)
EX [141]
(i-C3H7OH)Br
304.413.7
(2.32)
16.611.2
(69.415.0)
14.6 (61.1)
EX [141]
(n-C3H7OH)Br
303.811.3
(1.27)
16.810.4
(70.311.7)
(CH3CN)Br
315.313.2
(1.27)
13.410.9
(56.113.8)
12.7 (53.1)
12.9 (54.0)
EX [141]
EQ [138(b)]
11.0(45.9) MP4
[141]
(CH3CN)C|-
15.8 (66.1)
14.0 (58.6)
13.4 (56.0)
EX [137]
EQ [134]
EQ [131-133]
(CH3CN)f" 16.0(66.9) 3-21 G [129]
a. EQ = Gas-phase equilibrium (cf. eq 1.6 in the text)
EX = Bimolecular exchange (cf. eq 3.1 in the text)
mean scaling factor (cf. eq 3.11 in the text)

134
transition shown in eq 3.12 would be possible. The known PA's155 presented in
Table 3.1 follow the order: CH30 > C2H50' > i-C3H70' ~ n-C3H70' > F' > Cl' > Br >
I'. Since the halides all have lower PA's than the RO it is very likely that
photodetachment occurs to a repulsive state (eq 3.2).
Neumark et a/.156 studied the hydrogen abstraction reactions of halogens with
methanol and ethanol (F + HOR HF + OR) via photodetachment of ROFIF'
(R=CFI3, C2H5). Their calculated results for the potential energy surfaces are
illustrated in Figure 3.16. The solvated anions are photodetached to the rising
surface of the transition state, and subsequently revert to the dissociated neutrals.
Brauman et a/.91153 also used this model to interpret their photodetachment studies
of (ROH)X' {R=CH3l CgFIsCFI;;, X=F, OCFI3}. According to their ab initio calculations,
the structure of the solvated anion is essentially RO-FI-F, and vertical detachment
yields a neutral complex resembling ROH plus F. Consistent with the work of
Neumark, this species is localized on the rising potential energy surface between
the fully dissociated alcohol + halogen and the transition state.
Presuming this model to be correct, Figure 3.16 shows that the BE calculated
from the photodetachment threshold is higher than the true BE by the amount AE.
Typically these differences are very small and negligible for (CFI3OFI)F'.91/l57
Glauserand Koszykowski157 performed an ab initio calculation using the MP2/6-31G
(d,p) basis set for hydrogen abstraction of fluorine on methanol, and obtained a
barrier height of 1.89 kcal/mol for methyl-group attack and negligible height for
hydroxyl attack. Using these results, the upper limit of the absolute error of the

Figure 3.16. Potential energy curves for alcohol-solvated halide anions and
corresponding neutral species. The lower curve represents the interaction of the
solvent and halide from infinite separation to a stable bound state, a) The upper
solid curve represents the hydrogen abstraction reaction of ROH + X to form RO +
HX. (X = F and Cl). The arrows indicate the thermochemical cycle for the
photodetachment process, including the threshold (Ethresh0|d), electron affinity (EA)
of X, binding energy (BE) of (ROH)X', and the energy difference (AE) between the
reported and actual BE. b) Same as a) with dotted curve to indicate hydrogen
abstraction reaction for X= Br and I.

T.S.
136

137
CH3OHF' binding energy calculated from the photodetachment threshold is much
less than 0.5 kcal/mol. Thus, the measured BE for (CH3OH)F' is within experimental
error of the values obtained by McMahon's group137140 (Table 3.2), although the
experimental and calculated results of Hiraoka and Yamabe are several kcal/mol
lower.
As Table 3.2 shows, the BE's for the other methanol-solvated halides
obtained using this approach are about 1-3 kcal/mol higher than literature values.
This may be due to greater AE's for these systems, but preliminary calculations
indicate that the magnitude is very small for the alcohol solvates of F and Cl.
Because of the greater polarizabilities of the heavier halogens (Br and I), spin-orbit
coupling with the alcohol is expected. This would lead to a more gradually rising
potential energy surface, as shown by the dashed curve in Figure 3.16, and a
greater value of AE for complexes with these atoms. Measurement of the kinetic
energy of the separating neutrals, ROFI and X, at threshold, or extrapolations from
kinetic energies of photodetached electrons above threshold would allow AE to be
evaluated directly. Lacking such data, an estimation of the minimum height can be
obtained from available thermochemical data158 for neutral hydrogen abstraction
reactions (overall reaction depicted by the upper curve in Figure 3.16a and 3.16b):
Cl + CH3OH HCI + CH30 AH = 0.8 kcal/mol (3.13)
Br + CH3OH HBr + CH30
AH = 16.4 kcal/mol
(3.14)

138
I + CH3OH HI + CH30 AH = 32.6 kcal/mol (3.15)
The reaction barrier must be at least as high as the endothermicity of these
reactions. It is probable that the trend of the barrier heights may be similar to the
AH trend, i.e., much larger for Br and I than for F or Cl. It is generally true that, for
very endothermic reactions, the structure of the transition state is similar to that of
the products. However, the relative bond lengths (discussed below) indicate that
the neutral complex is very similar to the reactants. Thus, the complex must be
located very low on the rising surface, and the corresponding AE must be much less
than the overall barrier height. Given this argument, the value of AEs for Br and I
complexes should be greater than those for Cl complexes, but only by 1-2 kcal/mol.
As explained in the Introduction, there is some uncertainty in the BE values
obtained by equilibrium methods, although error bars are not given by the cited
authors. Examination of their data indicates a random uncertainty of about 1-2
kcal/mol in the reported BE's. Furthermore, binding energies determined by
equilibrium methods are subject to systematic error in the statistical mechanical
estimation of AS, which is based on the vibrational frequency of the H-X bond in
ROH-X'. As the halide becomes heavier, the vibrational frequency of the ROH-X'
bond decreases. It is possible that the vibrational frequency has not been
estimated correctly by previous authors. This would produce a systematic error in
the entropy change, which would subsequently affect the reported BE.
Results of ab initio calculations support the relative bond length predictions,

139
and also provide informative bond angle data. Del Bene159 studied the water
solvates of F' and Cl', HOHF' and HOFICIusing the Flartree-Fock 6-31 G(d) and
MP2/6-31+G(2d, 2p) basis sets for the optimized structures and stabilization
energies, respectively. The fluoride complex exhibits a conventional hydrogen-
bonded structure with an essentially linear O-FI-F system, and internuclear
distances of 1.038 and 1.379 for the O-H and H-F bonds, respectively.
Consistent with the lower proton affinity of Cl", the bond lengths in HOHCI" are 0.961
(O-FI) and 2.305 (H-CI). Flowever, the bond is bent with an O-H-CI angle of 18.6
degrees. The same trends were observed by Hiroaka and Yamabe130 in their
calculations on (CH3OH)X" complexes. In all cases the H-X bond is longer than the
O-FI bond. The H-X bond length increases in the order F < Cl < Br < I, while the
calculated BE's show the reverse order. In agreement with Del Bene's work, the
hydrogen bond is linear in (CFI3OFI)F" and non-linear in the other methanol solvated
halides. Results by Fliraoka and Mizuse138(a) on CI"(ROFI)n (R= OH, CFI3, C2H5, n-
C3H7, /-C3H7i f-C4FI9) using the ab initio 3-21G and 4-31 G+p (0.07) basis sets also
show that the RO-FI-CI bond is bent in these complexes. For the 3- and 4-carbon
alcohols, this allows the Cl" to interact with both the hydroxyl and methyl hydrogens
to form a "chelate" structure.
The possibility of photodetachment to a bound state for neutral (CFI3CN)Br
must still be addressed. Again, it is presumed that a vertical transition to CFI3CN +
Br is favored by a loose interaction of the bromide ion with CFI3CN. Yamabe and
Fliraoka optimized the geometry for (CFI3CN)F" and (CFI3CN)CI" using the 4-31 G+p

140
GTO basis set.129 The electronic distribution indicates that CH3CN is described by
H36+-C6'-Cs+=N, rather than CH 3C +=N, with the methyl hydrogens having more
positive character than the center carbon. According to their calculations, the most
stable structure (CH3CN)F has Cs symmetry with the F' bound to one of the three
hydrogens in the methyl group via a linear F-H-C bond with F-H and H-C bond
distances of 1.70 and 1.14 , respectively. The bond lengths in (CH3CN)Cr are
2.70-3.00 and 1.09 for the Cl-H and H-C bonds, respectively. Recently, similar
results were obtained by Riveros et a/.,141 who optimized the geometries for solvated
anions at the MP4(SDTQ) level. In (CH3CN)Br the Br is bonded collectively to the
methyl hydrogens, with Br-H and H-C bond distances of 3.27 and 1.1 ,
respectively. Thus, all the calculated results for (CH3CN)X" complexes indicate a
loose interaction of X' with the solvent. The photodetachment process should,
therefore, be similar to the situation depicted in Figure 3.16. The endothermicity of
the hydrogen abstraction reaction of Br with CH3CN:
Br + CH3CN HBr + CH2CN AH = 4.8 kcal/mol (3.16)
is much less than that for Br + CH3OH (eq 3.14). Considering the very loose
interaction in the neutral (CH3CN)Br complex, the AE is expected to be very small
for this system. This is supported by the close agreement of the BE with previous
results.
Although the threshold data in Table 3.2 were obtained at room temperature,

141
these results are also valid at 0 K. As argued above, the photodetachment process
produces a neutral of the same geometry as the initial ion. The occupied vibrational
levels and the energy distribution in the neutral should also correspond to those in
the ion (T= 298 K). This means that the usual temperature correction to the energy
change (using heat capacities calculated from vibrational frequencies) should be
negligible, and the reported thresholds should be the same at either 298 K or 0 K.
Since the halide electron affinities used in the data analysis are zero-kelvin
values,154 the binding energies in Table 3.2 are valid at 0 K. Because the usual
AnRT correction is zero at 0 K, the BE values also represent enthalpy changes at
0 K.
Conclusion
The results in Table 3.2 show that the binding energy trend is I' < Br < Cl'
F' for the methanol solvates. The BE difference for (CH3OH)F' compared to
(CH3OH)Cr is 11.1 kcal/mol, whereas the (CH3OH)CI'/(CH3OH)Br and (CH3OH)Br
/(CH3OH)l" differences are only 3.7 and 0.6 kcal/mol, respectively. The same order
and approximate differences were found by previous investigators. Referring to
Table 3.1, this is also the trend of the proton affinities of the halide ions. For the
complexes with bromide ion, the BE order is CH3OH < C2H5OH < /'-C3H7OH < n-
C3FI7OH, which is the same as the order of the alkyl-group size, but opposite to the
RO' proton affinity order. Thus, for the alcohol-solvated halides, the overall

142
observation is an increase in BE with decreasing RO' proton affinity and increasing
X' proton affinity. As the hydrogen becomes more loosely bound to the alcohol
oxygen, it is more available for bonding to X.
The results of this and previous work on solvated bromide anions, as well as
other studies of chloride complexes, show that complexes of these halides with
acetonitrile have BE's about 1-2 kcal/mol lower than the methanol solvates.
However, calculated results show that the BE for (CH3CN)F' is more than 7 kcal/mol
less than that for (CH3OH)F\ The difference in behavior of F' compared to Cl', Br',
and I' can be understood in terms of the calculated geometries discussed above.
The (CH3OH)CI', (CH3OH)Br, and (CH3OH)l' complexes have nonlinear CH30-H-X
hydrogen bonds, and the results indicate that there is additional interaction of X'
with the methyl hydrogens. However, because of its small size, the F" can not
interact with the methyl group in CH30-H-F', resulting in a linear hydrogen bond. As
stated above, the halide ion bonds via the methyl hydrogens in all the (CH3CN)X'
complexes. Thus the bonding is somewhat similar for the CH3OH and CH 3CN
solvates of Cl' and Br', and this is reflected in the closeness of the BE's of the
(CH3OH)X' and (CH3CN)X' complexes for these halides.
Although the BDEs of some halide complexes have been measured using
other techniques in different types of mass spectrometers, the photodetachment
technique has not previously been applied to measure BDEs in ICR mass
spectrometry. This work is the first to apply this technique for the BDE
measurements. The data obtained from this work in general confirm the results of

143
previous measurements which utilized entirely different experimental approaches.
Thus, we have shown that the photodetachment technique is a reliable approach
to obtain BDEs by ICR mass spectrometry. Since electrospray ionization (ESI) can
produce multiple-charged and solvent cluster-attached ions, the BDEs and/or step
wise solvation energetics should also be obtainable with photodetachment
techniques in the ICR instrument.

CHAPTER 4
BOND DISSOCIATION ENERGIES OF ALKALI AND HYDRONIUM CATIONS
BOUND TO CROWN ETHERS DETERMINED BY ON-RESONANCE
COLLISIONALLY-ACTIVATED DISSOCIATION TECHNIQUES
Background
Crown ethers constitute an important group of compounds and are especially
well known for their ability to form complexes with alkali-metal cations, which do not
bind favorably to other common ligands. Previous investigations of alkali-
metal/crown-ether complexes have provided information about the intrinsic nature
of the binding interactions and the size-selectivity of complexation.78'89,160'173 In
solution the order of the complexation equilibrium constants is generally K+ >
Rb+ > Cs+ > Na + > Li+ in solvents such as water or methanol. 166'171 However,
competitive equilibrium studies in the gas phase have indicated that Na+ binds to
crown ethers more strongly than k+.78'89160'161 The stability difference is attributed
to the greater solvation enthalpy of Na+ compared to K+. This behavior
demonstrates the importance of gas-phase studies to determine the fundamental
nature of the ion-molecule interaction in the absence of solvent.
Mass spectrometry (MS) can be used to obtain data on the binding of gas-
phase complexes, if there is a means to add a controlled amount of energy to
dissociate the complex. Photodissociation has been used successfully in several
144

145
such studies,120174'175 but it is not feasible, if, as in this investigation, there is no
absorption of the laser radiation near the threshold range of the ions. An alternative
method is collisionally-activated (formerly collisionally-induced) dissociation (CAD
or CID),88104'105'176'184 which can be applied over a wide energy range, and which
does not require laser apparatus and the associated alignment problems.
Several groups have used CAD and Fourier transform ion cyclotron
resonance mass spectrometry (FTICR-MS) for the measurement of bond
dissociation energies (BDE's).88'104'105,176 After formation of the gas-phase ions,
unwanted ionic species are ejected to isolate the precursor ion (alkali-metal or
hydronium/crown-ether complex in this study) with the neutral collision gas (Ar) in
the ICR cell. A short radiofrequency (rf) pulse of voltage Vpp and duration t is then
applied to the excite plates of the cell. This increases the center-of-mass
translational energy, E^,, of the ions according to the relationship given in equation
1.21.
The energetic ion is then allowed to interact with a neutral collision gas (Ar
or Xe) for a delay time, during which an ion/molecule collision converts translational
energy to internal vibrational and rotational modes. If the energy imparted (eq 1.21)
was sufficient, the complex dissociates to the successor ion and the neutral crown
ether, L:
M+-L - M+ + L (4.1)
The precursor and successor ion intensities are then measured by the usual ICR

146
technique. The procedure is repeated with varying rf pulse lengths to determine the
minimum energy for appearance of the successor ions. Unlike chemical processes
in liquids, gas-phase ion/molecule reactions usually do not have activation barriers.
Thus, the threshold for the dissociation process may be attributed entirely to the
BDE, provided that there is proper control of a number of parameters, as described
below.
The CAD method has been used to observe several alkali-metal/crown-ether
complexes,88-89 and in Eyler's group88 approximate bond dissociation energies were
obtained for the 18-crown-6 complexes of K\ Cs+ and Ba+ using laser desorption
(LD) to form the ions. Subsequent ab initio calculations by Thompson et a/.173 using
the RHF/3-21G, 6-31G, and MP2/6-31G basis sets produced results as much as 30
kcal/mol (1.3 eV) higher than the measured values. The discrepancy can be
attributed in part to the LD process, because shot-to-shot instability of the laser
power results in poor control of precursor ion intensities. Another problem was the
pressure of the argon collision gas (1-2 10-6 torr), which must be kept low (10 8
torr range) to prevent collisions during excitation, as well as multiple collisions
during the CAD delay time.
In this study the BDE's of complexes of 18-C-6, 15-C-5, and 12-C-4 with Na+,
K+, Rb+, Cs+ and H30+ (Figure 1.9) have been measured by CAD in FTICR-MS,
using electrospray ionization (ESI) to produce the gas-phase ions. This method
affords a very reproducible and direct way to form ion/neutral complexes or/and
solvent attached ion complexes from solution. In addition, compared to other

147
ionization methods, such as El, LD, and MALDI, ESI imparts much less internal
energy, which is an important factor in BDE measurements.
The accuracy of BDE's obtained by the CAD technique depends on the
model used to analyze the cross section of the ion/molecule reaction and the
corrections for various systematic effects that may otherwise distort the
thermodynamic data. Both McMahon104 and Eyler88105 have used linear-
extrapolation plots to obtain the observed thresholds, assuming that the cross-
section for dissociation increases linearly with energy above the threshold.
Assuming that the precursor ions were populated at room temperature, and that all
available energy was used for dissociation, the observed thresholds were
interpreted to equal the BDE's with the successor ions produced at zero Kelvin. To
obtain the true thresholds, these values were then deconvoluted with a trial function,
which includes Chantry's107 Doppler-broadening term to account for the thermal
spread of the neutral collision gas. The values were then converted to BDE's at
room temperature by estimating the internal energy of the successor ions and
adding to the 0 K result, (i.e., D298, as shown in Figure 4.1). There are assumptions
and uncertainties in both this type of fitting procedure and the estimation of the
internal energies.
In the present work, a much more careful data analysis was used. The
threshold energies were evaluated using the CRUNCH deconvolution program
developed by Armentrout and coworkers.178'179 This model corrects for the internal
energies of the precursor ions based on calculated frequencies of the normal

Figure 4.1. Potential energy curve for ion crown ether complexes and corresponding neutral species. The D298
represents the bond dissociation energy at room temperature, the D0 represents the bond dissociation energy at zero
Kelvin temperature.

149

150
vibrational modes, the kinetic energy distribution of the precursor ions, the Doppler
broadening mentioned above, and the lifetime of the dissociation process. The
vibrational frequencies for use in the CRUNCH program were obtained using the
HyperChem program for ab initio and ZNDO empirical calculations. All the
calculated vibrational frequencies were scaled to reasonable values by comparing
the calculated and experimental185 results on both cyclohexane and 1,4-dioxane
compounds. The best scaling factor found in this case is {0.89+0.745xsin[(n/2)(f¡-
fm¡ny(fmax-fmin)]}\ where f¡, fmin, and fmax are the frequencies of the ith mode, minimum
and maximum vibrational frequencies obtained from the HyperChem calculation.
The results are shown in Table 4.1. In addition to ground state vibrational
frequencies which were used to estimate the effect of vibrational energy in
CRUNCH program, the vibrational frequencies of the activated complexes were also
obtained using a single-point calculation, and the normal mode which corresponds
to the metal-crown ether stretch mode was removed from the transition state to
account for the dissociation lifetime, again using the CRUNCH program.
Experimental
Ion Formation
All crown ethers and alkali metal salts were obtained from Fisher Scientific,
Inc., and were used without further purification. Solutions were prepared by

Table 4.1. Vibrational frequencies of cyclohexane and 1,4-dioxane.

Calculated
values
C6Hi2
Scaled
values
Exp.
values185
Calculated
values
c4h8o2
Scaled
values
Exp.
values
245.ob
221.29
248.47
248
221.67
248.62
288
350.76
375.88
383
335.62
364.51
424
456.39
475.33
426
425.80
451.17
435
456.40
475.34
426
518.35
535.84
490
578.27
583.42
523
661.48
659.07
610
929.16
860.8
785
972.33
899.09
837
929.29
860.89
785
1060.59
961.39
853
1049.16
945.99
802
1170.16
1035.65
881
1049.19
946.01
863
1175.89
1039.44
889
1094.78
977.27
863
1359.01
1156.52
1015
1200.45
1047.58
907
1379.96
1169.43
1052
1221.44
1061.21
907
1391.58
1176.55
1086
1302.69
1112.96
1027
1409.73
1187.61
1110
1302.84
1113.06
1027
1471.34
1224.67
1128
1312.33
1119
1030
1491.88
1236.86
1136
1355.75
114 5.95
1057
1509.76
1247.41
1217
1355.90
1146.05
1057
1529.81
1259.17
1256
1441.88
1198.23
1157
1609.93
1305.43
1291
1442.08
1198.35
1157
1667.50
1338.00
1305
1442.15
1198.4
1261
1680.19
1345.11
1335
1450.94
1203.65
1261
1689.54
1350.33
1369
1451.09
1203.74
1266
1848.15
1436.98
1378
1535.38
1253.34
1266
1927.30
1478.97
1397
1541.21
1256.72
1347
2128.73
1582.71
1444
1683.27
1337.36
1347
2290.68
1663.43
1449
1683.31
1337.38
1355
2303.92
1669.95
1457
1702.06
1347.78
1355
2400.81
1717.27
1459
1703.53
1348.59
1383
4742.10
2900.54
2856
1703.59
1348.63
1437
4742.83
2900.98
2856
1816.87
1410.36
1437
4744.38
2901.92
2863
1949.07
1480.27
1443
4745.55
2902.63
2863
1949.29
1480.39
1443
4785.09
2926.66
2968
2190.75
1603.21
1457
4785.46
2926.88
2968
2190.86
1603.27
1457
4789.80
2929.54
2970
2282.09
1648.35
1465
4790.02
2929.67
2970
4750.45
2887.95
2852
4754.47
2890.37
2860
4754.65
2890.48
2863
4759.07
2893.14
2863
4759.28
2893.27
2897
4761.98
2894.89
2897
4781.02
2906.4
2915
4782.29
2907.17
2930
4782.46
2907.28
2930
4788.69
2911.05
2930
4788.91
2911.19
2933
4793.89
2914.21
2933
152

153
dissolving 1.0 mg of each sample in 10 mL of Me0H/H20 (1:1/v:v) for the metal
complexes or 10 mL 1% acetic acid/0.1% HCI in Me0H/H20 (1:1/v:v) for the
hydronium complexes. Samples were injected into the source through a needle at
a flow rate of 1 pL/min from a syringe pump. Referring to Figure 2.6, the
electrospray needle was held at ground potential, and a ca. -4 kV potential was
applied to the platinum coating on the inlet of the quartz capillary (0.5 mm i.d.), in
order to form charged droplets between the needle and the capillary. A potential of
either 75 or 25 volts was applied to the outlet of the capillary, and heated nitrogen
gas (175C for alkali metal complexes, 25C for H30+ complexes) was passed
through the capillary to desolvate the charged droplets and form the gas-phase
ions. The ions were then guided by a set of skimmers (0.8 and 1.0 mm diameters)
and lenses in the ESI source, prior to entering the transfer region. The first skimmer
was held at +5V for all experiments. Thus, the kinetic energy of the ions (all singly
charged) passing between the outlet and the skimmer was either 70 or 20 V,
depending on the capillary voltage.
Referring to Figure 2.9, the ions were then accelerated toward the analyzer
by a series of electrostatic lenses (3 kV maximum potential). After acceleration, the
ions passed through a series of deceleration lenses with potentials decreasing step
wise to ground. A set of deflection plates was used to gate the continuous ion
beam, so that ions entered the analyzer region only during the ion-accumulation
period (100 ms). The "side-kick" plates, with pulsed low voltages (1-5 V) during the
formation pulse, converted part of the translational energy into cyclotron motion.

154
After the ions entered the cell, the voltages on the side-kick plates were dropped
to the trapping potential (0.5-1 V) and ions were then ready to be analyzed.
Pulse Sequence
As shown in Figure 2.11, a quench pulse initiated the experiment by
removing all remaining ions from the ICR cell. As described above, the ions were
formed in the external ESI source, transferred, and collected in the analyzer cell for
100 ms. After a 2-second delay period to allow thermalization by collisions with
background neutrals and radiative relaxation, the precursor ions were isolated using
a combination of chirp-frequency and single-frequency ejection pulses. Then, an
rf pulse with frequency corresponding to the exact m/z of the precursor ion (on-
resonance excitation; cf., eq 1.21) was applied to the excite plates of the ICR cell
to increase the ions' translational energy. The activation pulse width was kept
below 50 ps to ensure that the ions did not undergo any collisions during excitation.
The imparted translational energy was varied by altering the duration of the pulse
over a range of zero to 50 ps at a constant rf amplitude (Vpp in eq 1.21), which was
measured with a Tektronix 2235 oscilloscope ( 3%). The isolated ions were then
allowed to interact with a static pressure (3-5 x 10'8 torr) of Ar gas for a period (10-
30 ms CAD delay time) which permitted the ion/molecule collisions necessary to
convert translational energy into internal energy (Xe collision gas should effect this
convertion more efficiently than Ar). During this time, the precursor ions dissociated

155
to the successor ions and neutral crown ether molecules if the activation energy
exceeded the threshold energy. The precursor and successor ions were
subsequently excited and detected simultaneously using broad-band chirp
excitation with a mass-detection range from 21 to 2000 m/z. Fifty 32K-point time-
domain transients were signal-averaged for all ions. For each rf pulse width, the
intensity measurements for both the precursor and successor ions were repeated
five times, and the respective means and standard deviations were calculated.
Data Analysis
Ion Appearance Curve178
The experimental total rate coefficient for fragmentation, ktot, is related to the
measured intensities, lp and ls, of the precursor and successor ions by the
equation,178
'p VP + z/s) exP( ktot n 0 (4.2)
where n is the number density of the argon collision gas, and t is the effective
reaction time. The rate coefficient for each successor ion, K. is subsequently given
by,

156
(4.3)
Since the total ion intensity, ltot= lp + X ls, is constant, the exponential term in eq 4.2
may be expanded in a power series to give
= ks nt
(4.4)
Providing that the Ar number density is sufficiently low. For each experiment, n and
t are constants. Thus, ls/ltot is proportional to the rate coefficient, and a plot of ls/llot
versus E cm (from equation 1.21) gives the experimental ion appearance curve as
shown in Figure 4.2 for (18-crown-6)Rb+.
As stated in the Introduction, the CRUNCH program developed by
Armentrout and cowokers178,179 was used to obtain the bond dissociation energies.
The CRUNCH calculates the BDE (plus other parameters) which produces the best
fit to the experimental appearance curve. In addition to ls/ltot and E^ values, the
program requires other input data and fitting procedure which are described in the
following sections.

Figure 4.2. Appearance curve of (18-C-6)Rb+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.

Relative rate coefficient
1
0.8
0.6
0.4
0.2
0
(18-crown-6)Rb
1.65
i r
0
2 4 6 8 10 12 14 16 18
Energy in center-of-mass frame/eV
20
158

159
Internal Energies
Rotational energy
Because the rotational energy distribution is relatively narrow, the average
rotational energy of the ions (3/2 kT = 0.039 eV for nonlinear ions at 298 K) was
added to the measured threshold in CRUNCH program.
Vibrational energies
The frequencies for all vibrational modes in both the vibrational ground and
excited precursor ions were calculated using the HyperChem program. The
program first optimizes precursor ion geometries using both ZNDO semi-empirical
and ab initio calculations and subsequently uses these geometries for the frequency
calculation. The ground state values are used by CRUNCH to calculate the
vibrational energy which is summed over the vibrational states i having energies E¡
and population g¡ of the precursor at room temperature. In this program, the Beyer-
Swinehart algorithm186 is used to evaluate the density of the ion vibrational states,
and then the relative populations, g¡, are calculated from an appropriate Maxwell-
Boltzmann distribution. The excited state values are needed to correct for the
probability that some ions may not dissociate until after the CAD delay time (RRKM
calculation).

160
Trial Function
The trial function accounts for the distribution of Ecm values of the precursor
ion, caused by the random thermal motion of the collision gas molecules. According
to Chantry107, this broadening has an isotropic Maxwellian velocity distribution, given
by:
(4.5)
In eq 4.5 e = E/ykbT and e0= Eo/ykbT, where E and E0 are the relative and nominal
center-of-mass energies of the precursor ions, respectively, y = M/(M+m) where M
is collision target molecular mass, m is the precursor ionic mass as shown in eq
1.21, kb is the Boltzmann constant, and T is the collision gas temperature. As a
result of the energy distribution, the observed appearance curve, oobs(E0), differs
from the true curve, o(E). Chantry has derived the form of the convolution of o(E)
as
oobs (E0) jA1/2 f(E, E0)a(E)dE
(4.6)
0
This equation is used by CRUNCH to calculate the corrected curve.
Considering eq 4.6, at high precursor ion energies, E0 kbT, this distribution
is relatively narrow, with a maximum at E = E0 and a width given approximately by
FWHM=(11.1ykbTE0)1/2
(4.7)

161
This was the situation in the threshold region of the experimental data. If the
energies are very low (kbT < E0< 1 eV), the distribution function becomes broad
relative to the center-of-mass ion energy. This corresponds to a breakdown of the
stationary target approximation at energies where the velocities of the Ar gas
molecules are comparable to or larger than the ion velocity. At the lowest ion
energies, there is a thermal distribution of the translational energies, and the
interaction energies are determined primarily by the thermal velocities of the target
molecules.
Deconvolution
The form of the convolution for Doppler broadening (eq. 4.6) is such that the
Langevin-Gioumousis-Stevenson (LGS) cross section187, oLGS(E) is directly
proportional to E'1/2 and is unchanged by the broadening effects. The simplest form
of the model threshold law97 is given by equation 1.19. The threshold energy
determined in this manner corresponds to dissociation of room-temperature (298
K) precursor ions to room-temperature successors, i.e. BDE298 (cf., Fig.4.2), and
does not account for the internal energies of precursor ion and probability of
dissociation after the CAD delay time. To correct for these effects the CRUNCH
program uses the relation178
ks -
o0 Igi P(E,E,.,7)
(E*Ei*Erof-ET-)n
E
(4.8)

162
where E is the relative translational energy of the precursor ions, Erot is the rotational
energy (3kT/2 = 0.039 eV), ET is the zero-kelvin BDE (D0 as shown in Figure 4.2),
n is an adjustable parameter, E¡ and g¡ represent the energy and population of the
ith vibrational state, respectively, and P is the probability that metastable ions
formed with initial translational energy, E, and internal energy, E¡+Erot, will dissociate
within the experimental time window (10-30 msec in this case).
The CRUNCH program uses nonlinear least-squares analysis to optimize the
parameters o0, n, and E T to give the best fit to the data without including RRKM
caculations. After obtaining optimized o0, n, and E T, the RRKM calculations were
included. In this manner, we can observe the shift of the threshold with RRKM
calculations. The RRKM calculation lowered the ET by approximately 0.3 to 0.8 eV
(7 to 18 kcal/mol) in this study and the effect is more pronounced for larger
complexes. As the size of complexes increases, the number of vibrational degrees
of freedom increases. Therefore, the dissociation lifetime becomes more important
and dissociation may not be complete within the CAD delay time. Although this
problem is not as severe in ICR experiments (10-30 msec) as it is in ion-beam
experiments (microseconds), in some cases the intermediates may require
milliseconds to several seconds for dissociation.188'190 In the optimization procedure,
the parameters are very sensitive to the fitting of the appearance curve; in other
words, a slight change in the number can cause tremendous distortion to the fitting.
Therefore, the zero-kelvin bond dissociation energy, D0, can be obtained with very
small uncertainty, 0.01 to 0.05 eV ( 0.2 to 1.0 kcal/mol) without RRKM

163
calculation; the uncertainty is about 0.1 eV ( 2.3 kcal/mol) when RRKM
calculations were included.
Results
Collision Conditions
In order to control the amount of energy imparted by the rf excitation, the
precursor ions can suffer no collisions during the excitation process. In this work,
rf pulse lengths of 0-50 psec were used. At a total pressure of ca. 3-5 *10'8 mbar
(2-4 xio 8 torr), the average time between ion/molecule collisions (using the
Langevin rate constant191 of 1.24x1 O'9 cm3/s at 300 K) is 2.13 s, which is much
longer than the excitation time. Thus, the average number of collisions in 50 psec
is 2.7x10'6. It is apparent that during the 50 psec excitation time most of ions
undergo no collisions.
It is also essential that each precursor ion collide with only one Ar molecule
within the CAD delay time, to avoid broadening during the conversion of
translational to internal energy. The probability, qn, that an ion will encounter n
molecules during the CAD delay time is approximated by the Poisson distribution192
(4.9)

164
where a = nDol represents the average number of target molecules. The ratio of
double to single collisions, qj/q,, is approximately 0.001 with an Ar pressure of 5
x10 8 mbar (3.8 xio8torr) and a 30 msec collision time. In other words, fewer than
0.1 % of the precursor ions should experience multiple collisions with this pressure
and delay time.
To ensure that the single-collision condition was obeyed, the dissociation of
(18-crown-6) Rb+ was studied using different Ar pressures (3.0, 5.5, 7.3, and 9.2 x
10'8 mbar) with a constant CAD delay time of 50 ms and, in a separate series of
experiments, using different CAD delay times (0.05, 0.08, 0.1, and 0.15 s) with the
Ar pressure held at 3.3x1 O'8 mbar. The resulting appearance curves are shown in
Figures 4.3 and 4.4. The curves obtained with pressures of 3.0 and 5.5x 10'8 mbar
and CAD delay times of 0.05 and 0.08 sec are almost identical, and the thresholds
are about the same. With higher pressures and CAD delay times, the thresholds
shifted to lower energy due to multiple collisions. All subsequent appearance
curves were obtained with the pressure and delay times in the acceptable ranges.
Fragmentation Pathways
Table 4.2 presents a compilation of all ions observed in the mass spectra,
both prior to isolation of the precursor ions and subsequent to the CAD process.
Referring to the columns showing the ions present prior to isolation, only the
molecular ions (LM+) were formed by 18-C-6 and 15-C-5 with M+ (M=Na, K, Rb, Cs,

Figure 4.3. Appearance curves of (18-C-6)Rb+ at different pressures (50 ms CAD delay time); 3.0 10'8 torr, -
5.5 x 10'8 torr, 7.3 x 10'8 torr, A 9.2 x 10'8 torr.

Relative rate coefficient
166

Figure 4.4. Appearance curves of (18-C-6)Rb+ at different CAD delay times (P = 3.3 10 8 torr); 0.05 s, A 0.08
s, 0.1 s, T 0.15 s.

Relative rate coefficient
+
0.4 Delay Time Dependence on (18-C-6)Rb
0123456789
Energy in center-of-mass frame/eV

Table 4.2. Ions observed in mass spectra. The capillary potential was 25 V for ions
labelled with an asterisk, 75 V for all other ions; (s) strong signal, (w) weak signal.

Ions Observed in Mass Spectra
170
ions formed by ESI prior to isolation
isolated precursor
ion
successor ions
(12-C-4)Na* (s),
m/z=199
(12-C-4)2Na+ (w)
m/z=375
(12-crown-4)Na*
Na*
(12-C-4)K* (s),
m/z=215
(12-C-4)2K+ (w)
m/z=391
(12-crown-4)K+
K*
(12-C-4)Rb+ (s),
m/z=261, 263,
(12-C-4)2Rb+ (s),
m/z=437, 439, 441
Rb* (w)
m/z=85, 87
(12-crown-4)Rb*
Rb*
(12-C-4)Cs+(s),
m/z=309
(12-C-4)2Cs+(s),
m/z=485
Cs* (w)
m/z=133
(12-crown-4)Cs+
Cs+
(12-C-4)H+ (s),
m/z=177
(12-C-4)H* (s)
(12-C-4)2H+(s)\
m/z=353
(12-C-4)2H+ (w)
(12-C-4)2H30+(w)
m/z=371
(12-crown-4)H+
(C2H40)nH+
(n=3,2,1)
m/z=133,89,45
(15-C-5)Na+(s),
m/z=243
(15-C-5)2Na+ (w)'
m/z=463
(15-crown-5)Na+
Na+
(15-C-5)K+(s),
m/z=259
(15-C-5)2K+ (w)
m/z=479
(15-crown-5)K*
K*
(15-C-5)Rb+(s),
m/z=305, 307
(15-C-5)2Rb+ (w)'
m/z=525, 527, 529
(15-crown-5)Rb+
Rb+
(15-C-5)Cs+(s),
m/z=353
(15-C-5)2Cs+ (w)"
m/z=573
(15-crown-5)Cs+
Cs+
(15-C-5)H+ (s),
m/z=221
(15-C-5)H30+ (s)'
m/z=239
(15-crown-5)H30+
(15-C-5)H+
(C2H40)nH+
(n=3,2)
(18-C-6)Na+
m/z=287
(18-crown-6)Na+
Na+
(18-C-6)K*
m/z=303
(18-crown-6)K+
K*
(18-C-6)Rb*
m/z=349, 351
(18-crown-6)Rb+
Rb*
(18-C-6)Cs+
m/z=397
(18-crown-6)Cs*
Cs+
(18-C-6)IT (s)",
m/z=265
(18-C-6)H30+ (s)*
m/z=283
(18-crown-6)H30*
(18-C-6)hT
The capillary potential was 25 V for ions labelled with an asterisk, 75 V for all other ions.

171
H) using a voltage difference, AV, between capillary and skimmer, equal to 70 V.
In addition to the simple adduct ions, the sandwich-type complexes (L2M+) were
also observed in complexes of 12-C-4 with M+ at AV=70. With a capillary voltage
of 25 V (AV=20 V), the sandwich-type complexes (L2M+) were observed for both 12-
C-4 and 15-C-5 with M+ as shown in Figures 4.5 and 4.6 for (12-C-4)2H30+, (12-C-
4)2H+ and (12-C-4) 2Rb +, but not for 18-C-6. This behavior indicates that the
stabilities of the L2M+ complexes increase as the size of the crown ether decreases.
As the capillary voltage was increased from 25 to 75 volts (AV from 20 to 70
V), the intensity of sandwich-type complexes decreased and the intensity of the
molecular ions increased, as shown in Figure 4.7 and 4.8 for Fl30+ with 12-crown-4.
This indicates that the binding in the sandwich-type complexes is much weaker than
that in the molecular ions. As the capillary voltage was increased, the more weakly
bound sandwich-type complexes dissociated between the capillary and skimmer as
a result of collision with background neutrals.
The sandwich complexes of 15-C-5 with Na+, K+, FT and Fl30+ and 12-C-4
with K+ were observed in previous work with an internal ESI source and infrared
multiple photon dissociation (IRMPD)160. The (15-C-5)2K+ and (15-C-5) 2Na +
complexes were fragmented to the molecular ions, (15-C-5)K+ and (15-C-5)Na+, by
the laser radiation. This fragmentation behavior is similar to that observed in the
present study.
For the metal-ion complexes, the successor ions of the CAD process were
the bare metals in all cases,

Figure 4.5. Mass spectrum of (12-C-4)Rb+ at AV=20 volts. Peaks corresponding to (12-C-4)Rb+ and (12-C-4)2Rb+ ions
were observed.

80
60
40
20
0
AV=20 volts
261.1
(12-crown-4)Rb+
263.0
437.3
(12-crown-4)2Rb+
439.2
ll
.j i
, J
v Li I
l
30.0
140.0
230.0
320.0 410.0
500
Mass-to-charge ratio
173

Figure 4.6. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peaks corresponding to (15-C-5)H30+, (15-C-5)2H+, and
(15-C-5)2H30+ ions were observed.

AV=20 volts
Mass-to-charge ratio
175

Figure 4.7. Mass spectrum of (12-C-4)H30+ at AV=20 volts. Peaks corresponding to (12-C-4)H+, (12-C-4)2H+, and (12-
C-4)2H30 + ions were observed. The peak intensity of (12-C-4)2H+ is higher than that of (12-C-4)H+.

100
80
60 -
40 -
20 -
0
30.0
AV=20 volts
177.1
(12-crown-4)H+
209.1
a JL
353.2
(12-crown-4)2H+
371.2
(12-crown-4)2H3G
124.0
218.0 312.0
Mass-to-charge ratio
406.0
500.0
177

Figure 4.8. Mass spectrum of (12-C-4)H30+ at AV=70 volts. Peaks corresponding to (12-C-4)H+ and (12-C-4)2H+ ions
were observed. The peak intensity of (12-C-4)2H+ is lower than that of (12-C-4)H+.

100
80
60
40
20
0
AV=70 volts
177.1
(12-crown-4)H+
(12-crown-4)2H+
1
:
353.2
u
7.0
113.6
210.2
306.8
403.4
500
Mass-to-charge ratio
179

180
M+L-> M+ + L,
M= Na, K, Rb and Cs
(4.10)
as shown in Figures 4.9 and 4.10 for (12-C-4)Rb+. The (18-C-6)H30 + complex
underwent desolvation and formed (18-C-6)H+ (Figs. 4.11 and 4.12).
(4.11)
(18-C-6)H30+ (18-C-6)H+ + H20
as observed in previous IRMPD studies at low laser fluence.160
The (15-C-5)FI30+ underwent desolvation, followed by fragmentation of the
crown ether. At low rf excitation energies, it desolvated into (15-C-5)H+ (Figs. 4.13
and 4.14), but at high energies, desolvation was followed sequentially by
dissociation into (C2FI40)3FI+ and (C2FI40)2H+ fragments (Fig. 4.15). The absolute
intensities plot versus center-of-mass translational energy, Ecm, for (15-C-5)FI30+
complex including precursor, successor, and all the fragment ions is shown in
Figure4.16. The intensity of precursor ion, (15-C-5)FI30+, decays exponentially with
Ecm, the intensity of successor ion, (15-C-5)FT, decreases with Ecm at low
translational energy. At high translational energy, the signal of (15-C-5)FT
decreases as the signal of the fragments increases.
(15-C-5)H30+ -+ (15-C-5)H+ (C2H40)3H+
(4.12)
(C2H40)2H+

Figure 4.9. Mass spectrum of (12-C-4)Rb+ at AV=70 volts. Peak corresponding to (12-C-4)Rb+ ion was isolated before
excitation.

100
80
60
40
20
0
AV=70 volts
261.1
(12-crown-4)Rb+
263.0
J
.0 140.0 230.0 320.0 410.0 500.0
Mass-to-charge ratio
182

Figure 4.10. Mass spectrum of (12-C-4)Rb+ after CAD at AV=70 volts. The two isotopes of Rb+ were the only
successor ions observed.

100
80
60
40
20
0
Upon CAD
261.1
(12-crown-4)Rb+
84.9
Rb+
86.9
263.0
1
iO.O
140.0
230.0 320.0
Mass-to-charge ratio
i 1
410.0 500.0
184

Figure 4.11. Mass spectrum of (18-C-6)H30+ at AV=20 volts. Peak corresponding to (18-C-6)H30+ ion was isolated
before excitation.

100
80
60
40
20
0
AV=20 volts
283.2
(18-crown-6)H30+
t i 1 1 r
17.1 113.7 210.3 306.9 403.4
Mass-to-charge ratio
186

Figure 4.12. Mass spectrum of (18-C-6)H30+ after CAD at AV=20 volts. Peak corresponding to (18-C-6)H+ ion was the
only successor ion observed.

100
80
60
40
20
0
Upon CAD
283.2
(18-crown-6)H30+
(18-crown-6)H+
26
5.2
17.1
113.7
210.3 306.9
Mass-to-charge ratio
403.4
500.0
188

Figure 4.13. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peak corresponding to (15-C-5)H30+ ion was isolated
before excitation.

100
80
60
40
20
0
AV=20 volts
239.2
(15-crown-5)H30+
7.1
113.7 210.3 306.9 403.4 500.0
Mass-to-charge ratio
190

Figure 4.14. Mass spectrum of (15-C-5)H30+ after CAD at low activation energy and AV=20 volts. Peak corresponding
to (15-C-5)H+ ion was the only successor ion observed.

100
80
60
40
20
0
Upon CAD Low activation energy
239.2
(15-crown-5)H30+
(15-crown-5)H+
22
1. 1
17.1
113.7
210.3 306.9
Mass-to-charge ratio
403.4
500.0
192

Figure 4.15. Mass spectrum of (15-C-5)H30+ after CAD at high activation energy and AV=20 volts,
corresponding to (15-C-5)H\ (C2H40)nH+ (n = 2 and 3) were the succesor ions observed.
Peaks

100
80
60
40
20
0
Upon CAD High activation energy
(15-crown-5)H+
(C2H40)2H+ (C2H40)3H+
22
239.2
(15-crown-5)H30+
i. i
JL~v
[7.1
113.7
210.3 306.9
Mass-to-charge ratio
403.4
500.0
194

Figure 4.16. Absolute intensity plots versus Ecm of (15-C-5)H30+, (15-C-5)H+, and (C2H40)nH+ (n = 2 and 3).

On Resonance CAD of (15C-5)H30+
i 1 1 1 1 1 1
0.00 0.80 1.60 2.40 3.20
E,m/eV
196

197
The latter two species were the only fragments observed in the IRMPD studies of
(15-C-5)H30+ at high laser fluences. In the case of 12-crown-4, only (12-C-4)H+,
rather than (12-C-4)H30+, was formed during ESI at either high or low applied
capillary voltages. After rf excitation the (12-C-4)H+ underwent fragmentation
completely to (C2H40)3H+ and (C2H40)2H+ (Figure 4.17 and 4.18).
Bond Dissociation Energies
Plots of the relative cross section (ls/ltot) vs. center-of-mass translational
energies and the best fits using eq 4.8 are shown in Figures 4.2, 4.19, 4.20 and
4.21, for (18-C-6)Rb+, (15-C-5)Rb+, (12-C-4)K+, and (12-C-4)Cs+, respectively. The
BDE's determined in this work, as well as values obtained in previous experimental
and theoretical investigations, are presented in Table 4.3.
For all three crown ethers the BDE trend is Na > K > Rb > Cs. The same
stability order was observed in the previous IRMPD work. The results of
calculations on 18-C-6 complexes (with Li+, Na+, K+, Rb+, and Cs+) by Thompson
and coworkers173 showed the same trend, but their values were systematically
higher. For (18-C-6)K+ Wipff et a/.193 reported a calculated BDE of 61.8 kcal/mol for
a D3d structure using an earlier version of the AMBER force field. Later, Dang and
Kollman194 reported a BDE of 57 kcal/mol, also using AMBER. Thompson195 used
the hybrid quantum mechanical/molecular mechanical (QM/MM) molecular
dynamics method to study (18-C-6)K+ and obtained 70.0 kcal/mol vs. 72.0 kcal/mol

Figure 4.17. Mass spectrum of (12-C-4)H+ at AV=70 volts. Peak corresponding to (12-C-4)H+ ion was isolated before
excitation.

100
80
60
40
20
0
AV=70 volts
177.2
(12-crown-4)H+
J
,
.0 124.0 218.0 312.0 406.0 500.0
Mass-to-charge ratio
199

Figure 4.18. Mass spectrum of (12-C-4)H+ after CAD at AV=70 volts. Peaks corresponding to (C2H40)nH+ (n = 2 and
3) were ions observed.

100
80
60
40
20
0
Upon CAD
177.2
(12-crown-4)H+
89.1
(C2H40)2H+
- I
.0 124.0 218.0 312.0 406.0 500.0
Mass-to-charge ratio
ro
o

Figure 4.19. Appearance curve of (15-C-5)Rb+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.

Relative rate coefficient
0.5
0.4
0.3
0.2
0.1
Energy in center-of-mass frame/(eV
203

Figure 4.20. Appearance curve of (12-C-4)K+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.

Relative rate coefficient
Energy in center-of-mass frame/eV

Figure 4.21. Appearance curve of (12-C-4)Cs+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.

Relative rate coefficient
0.5
0.4
0.3
0.2
0.1
0
(12-crown-4)Cs
1.71
i 1 1 r
i 1 r
mmmA
0 2 4 6 8 10 12 14 16 18 20
Energy in center-of-mass frame/eV
207

Table 4.3. Bond dissociation energies for alkali-metal and hydronium ion crown ether complexes including present
work, previous work, and theoretically calculated values.

Bond Dissociation Energies for Alkali-Metal and Hydronium Ion Crown Ether Complexes
BDE
Na
K
Rb
Cs
h3o+
12-crown-4
82.1(3.52)
60.6(2.60)
43.2(1.85)
39.9(1.71)
15-crown-5
79.5(3.41)
55.9(2.41)
40.1(1.72)
37.5(1.61)
24.7(1.06)'
This work
18-crown-6
75.1(3.22)
52.2(2.25)
38.5(1.65)
36.1(1.55)
36.4(1.56)'
Previous work
[ref. 88]
18-crown-6
40
32
Calculation
[ref. 193]
18-crown-6
71.1
62.3,52.5
[ref. 194]
18-crown-6
57.0
[ref. 173]
18-crown-6
81. T
68.1a
52.8a
30.4a
[ref. 173]
18-crown-6
83.4b
72.0b
59.2C
49.6
[ref. 195]
[ref. 197]
12-crown-4
70.0
55.2d
44.6d
66.49
[ref. 197]
15-crown-5
77.09
fref. 1971
18-crown-6
79.89
All the values are in kcal/mol unit with uncertainty 2.3 kcal/mol ( 0.1 eV) (except eV unit in parenthesis),
a RHF/6-31+G* at D3d symmetry, b MP2/6-31+G* at D3d symmetry, c -MP2/6-31+G' at C3v symmetry,
d RHF/6-31+G' at C3v symmetry,
e Hybrid quantum QM/MM molecular dynamics method,
f BDE for the process LH30+ LH+ + H20, RRKM calculation not included,
g aH for the process LH30+ L + H30+.
209

210
for the ab initio result. In spite of this good agreement between the QM/MM and ab
initio results, he reported that the QM/MM value is likely to be high, since the lower
AM1 OCCO torsion barrier allows for shorter K+-oxygen distances, and hence
stronger electrostatic stabilization, without a sufficiently large torsion barrier penalty.
Thompson also pointed out that both Wipffs and Kollman's results should be
decreased by an additional 2-4 kcal/mol, because of the different geometries used
in the calculations. These arguments predict a BDE range for (18-C-6)K+ of 53-60
kcal/mol, which overlaps the experimental value of 52.2 2.3 kcal/mol obtained in
this work.
Discussion
Previous investigations78'89 of gas-phase crown ether complexes have
indicated that the stabilities depend on the relative sizes of the metal ions and the
crown-ether cavities. However, other factors must also be considered.
The full long-range potential energy interaction, V(r, 0), for the CAD process
is given by:96
V (r,0) (3cos20-1)x[-
Or*
2 r
92(graJ Qq,
6r4 2r3
(4.13)
where r is the distance between the metal ion nucleus and the center of the crown
ether, 0 is the angle between r and the quadrupolar axis of the complex, q is the ion

211
charge, a is the polarizability of the neutral, and a, and ax are the parallel and
perpendicular polarities of the complex ion. The first two terms in eq 4.13 give the
charge-dipole (or charge-induced dipole) potential, and the third term is the ion-
quadrupole potential. Since ion-dipole (if the ligand has a dipole moment) and ion-
induced dipole interactions are the major factors which affect the BDE, the BDE is
related to the distance between the metal and the ligand, the charge density of the
metal ion, and the polarizability of the ligand. The size-matching of the metal ion
and the crown-ether is included in the internuclear distance, which is very small for
ions which fit well in the cavity. In their QM/MM molecular dynamics study of
dimethyl ether and 18-C-6 with K\ Thompson et a/.196 found that polarizability
effects are very important. Likewise, Howard's197 Amber MM calculation showed
that polarizabilities are significant in complexes of NH4+ and H30+ with crown ethers.
The results for which highly polarizable atoms or ionic interactions are treated
provide the relative agreement to the experimental values on the calculation of
relative conformation energies and cation complexation enthalpies.
The sizes of the metal ions198 and crown ether cavities199 are given in Table
1.1. For all three crown ethers the BDE trend is Na+ > K+ > Rb+ > Cs+. The BDE
difference is most pronounced for Na+ compared to K+, with decreasing differences
for successive metal ions. The BDE trend correlates with the inverse of the M+ radii,
or, correspondingly with the charge density of the metal ion.
Previous investigators78 79 83'85161 have proposed that the major factor affecting
the stability of the crown ether complexes is the rM7rL ratio (cf., Table 1.1, rM+ is the

212
size of metal ions, rL is cavity size of the crown ethers). Indeed, for Na+, K+, Rb+,
and Cs+the BDE order is 12-C-4 > 15-C-5 > 18-C-6 when RRKM corrections are
included, which follows the rM7rL order. However, for complexes with similar rM+/rL
((12-C-4)Na+ and (15-C-5)K+; (15-C-5)Na+ and (18-C-6)K+ have approximately the
same size ratio, 0.76 and 0.71; 0.52 and 0.48, respectively), the BDE's are much
greater for Na+ compared to K+. Thus, the charge density on the metal ion appears
to be an important effect, as shown above.
Another criterion has been the closeness of the rM7rL ratio to unity,78-79'83-85'161
i.e. a matching of ionic radius and the crown cavity radius. However, the data only
partially support that argument. For rM7rL significantly greater than one (12-C-4
with Rb+ and Cs+) the BDE's are quite low, especially for Cs+. For these complexes
the metal ion must sit above the cavity, thus increasing the M+-L internuclear
distance. Otherwise, having a rM7rL ratio close to one does not seem to favor a
large BDE. The ratio for (12-C-4)K+ is essentially equal to one, but the BDE is
considerably less than that for the Na+ complex. Likewise, the BDE's for 15-C-5 and
18-C-6 complexes decreases as the rM7rL values increase towards one.
Bowers et a/.161 measured collision cross sections for complexes of this
ligand with Li+, Na+, K+, and Cs+ by gas-phase ion chromatography (IC), and used
the experimental results to calculate gas-phase conformations by the Amber MM
procedure. The theoretical results indicated that all alkali metal ions are centered
on the axis of the crown cavity. However, only for K+ is the complex quasi-planar.
The small Li+ ion induces substantial distortion in the crown ether in order to better

213
coordinate with the partially charged oxygen atoms. When complexed to Na+ (or
Li+) the ligand adopts a cup-like structure to bring oxygens closer to the metal ion,
and the cation is cradled by the ligand. The Cs+ ion is too large for the crown cavity
and is forced out of plane. In (18-C-6)K+ the metal ion is coplanar with the ligand
and is exposed to the environment above and below the ring plane. The results
showed that the Li+ complex is most compact, which reflects relatively high charge
density on Li+. In solution, Li+ is essentially fully solvated by the crown ethers; it
benefits only in a minor way from further solvation, even though it has the largest
solvation energy of any alkali ions. The same holds to a lesser extent for Na+.
Hence, these two systems have the weakest BDE in solution. In contrast, K+ and
Cs+ are exposed and can benefit from solvation. Theory193'196 suggests that water
as a solvent displaces the K+ ion an average of 0.25 from crown center, resulting
in coordination with two water molecules in the first solvation shell. However, the
K+ ion is still primarily solvated by the crown in competition with water as a solvent
because K+ is bound much more strongly to small ethers than to H20 due to the
contribution of larger dispersion forces for the ethers.202'203 Similar reasoning
applies for the remaining alkali ions. Alkali ions strongly prefer association with
small ethers over association with water in gas phase. Theory indicates that
solvating a K+ complex in H20 has very little effect on the KVcrown structure.
The BDE's for (15-C-5)H30+ and (18-C-6)H30+ correspond to loss of neutral
water, i.e. the solvation energies for these complexes. Brodbelts group89 has
reported CID experiments on ammonium/crown ethers. In their work, the (18-

214
crown-6)NH4+ complex fragmented into the protonated crown ether by the loss of
NH3 which is analogous to the loss of H20 in our case. Izatt et al.200 showed that
H30+ can form a very stable complex with 18-crown-6, but not with 15-crown-5, in
agreement with the higher BDE for H30718-crown-6.
Attempts to measure the LH+ binding energies were unsuccessful, but the
observations indicate that H+ must be bound much more strongly than M+. The (18-
C-6)H30+ dissociated to (18-C-6)H+, which did not dissociate further. The (15-C-
5)H30+ complex underwent desolvation followed by fragmentation of the crown
ether with the H+ remaining attached to the fragments ((C2H40)nH+). Only the
protonated complex was formed by 12-C-4, which also fragmented with the H+
attached. Because no simple dissociation to L and H+ occurred and fragmentation
occurred without loss of H+, it may be concluded that H+ binds very strongly to the
crown ethers. The fragmentation trend indicates that the protonation strength
increases as the size of the crown ether decreases.
As shown in Table 4.3, the solvation energies of the hydronium ion
complexes are much smaller than the BDE's of metal ion complexes, especially for
(15-C-5)H30+. Loss of H20 from the hydronium/crown ethers indicates that the H+
is bound strongly to the crown ether cavity, with a much weaker solvation of the
entire complex.
Fragmentation also suggests that H307crown ethers have different structures
than MVcrown ethers. Studies of the protonation of crown ethers in condensed
phases have been performed by Jagur-Grodzinski and co-workers.201 The results

215
indicated that the proton coordinates to only two oxygens in the crown ether. The
protonation of polyethers, glymes and crown ethers in the gas phase has been
studied by Kebarle,40who observed that the protonated base, BH+, reacts with water
to form (BH+)H20. He interpreted the behavior of H+ with the crown ether as
analogous to H7(H20)n complexation, in which a cluster of H20's surrounds the
proton. According to Kebarle, the proton coordinates to two oxygens in the crown
ether (two strong hydrogen bonds), and an H20 molecule contributes to the
stabilization of the complex via its permanent dipole and more weakly through its
polarizability (one partial hydrogen bond).
Based on the calculations of Kollman192 et al., the complexation enthalpies
including polarizability of H30+ with crown ethers are about 20 kcal/mol higher than
the values obtained without including polarizability. They found one strong and two
weaker hydrogen bonds in this complex with C4 symmetry. Calculated results on
the H30+ complex do not agree with experimental values.43,202 They suspected the
experimental values corresponded to the complex H20 H+-ether and not
H30+-ether. Further calculations203 using single point 6-31G basis sets based on
the H20 H+-etherstructure were found to be in better agreement with experimental
values. Hence, it is most likely that the H30+ complex is essentially H20 H+-ether
and not H30+ ether.

216
Conclusion
For 12-crown-4, 15-crown-5, and 18-crown-6 the trend of the alkali-metal
BDE's is Na+ > K+ > Rb+ > Cs +. The BDE's for Na\ K +, Rb+ and Cs+ complexes
follow the order 12-C-4 > 15-C-5 > 18-C-6. The charge density on the metal ion is
a major factor in the observed trends. For the crown ether complexes with H30+,
loss of H20 and ion fragmentation were observed instead of dissociation of H30+.
Although the BDE's for LFT complexes could not be measured, the observations
indicate that H+ binds very strongly. The solvation energies for (15-C-5)H+ and (18-
C-6)H+ are much smaller than the metal ion BDE's, indicating that H+ is bound
strongly to the crown ether cavity, with weak solvation of the entire complex.
This investigation has shown that CAD with FTICR-MS can be used
effectively for measurements of bond dissociation energies. Quantitative results
can be obtained if experimental conditions are properly controlled, and the data
analysis accounts for systematic errors due to the internal energy and kinetic energy
distribution of precursor ion, thermal spread of the collision argon gas and the
lifetime of dissociation.
Although the CAD technique has been widely applied to measure BDEs in
several groups using different types of instruments, such as guided-ion beam and
quadrupole mass spectrometers, this study represents the first time it has been
applied in an ICR mass spectrometer. Different models have been applied to obtain
quantitative BDEs using different instruments; we have successfully applied the

217
CRUNCH model, which is a more thoughtful and accurate fitting procedure to obtain
BDEs for the first time in ICR mass spectrometry. Therefore, this technique can be
extensively utilized by ICR researches in the future to obtain BDE's for a variety of
compounds.

CHAPTER 5
CONCLUSION
Photodetachment was first used to measure the binding energies of
(CH3OH)X- (X = F, Cl, Br, I), (CH 3CN)Br, and (ROH)Br' (R= CH3, C2H5, /-C3H 7, n-
C3H 7), in a Fourier transform ion cyclotron resonance (FT-ICR) mass spectrometer.
The methanol complexes of F and Cl were prepared by sequential ion-molecule
reactions, while equilibrium solvent exchange was used for preparation of the Br
and I' complexes. The photodetachment process was observed over the
wavelength range of 260-350 nm using a tunable dye laser with a frequency
doubling accessory. The laser can be tuned over a wide wavelength range, and
allows an accurate threshold to be determined due to its very narrow bandwidth.
Assuming that the photodetachment process leads to complete dissociation
to the neutral solvent plus halogen atom, the binding energies were calculated by
subtracting the known electron affinities of halogens from the measured
photodetachment thresholds. The binding energy values obtained for (CH3OH)F',
(CH3OH)C|-, (CH3OH)Br, (CH3OH)l, (C2H5OH)Br, (/-C3H7OH)Br\ (n-C3H7OH)Br\
and (CH3CN)Br are 30.2, 18.9, 15.2, 14.6, 15.7, 16.6, 17.0, and 13.5 kcal/mol,
respectively.
218

219
Based on previously-published theoretical potential surfaces for the neutral
halogen-solvent system, the reported BE's may be high. For (CH3OH)F' and
(CH3OH)C|- the error should be only about 0.5 kcal/mol of the upper limit, but it may
be 1-2 kcal/mol for the Br and I' complexes due to the higher endothermicity for
neutral hydrogen abstraction compared to F' and Br complexes. Examination of
previously reported data indicates a random uncertainty of about 1-2 kcal/mol in
the reported BE's. Furthermore, binding energies determined by equilibrium
methods are subject to systematic error in the statistical mechanical estimation of
AS, which is based on the vibrational frequency of the H-X bond in ROFI-X'. As the
halide becomes heavier, the vibrational frequency of the ROFI-X' bond decreases.
It is possible that the vibrational frequency has not been estimated correctly by
previous authors. This would produce a systematic error in the entropy change,
which would subsequently affect the reported BE. For the alcohol-solvated halides,
the overall observation is an increase in binding energy with decreasing RO proton
affinity (CFI30 > C2H50' > i-C3H70 ~ n-C3FI70') and increasing X' proton affinity (
I > Br> C|-> F ).
This is the first study which has obtained BDE's using photodetachment in
ICR mass spectrometry. The ability to form solvated halides, to trap ions for long
periods of time, and to successfully measure BE's using the photodetachment
technique in FTICR-MS has shown that photodetachment of these complexes is a
reliable approach to determining binding energies. With these advantages and
success, photodetachment techniques can be extensively applied for quantitative

220
measurements such as electron affinities (EA's) and bond dissociation energies
(BDE's) of molecular/solvated anions and/or van der Waals forces of weakly bound
cluster anions formed by either ESI and/or ion-molecule reactions.
On-resonance CAD has been applied to obtain BDE's of alkali-metal ions
with crown ethers using appropriate threshold fit model. The use of an electrospray
ionization (ESI) source for soft ionization that produces intact, multiple charged and
solvated gas-phase ions directly from various compounds in solution has been
critical for this work. In addition, the ions formed using ESI carry little internal
energy compared to the other ionization techniques such as electron impact (El),
laser desorption (LD) and matrix-assisted laser desorption ionization (MALDI).
The bond dissociation energy trend is Cs+ < Rb+ < K + < Na + for all three
crown ethers. The BDE's for Na+ and K+ follow the same trend as the rM7rL ratios
(12-C-4 > 15-C-5 > 18-C-6), an indication that the size effect plays a major role in
the binding of these complexes. However, this cannot be the only factor. For
similar ratios ((12-C-4)Na+, (15-C-5)K+, (15-C-5)Na+, and (18-C-6)K+ are 0.76,
0.71, 0.52, and 0.48, respectively) the K+ complexes have much lower BDE's. It is
expected that the charge density of the metal ion is important in these cases. As
the size of the crown ether increases, the BDEs trend decreases for all metal ions.
The different dissociation pathways for metal and hydronium ion complexes
indicate that the hydronium ion complexes have different structure than the metal
complexes. Loss of water indicates that the hydronium ion complexes have the
structures of protonated crown ether complexes which is solvated with water. The

221
BDE's for (15-C-5)H30+ and (18-C-6)H30+ correspond to loss of neutral water, i.e.
the solvation energies for these complexes. Loss of H20 and fragmentation from
the hydronium/crown ethers indicates that the H+ is bound strongly to the crown
ether cavity, with a much weaker solvation of the entire complex. The observations
of higher BDE of (18-C-6)H30+ than (15-C-5)H30+ and the fragmentation of crown
ethers with proton attached to it for (15-C-5)H30+, but not for (18-C-6)H30+, indicate
that the proton bonds to 15-C-5 stronger than 18-C-6.
As mentioned in the Conclusions of Chapters 3 and 4, this is the first study
which has applied these two techniques to obtaining reliable BDE's using ICR mass
spectrometry. The ability to obtain quantitative measurements of bond dissociation
energies of metal and hydronium ions with crown ethers by improving threshold
determinations to account for a variety of systematic effects that would otherwise
distort the thermochemistry was demonstrated. Improvements in the accuracy of
measuring the thresholds and providing more precise quantitative measurements
in the gas phase by accounting for the effects of multiple ion-molecule collisions,
internal energy of precursor ions, kinetic energy distribution of precursor ions and
dissociation lifetime have also been achieved for the first time. The CRUNCH
program allows one to choose several options for different application purposes.
For example, one can select an appropriate fitting model, such as a simple fitting
model described in equations 1.18 and 1.19, without reactant's internal energies
and RRKM calculation, complicated model with reactant's internal energies summed
over all the vibrational states or/and RRKM calculations (eq 1.20). Also one can

222
select different types of Doppler broadening functions, such as Tiernan or Chantry
functions with or without ion energy distribution. With these abilities, the on-
resonance CAD technique can be comprehensively utilized to measure BDE's of
many types of ions, such as organic, inorganic, organometallic and
neutrals/solvents attached ions in our laboratory. The RRKM calculations in
CRUNCH also give other information such as dissociation rate constant (k),
probability (l-exp(-kt)), sum of states and density of states, which can be used for
the investigation of slow unimolecular reactions.

REFERENCE LIST
1. Sommer, H.; Thomas, H. A.; Hippie, J. A. Phys. Rev. 1950, 78, 806.
2. Sommer, H.; Thomas, H. A.; Hippie, J. A. Phys. Rev. 1951, 82, 697.
3. Wobschall, D.; Graham, J. R.; Malone, D. P. Phys. Rev. 1963, 131, 1565.
4. Wobschall, D. Rev. Sci. Instrum. 1965, 36, 466.
5. Anders, L. R.; Beauchamp, J. L.; Dunbar, R. C.; Baldeschwieler, J. D. J.
Chem. Phys. 1966, 45, 1062.
6. Mclver, R. T., Jr. Rev. Sci. Instrum. 1970, 41, 555;
7. Mclver, R. T., Jr. Int. J. Mass Spectrom. Ion Phys. 1971, 7, 471.
8. Beauchamp, J. L.; Anders, L. R.; Baldeschwieler, J. D. J. Am. Chem. Soc.
1967, 89, 4569.
9. Comisarow, M. B.; Marshall A. G. Chem. Phys. Lett. 1974, 4, 282.
10. Comisarow, M. B.; Marshall A. G. Chem. Phys. Lett. 1974, 26, 489.
11. Comisarow, M. B.; Marshall A. G. Can. J. Chem. 1974, 52, 1997.
12. Allemann, M.; Kellerhals Hp.; Wanczek, K. P. Chem. Phys. Lett. 1980, 75,
328.
13. Cody, R. B.; Burn, R. C.; Freiser, B.S. Anal. Chem. 1982, 54, 96.
14. McCrery, D. A.; Ledford, E. B., Jr.; Gross, M. L. Anal Chem. 1982, 54, 1435.
15. Mclver, R. T., Jr.; Bower, W. D. Tandem Mass Spectrometry, McLafferty, F.
W. Ed.; Wiley: New York, 1983, 287.
16. Cody, R. B.; Kinsinger, J. A.; Ghaderi, S.; Amster, I. J.; McLafferty, F. W.;
Brown, C. E. Anal. Chim. Acta. 1985, 178, 43.
17. Mclver, R. T.; Jr, Hunter, R. L.; Syka, M. S.; Labunsky, M. Paper presented
at the 31st Annual Conference on Mass Spectrometry and Allied Topics, Boston,
MA., May 8-13, 1983.
223

224
18. Marshall, A. G.; Wang, T.-C. L; Ricca, T. L. J. Am. Chem. Soc. 1985, 107,
7893.
19. Hunt, D. F.; Shabanowitz, J.; Yates, J. R., Ill; Zhu, N.-Z.; Russell, D. H.;
Castro, M. E. Proc. Natl. Acad. Sci. USA. 1987, 84, 620.
20. Henry, K. D.; Williams, E. R.; Wang, B. H.; McLafferty, F. W.; Shabnowitz, J.;
Hunt, D. F. Proc. Natl. Acad. Sci. USA. 1989, 86, 9075.
21. Castoro, J. A.; Koster, C.; Wilkins, C. L. Rapid Commun. in Mass Spec.
1992, 6, 239.
22. Schweikhard, L; Guan, Z.; Marshall, A. G. J. Mass Spec. Ion Proc. 1992,
120, 71.
23. Bruce, J. E.; Cheng, X.; Bakhiar, R.; J. Am. Chem. Soc. 1994, 116, 7839.
24. Buchanan, M. V. Fourier Transform Mass Spectrometry: Evolution,
innovation, and Applications. Buchanan, M. V. Ed.; ACS Symposium series, 359,
Washington, D.C. 1987
25. Dunbar, R. C. Analitical Applications of Fourier Transform Ion Cyclotron
Resonance Mass Spectrometry, Asamoto, B. Ed.; VCH publishers, Inc.: New
York, 1991.
26. Baykut, G.; Eyler, J. R., Trends in Anal. Chem., 1986, 5, 44.
27. Marshall, A. G.; Grosshans, P. B., Anal. Chem., 1991, 63, 215A.
28. Wang, M.; Marshall, A. G. Int. J. Mass Spec. Ion Proc., 1990, 100, 323.
29. Grosshans, P. B.; Marshall, A. G. Int. J. Mass Spec. Ion Proc., 1990, 100,
346.
30. Marshall, A. G.; Wang, T.-C. L; Ricca, T. L. Chem. Phys. Lett. 1984, 105,
233.
31. Tomlinson, B. L; Hill, H. D. W. J. Chem. Phys. 1973, 59, 1775.
32. Comisarow, M. B.; Melka, J. D. Anal. Chem. 1979, 51, 2198.
33. Comisarow, M. B. J. Chem. Phys. 1978, 69, 4097.

225
34. Comisarow, M. B. Adv. Mass Spec. 1978, 7, 1042.
35. Comisarow, M. B. Adv. Mass Spec. 1980, 9, 1698.
36. Keesee, R. G.; Castleman, A. W., Jr. J. Phys. Chem. Ref. Data, 1986, 15,
1011.
37. Jennings, K. R. Ed. Fundamentals of Gas Phase Ion Chemistry, NATO ASI
Series, Kluwer: Dordrecht, 1991, Vol. 347.
38. Miller, T. M. Adv. Electron. Electron Phys. 1981, 55, 119.
39. Janousek, B. K.; Brauman, J. I.; Simons, J. J. Chem. Phys. 1979, 71, 2057.
40. Sharma, R. B.; Blades, A. T.; Kebarle, P. J. Am. Chem. Soc, 1984, 106, 510.
41. Meot-Ner (Mautner), M. J. Am. Chem. Soc, 1983, 105, 4906.
42. Blades, A. T.; Jayaweera, P.; Ikonomou, M. G.; Kebarle, P. Int. J. of Mass
Spec. Ion Proc. 1990, 101, 325.
43. Meot-Ner (Mautner), M. J. Am. Chem. Soc, 1983, 105, 4912.
44. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc, 1985, 107, 2612.
45. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc, 1988, 110, 1087.
46. Jarrold, M. F.; lilies, A. J.; Bowers, M. T. J. Am. Chem. Soc, 1985, 107, 7339.
47. Freiser, B. S.; Hettich, R. L.; J. Am. Chem. Soc, 1987, 109, 3537.
48. Brucat, J. P.; Zheng, L.-S.; Pettiette, C. L.; Yang, S. and Smalley, R. E. J.
Chem. Phys. 1986, 84, 3078.
49. Geusic, M. E.; Jarrold, M. F.; Mcllrath, J. J.; Freeman, R. R.; Brown, W. L.; J.
Chem. Phys. 1987, 86, 3862.
50. Moseley, J. T.; Cosby, P. C.; Peterson, J. R. J. Chem. Phys. 1976, 65, 2512.
51. Bowen, K. H. J. Chem. Phys. 1985, 83, 3169.
52. Stevens, A. E.; Feigerle, C. S.; Lineberger, W. C. J. Chem. Phys. 1983, 78,
5420.

226
53. Pozniak, B.; Dunbar, R. C. Int. J. Mass Spec.; Ion Proc., 1994, 133, 97.
54. Schultz, R. H.; Crellin, K. C.; Armentrout, P. B. J. Am. Chem. Soc. 1991, 113,
8590.
55. Dalleska, N. F.; Honma, K..; Armentrout, P. B. J. Am. Chem. Soc. 1993, 115,
12125.
56. Magnera, T. F.; David, D. E.; Stulik, D.; Orth, R. G.; Jonkman, H. T.; Michl, J.
J. Am. Chem. Soc. 1989, 111, 5036.
57. (a) Lane, K. R.; Sallans, L.; Squires, R. R. J. Am. Chem. Soc., 1986, 108,
4368. (b) Sunderlin, L. S.; Wang, D.; Squires, R. R. J. Am. Chem. Soc., 1993,
115, 12060.
58. (a) Mclver, R. T, Jr. Kinetics of Ion-Molecule Reactions, Ausloos, P. Ed.;
Plenum: New York, 1978. (b) Mclver, R. T, Jr. Sci. Am. 1980, 243, 186.
59. Kebarle, P. Ion-Molecule Reactions, Franklin, J. F. Ed.; Butterworth: London,
1972, Vol. 1, 315.
60. Pauling, L. The Nature of the Chemical Bond, 3rd ed.\ Cornell University
Press: Ithaca, NY, 1960.
61. Murray-Rust, P.; Stallings, W. C.; Monti, C. T.; Preston, R. K.; Glusker, J. P.
J. Am. Chem. Soc., 1983, 105, 3206.
62. Scheiner, S. Reviews in Computational Chemistry, Lipkowitz, K. B.; Boyd, D.
B. Eds.; VCFI publishers: New York, 1991.
63. Latimer, W. M.; Rodebush, W. H. J. Am. Chem. Soc., 1920, 42, 1419.
64. Perutz, M. F.; Rich, A.; Crick, F. The Chemical Bond: Structure and
Dynamics', Zewail, A. Ed.; Academic Press: San Diego, 1992, ch. 1.
65. Luettringhaus, A.; Zeigler, K. Liebigs Ann. Chem. 1937, 528, 155.
66. Luettringhaus, A. Liebigs Ann. Chem. 1937, 528, 181.
67. Luettringhaus, A. Liebigs Ann. Chem. 1937, 528, 211.
68. Cram, D. J.; Trueblood, K. N. Top. Curr. Chem. 1981, 98, 43.

227
69. Weber, E.¡ Vogtle, F. Inorg. Chim. Acta 1980, 45, L65.
70.Parker, D. Adv. Inorg. Chem. Radiochem. 1983, 27, 1.
71. Samec, Z.; Papoff, P. Anal. Chem. 1990, 62, 1010.
72. De Jong, F.; Reinhoudt, D. N. Adv. Phys. Org. Chem. 1980, 17, 279.
73. Pedersen, C. J. J. Am. Chem. Soc. 1967, 89, 2495.
74. Pedersen, C. J. J. Am. Chem. Soc. 1967, 89, 7017.
75. Horwitz, E. P.; Dietz, M. L; Fisher, D. E. Solvent Extr. Ion Exch. 1991, 9, 1.
76. Chritensen, J.; Eatough, D. J.; Izatt, R. M. Chem. Rev. 1974, 74, 351.
77. Christinsen, J.; Eatough, D. J.; Izatt, R. M. Chem. Rev. 1974, 74, 251.
78. I. H. Hou, H. Zang, and D. V. Dearden, J. Am. Chem. Soc. 1993, 115, 5736.
79. Maleknia, S.; Brodbelt, J. S. J. Am. Chem. Soc. 1993, 115, 2837.
80. Dang, L. X.; Kollman, P. A. J. Am. Chem. Soc. 1990, 112, 5716.
81. Thompson, M. A.; Glendening, E. D.; Feller, D. J. Phys. Chem. 1994, 98,
10465.
82. Miller, J. M.; Balasanmugam, K.; Fulcher, A. Org. Mass Spectrom. 1989, 24,
497.
83. Zhang, H.; Chu, I. H.; Leming, S.; Dearden, D. V. J. Am. Chem. Soc. 1991,
113,
7415.
84. Zhang, H.; Dearden, D. V. J. Am. Chem. Soc. 1992, 114, 2754.
85. Maleknia, S.; Brodbelt, J. J. Am. Chem. Soc. 1992, 114, 4295.
86. Ammann, D. Ion and Enzyme Electrodes in Biology and Medicine, Kessler,
M. Ed.; University Park Press: Baltimore, 1976, p.22.
87. Thoma, A. P.; Simon, W. Metal-Ligand Interactions in Organic Chemistry and
Biochemistry. Pullman, B.; Goldblum, N. Eds.; D. Reidel Pulbishing Co.

228
Dordrecht: Holland, 1977, part 2.
88. Katritzky, A. R.; Malhotra, N.; Ramanathan, R.; Kemerait, R. C., Jr;
Zimmerman, J. A.; Eyler, J. R. Rapid. Commun. Mass Spectrom. 1992, 6, 25.
89. C. Liou and J. S. Brodbelt, J. Am. Chem. Soc. 1992, 114, 6761.
90. (a) Branscomb, L. M. Atomic and Molecular Processes. Bates, D. R. Ed.;
Academic: New York, 1962, 112. (b) Smith, S. J. Methods Exp. Phys. 1972, 7a,
179.
91. Moylan, C. R.; Dodd, J. A.; Han, C.; Brauman, J. I. J. Chem. Phys. 1987, 86,
5350.
92. Steiner, B. J. J. Chem. Phys. 1968, 49, 5097.
93. Wigner, E. P. Phys. Rev. 1948, 73, 1002.
94. Reed, K. J.; Zimmerman, A. H.; Andersen, H. C.; Brauman, J. I. J. Chem.
Phys. 1976, 64, 1368.
95. Gioumousis, G.; Stevenson, D. P. J. Chem. Phys. 1958, 29, 294.
96. Su, T; Bower, M. T. Gas Phase Ion Chemistry, Franklin, J. L. Ed.; Plenum:
New York, 1972, Vol. 1, 101.
97. Levine, R. D.; Bernstein, R. B. J. Chem. Phys. 1972, 56, 281.
98. Light, J. C. J.Chem. Phys. 1964, 40, 3221.
99. Weber, M. E.; Elkind, J. L.; Armentrout, P. B. J. Chem. Phys. 1986, 84, 6750.
100. Ervin, K. M.; Armentrout, P. B. J. Chem. Phys. 1987, 86, 2659.
101. Ervin, K. M.; Armentrout, P. B. J. Chem. Phys. 1986, 84, 6750.
102. Elkind, J. L.; Armentrout, P. B. J. Phys. Chem. 1987, 91, 2037.
103. Aristov, N.; Armentrout, P. B. J. Am. Chem. Soc. 1986, 108, 1806.
104. Hop, C.E.C.A.; McMahon, T. B.; Willett, G. D. Int. J. Mass Spec.; Ion Proc.,
1990, 101, 191.

229
105. Katritzky, A. R.; Watson, C. H.; Dega-Szafran, Z.; Eyler, J. R. J. Am. Chem.
Soc. 1990, 112, 2471.
106. P. Caravatti, H. Fr. Grutzmacher and H. L. Sievers. ICR Newsletter. 1994,
spring.
107. Chantry, P. J. J. Chem. Phys. 1971, 55, 2746.
108. Chesnavich, W. J.; Bowers, M. T. J. Phys. Chem. 1979, 83, 900.
109. Moini, M.; Eyler, J. R. J. Chem. Phys. 1988, 88, 5512. The KIM
microcomputer described in this reference has now been replaced by a (slightly)
upgraded computer system which is an Apple II computer with a John Bell Co.
interface card.
110. Mann, M.; Fenn, J. B.; Meng, C.-K.; Wong, S. F.; Whitehouse, C. M. Mass
Spec. Rev. 1990, 9, 37.
111. Fenn, J. B.; Yamashita; M. J. Phys. Chem. 1984, 88, 4451.
112. Fenn, J. B.; Yamashita; M. J. Phys. Chem. 1984, 88, 4671.
113. Whitehouse, C. M.; Dreyer, R. N.; Yamashita, M.; Fenn, J. B. Anal. Chem.
1985, 57, 675.
114. Mann, M.; Fenn, J. B.; Meng, C.-K.; Wong, S. F.; Whitehouse, C. M.
Science 1989, 246, 64.
115. Iribarne, J. V.; Thompson, B. A. J. Chem. Phys., 1976, 64, 2287.
116. (a) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1991, 63,
1989. (b) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1990, 62,
957.
117. Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1991, 113, 8534.
118. Smith, R. D.; Loo, J. A.; Edmonds, C. G.; Barinaga, C. J.; Udseth, H. R.
Anal. Chem. 1990, 62, 882.
119. Tang, L.; Kebarle, P. Anal. Chem. 1991, 63, 882.
120. Misaizu, F.; Sanekata, M.; Tsukamoto, K.; Fuke, K.; Iwata, S. J. Phys.
Chem. 1992, 96, 8259.

230
121. Yeh, C. S.¡ Willey, K. F.; Robbins, D. L; Pilgrim, J. S.; Duncan, M. A. Chem.
Phys. Lett. 1992, 196, 233.
122. Lessen, D. E.; Asher, R. L.; Brucat, P. J. J. Chem. Phys. 1990, 93, 6102.
123. Bradforth, S. E.; Arnold, D. W.; Metz, R. B.; Weaver, A.; Neumark, D. E. J.
Phys. Chem. 1991, 95, 8066.
124. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1985, 107, 766.
125. Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1988, 110, 7604.
126. Caldwell, G.; Rozeboom, M. D.; Kiplinger J. P.; Bartmess, J. E. J. Am.
Chem. Soc. 1984, 106, 4660.
127. Posey, L. A.; Campagnola, P. J.; Johnson, M. A. J. Chem. Phys. 1989, 91,
6536.
128. Dallestka, N. F.; Honma, K.; Sunderlin, L. S.; Armentrout, P. B. J. Am.
Chem. Soc. 1994, 116, 3519.
129. Yamabe, S.; Hirao K. Chem. Phys. Lett. 1981, 84, 598.
130. Hiraoka, K.; Yamabe, S. Int. J. Mass Spec. Ion Proc. 1991, 109, 133.
131. (a) Yamdagni, R.; Payzant, J. D.; Kebarle, P. Can. J. Chem. 1973, 51,
2507; (b) Arshadi, M.; Yamdagni, R.; Kebarle, P. J. Phys. Chem. 1970, 74, 1475;
132. Yamdagni, R.; Kebarle, P. J. Am. Chem. Soc. 1972, 94,2940.
133. (a) Caldwell, G.; Kebarle, P. J. Am. Chem. Soc. 1984, 106, 967; (b)
Caldwell, G.; Kebarle, P. J. Am. Chem. Soc. 1985, 107, 2612.
134. (a) Meot-Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 3858; (b) Meot-
Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 3854.
135. Meot-Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 7604.
136. (a) Meot-Ner (Mautner), M. J. Am. Chem. Soc. 1986, 108, 7525; (b) Meot-
Ner (Mautner), M. J. Am. Chem. Soc. 1988, 110, 3071.
137.(a) Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1983, 105, 2944.; (b)
Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1984, 106, 517.; (c) Larson, J.

231
W.; McMahon, T. B. Can. J. Chem. 1984, 62, 675.
138. (a) Hiraoka, K.; Mizuse, S. Chem. Phys. 1987, 118, 457; (b) Hiraoka, K.;
Mizuse, S.; Yamabe, S. J. Phys. Chem. 1988, 92, 3934.
139. Schuster, P. Electron-Solvent and Anion-Solvent Interactions, Kevan, L.
and Webster, B. C. Ed.; Elsevier Science: New York, 1976; Ch. 8.
140. Szulejko, J. E.; Wilkinson, F. E.; McMahon, T. B. Presented at 37th ASMS
Conference on Mass Spectrometry and Allied Topics. May 21-26, 1989, Miami,
FL, 333.
141. Tanabe, F. K. J.; Morgon, N. H.; Riveros, J. M. J. Phys. Chem. (in press).
142. Janousek, B. K.; Zimmerman, A. H.; Reed, K. J.; Brauman, J. I. J. Am.
Chem. Soc. 1978, 100, 6142.
143. Berger, S.; Brauman, J. I. J. Am. Chem. Soc. 1992, 114, 4737.
144. (a) Blair, L. K.; Isolani, P. C.; Riveros, J. M. J. Am. Chem. Soc. 1973, 95,
1057; (b) Isolani, P. C.; Riveros, J. M. Chem. Phys. Lett. 1975, 33, 362.
145. Riveros, J. M.; Breda, A. C.; Blair, L. K. J. Am. Chem. Soc. 1973, 95, 4066.
146. Faigle, J. F. G.; Isolani, P. C.; Riveros, J. M. J. Am. Chem. Soc. 1976, 98,
2049.
147. Riveros, J. M.; Ingemann, S.; Nibbering, N. M. M. J. Am. Chem. Soc. 1991,
113, 1053.
148. Peiris, D; Riveros, J. M.; Eyler, J. R. submitted.
149. Wetzel, D. M.; Brauman, J. I. Chem. Review. 1987, 87, 607.
150. Rosenfeld, R. N.; Jasinski, J. M.; Brauman, J. I. J. Am. Chem. Soc. 1979,
101, 3999.
151. Brauman, J. I., Gas Phase Ion Chemistry: Ions and Light, Bowers, M. T.
Ed.; Academic: New York, 1984; Vol. 3, Ch. 21.
152. Bevington, P. R., Data Reduction and Error Analysis for the Physical
Sciences, McGraw-Flill: New York, 1969, Ch.6.

232
153. Wladkowski, B. D.; East, A. L. L; Mihalick, J. E.; Allen, W. D.; Brauman, J.
J. Chem. Phys. 1994, 100, 2058.
154. Hotop, H.; Lineberger, W. C. J. Phys. and Chem. Ref. Data, 1975, 4, 539.
155. Gas Phase Ion and Neutral Thermochemistry, J. Phys. and Chem. Ref.
Data, Lias, S. G. Ed.; 1988, Vol. 17, No.1.
156. Bradforth, S. E.; Arnold, D. W.; Metz, R. B.; Weaver, A.; Neumark, D. M. J.
Phys. Chem. 1991, 95, 8066.
157. Glauser, W. A.; Koszykowski, M. L. J. Phys. Chem. 1991, 95, 1070.
158. Lide, D. R. Ed. CRC Handbook of Chemistry and Physics, Ref. Data, 75th
ed.; CRC press, 1995.
159. Del Bene, J. E. J. Phys. Chem. 1988, 92, 2874.
160. Peiris, D.; Yang, Y.; Ramanathan, R.; Williams, K. R.; Watson, C. H.; Eyler,
J. R. submitted.
161. Lee, S.; Wyttenbach, T; von Helden, G.; Bowers, M. T. J. Am. Chem. Soc.
1995, 117, 10159.
162. Lee, Y. C.; Allison, Int. J. Mass Sped, and Ion Phys. 1983, 51, 267.
163. (a) Lybrand, T. P.; Kollman, P. A. J. Chem. Phys. 1985, 83, 2923. (b)
Auffinger, P.; Wipff, G. J. Am. Chem. Soc. 1991, 113, 5976.
164. (a) Man, V. F.; Lin, J. D.; Cook, K. D. J. Am. Chem. Soc. 1985, 107, 4635.
(b) Bonas, G,; Bosso, C.; Vignon, M. R. J. Inclusion Phenom. Mol. Recognit.
Chem. 1989, 7, 637.
165. Cram. D. J.; Ho, S. P. J. Am. Chem. Soc. 1986, 108, 2998.
166. Inoue, Y.; Gokel, G. W.; Eds. Cation Binding by Macrocycles: Complexation
of Cationic Species by Crown Ethers. Marcel Dekker: New York. 1990.
167. Hilgenfeld, R.; Saenger, W. Host Guest Complex Chemistry Macrocycles:
Synthesis, Structures, Applications. Vgtle, F.; Weber, E. Eds. Springer-Verlag:
New York, 1985. Ch.5.
168. Frensdorff, H. K. J. Am. Chem. Soc. 1971, 93, 600.

233
169. Li, H.; Jiang, T; Butler, I. S. J. Raman Sped. 1989, 20, 569.
170. Takeuchi, H.; Arai, T; Harada, I. J. Mol. Strudure 1986, 146, 197.
171. Izatt, R. M.; Pawlak, K.; Bradshaw, J. S.; Bruening, R. L. Chem. Rev. 1991,
91, 1721.
172. Izatt, R. M.; Bradshaw, J. S.; Nielsen, S. A.; Lamb, J. D.; Christinsen, J.;
Sen, D. Chem. Rev. 1985, 85, 271.
173. Glendening, E. D.; Feller, D.; Thompson, M. A. J. Am. Chem. Soc. 1994,
116, 10657.
174. Yeh, C. S.; Willey, K. F.; Robbins, D. L.; Pilgrim, J. S.; Duncan, M. A. Chem.
Phys. Lett. 1992, 196, 233.
175. Pozniak, B.; Dunbar, R. C. Int. J. Mass Sped.; Ion Processes, 1994, 133,
97.
176. Dawson, P. H. Int. J. Mass Sped., Ion Processes. 1982, 43, 195.
177.Gentry, W. R. Gas Phase Ion Chemistry, Bowers, M. T. Ed. Academic
Press: New York, 1979, Vol. 2, p. 221.
178. Ervin, K. M.; Armentrout, P. B. J. Chem. Phys. 1985, 83, 166.
179. (a) Meyer, F.; Chen, Y.; Armentrout, P. B. J. Am. Chem. Soc. 1995, 117,
4071. (b) Hales, D. A.; Su, C.; Lian, L.; Armentrout, P. B. J. Chem. Phys. 1994,
100, 1049.
180. Dalleska, N. F.; Tjelta, B. L.; Armentrout, P. B. J. Phys. Chem. 1994, 98,
4191.
181. Marinelli, P. J.; Squires, R. R. J. Am. Chem. Soc. 1989, 111, 4101.
182. Wenthold, P. G.; Squires, R. R. J. Am. Chem. Soc. 1994, 116, 6401.
183. Magnera, T. F.; David, D. E.; Michl, J. J. Am. Chem. Soc. 1989, 111, 4100.
184. Nonose, S.; Tanaka, H.; Nagata, T.; Kondow, T. J. Phys. Chem. 1994, 98,
8866.
185. Shimanouchi, T.; Tables of Molecular Vibrational Frequencies, Part I.
National Standard Reference Data Series-NBS 6: Washington D.C. 1967.

234
186. Khan, F. A.; Clemmer, D. E.; Schultz, R. H.; Armentrout, P. B. J. Phys.
Chem. 1993, 97, 7978.
187. Gioumousis, G.; Stevenson, D. P. J. Chem. Phys. 1958, 29, 294.
188. Faulk, J. D.; Dunbar, R. C.; Lifshitz, C. J. Am. Chem. Soc. 1990, 112, 7893.
189. Fluang, F.-S.; Dunbar, R. C. J. Am. Chem. Soc. 1990, 112, 8167.
190. So, H. Y.; Dunbar, R. C. J. Am. Chem. Soc. 1988, 110, 3080.
191. Kim, M. S. Int. J. Mass Spect. and Ion Physics, 1983, 50, 189.
192. Parzen, E. Modern Probability Theory and It's Applications, Wiley: New
York. 1960, p.251.
193. Wipff, G.; Weiner, P.; Kollman, P. J. Am. Chem. Soc. 1982, 104, 3249.
194. Dang, L. X.; Kollman, P. A. J. Am. Chem. Soc. 1990, 112, 5716.
195. Thompson, M. A. J. Phys. Chem. 1995, 98, 4794.
196. Thompson, M. A.; Glendening, E. D.; Feller, D. J. Phys. Chem. 1994, 98,
10465.
197. Howard, A. E.; Singh, U. C.; Billeter, M.; Kollman, P. A. J. Am. Chem. Soc.
1988, 110, 6984.
198. Dailey, N. K. Synthetic Multidentate Macrocyclic Compounds, Izatt, R.M.;
Christensen, J.J., Eds. Academic Press: New York 1978. p 207.
199. Shannon, R.D. Acta Crystallogr., Sect. A: Found Crystallogr. 1976, 32, 751.
200. Christinsen, J.; Eatough, D. J.; Izatt, R. M. Chem. Rev. 1974, 74, 251.
201. (a) Schori, E.; Jagur-Grodzinski, J. J. Am. Chem. Soc. 1972, 94, 7957. (b)
Nae, N.; Jagur-Grodzinski, J. J. Am. Chem. Soc. 1972, 94, 7957.
202. (a) Lau, Y. K.; Ikuta, S.; Kebarle, P. J. Am. Chem. Soc. 1982, 104, 1462. (b)
Lau, Y. K.; Saluja, P. P. S.; Kebarle, P. J. Am. Chem. Soc. 1980, 102, 7429.
203. Singh, U. C.; Kollman, P. A. QCEP 1982, 2, 17.

BIOGRAPHICAL SKETCH
Yar-Jing Yang was born in Taiwan, 1964. She grew up in a big family with
two brothers and two sisters. In 1984, she attended National Chung-Hsiung
University at Taichung after passing the national college entrance exam and
received her B.S. degree in chemistry four years later. Upon graduation, she
passed the entrance exam for the graduate program at National Chung-Hsiung
University and enrolled in the graduate program in 1988. During graduate school,
she joined a program in the Institute of Atomic and Molecular Sciences, Academia
Sinica at Taipei. Her research area was photodissociation of CCI3N02 and CBrCI3
by translational spectroscopy using a time-of-flight (TOF) mass spectrometer. In
1988, she met Chen-Chan Hsueh, a wonderful and smart man, who was a research
assistant at the national laboratory. In 1990, she received her M.S. degree in
physical chemistry in May, married Chen-Chan Hsueh in June, and came to the
United States with her new husband in August.
One year later, she began her Ph.D. studies at the University of Florida. She
was involved with the activities of the Chinese Association Club as an associate
chairperson. In 1994, she went to the University of Sao Paulo in Brazil as a short
term exchange student to learn how to form gas-phase solvated anions by
sequential ion-molecule reactions. She received an award for an outstanding
235

236
presentation at the Florida Section ACS Meeting in May and was awarded a Russell
Dissertation Fellowship in December, 1995. In 1996, Mrs. Yang completed her
Ph.D. in physical chemistry under the guidance of Professor John R. Eyler. Her
current research involves determination of bond dissociation energies using
photodetachment, on-resonance collision activated dissociation (CAD) and
photodissociation techniques in FT-ICR mass spectrometry.

I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
C.
Eyler, Chai
ssor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Associate Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Willis B. Person
Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
lk s.
Kirk S. Schanze
Associate Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Associate Professor of Materials
Science and Engineering

This dissertation was submitted to the Graduate Faculty of the Department
of Chemistry in the College of Liberal Art and Sciences and to the Graduate School
and was accepted as partial fulfillment of the requirements for the degree of Doctor
of Philosophy.
May, 1996
Dean, Graduate School



100
80
60
40
20
0
Upon CAD
261.1
(12-crown-4)Rb+
84.9
Rb+
86.9
263.0
1
iO.O
140.0
230.0 320.0
Mass-to-charge ratio
i 1
410.0 500.0
184


-c-
o


Table 3.2. Photdetachment thresholds (ET) and binding energies (BE's) of solvated
halides.


Figure 3.10. Photodetachment spectrum of (CH3OH)l': experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.


111
o(E) (E-Et)* (3.10)
or In (N0/N) = a(E-ET)% (3.11)
in which a is a constant. To evaluate ET an iterative procedure was used. First, a
plot of ln(N0/N) versus laser energy, E, was prepared, as shown in Figures 3.8,
3.9, 3.10, and 3.11 for (C2H5OH)Br, (CH 3CN)Br \ (CH 3OH)l\ and (CH 3OH)F \
respectively. An approximate ET (ET1), was chosen from the sloping part of the plot,
and the experimental data points (E, ln(N0/N)) were substituted into eq 3.11 to
obtain a series of initial values of a. The mean (a,) and standard deviations of the
points about the mean for the initial values of a were calculated. This procedure
was repeated for decreasing ETs (at intervals of 0.01 kcal/mol) until the minimum
standard deviation about an was obtained. The reported ET was taken as the ETn
producing this minimum standard deviation.
The curves shown in Figures 3.8 to 3.11 are the fit to the data using the ET
obtained in this manner. For each of these plots a least-squares fit was also made
for the points in the near threshold region, and the threshold was taken as the
energy corresponding to the x-intercept. The results from the least-squares fit were
very close to those of obtained using the iterative procedure described above. The
uncertainty in ET was evaluated from the standard deviations of the least-squares
slope and y-intercept.152


Table 4.1. Vibrational frequencies of cyclohexane and 1,4-dioxane.


79
compounds having finite solubilities are used. The analyte is actually precipitating
out in solution at higher concentrations, thus forming clusters on the original
species. The concentration of sample may actually lead to decreased ion signal in
the MS for the analyte of interest.
Solubility
As solubility decreases, the likelihood of precipitation, and thus the formation
of clusters, in solution increases. As shown above, this may lead to reduced ion
signal. For this reason, it is necessary to consider the solubility of the compound
of interest when choosing solvents and sample concentrations. Generally, ESI-MS
spectra cannot be obtained for very large proteins (especially membrane-bound
types) due to solubility issues. In these and other cases, usually a better solvent
can be used to increase signal.
Molecular conformation
Most evidence concerning ESI leads to the conclusion that the efficiency of
generating protonated ion is related to the pKa, as shown above. Also important is
the accessibility of basic sites which attract the charges. This feature is important
in larger macromolecules, such as proteins, which have tertiary and even
quaternary structural components. In such species, the available basic sites may
be "hidden" from the protons they seek.117 For this reason, it is sometimes
necessary to change the solvent system, by adding dilute acid for example, in order
to change the structure of the protein. In severe cases, the protein may be
subjected to enzymatic or chemical degradation in order to increase ion signal.


1.8.The structures of polyether macrocycles.
30
1.9.The structures of metal/hydronium ion-crown ether complexes. . 33
1.10.Photodetachment process: (a) A' + hv A + e', (b) A + hv ET
+ C, (c) A' + hv B + C + e 40
2.1. Simplified block diagram of FTICR mass spectrometer 55
2.2. 2 tesla ICR instrument (a) 2 T magnet, (b) Inlet source, (c) Three
window flange, (d) One window flange, (e) ICR cell, (f)
Diffusion pump, (g) Ion gauge 58
2.3.Laser set-up for photodetachment experiment (1) Apple II
microcomputer controls "charge" and "fire" commands of YAG
laser, (2) Console sends external TTL pulse to trigger the
laser, (3) Power control from console to the cell; (a) Quartz
prism, (b) Laser quartz windows, (c) ICR cell, (d) 2 T magnet,
(e) Apple II computer, (f) Console, (g) Photodiode 60
2.4. Pulse sequence for photodetachment experiments 63
2.5. Cylindrical cell 65
2.6.7 T ICR instrument (a) 7 T magnet, (b) Infinity cell, (c) 400 L/s
cryopumps, (d) 800 L/s cryopump, (e) Mechanical pumps, (f)
Capillary coated with platinum at entrance and exit ends, (g)
Needle, (h) Syringe pump 70
2.7. Ion formation in ESI 73
2.8. ESI design 76
2.9. Electrostatic lenses in transfer region 83
2.10.Infinity cell with three "side-kick" plates: (a) PVX1, (b) PVX3,
(c) PVX4 plates, (d) the macor plate which is attached to
trapping plates to hold the "side-kick" plates and trapping
plates, (e) the rf-shimmed trapping plate, (f) the Infinity cell
with four brass plates for excition and detection 85
2.11. Pulse sequence for CAD experiments 87
viii


Table 1.1. Size of alkali-metal ions and cavity size of crown ethers.


24
intact, multiple charged and solvated gas-phase ions directly from various
compounds in solution has been critical for this work. In addition, the ions formed
using ESI carry little internal energy compared to other ionization techniques such
as electron impact (El), laser desorption (LD) and matrix-assisted laser desorption
ionization (MALDI).
Solvated Anions
Living organisms and nonliving matter are affected by interactions with
organic chemicals, which are highly dependent upon how molecular structures
produce site-specific interactions, such as hydrogen bonds. In general, in the life
sciences and in chemistry, the hydrogen-bond (HB) is perhaps the most important
kind of specific molecular interaction between molecular species. Examples include
solubility, bodily transport and docking to active sites. Hydrogen bonding was first
recognized in the physical properties of liquids, and was thought to involve only
fluorine, oxygen or nitrogen atoms bridged by an intervening hydrogen.60 Since
then structural studies have indicated that atoms such as chlorine, bromine and
sulfur can serve as hydrogen bond acceptors,6162 and that hydrogens bonded to
carbon and sulfur can serve as hydrogen bond donors.63 Another significant type
of site-specific noncovalent interaction has been identified as "halogen bonding".63
Therefore, study of alcohol solvated halogen ions through hydrogen bonding has
been an important issue.
Ion solvation has been studied extensively for many years by both theoretical


22
A few years ago the development of high pressure mass spectrometry
reached a stage where it could be applied successfully to chemical problems. The
technique and its applications have been summarized by Kebarle.59 An impressive
number of clustering equilibria of the type in eq. 1.6 have been studied,
X Sn X S^ + S, In Kntrv1= AS/R AH/(RT) (1.6)
where X represents a central ion, S a solvent molecule, and X Sn and X S^ are two
cluster ions of different stoichiometry. Based on the accessibility of suitable ion
sources, single ion solvation can be studied by mass spectrometry provided
equilibrium is established in such a system.
Photodissociation thresholds can be interpreted as reflecting at least an
upper limit to the thermochemical dissociation threshold. Moseley et al.50 measured
the vibrationally resolved photodissociation spectrum of C03', which was analyzed
to give a reliable limit of 1.85 eV for D(C02-0).
Energy-resolved collision-induced dissociation of metal carbonyl anions
M(CO)n' (M=V, Cr, Mn, and Co) were investigated to determine sequential metal-
carbonyl bond energies by Squire's group.57b They used the model described in the
Methodology Section of this Chapter to obtain fairly accurate BDE's of 30.8 3.5,
40.6 3.5, 40.6 3.9, and 39.7 3.7 kcal/mol for (CO)5V-CO, (CO)4Cr-CO,
(CO)4Mn'-CO, and (CO)3Co -CO, respectively.
Electron photodetachment spectroscopy is one of the most useful methods


48
M
(eV)
(1.21)
8m M+m
where Ecm and E,ab are the energies in the center-of-mass and laboratory frame,
respectively, M is the collision target molecular mass, m is the precursor ionic mass,
t is the r.f. pulse length, Erf is the electric field (which is giving by Vpp Id for an ideal
infinitely long parallel plate cell, where Vpp is the applied r.f. voltage (peak-to-peak))
and d is the plate spacing diameter of the ICR cell. Unfortunately, the measured Erf
for cells of finite dimensions is only a fraction of the ideal value, so it is calculated
using aVpp Id, where a is a geometry factor for the ICR cell (a=0.90 for the Infinity
cell used in this work).106 The activated ion is then allowed to interact with the
collision gas (Ar or Xe) for a short period of time (CAD delay time). The translational
energy is converted into internal energies such as rotational and vibrational
energies via ion/molecule collisions, allowing dissociation to occur during the CAD
delay time.
- L + T (ML)" + T->M' + L+ T
(1.22)
where M+-L is the ion-ligand complex, T is the collision target gas (Ar), M+ is the ion,
and L is the neutral ligand. After dissociation, the intensities of the fragment ions
are measured. The ratio of the intensity of the product ion to the total ion intensity
is determined. From plots of the intensity ratio (lM7(lM++ lML+)) vs. the center-of-mass


Quench pulse
Electron impact pulse
Ion-molecule reactions
n
Ejections
Laser fire
Excitation pulse
Detection
Time


100
80
60
40
20
0
(CH3OH)F-
51.029
Mfa
HMMu
17.0
t
21.1
27.7
40.4
74.8
499.7
Mass-to-charge ratio
CD
CD


Relative rate coefficient
0.5
0.4
0.3
0.2
0.1
Energy in center-of-mass frame/(eV
203


162
where E is the relative translational energy of the precursor ions, Erot is the rotational
energy (3kT/2 = 0.039 eV), ET is the zero-kelvin BDE (D0 as shown in Figure 4.2),
n is an adjustable parameter, E¡ and g¡ represent the energy and population of the
ith vibrational state, respectively, and P is the probability that metastable ions
formed with initial translational energy, E, and internal energy, E¡+Erot, will dissociate
within the experimental time window (10-30 msec in this case).
The CRUNCH program uses nonlinear least-squares analysis to optimize the
parameters o0, n, and E T to give the best fit to the data without including RRKM
caculations. After obtaining optimized o0, n, and E T, the RRKM calculations were
included. In this manner, we can observe the shift of the threshold with RRKM
calculations. The RRKM calculation lowered the ET by approximately 0.3 to 0.8 eV
(7 to 18 kcal/mol) in this study and the effect is more pronounced for larger
complexes. As the size of complexes increases, the number of vibrational degrees
of freedom increases. Therefore, the dissociation lifetime becomes more important
and dissociation may not be complete within the CAD delay time. Although this
problem is not as severe in ICR experiments (10-30 msec) as it is in ion-beam
experiments (microseconds), in some cases the intermediates may require
milliseconds to several seconds for dissociation.188'190 In the optimization procedure,
the parameters are very sensitive to the fitting of the appearance curve; in other
words, a slight change in the number can cause tremendous distortion to the fitting.
Therefore, the zero-kelvin bond dissociation energy, D0, can be obtained with very
small uncertainty, 0.01 to 0.05 eV ( 0.2 to 1.0 kcal/mol) without RRKM


Figure 1.2 Standard pulse sequence for ICR experiment.


Structures of CH3OHX-
0.97A
H ,1.44A
1.47A 1.02
F0.
174.8
O
v1.42A
C
H
H
o
H
o
H
C
H
2.39A *98A o
0 1.43A
Br-""
162.4
H
H
C
H
Cl
H
0 0.98A
2-26A_ H v O.
v1.43A
163.2 'C H
yj
H
o 0.97A
2..82A 0.143X
160.4 XC H
H
/j
H
ro
co


Figure 2.7. Ion Formation in ESI.


Figure 2.5. Cylindrical cell.


0 35 CoH.OHBr"


Table 4.3. Bond dissociation energies for alkali-metal and hydronium ion crown ether complexes including present
work, previous work, and theoretically calculated values.


171
H) using a voltage difference, AV, between capillary and skimmer, equal to 70 V.
In addition to the simple adduct ions, the sandwich-type complexes (L2M+) were
also observed in complexes of 12-C-4 with M+ at AV=70. With a capillary voltage
of 25 V (AV=20 V), the sandwich-type complexes (L2M+) were observed for both 12-
C-4 and 15-C-5 with M+ as shown in Figures 4.5 and 4.6 for (12-C-4)2H30+, (12-C-
4)2H+ and (12-C-4) 2Rb +, but not for 18-C-6. This behavior indicates that the
stabilities of the L2M+ complexes increase as the size of the crown ether decreases.
As the capillary voltage was increased from 25 to 75 volts (AV from 20 to 70
V), the intensity of sandwich-type complexes decreased and the intensity of the
molecular ions increased, as shown in Figure 4.7 and 4.8 for Fl30+ with 12-crown-4.
This indicates that the binding in the sandwich-type complexes is much weaker than
that in the molecular ions. As the capillary voltage was increased, the more weakly
bound sandwich-type complexes dissociated between the capillary and skimmer as
a result of collision with background neutrals.
The sandwich complexes of 15-C-5 with Na+, K+, FT and Fl30+ and 12-C-4
with K+ were observed in previous work with an internal ESI source and infrared
multiple photon dissociation (IRMPD)160. The (15-C-5)2K+ and (15-C-5) 2Na +
complexes were fragmented to the molecular ions, (15-C-5)K+ and (15-C-5)Na+, by
the laser radiation. This fragmentation behavior is similar to that observed in the
present study.
For the metal-ion complexes, the successor ions of the CAD process were
the bare metals in all cases,


228
Dordrecht: Holland, 1977, part 2.
88. Katritzky, A. R.; Malhotra, N.; Ramanathan, R.; Kemerait, R. C., Jr;
Zimmerman, J. A.; Eyler, J. R. Rapid. Commun. Mass Spectrom. 1992, 6, 25.
89. C. Liou and J. S. Brodbelt, J. Am. Chem. Soc. 1992, 114, 6761.
90. (a) Branscomb, L. M. Atomic and Molecular Processes. Bates, D. R. Ed.;
Academic: New York, 1962, 112. (b) Smith, S. J. Methods Exp. Phys. 1972, 7a,
179.
91. Moylan, C. R.; Dodd, J. A.; Han, C.; Brauman, J. I. J. Chem. Phys. 1987, 86,
5350.
92. Steiner, B. J. J. Chem. Phys. 1968, 49, 5097.
93. Wigner, E. P. Phys. Rev. 1948, 73, 1002.
94. Reed, K. J.; Zimmerman, A. H.; Andersen, H. C.; Brauman, J. I. J. Chem.
Phys. 1976, 64, 1368.
95. Gioumousis, G.; Stevenson, D. P. J. Chem. Phys. 1958, 29, 294.
96. Su, T; Bower, M. T. Gas Phase Ion Chemistry, Franklin, J. L. Ed.; Plenum:
New York, 1972, Vol. 1, 101.
97. Levine, R. D.; Bernstein, R. B. J. Chem. Phys. 1972, 56, 281.
98. Light, J. C. J.Chem. Phys. 1964, 40, 3221.
99. Weber, M. E.; Elkind, J. L.; Armentrout, P. B. J. Chem. Phys. 1986, 84, 6750.
100. Ervin, K. M.; Armentrout, P. B. J. Chem. Phys. 1987, 86, 2659.
101. Ervin, K. M.; Armentrout, P. B. J. Chem. Phys. 1986, 84, 6750.
102. Elkind, J. L.; Armentrout, P. B. J. Phys. Chem. 1987, 91, 2037.
103. Aristov, N.; Armentrout, P. B. J. Am. Chem. Soc. 1986, 108, 1806.
104. Hop, C.E.C.A.; McMahon, T. B.; Willett, G. D. Int. J. Mass Spec.; Ion Proc.,
1990, 101, 191.


Calculated
values
C6Hi2
Scaled
values
Exp.
values185
Calculated
values
c4h8o2
Scaled
values
Exp.
values
245.ob
221.29
248.47
248
221.67
248.62
288
350.76
375.88
383
335.62
364.51
424
456.39
475.33
426
425.80
451.17
435
456.40
475.34
426
518.35
535.84
490
578.27
583.42
523
661.48
659.07
610
929.16
860.8
785
972.33
899.09
837
929.29
860.89
785
1060.59
961.39
853
1049.16
945.99
802
1170.16
1035.65
881
1049.19
946.01
863
1175.89
1039.44
889
1094.78
977.27
863
1359.01
1156.52
1015
1200.45
1047.58
907
1379.96
1169.43
1052
1221.44
1061.21
907
1391.58
1176.55
1086
1302.69
1112.96
1027
1409.73
1187.61
1110
1302.84
1113.06
1027
1471.34
1224.67
1128
1312.33
1119
1030
1491.88
1236.86
1136
1355.75
114 5.95
1057
1509.76
1247.41
1217
1355.90
1146.05
1057
1529.81
1259.17
1256
1441.88
1198.23
1157
1609.93
1305.43
1291
1442.08
1198.35
1157
1667.50
1338.00
1305
1442.15
1198.4
1261
1680.19
1345.11
1335
1450.94
1203.65
1261
1689.54
1350.33
1369
1451.09
1203.74
1266
1848.15
1436.98
1378
1535.38
1253.34
1266
1927.30
1478.97
1397
1541.21
1256.72
1347
2128.73
1582.71
1444
1683.27
1337.36
1347
2290.68
1663.43
1449
1683.31
1337.38
1355
2303.92
1669.95
1457
1702.06
1347.78
1355
2400.81
1717.27
1459
1703.53
1348.59
1383
4742.10
2900.54
2856
1703.59
1348.63
1437
4742.83
2900.98
2856
1816.87
1410.36
1437
4744.38
2901.92
2863
1949.07
1480.27
1443
4745.55
2902.63
2863
1949.29
1480.39
1443
4785.09
2926.66
2968
2190.75
1603.21
1457
4785.46
2926.88
2968
2190.86
1603.27
1457
4789.80
2929.54
2970
2282.09
1648.35
1465
4790.02
2929.67
2970
4750.45
2887.95
2852
4754.47
2890.37
2860
4754.65
2890.48
2863
4759.07
2893.14
2863
4759.28
2893.27
2897
4761.98
2894.89
2897
4781.02
2906.4
2915
4782.29
2907.17
2930
4782.46
2907.28
2930
4788.69
2911.05
2930
4788.91
2911.19
2933
4793.89
2914.21
2933
152


Figure 1.9 The structures of metal/hydronium ion-crown ether complexes.


Figure 4.7. Mass spectrum of (12-C-4)H30+ at AV=20 volts. Peaks corresponding to (12-C-4)H+, (12-C-4)2H+, and (12-
C-4)2H30 + ions were observed. The peak intensity of (12-C-4)2H+ is higher than that of (12-C-4)H+.


131
Electron Affinities3 (EA) and Proton Affinities6 (PA) of Species Studied in this Work
Halogen
EA/kcal-mol'1
()=/kJ-mol1
Anion
PA/kcal-mol'1
()=/kJmol'1
F
78.37 (327.8)
OH
390.8 (1635)
Cl
83.44 (349.0)
ch3o-
381.4 (1595)
Br
77.45 (323.9)
c2h5o-
377.8 (1580)
I
70.53 (295.0)
i-C3H70"
376.0 (1573)
n-C3H70"
376.4 (1574)
t- C3H70
374.9 (1564)
ch2cn-
372.2 (1557)
f-
371.3 (1553)
ci-
333.3 (1394)
Br
323.6 (1353)
I
314.3(1315)
a-Ref. 154
b- Ref. 155


Figure 2.9. Electrostatic lenses in transfer region.


6
(LD), fast atom bombardment (FAB), electrospray ionization (ESI), and matrix
assisted laser desorption ionization (MALDI). The basic FTICR-MS experiment is
conducted using a series of computer-controlled pulses, as depicted in Figure 1.2.
Initially a quench pulse is applied to the trapping plate(s) to remove all the ions from
the cell prior to starting the next pulse sequence. After ion formation, ions are
trapped in the cell by constraining their z-axis motion (along the magnetic field) by
applying low voltages (0.5 to 5 V typically) to the trapping plates. Either positive or
negative ions can be trapped in the cell by changing the polarity of the voltage
applied to the cell plates.
Ion excitation
Three types of ion excitation are generally used for ion detection, as shown
in Figure 1.3. (1) In impulse excitation,18 an approximation to an ideal delta function
pulse (infinite amplitude, zero width) is applied. This results in a flat excitation
spectrum and should excite all ions equally. The angular cyclotron frequency is of
the order of wcmax = 1/At, where At is the pulse width. (2) With chirp excitation,10 a
radio frequency (r.f.) pulse is applied whose frequency sweeps rapidly over a range
from the lowest to highest frequency desired. The disadvantage of chirp excitation
is the nonuniform excitation of the ions, which becomes pronounced for ions near
the edge of the swept frequency range. The final cyclotron radius is limited to less
than the full dimensions of the cell since ions absorb different amounts of power
during the chirp. This introduces peak height distortion and the frequency chirp
excitation also contributes to z-axis ejection. (3) For stored waveform inverse


100
80
60
40
20
0
Upon CAD
283.2
(18-crown-6)H30+
(18-crown-6)H+
26
5.2
17.1
113.7
210.3 306.9
Mass-to-charge ratio
403.4
500.0
188


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
C.
Eyler, Chai
ssor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Associate Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Willis B. Person
Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
lk s.
Kirk S. Schanze
Associate Professor of Chemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and
quality, as a dissertation for the degree of Doctor of Philosophy.
Associate Professor of Materials
Science and Engineering


100
80
60
40
20
0
AV=20 volts
283.2
(18-crown-6)H30+
t i 1 1 r
17.1 113.7 210.3 306.9 403.4
Mass-to-charge ratio
186


51
halide compounds which is described detail in Chapter 3 is excessively complicated
by CCI4 and SF6. On the other hand, photoelectron spectroscopy (PES) can obtain
the same information by measuring the electron kinetic energy at a fixed laser
wavelength.
Photodetachment threshold spectroscopy and photoelectron spectroscopy
have been analyzed to obtain electron affinities and molecular structural
information, and each has its particular strength and weaknesses.
Photodetachment can be utilized to determine electron affinities with an accuracy
limited essentially by the wavelength resolution of the light source, while similar
determinations using photoelectron spectroscopy studies require only that the
wavelength of the intense light source be shorter than the wavelength
corresponding to the photodetachment threshold. As compared to CAD and
thermal-equilibrium techniques, detachment thresholds obtained using
photodetachment are much sharper and can be measured with greater accuracy.
CAD
The use of CAD techniques in FTICR-MS to measure thermodynamic
quantities for a variety of ionic species provides reaction cross section (or rate
constants) over a wide range of kinetic energies. This technique does not require
a laser apparatus and associated laser alignment compared to photodissociation
and PES techniques. CAD is an alternate approach to determining BDE when the
ions of interest do not absorb or absorb little laser energy at near threshold ranges.
However, the accuracy of such information is dependent on the models used to


220
measurements such as electron affinities (EA's) and bond dissociation energies
(BDE's) of molecular/solvated anions and/or van der Waals forces of weakly bound
cluster anions formed by either ESI and/or ion-molecule reactions.
On-resonance CAD has been applied to obtain BDE's of alkali-metal ions
with crown ethers using appropriate threshold fit model. The use of an electrospray
ionization (ESI) source for soft ionization that produces intact, multiple charged and
solvated gas-phase ions directly from various compounds in solution has been
critical for this work. In addition, the ions formed using ESI carry little internal
energy compared to the other ionization techniques such as electron impact (El),
laser desorption (LD) and matrix-assisted laser desorption ionization (MALDI).
The bond dissociation energy trend is Cs+ < Rb+ < K + < Na + for all three
crown ethers. The BDE's for Na+ and K+ follow the same trend as the rM7rL ratios
(12-C-4 > 15-C-5 > 18-C-6), an indication that the size effect plays a major role in
the binding of these complexes. However, this cannot be the only factor. For
similar ratios ((12-C-4)Na+, (15-C-5)K+, (15-C-5)Na+, and (18-C-6)K+ are 0.76,
0.71, 0.52, and 0.48, respectively) the K+ complexes have much lower BDE's. It is
expected that the charge density of the metal ion is important in these cases. As
the size of the crown ether increases, the BDEs trend decreases for all metal ions.
The different dissociation pathways for metal and hydronium ion complexes
indicate that the hydronium ion complexes have different structure than the metal
complexes. Loss of water indicates that the hydronium ion complexes have the
structures of protonated crown ether complexes which is solvated with water. The


141
these results are also valid at 0 K. As argued above, the photodetachment process
produces a neutral of the same geometry as the initial ion. The occupied vibrational
levels and the energy distribution in the neutral should also correspond to those in
the ion (T= 298 K). This means that the usual temperature correction to the energy
change (using heat capacities calculated from vibrational frequencies) should be
negligible, and the reported thresholds should be the same at either 298 K or 0 K.
Since the halide electron affinities used in the data analysis are zero-kelvin
values,154 the binding energies in Table 3.2 are valid at 0 K. Because the usual
AnRT correction is zero at 0 K, the BE values also represent enthalpy changes at
0 K.
Conclusion
The results in Table 3.2 show that the binding energy trend is I' < Br < Cl'
F' for the methanol solvates. The BE difference for (CH3OH)F' compared to
(CH3OH)Cr is 11.1 kcal/mol, whereas the (CH3OH)CI'/(CH3OH)Br and (CH3OH)Br
/(CH3OH)l" differences are only 3.7 and 0.6 kcal/mol, respectively. The same order
and approximate differences were found by previous investigators. Referring to
Table 3.1, this is also the trend of the proton affinities of the halide ions. For the
complexes with bromide ion, the BE order is CH3OH < C2H5OH < /'-C3H7OH < n-
C3FI7OH, which is the same as the order of the alkyl-group size, but opposite to the
RO' proton affinity order. Thus, for the alcohol-solvated halides, the overall


Figure 4.2. Appearance curve of (18-C-6)Rb+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.


80
Molecular weight
The possible effect of molecular weight on ionization is the mass
discrimination118 which may occur in both the mass analyzer and the ion optics.
Usually by adjusting the voltage parameters of the ion optics one can observe the
effect of charge states for multiple charged ions.
Solvent properties
High conductivity solutions will not electrospray well under "normal"
conditions since the formation of a Taylor cone is prevented in this case.119
Solutions which are high in conductivity usually have acids, bases or salts added
to them for various reasons. When using these solvents for electrospray, it is
necessary to remove these additives to the sample, as the ion signal and stability
are strongly dependent on the quality of the spray generated and the formation of
the Taylor cone.
In a similar manner, solutions with a high surface tension are also difficult to
use with ESI. It is for this reason that pure water does not electrospray well. This
problem is generally overcome through the addition of an organic modifier, such as
methanol, to the sample solution. The supporting solution is at least 20% in organic
makeup to spray well. The accepted standard for optimum performance is 50%.
Finally, the solvent volatility may also slightly affect the ion signal observed in ESI
for a particular analyte. After a droplet has been formed by ESI, it must evaporate
substantially before the electric field on the droplet surface is high enough to induce
a coulombic explosion and subsequent ion evaporation of the analyte. This rate of


41
frequency light and an ion beam. Photodetachment spectroscopy has been used
to determine a number of thermochemical and spectroscopic parameters for a
variety of atoms and molecules, including the electron affinity of the neutral
photoproduct, the energies and transition moments of electronically excited states
of anions, spin-orbit splitting in radicals, and vibrational and rotational spacings in
both neutrals and anions. Most anions of organic molecules exhibit only
photodetachment; a number of inorganic anions can also photodissociate upon
irradiation with visible light. Photodetachment experiments are useful in probing
excited states of anions, as well as providing information important in determining
binding strengths.
The use of electron affinities derived from either photodetachment
experiments or literature data to determine bond dissociation energies is applied in
this work. By measuring the depletion of the parent ion as a function of photon
energy, the photodetachment threshold, related directly to the binding strength of
solvated ions, can be determined.
The raw data in a photodetachment experiment are the measurements of the
precursor ion signal with and without laser irradiation. At long wavelengths, the
photon energy is not high enough to detach an electron from the precursor anion
so the intensity of the precursor ion remains unchanged, i.e. the fraction of
photodetachment is zero. When the photon energy is high enough, the intensity of
the precursor ion decreases and the fraction of photodetachment is nonzero,
increasing linearly near threshold. In order to convert these data into a


Figure 4.21. Appearance curve of (12-C-4)Cs+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.


149


217
CRUNCH model, which is a more thoughtful and accurate fitting procedure to obtain
BDEs for the first time in ICR mass spectrometry. Therefore, this technique can be
extensively utilized by ICR researches in the future to obtain BDE's for a variety of
compounds.


Figure 2.1. Simplified block diagram of FTICR mass spectrometer.


Figure 4.5. Mass spectrum of (12-C-4)Rb+ at AV=20 volts. Peaks corresponding to (12-C-4)Rb+ and (12-C-4)2Rb+ ions
were observed.


156
(4.3)
Since the total ion intensity, ltot= lp + X ls, is constant, the exponential term in eq 4.2
may be expanded in a power series to give
= ks nt
(4.4)
Providing that the Ar number density is sufficiently low. For each experiment, n and
t are constants. Thus, ls/ltot is proportional to the rate coefficient, and a plot of ls/llot
versus E cm (from equation 1.21) gives the experimental ion appearance curve as
shown in Figure 4.2 for (18-crown-6)Rb+.
As stated in the Introduction, the CRUNCH program developed by
Armentrout and cowokers178,179 was used to obtain the bond dissociation energies.
The CRUNCH calculates the BDE (plus other parameters) which produces the best
fit to the experimental appearance curve. In addition to ls/ltot and E^ values, the
program requires other input data and fitting procedure which are described in the
following sections.


Relative rate coefficient
1
0.8
0.6
0.4
0.2
0
(18-crown-6)Rb
1.65
i r
0
2 4 6 8 10 12 14 16 18
Energy in center-of-mass frame/eV
20
158


Reference
Mixer
Sum and
Difference Signal
Low Pass Filter
~1,
Difference signal


Relative rate coefficient
0.5
0.4
0.3
0.2
0.1
0
(12-crown-4)Cs
1.71
i 1 1 r
i 1 r
mmmA
0 2 4 6 8 10 12 14 16 18 20
Energy in center-of-mass frame/eV
207


211
charge, a is the polarizability of the neutral, and a, and ax are the parallel and
perpendicular polarities of the complex ion. The first two terms in eq 4.13 give the
charge-dipole (or charge-induced dipole) potential, and the third term is the ion-
quadrupole potential. Since ion-dipole (if the ligand has a dipole moment) and ion-
induced dipole interactions are the major factors which affect the BDE, the BDE is
related to the distance between the metal and the ligand, the charge density of the
metal ion, and the polarizability of the ligand. The size-matching of the metal ion
and the crown-ether is included in the internuclear distance, which is very small for
ions which fit well in the cavity. In their QM/MM molecular dynamics study of
dimethyl ether and 18-C-6 with K\ Thompson et a/.196 found that polarizability
effects are very important. Likewise, Howard's197 Amber MM calculation showed
that polarizabilities are significant in complexes of NH4+ and H30+ with crown ethers.
The results for which highly polarizable atoms or ionic interactions are treated
provide the relative agreement to the experimental values on the calculation of
relative conformation energies and cation complexation enthalpies.
The sizes of the metal ions198 and crown ether cavities199 are given in Table
1.1. For all three crown ethers the BDE trend is Na+ > K+ > Rb+ > Cs+. The BDE
difference is most pronounced for Na+ compared to K+, with decreasing differences
for successive metal ions. The BDE trend correlates with the inverse of the M+ radii,
or, correspondingly with the charge density of the metal ion.
Previous investigators78 79 83'85161 have proposed that the major factor affecting
the stability of the crown ether complexes is the rM7rL ratio (cf., Table 1.1, rM+ is the


3.1. Mass spectrum of F" 94
3.2. Mass spectrum of COCI' 96
3.3. Mass spectrum of (CH3OH)F, m/z = 51 99
3.4. Mass spectrum of (CH3OH)Cr, m/z = 67, 69 101
3.5. Mass spectrum of OH', m/z =17 104
3.6. Mass spectrum of (H20)Br\ m/z = 97, 99 106
3.7. Mass spectrum of (CH3OH)Br, m/z =111,113 108
3.8. Photodetachment spectrum of (C2H5OH)Br: experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2- the arrow points to the
onset threshold 113
3.9. Photodetachment spectrum of (CH3CN)Br: experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2 the arrow points to the
onset threshold 115
3.10. Photodetachment spectrum of (CH3OH)l': experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2 the arrow points to the
onset threshold 117
3.11. Photodetachment spectrum of (CH3OH)F': experimental
data; curve best fit to the threshold region of
photodetachment effect vs. (E-ET)1/2- the arrow points to the
onset threshold 119
3.12. Photodetachment spectra: (CH3CN)Br; (CH3OH)Br; -
(C2H5OH)Br; (i-C3H7OH)Br; X (n-C3H7OH)Br 122
3.13. Photodetachment spectra of (CH3OH)F'obtained using different
thermalization times (t): -1=2.34 sec.; X -1=3.20 sec.; -
t=3.77 sec.; X -1=4.31 sec.; -1=6.05 sec 124
3.14. Photodetachment spectra of (CH3OH)CI' obtained using
different thermalization times (t): -1=2.05 sec.; -
IX


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EHFH2DPQN_EWV4VW INGEST_TIME 2014-10-06T22:15:13Z PACKAGE AA00025777_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


92
as well as (CH3CN)Br. Riveros et a/.144'147 have developed methods for producing
these adducts (as shown for specific cases in the Procedure section). These
preparation procedures have been used successfully in our laboratory148 to obtain
IR spectra of solvated halides by IR multiphoton dissociation (IRMPD)
measurements. In this work the binding energies of (ROH)X' and (CH3CN)Br have
been determined by the photodetachment method in the 260-350 nm wavelength
range.
Experimental Procedures and Data Analysis
Chemicals and Ion-Molecule Reactions
All chemicals were commercial products of spectral-grade purity, and were
used without further purification, except for several freeze-pump-thaw cycles. NF3
was purchased from Air Products and Chemicals, Inc., phosgene was obtained from
Matheson, optima-grade methanol was purchased from Fisher Scientific, and the
rest of the samples were obtained from Aldrich.
The solvated halide ions were formed by the methods of Riveros.144147 A
two-step procedure was used to produce the methanol-solvated fluoride and
chloride ions.144146 In the first step the pertinent reagent gas (NF3 for the F' adduct;
phosgene for the Cl adduct) underwent dissociative electron capture of 6-12 volt
electrons (Figures 3.1 and 3.2 for F' and COCI ):


46
For endothermic reactions, microscopic reversibility can be applied to eq 1.17
to yield,97
o(E) = o0(E-Et)*/E, (1.18)
where ET is the reaction endothermicity. The scaling parameter is defined as o0 =
ne (2a)/j (psgs / p pg p), where p is the reduced mass and g is the electronic
degeneracy of the precursor (p) or successor (s). Although it would be nice if eq
1.18 accurately predicted the energy dependence of all endothermic ion-molecule
reactions, this is not observed experimentally.54'55 57a Indeed, reactions that conform
to eq 1.18 are exception rather than rule, as described later.
Alternate procedures for determining the energy dependence of reactions
include trajectory calculations, statistical theories, and empirical theories. Trajectory
calculations are not generally useful because they require a potential energy
surface, which is rarely available. Statistical theories such as phase space theory
(PST) in which energy and angular momentum are explicitly observed,98 are much
more useful.99'102 Empirical models are the most general means of describing the
energy dependence of reactions, and in this work, we use,
o(E) = o0(E-ET)7Em, (1.19)
where the scaling parameter o0 (with units of area-energy"1") need not be given by


Figure 2.6. 7 T ICR instrument (a) 7 T magnet, (b) Infinity cell, (c) 400 L/s cryopumps, (d) 800 L/s cryopump, (e)
Mechanical pumps, (f) Capillary coated with platinum at entrance and exit ends, (g) Needle, (h) Syringe pump.


77
the vacuum or exit side of the capillary, the ions are transferred through several
skimmers (0.8 and 1.0 mm diameters) and lenses with pumping stages in the ESI
source prior to entering the transfer region. The manipulation of all these voltages
is important as they shape the electric field which causes the electrospray to form
and guide the ions toward the capillary so they may be transported into the vacuum
region of the source.
At the back end of the capillary where the ions emerge, a supersonic
expansion takes place between the capillary exit and the skimmer. The capillary
exit itself acts as a lens (L1) and so has a variable potential. The region between
the capillary exit and first skimmer (L2) is differentially pumped to allow for the
passage of as many ions as possible while reducing the gas pressure as well. After
this skimmer, there is another differential pumping region, where another lens
element, L3, and a second skimmer, L4, are present. The final lens elements were
L5 and L6. This pressure region is "open" and therefore pumped by the pump in the
source region of the mass spectrometer vacuum system. For this reason, L5 and
L6 are visible from the outside of the source. L6 is the last controllable lens before
the mass analyzer optics. Once the ion has been formed, it is transferred from
atmospheric pressure, where the ionization process occurs, to the mass analyzer
which is under vacuum through differential pumping stages.
Introduction of ESI Features
The efficiency of ionization in ESI is greatly affected by the chemistry of the
analyte with its supporting solution. The reason for this is the characteristic relation


100
80
60
40
20
0
(CH3OH)Br
110.923
112.957
.5.0
18.6
24.5 35.9
Mass-to-charge ratio
67.0
500.5
o
00


159
Internal Energies
Rotational energy
Because the rotational energy distribution is relatively narrow, the average
rotational energy of the ions (3/2 kT = 0.039 eV for nonlinear ions at 298 K) was
added to the measured threshold in CRUNCH program.
Vibrational energies
The frequencies for all vibrational modes in both the vibrational ground and
excited precursor ions were calculated using the HyperChem program. The
program first optimizes precursor ion geometries using both ZNDO semi-empirical
and ab initio calculations and subsequently uses these geometries for the frequency
calculation. The ground state values are used by CRUNCH to calculate the
vibrational energy which is summed over the vibrational states i having energies E¡
and population g¡ of the precursor at room temperature. In this program, the Beyer-
Swinehart algorithm186 is used to evaluate the density of the ion vibrational states,
and then the relative populations, g¡, are calculated from an appropriate Maxwell-
Boltzmann distribution. The excited state values are needed to correct for the
probability that some ions may not dissociate until after the CAD delay time (RRKM
calculation).


Figure 2.2. 2 tesla ICR instrument (a) 2 T magnet, (b) Inlet source, (c) Three window flange, (d) One window flange,
(e) ICR cell, (f) Diffusion pump, (g) Ion gauge.


Figure 3.14. Photodetachment spectra of (CH3OH)Cr obtained using different thermalization times (t): -1=2.05 sec.
A -1=3.20 sec.; -1=4.34 sec.


Photodetachment 50
CAD 51
2. INSTRUMENTATION 53
Introduction 53
Nicolet 2 Tesla ICR Instrument 56
Bruker 7 Tesla ICR Instrument 66
Electrospray Ionization (ESI) Source Region 71
Introduction of ESI Features 77
pKa or pKb 78
Concentration 78
Solubility 79
Molecular conformation 79
Molecular weight 80
Solvent properties 80
Ion Transfer Region 81
Ion Analyzer region 81
3. DETERMINATION OF BINDING ENERGIES OF
SINGLY-SOLVATED HALIDE IONS USING
PHOTODETACHMENT 88
Background 88
Experimental Procedures and Data Analysis 92
Chemicals and Ion-Molecule Reactions 92
Pulse Sequence 109
Data Analysis 109
Results and Discussion 120
Conclusion 141
4.BOND DISSOCIATION ENERGIES OF ALKALI AND
HYDRONIUM CATIONS BOUND TO CROWN ETHERS
DETERMINED BY ON-RESONANCE COLLISIONALLY-
ACTIVATED DISSOCIATION (CAD) 144
Background 144
Experimental 150
Ion formation 150
Pulse sequence 154
Data Analysis 155
Ion Appearance Curve 155
Internal Energies 159
IV


REFERENCE LIST
1. Sommer, H.; Thomas, H. A.; Hippie, J. A. Phys. Rev. 1950, 78, 806.
2. Sommer, H.; Thomas, H. A.; Hippie, J. A. Phys. Rev. 1951, 82, 697.
3. Wobschall, D.; Graham, J. R.; Malone, D. P. Phys. Rev. 1963, 131, 1565.
4. Wobschall, D. Rev. Sci. Instrum. 1965, 36, 466.
5. Anders, L. R.; Beauchamp, J. L.; Dunbar, R. C.; Baldeschwieler, J. D. J.
Chem. Phys. 1966, 45, 1062.
6. Mclver, R. T., Jr. Rev. Sci. Instrum. 1970, 41, 555;
7. Mclver, R. T., Jr. Int. J. Mass Spectrom. Ion Phys. 1971, 7, 471.
8. Beauchamp, J. L.; Anders, L. R.; Baldeschwieler, J. D. J. Am. Chem. Soc.
1967, 89, 4569.
9. Comisarow, M. B.; Marshall A. G. Chem. Phys. Lett. 1974, 4, 282.
10. Comisarow, M. B.; Marshall A. G. Chem. Phys. Lett. 1974, 26, 489.
11. Comisarow, M. B.; Marshall A. G. Can. J. Chem. 1974, 52, 1997.
12. Allemann, M.; Kellerhals Hp.; Wanczek, K. P. Chem. Phys. Lett. 1980, 75,
328.
13. Cody, R. B.; Burn, R. C.; Freiser, B.S. Anal. Chem. 1982, 54, 96.
14. McCrery, D. A.; Ledford, E. B., Jr.; Gross, M. L. Anal Chem. 1982, 54, 1435.
15. Mclver, R. T., Jr.; Bower, W. D. Tandem Mass Spectrometry, McLafferty, F.
W. Ed.; Wiley: New York, 1983, 287.
16. Cody, R. B.; Kinsinger, J. A.; Ghaderi, S.; Amster, I. J.; McLafferty, F. W.;
Brown, C. E. Anal. Chim. Acta. 1985, 178, 43.
17. Mclver, R. T.; Jr, Hunter, R. L.; Syka, M. S.; Labunsky, M. Paper presented
at the 31st Annual Conference on Mass Spectrometry and Allied Topics, Boston,
MA., May 8-13, 1983.
223


Figure 2.8. ESI design.


Relative rate coefficient
166


100
80
60
40
20
0
17.005
OH'
16.4
28.1
43.9
100.0
Mass-to-charge ratio
104


BIOGRAPHICAL SKETCH
Yar-Jing Yang was born in Taiwan, 1964. She grew up in a big family with
two brothers and two sisters. In 1984, she attended National Chung-Hsiung
University at Taichung after passing the national college entrance exam and
received her B.S. degree in chemistry four years later. Upon graduation, she
passed the entrance exam for the graduate program at National Chung-Hsiung
University and enrolled in the graduate program in 1988. During graduate school,
she joined a program in the Institute of Atomic and Molecular Sciences, Academia
Sinica at Taipei. Her research area was photodissociation of CCI3N02 and CBrCI3
by translational spectroscopy using a time-of-flight (TOF) mass spectrometer. In
1988, she met Chen-Chan Hsueh, a wonderful and smart man, who was a research
assistant at the national laboratory. In 1990, she received her M.S. degree in
physical chemistry in May, married Chen-Chan Hsueh in June, and came to the
United States with her new husband in August.
One year later, she began her Ph.D. studies at the University of Florida. She
was involved with the activities of the Chinese Association Club as an associate
chairperson. In 1994, she went to the University of Sao Paulo in Brazil as a short
term exchange student to learn how to form gas-phase solvated anions by
sequential ion-molecule reactions. She received an award for an outstanding
235


66
plate. Because the volume of the cylindrical cell is much larger than that of
conventional cubic cells used in our laboratories and elsewhere, space charge
effects are reduced. The trapping potential was kept as low as possible (1.2-1.4 V)
to prevent high velocity ions from being trapped.
Radiation in the 260-350 nm range after frequency doubling was provided by
a tunable dye laser (Quantel TDL 50) with a frequency-doubling accessory. This
dye laser was pumped by 532 nm (second harmonic of YAG fundamental) output
from a Nd3+:YAG laser (Quantel YG 581C), which was synchronized with the mass
spectrometer using laboratory-built interface electronics.109 The wavelength scan
was controlled by commercial software provided with the dye laser. The wavelength
accuracy of this laser system is 0.5 A, and the resetting stability is 0.02 A; the
laser linewidth is 0.08 A. Five different dyes (Fluorescein, Rhodamine 590, 610,
and 640, and DCM) were used and the provided wavelength range for each dye is
listed in Table 2.1. The laser output power (ca. 8-15 mJ/pulse) was kept constant
for each photodetachment spectrum.
Bruker 7 Tesla ICR Instrument
All the CAD experiments were performed on a Bruker FTICR mass
spectrometer (Bruker Analytical Systems, Billerica MA) equipped with a Spectrospin
(Fllanden, Switzerland), 7.0 T superconducting magnet, an RF-shimmed infinity


231
W.; McMahon, T. B. Can. J. Chem. 1984, 62, 675.
138. (a) Hiraoka, K.; Mizuse, S. Chem. Phys. 1987, 118, 457; (b) Hiraoka, K.;
Mizuse, S.; Yamabe, S. J. Phys. Chem. 1988, 92, 3934.
139. Schuster, P. Electron-Solvent and Anion-Solvent Interactions, Kevan, L.
and Webster, B. C. Ed.; Elsevier Science: New York, 1976; Ch. 8.
140. Szulejko, J. E.; Wilkinson, F. E.; McMahon, T. B. Presented at 37th ASMS
Conference on Mass Spectrometry and Allied Topics. May 21-26, 1989, Miami,
FL, 333.
141. Tanabe, F. K. J.; Morgon, N. H.; Riveros, J. M. J. Phys. Chem. (in press).
142. Janousek, B. K.; Zimmerman, A. H.; Reed, K. J.; Brauman, J. I. J. Am.
Chem. Soc. 1978, 100, 6142.
143. Berger, S.; Brauman, J. I. J. Am. Chem. Soc. 1992, 114, 4737.
144. (a) Blair, L. K.; Isolani, P. C.; Riveros, J. M. J. Am. Chem. Soc. 1973, 95,
1057; (b) Isolani, P. C.; Riveros, J. M. Chem. Phys. Lett. 1975, 33, 362.
145. Riveros, J. M.; Breda, A. C.; Blair, L. K. J. Am. Chem. Soc. 1973, 95, 4066.
146. Faigle, J. F. G.; Isolani, P. C.; Riveros, J. M. J. Am. Chem. Soc. 1976, 98,
2049.
147. Riveros, J. M.; Ingemann, S.; Nibbering, N. M. M. J. Am. Chem. Soc. 1991,
113, 1053.
148. Peiris, D; Riveros, J. M.; Eyler, J. R. submitted.
149. Wetzel, D. M.; Brauman, J. I. Chem. Review. 1987, 87, 607.
150. Rosenfeld, R. N.; Jasinski, J. M.; Brauman, J. I. J. Am. Chem. Soc. 1979,
101, 3999.
151. Brauman, J. I., Gas Phase Ion Chemistry: Ions and Light, Bowers, M. T.
Ed.; Academic: New York, 1984; Vol. 3, Ch. 21.
152. Bevington, P. R., Data Reduction and Error Analysis for the Physical
Sciences, McGraw-Flill: New York, 1969, Ch.6.


219
Based on previously-published theoretical potential surfaces for the neutral
halogen-solvent system, the reported BE's may be high. For (CH3OH)F' and
(CH3OH)C|- the error should be only about 0.5 kcal/mol of the upper limit, but it may
be 1-2 kcal/mol for the Br and I' complexes due to the higher endothermicity for
neutral hydrogen abstraction compared to F' and Br complexes. Examination of
previously reported data indicates a random uncertainty of about 1-2 kcal/mol in
the reported BE's. Furthermore, binding energies determined by equilibrium
methods are subject to systematic error in the statistical mechanical estimation of
AS, which is based on the vibrational frequency of the H-X bond in ROFI-X'. As the
halide becomes heavier, the vibrational frequency of the ROFI-X' bond decreases.
It is possible that the vibrational frequency has not been estimated correctly by
previous authors. This would produce a systematic error in the entropy change,
which would subsequently affect the reported BE. For the alcohol-solvated halides,
the overall observation is an increase in binding energy with decreasing RO proton
affinity (CFI30 > C2H50' > i-C3H70 ~ n-C3FI70') and increasing X' proton affinity (
I > Br> C|-> F ).
This is the first study which has obtained BDE's using photodetachment in
ICR mass spectrometry. The ability to form solvated halides, to trap ions for long
periods of time, and to successfully measure BE's using the photodetachment
technique in FTICR-MS has shown that photodetachment of these complexes is a
reliable approach to determining binding energies. With these advantages and
success, photodetachment techniques can be extensively applied for quantitative


Figure 3.9. Photodetachment spectrum of (CH3CN)Br: experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.


Figure 3.7. Mass spectrum of (CH3OH)Br, m/z = 111, 113.


Relative rate coefficient
Energy in center-of-mass frame/eV


226
53. Pozniak, B.; Dunbar, R. C. Int. J. Mass Spec.; Ion Proc., 1994, 133, 97.
54. Schultz, R. H.; Crellin, K. C.; Armentrout, P. B. J. Am. Chem. Soc. 1991, 113,
8590.
55. Dalleska, N. F.; Honma, K..; Armentrout, P. B. J. Am. Chem. Soc. 1993, 115,
12125.
56. Magnera, T. F.; David, D. E.; Stulik, D.; Orth, R. G.; Jonkman, H. T.; Michl, J.
J. Am. Chem. Soc. 1989, 111, 5036.
57. (a) Lane, K. R.; Sallans, L.; Squires, R. R. J. Am. Chem. Soc., 1986, 108,
4368. (b) Sunderlin, L. S.; Wang, D.; Squires, R. R. J. Am. Chem. Soc., 1993,
115, 12060.
58. (a) Mclver, R. T, Jr. Kinetics of Ion-Molecule Reactions, Ausloos, P. Ed.;
Plenum: New York, 1978. (b) Mclver, R. T, Jr. Sci. Am. 1980, 243, 186.
59. Kebarle, P. Ion-Molecule Reactions, Franklin, J. F. Ed.; Butterworth: London,
1972, Vol. 1, 315.
60. Pauling, L. The Nature of the Chemical Bond, 3rd ed.\ Cornell University
Press: Ithaca, NY, 1960.
61. Murray-Rust, P.; Stallings, W. C.; Monti, C. T.; Preston, R. K.; Glusker, J. P.
J. Am. Chem. Soc., 1983, 105, 3206.
62. Scheiner, S. Reviews in Computational Chemistry, Lipkowitz, K. B.; Boyd, D.
B. Eds.; VCFI publishers: New York, 1991.
63. Latimer, W. M.; Rodebush, W. H. J. Am. Chem. Soc., 1920, 42, 1419.
64. Perutz, M. F.; Rich, A.; Crick, F. The Chemical Bond: Structure and
Dynamics', Zewail, A. Ed.; Academic Press: San Diego, 1992, ch. 1.
65. Luettringhaus, A.; Zeigler, K. Liebigs Ann. Chem. 1937, 528, 155.
66. Luettringhaus, A. Liebigs Ann. Chem. 1937, 528, 181.
67. Luettringhaus, A. Liebigs Ann. Chem. 1937, 528, 211.
68. Cram, D. J.; Trueblood, K. N. Top. Curr. Chem. 1981, 98, 43.


Figure 4.12. Mass spectrum of (18-C-6)H30+ after CAD at AV=20 volts. Peak corresponding to (18-C-6)H+ ion was the
only successor ion observed.


133
Photodetachment Thresholds (ET) and Binding Energies (BE) of Solvated Halides
This work
Other experimental results
Calculated results
ET/nm BE/kcal-mol'1
() ()=/kJ-mol'1
BE/kcal-mol-1
0=/kJ-mol'1
Method3
[Ref.]
BE/kcal-mol'1 Method [Ref.]
0=/kJ-mol'1 (basis set)
(CH3OH) F-
265.11.2
(1.71)
29.910.5
(125.112.1)
29.6 (123.8)
30.6 (128.0)
23.3 (97.5)
EX [137]
EQ [140]
EQ [130]
26.2 (109.6) 3-21 G
[130]
(CH3OH) Cl-
280.51.4
(2.18)
18.810.5
(78.712.1)
16.8 (70.2)
16.3 (68.2)
14.2 (59.4)
17.4 (72.8)
EX [137]
EQ [134]
EQ [131-133]
EQ [130,138]
16.2(67.8) 3-21 G
[130]
(CH3OH)Br
309.01.3
(1.82)
15.210.4
(63.611.7)
13.9 (58.1)
EQ [130]
16.0(66.9) 3-21 G [130]
12.9(54.0) MP4/SDTQ[141]
(CH3OH)|-
337.21.6
(1.31)
14.410.4
(60.311.7)
11.3 (47.3)
11.2 (46.8)
EQ [131-133]
EQ [130]
10.8 (45.2) 3-21 G
[130]
(C2H5OH)Br
308.811.9
(2.31)
15.410.6
(64.412.5)
14.4 (60.2)
EX [141]
(i-C3H7OH)Br
304.413.7
(2.32)
16.611.2
(69.415.0)
14.6 (61.1)
EX [141]
(n-C3H7OH)Br
303.811.3
(1.27)
16.810.4
(70.311.7)
(CH3CN)Br
315.313.2
(1.27)
13.410.9
(56.113.8)
12.7 (53.1)
12.9 (54.0)
EX [141]
EQ [138(b)]
11.0(45.9) MP4
[141]
(CH3CN)C|-
15.8 (66.1)
14.0 (58.6)
13.4 (56.0)
EX [137]
EQ [134]
EQ [131-133]
(CH3CN)f" 16.0(66.9) 3-21 G [129]
a. EQ = Gas-phase equilibrium (cf. eq 1.6 in the text)
EX = Bimolecular exchange (cf. eq 3.1 in the text)
mean scaling factor (cf. eq 3.11 in the text)


Figure 4.17. Mass spectrum of (12-C-4)H+ at AV=70 volts. Peak corresponding to (12-C-4)H+ ion was isolated before
excitation.


Figure 1.10. Photodetachment process: (a) A' + hv A + e', (b) A' + hv B' + C, (c) A" + hv B + C + e\


Size of Alkali-Metal Ions and Cavity Size of Crown Ethers
rL(A) [ref. 199]
Na
K
Rb
Cs
rm*(A) [ref 198]
1.02
1.38
1.52
1.67
C/rL
12-crown-4
1.2-1.5
0.76
1.02
1.13
1.24
15-crown-5
1.7-2.2
0.52
0.71
0.78
0.86
18-crown-6
2.6-3.2
0.35
0.48
0.52
0.58
rM+- metal ion radius; rL cavity size of crown ethers
CO
CD


In (N/N)
302.0 304.0 306.0 308.0 310.0 312.0
Wavelength/nm
128


100
80
60
40
20
0
Upon CAD
177.2
(12-crown-4)H+
89.1
(C2H40)2H+
- I
.0 124.0 218.0 312.0 406.0 500.0
Mass-to-charge ratio
ro
o


100
80
60
40
20
0
Upon CAD High activation energy
(15-crown-5)H+
(C2H40)2H+ (C2H40)3H+
22
239.2
(15-crown-5)H30+
i. i
JL~v
[7.1
113.7
210.3 306.9
Mass-to-charge ratio
403.4
500.0
194


38
that have been developed for the study of electron photodetachment are ion beams,
trapped ion cells, and drift tubes.38'40 The advent of tunable laser systems has
increased the precision of measurements made in photodetachment experiments,
allowing for the detection of vibrational and even rotational transitions in some
systems.52 In addition, the wide wavelength range and narrow bandwidth of the
tunable output provide extensive new study areas.
Upon irradiation with light, a molecular anion may either detach an electron
(1.7), dissociate (1.8), or undergo photodetachment followed by dissociation (1.9):
A' + hv-A + e' (1.7)
A' + hv-B' + C (1.8)
A' + hv-B + C + e' (1.9)
Processes (1.7), (1.8), and (1.9) are illustrated in Figure 1.10 (a), (b), and (c),
respectively. In photodetachment spectroscopy, these processes are monitored as
function of the wavelength of the incident light. This can be accomplished by
detecting the decrease in the intensity of precursor ion, A' (processes 1.7, 1.8, and
1.9) or an increase of successor B' for process 1.8 or the detached electrons for all
three processes. The related technique of negative ion photoelectron spectroscopy
analyzes the energy of detached electrons produced by the interaction of a fixed


143
previous measurements which utilized entirely different experimental approaches.
Thus, we have shown that the photodetachment technique is a reliable approach
to obtain BDEs by ICR mass spectrometry. Since electrospray ionization (ESI) can
produce multiple-charged and solvent cluster-attached ions, the BDEs and/or step
wise solvation energetics should also be obtainable with photodetachment
techniques in the ICR instrument.


Figure 4.20. Appearance curve of (12-C-4)K+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.


DETERMINATION OF BOND DISSOCIATION ENERGIES USING FOURIER
TRANSFORM ION CYCLOTRON RESONANCE
(FT-ICR) MASS SPECTROMETRY
By
YARJING YANG
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1996


212
size of metal ions, rL is cavity size of the crown ethers). Indeed, for Na+, K+, Rb+,
and Cs+the BDE order is 12-C-4 > 15-C-5 > 18-C-6 when RRKM corrections are
included, which follows the rM7rL order. However, for complexes with similar rM+/rL
((12-C-4)Na+ and (15-C-5)K+; (15-C-5)Na+ and (18-C-6)K+ have approximately the
same size ratio, 0.76 and 0.71; 0.52 and 0.48, respectively), the BDE's are much
greater for Na+ compared to K+. Thus, the charge density on the metal ion appears
to be an important effect, as shown above.
Another criterion has been the closeness of the rM7rL ratio to unity,78-79'83-85'161
i.e. a matching of ionic radius and the crown cavity radius. However, the data only
partially support that argument. For rM7rL significantly greater than one (12-C-4
with Rb+ and Cs+) the BDE's are quite low, especially for Cs+. For these complexes
the metal ion must sit above the cavity, thus increasing the M+-L internuclear
distance. Otherwise, having a rM7rL ratio close to one does not seem to favor a
large BDE. The ratio for (12-C-4)K+ is essentially equal to one, but the BDE is
considerably less than that for the Na+ complex. Likewise, the BDE's for 15-C-5 and
18-C-6 complexes decreases as the rM7rL values increase towards one.
Bowers et a/.161 measured collision cross sections for complexes of this
ligand with Li+, Na+, K+, and Cs+ by gas-phase ion chromatography (IC), and used
the experimental results to calculate gas-phase conformations by the Amber MM
procedure. The theoretical results indicated that all alkali metal ions are centered
on the axis of the crown cavity. However, only for K+ is the complex quasi-planar.
The small Li+ ion induces substantial distortion in the crown ether in order to better


87
Quench 5 msec
Injection of Ions Formed by ESI -100 msec
Thermalizafjon 2 sec
nn
Ejections
Activation 0-50 psec
CAD delay -10-30 msec
Excite
Detect
Time


Figure 3.8. Photodetachment spectrum of (C2H5OH)Br: experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.


142
observation is an increase in BE with decreasing RO' proton affinity and increasing
X' proton affinity. As the hydrogen becomes more loosely bound to the alcohol
oxygen, it is more available for bonding to X.
The results of this and previous work on solvated bromide anions, as well as
other studies of chloride complexes, show that complexes of these halides with
acetonitrile have BE's about 1-2 kcal/mol lower than the methanol solvates.
However, calculated results show that the BE for (CH3CN)F' is more than 7 kcal/mol
less than that for (CH3OH)F\ The difference in behavior of F' compared to Cl', Br',
and I' can be understood in terms of the calculated geometries discussed above.
The (CH3OH)CI', (CH3OH)Br, and (CH3OH)l' complexes have nonlinear CH30-H-X
hydrogen bonds, and the results indicate that there is additional interaction of X'
with the methyl hydrogens. However, because of its small size, the F" can not
interact with the methyl group in CH30-H-F', resulting in a linear hydrogen bond. As
stated above, the halide ion bonds via the methyl hydrogens in all the (CH3CN)X'
complexes. Thus the bonding is somewhat similar for the CH3OH and CH 3CN
solvates of Cl' and Br', and this is reflected in the closeness of the BE's of the
(CH3OH)X' and (CH3CN)X' complexes for these halides.
Although the BDEs of some halide complexes have been measured using
other techniques in different types of mass spectrometers, the photodetachment
technique has not previously been applied to measure BDEs in ICR mass
spectrometry. This work is the first to apply this technique for the BDE
measurements. The data obtained from this work in general confirm the results of


213
coordinate with the partially charged oxygen atoms. When complexed to Na+ (or
Li+) the ligand adopts a cup-like structure to bring oxygens closer to the metal ion,
and the cation is cradled by the ligand. The Cs+ ion is too large for the crown cavity
and is forced out of plane. In (18-C-6)K+ the metal ion is coplanar with the ligand
and is exposed to the environment above and below the ring plane. The results
showed that the Li+ complex is most compact, which reflects relatively high charge
density on Li+. In solution, Li+ is essentially fully solvated by the crown ethers; it
benefits only in a minor way from further solvation, even though it has the largest
solvation energy of any alkali ions. The same holds to a lesser extent for Na+.
Hence, these two systems have the weakest BDE in solution. In contrast, K+ and
Cs+ are exposed and can benefit from solvation. Theory193'196 suggests that water
as a solvent displaces the K+ ion an average of 0.25 from crown center, resulting
in coordination with two water molecules in the first solvation shell. However, the
K+ ion is still primarily solvated by the crown in competition with water as a solvent
because K+ is bound much more strongly to small ethers than to H20 due to the
contribution of larger dispersion forces for the ethers.202'203 Similar reasoning
applies for the remaining alkali ions. Alkali ions strongly prefer association with
small ethers over association with water in gas phase. Theory indicates that
solvating a K+ complex in H20 has very little effect on the KVcrown structure.
The BDE's for (15-C-5)H30+ and (18-C-6)H30+ correspond to loss of neutral
water, i.e. the solvation energies for these complexes. Brodbelts group89 has
reported CID experiments on ammonium/crown ethers. In their work, the (18-


47
the expression above. This model includes the model described by eq 1.18 as well
as most other theoretical predictions for endothermic cross sections.103
Unfortunately, even for atom-diatom reactions, there are many theoretical predicted
values of n and m, with no single form that describes the energy dependence of all
reactions.
Although eq 1.19 can adequately represent the kinetic energy dependence
of the reaction cross sections, it does not explicitly include the internal energies of
the ion or neutral reactants. This can be accomplished by modifying eq 1.19 to,
(E) = Eo0g¡ (E+Er+ErET)n/Em, (1.20)
where the subscripts r and i refer to a rotational and vibrational states; Er or E¡ is the
energy of that state; g, is the statistical population of the ith state; and the sum is
over all states. ET is thus the reaction endothermicity corresponding to 0 K. Eq
1.20 is the model which is used to determine the CAD threshold in Chapter 4.
On-resonance CAD
On-resonance CAD techniques have been demonstrated to measure BDE's
in FTICR mass spectrometry in several groups.89104 105 A short radio frequency (r.f.)
pulse is applied to the excite plates on an ICR cell to increase the translational
energy of a mass-selected precursor ion. The center-of-mass (c.m.) translational
energy imparted to an ion during the excitation in FTICR is given in eq. 1.21


21
drift tubes, and ion traps.38-40 Each device offers its own advantages for different
types of ions. For the study of ion-neutral reactions, ion cyclotron resonance (ICR)
mass spectrometry has been the most extensively applied method. The
advantages offered by ICR mass spectrometry include the ability to generate a large
variety of molecular ions and high sensitivity. The technique of ICR-MS allows one
to generate gas phase ions and trap them for periods of millisecond to several
seconds, sometimes even for hours.
The goal of this work was to apply different FTICR-MS techniques including
photodetachment and collisionally activated dissociation (CAD), to obtain bond
dissociation energies for solvated anions and host-guest complexes. A secondary
goal was obtaining quantitative measurements of bond dissociation energies
(BDE's) by improving the threshold determination to account for a variety of
systematic effects that would otherwise distort the thermochemistry obtained.
Several methods have been used to measure the bond dissociation energies for
various ion-neutral complexes, including equilibrium measurements in both high
pressure41^13 and ICR44-45 mass spectrometers, photodissociation measurements in
a time-of-flight (TOF) mass spectrometer,46-50 photodetachment measurements in
photoelectron spectroscopy,51-53 collisionally activated dissociation (CAD) in both
quadrupole or octopole-guided ion beam54-57 and ICR58 mass spectrometers. The
measurement of BDE's by FTICR mass spectrometry has not been widely applied
and there is still some uncertainty as to the quality of the measurements. In this
section, some recent experimental studies of BDE measurements using different
types of techniques are briefly discussed.


81
evaporation will depend on the solvent composition of the droplet. For this reason,
solutions with higher organic content may give higher signal since the droplets
formed will evaporate faster, leaving more time for analyte ionization to occur.
Ion Transfer Region
After formation ions are then accelerated toward the analyzer by a series of
electrostatic lenses (3 kV maximum potential between PL5 and PL 7) as shown in
Figure 2.9. After acceleration the ions pass x,y-deflection, flight tubes (FL1, FL2),
and PL9 at decreasing potential to relative ground prior to entering the cell region.
A set of deflection lenses is used to gate the continuous ion beam, so that ions
enter the analyzer region only during the ion accumulation period. Finally, the ions
are transported into the mass analyzer where their mass is determined. The gate
valve is used to isolate the analyzer region from the ionization source region for
more rapid changes of the ionization source to El, MALDI, or GD.
Ion Analyzer Region
The trapping cell used for CAD work is a rf-shimmed infinity cell as shown
in Figure 2.10. At the front of the ICR cell there were two "side-kick" plates with
pulsed low voltages (1-5 V) during the ion formation pulse. Ions entered the cell
with a kinetic energy identical to the first "side-kick" plate potential. After the ions
entered the cell, the voltages on the side-kick plates were dropped to the trapping
potential (0.5 -1 V). To help trap the ions, the second set of side-kick plates (with
two halves) is designed to convert the z-axis translational energy into the xy-plane
cyclotron motion. Once the ions were trapped inside the cell, the CAD experiments
were performed using a pulse sequence illustrated in Figure 2.11.


163
calculation; the uncertainty is about 0.1 eV ( 2.3 kcal/mol) when RRKM
calculations were included.
Results
Collision Conditions
In order to control the amount of energy imparted by the rf excitation, the
precursor ions can suffer no collisions during the excitation process. In this work,
rf pulse lengths of 0-50 psec were used. At a total pressure of ca. 3-5 *10'8 mbar
(2-4 xio 8 torr), the average time between ion/molecule collisions (using the
Langevin rate constant191 of 1.24x1 O'9 cm3/s at 300 K) is 2.13 s, which is much
longer than the excitation time. Thus, the average number of collisions in 50 psec
is 2.7x10'6. It is apparent that during the 50 psec excitation time most of ions
undergo no collisions.
It is also essential that each precursor ion collide with only one Ar molecule
within the CAD delay time, to avoid broadening during the conversion of
translational to internal energy. The probability, qn, that an ion will encounter n
molecules during the CAD delay time is approximated by the Poisson distribution192
(4.9)


210
for the ab initio result. In spite of this good agreement between the QM/MM and ab
initio results, he reported that the QM/MM value is likely to be high, since the lower
AM1 OCCO torsion barrier allows for shorter K+-oxygen distances, and hence
stronger electrostatic stabilization, without a sufficiently large torsion barrier penalty.
Thompson also pointed out that both Wipffs and Kollman's results should be
decreased by an additional 2-4 kcal/mol, because of the different geometries used
in the calculations. These arguments predict a BDE range for (18-C-6)K+ of 53-60
kcal/mol, which overlaps the experimental value of 52.2 2.3 kcal/mol obtained in
this work.
Discussion
Previous investigations78'89 of gas-phase crown ether complexes have
indicated that the stabilities depend on the relative sizes of the metal ions and the
crown-ether cavities. However, other factors must also be considered.
The full long-range potential energy interaction, V(r, 0), for the CAD process
is given by:96
V (r,0) (3cos20-1)x[-
Or*
2 r
92(graJ Qq,
6r4 2r3
(4.13)
where r is the distance between the metal ion nucleus and the center of the crown
ether, 0 is the angle between r and the quadrupolar axis of the complex, q is the ion


45
I' = 2 if it is labeled by xy, xz, yz, x2-y2, or z2. The allowed I* values for
photodetachment are then these I' 1. Since the A" representation transforms as
Z in the Cs point group for alcohol solvated halides, the HOMO has I = 1 and the
selection rule Al = 1 gives I* = 0. Therefore, from eq 1.15 we obtain,
(E) Ek1/2 or In (N0/N) = a(E-ET)y>, (1.16)
The threshold, ET, can be obtained using an iterative procedure in which the
parameters, a (a scaling factor) and ET are optimized, to obtain the best fit to the
photodetachment spectra.
CAD
Energy behavior of reaction cross section
The maximum reaction cross section for an exothermic ion-molecule reaction
has an energy dependence that is determined by the long-range interaction
potential. For polarizable molecules without dipole moments, this is given by the
Langevin-Gioumousis-Stevenson (LGS) expression,95
Olgs(e) = TTe (2a/E),/j, (1.17)
where e is the electron charge, a is the polarizability of the neutral reactant, and E
is the relative kinetic energy of the reactants. Higher order terms in the long range
interaction potential can also be included.96


Figure 3.15. Photodetachment spectra of (CH3OH)Br obtained using different thermalization times (t): -1=2.05 sec.
A -1=3.77 sec.; -1=4.34 sec.


Figure 4.13. Mass spectrum of (15-C-5)H30+ at AV=20 volts. Peak corresponding to (15-C-5)H30+ ion was isolated
before excitation.


Figure 3.2. Mass spectrum of COCI'.


110
to the unirradiated sample, must be related to the photodetachment cross section,
o(X), which is the ordinate of an absolute photodetachment spectrum.
The probability, P(A), of removing an electron from a negative ion by laser
irradiation is given by equation 1.10.90 The equations, 1.10 and 1.11, yield an
expression for the cross section in terms of the measured quantities:91
In
V
o(A)
N
/
(1.13)
tff-p(A)dA
For a given photodetachment spectrum, all parameters in the denominator of eq
1.13 are constant, if the laser flux is kept constant. Thus, the relative
photodetachment cross-section spectrum, which is sufficient for determination of the
photodetachment threshold, may be obtained by plotting ln(No/N) versus the laser
wavelength. At each wavelength setting the N and N0 measurements were
repeated 4 to 6 times, and the respective mean and standard deviation were
calculated. The standard deviations of the ln(N0/N) values were propagated from
the standard deviations of the data points, and these were used as error bars for the
photodetachment spectra.
At the threshold the cross section for optically detaching an electron depends
only on the long-range interaction between the two product particles. For
polyatomic anions151 with Cs symmetry, the cross-section is directly proportional to
the square root of the energy above the threshold, (E-ET),/2:


In (N0/N)
260.0 264.0 268.0 272.0 276.0
Wavelength/nm
124


Figure 4.19. Appearance curve of (15-C-5)Rb+. The solid circles are the experimental data (mean of five measurements
with error bars corresponding to 1 standard deviation); the curve is the best fit to the data from the CRUNCH
program.


Figure 3.11. Photodetachment spectrum of (CH3OH)F': experimental data; curve best fit to the threshold region
of photodetachment effect vs. (E-ET)1/2- the arrow points to the onset threshold.


Ions Observed in Mass Spectra
170
ions formed by ESI prior to isolation
isolated precursor
ion
successor ions
(12-C-4)Na* (s),
m/z=199
(12-C-4)2Na+ (w)
m/z=375
(12-crown-4)Na*
Na*
(12-C-4)K* (s),
m/z=215
(12-C-4)2K+ (w)
m/z=391
(12-crown-4)K+
K*
(12-C-4)Rb+ (s),
m/z=261, 263,
(12-C-4)2Rb+ (s),
m/z=437, 439, 441
Rb* (w)
m/z=85, 87
(12-crown-4)Rb*
Rb*
(12-C-4)Cs+(s),
m/z=309
(12-C-4)2Cs+(s),
m/z=485
Cs* (w)
m/z=133
(12-crown-4)Cs+
Cs+
(12-C-4)H+ (s),
m/z=177
(12-C-4)H* (s)
(12-C-4)2H+(s)\
m/z=353
(12-C-4)2H+ (w)
(12-C-4)2H30+(w)
m/z=371
(12-crown-4)H+
(C2H40)nH+
(n=3,2,1)
m/z=133,89,45
(15-C-5)Na+(s),
m/z=243
(15-C-5)2Na+ (w)'
m/z=463
(15-crown-5)Na+
Na+
(15-C-5)K+(s),
m/z=259
(15-C-5)2K+ (w)
m/z=479
(15-crown-5)K*
K*
(15-C-5)Rb+(s),
m/z=305, 307
(15-C-5)2Rb+ (w)'
m/z=525, 527, 529
(15-crown-5)Rb+
Rb+
(15-C-5)Cs+(s),
m/z=353
(15-C-5)2Cs+ (w)"
m/z=573
(15-crown-5)Cs+
Cs+
(15-C-5)H+ (s),
m/z=221
(15-C-5)H30+ (s)'
m/z=239
(15-crown-5)H30+
(15-C-5)H+
(C2H40)nH+
(n=3,2)
(18-C-6)Na+
m/z=287
(18-crown-6)Na+
Na+
(18-C-6)K*
m/z=303
(18-crown-6)K+
K*
(18-C-6)Rb*
m/z=349, 351
(18-crown-6)Rb+
Rb*
(18-C-6)Cs+
m/z=397
(18-crown-6)Cs*
Cs+
(18-C-6)IT (s)",
m/z=265
(18-C-6)H30+ (s)*
m/z=283
(18-crown-6)H30*
(18-C-6)hT
The capillary potential was 25 V for ions labelled with an asterisk, 75 V for all other ions.


56
in half an hour or less. This chapter will describe two ICR instruments with internal
and external ionization sources. The cell and ESI design used for this work will also
be described. The experimental configurations will be briefly introduced and then
discussed in detail in the Experimental sections of Chapter 3 and Chapter 4 for both
photodetachment and CAD techniques, respectively.
Nicolet 2 Tesla ICR Instrument
The FT-ICR mass spectrometer used for photodetachment studies (Chapter
3) was equipped with a 2 tesla superconducting magnet, a vacuum system (Figure
2.2) and a Bruker Aspect 3000 console with 512K memory. The vacuum chamber
had 8" flanges on either end, one with three optical windows for transmission of
laser radiation, the other with one window which allowed laser access to the
filament side of the cell. This arrangement permitted easy alignment of the laser
beam.
Figure 2.3 shows the laser set-up for this work. The laser beam was
reflected into the center window of the three window flange by a quartz prism. The
laser was externally triggered from the console. The console sent a TTL trigger
pulse just before the electron beam pulse to an Apple II computer with a John Bell
Co. interface card. Unless triggered, the computer and interface card sent both
"charge" and "fire" commands to the Nd3+:YAG laser so that the flashlamps fired at
a rate of ca. 9 Hz. The computer shifted to an interrupt routine which sent not only


160
Trial Function
The trial function accounts for the distribution of Ecm values of the precursor
ion, caused by the random thermal motion of the collision gas molecules. According
to Chantry107, this broadening has an isotropic Maxwellian velocity distribution, given
by:
(4.5)
In eq 4.5 e = E/ykbT and e0= Eo/ykbT, where E and E0 are the relative and nominal
center-of-mass energies of the precursor ions, respectively, y = M/(M+m) where M
is collision target molecular mass, m is the precursor ionic mass as shown in eq
1.21, kb is the Boltzmann constant, and T is the collision gas temperature. As a
result of the energy distribution, the observed appearance curve, oobs(E0), differs
from the true curve, o(E). Chantry has derived the form of the convolution of o(E)
as
oobs (E0) jA1/2 f(E, E0)a(E)dE
(4.6)
0
This equation is used by CRUNCH to calculate the corrected curve.
Considering eq 4.6, at high precursor ion energies, E0 kbT, this distribution
is relatively narrow, with a maximum at E = E0 and a width given approximately by
FWHM=(11.1ykbTE0)1/2
(4.7)


Figure 3.13. Photodetachment spectra of (CH3OH)F" obtained using different thermalization times (t): -1=2.34 sec.
X -1=3.20 sec.; -1=3.77 sec.; X -1=4.31 sec.; -1=6.05 sec.


58


221
BDE's for (15-C-5)H30+ and (18-C-6)H30+ correspond to loss of neutral water, i.e.
the solvation energies for these complexes. Loss of H20 and fragmentation from
the hydronium/crown ethers indicates that the H+ is bound strongly to the crown
ether cavity, with a much weaker solvation of the entire complex. The observations
of higher BDE of (18-C-6)H30+ than (15-C-5)H30+ and the fragmentation of crown
ethers with proton attached to it for (15-C-5)H30+, but not for (18-C-6)H30+, indicate
that the proton bonds to 15-C-5 stronger than 18-C-6.
As mentioned in the Conclusions of Chapters 3 and 4, this is the first study
which has applied these two techniques to obtaining reliable BDE's using ICR mass
spectrometry. The ability to obtain quantitative measurements of bond dissociation
energies of metal and hydronium ions with crown ethers by improving threshold
determinations to account for a variety of systematic effects that would otherwise
distort the thermochemistry was demonstrated. Improvements in the accuracy of
measuring the thresholds and providing more precise quantitative measurements
in the gas phase by accounting for the effects of multiple ion-molecule collisions,
internal energy of precursor ions, kinetic energy distribution of precursor ions and
dissociation lifetime have also been achieved for the first time. The CRUNCH
program allows one to choose several options for different application purposes.
For example, one can select an appropriate fitting model, such as a simple fitting
model described in equations 1.18 and 1.19, without reactant's internal energies
and RRKM calculation, complicated model with reactant's internal energies summed
over all the vibrational states or/and RRKM calculations (eq 1.20). Also one can


Bond Dissociation Energies for Alkali-Metal and Hydronium Ion Crown Ether Complexes
BDE
Na
K
Rb
Cs
h3o+
12-crown-4
82.1(3.52)
60.6(2.60)
43.2(1.85)
39.9(1.71)
15-crown-5
79.5(3.41)
55.9(2.41)
40.1(1.72)
37.5(1.61)
24.7(1.06)'
This work
18-crown-6
75.1(3.22)
52.2(2.25)
38.5(1.65)
36.1(1.55)
36.4(1.56)'
Previous work
[ref. 88]
18-crown-6
40
32
Calculation
[ref. 193]
18-crown-6
71.1
62.3,52.5
[ref. 194]
18-crown-6
57.0
[ref. 173]
18-crown-6
81. T
68.1a
52.8a
30.4a
[ref. 173]
18-crown-6
83.4b
72.0b
59.2C
49.6
[ref. 195]
[ref. 197]
12-crown-4
70.0
55.2d
44.6d
66.49
[ref. 197]
15-crown-5
77.09
fref. 1971
18-crown-6
79.89
All the values are in kcal/mol unit with uncertainty 2.3 kcal/mol ( 0.1 eV) (except eV unit in parenthesis),
a RHF/6-31+G* at D3d symmetry, b MP2/6-31+G* at D3d symmetry, c -MP2/6-31+G' at C3v symmetry,
d RHF/6-31+G' at C3v symmetry,
e Hybrid quantum QM/MM molecular dynamics method,
f BDE for the process LH30+ LH+ + H20, RRKM calculation not included,
g aH for the process LH30+ L + H30+.
209


134
transition shown in eq 3.12 would be possible. The known PA's155 presented in
Table 3.1 follow the order: CH30 > C2H50' > i-C3H70' ~ n-C3H70' > F' > Cl' > Br >
I'. Since the halides all have lower PA's than the RO it is very likely that
photodetachment occurs to a repulsive state (eq 3.2).
Neumark et a/.156 studied the hydrogen abstraction reactions of halogens with
methanol and ethanol (F + HOR HF + OR) via photodetachment of ROFIF'
(R=CFI3, C2H5). Their calculated results for the potential energy surfaces are
illustrated in Figure 3.16. The solvated anions are photodetached to the rising
surface of the transition state, and subsequently revert to the dissociated neutrals.
Brauman et a/.91153 also used this model to interpret their photodetachment studies
of (ROH)X' {R=CH3l CgFIsCFI;;, X=F, OCFI3}. According to their ab initio calculations,
the structure of the solvated anion is essentially RO-FI-F, and vertical detachment
yields a neutral complex resembling ROH plus F. Consistent with the work of
Neumark, this species is localized on the rising potential energy surface between
the fully dissociated alcohol + halogen and the transition state.
Presuming this model to be correct, Figure 3.16 shows that the BE calculated
from the photodetachment threshold is higher than the true BE by the amount AE.
Typically these differences are very small and negligible for (CFI3OFI)F'.91/l57
Glauserand Koszykowski157 performed an ab initio calculation using the MP2/6-31G
(d,p) basis set for hydrogen abstraction of fluorine on methanol, and obtained a
barrier height of 1.89 kcal/mol for methyl-group attack and negligible height for
hydroxyl attack. Using these results, the upper limit of the absolute error of the


31
cyclic oligoethers exclusively containing oxygen as donor atom (see Figure 1.8).
The ability to form host-guest type complexes with metal ions is one of the
most important properties of crown ethers and related compounds. These
macrocyclic polyethers show a remarkable range of specificity for a wide variety of
cations that depends, in part, on the size of the ether, the type of donor atoms (e.g.
oxygen, nitrogen, sulfur), and the polarity of the solvent.70'72 Since their discovery
by Pedersen73 74 in 1967, crown ethers have generated a tremendous amount of
interest because of their ability to selectively bind guests such as alkali metal ions.
This crown ether/ion specificity can also serve as the basis for many potential
practical applications. For example, 90Sr and 137Cs are the two major generators of
radiation in nuclear waste, and thus complicate disposal efforts. Horowitz, Dietz,
and Fisher75 have used di-tert-butylcyclohexano-18-crown-6 for recovering 90Sr2+
from acidic solutions. A more thorough understanding of cation/crown ether
chemistry can provide the basis for rational design of new ligands useful in
separation of these and other radionuclides from complex waste streams and
hazardous waste storage facilities.
The 3-dimensional structure of a crown ether complex is typified by the
symmetrical coplanar array of the oxygen atoms. These atoms contact the cation
located in the center at an arithmetically calculated interatomic distance as shown
in Figure 1.9. Cations which fit less well spatially are tolerated within certain limits,
e.g. by deformation of the ligand skeleton or by movement of the cation out of the
ring plane. If the cation diameter is much larger, sandwich complexes with 2:1-


Figure 4.1. Potential energy curve for ion crown ether complexes and corresponding neutral species. The D298
represents the bond dissociation energy at room temperature, the D0 represents the bond dissociation energy at zero
Kelvin temperature.


100
80
60
40
20
0
AV=20 volts
239.2
(15-crown-5)H30+
7.1
113.7 210.3 306.9 403.4 500.0
Mass-to-charge ratio
190


External ESI Source
M. P.
~vl
O)


109
Pulse Sequence
As shown in Figure 2.4, after a quench pulse removed all the ions from the
cell, the TTL pulse was sent to externally trigger the laser. A 100-150 ms pulse of
6-12 volt electrons was generated for dissociative capture by the first reagent gas
(eqs 3.3, 3.4, and 3.7). This was followed by a 2.5-4 second delay for the ion-
molecule reactions (eqs 3.5, 3.6, and 3.9), as well as thermal relaxation of the
product solvated ions, as described above. After the delay, a series of single
frequency ejections was generated to remove undesired ions. The laser was then
fired, followed by detection of the solvated ions using a standard chirp excitation
sweep. Data were acquired over a broadband frequency range corresponding to
m/z=15 to m/z=1000. One-hundred 32K point time-domain transients were signal-
averaged for all ions except (CH3OH)F' and (CF^OHJCI" (50 transients each). The
absolute intensities of the solvated ions were recorded under the same conditions
with and without blocking the laser beam.
Data Analysis
In a photodetachment experiment the ion signals, both with and without laser
irradiation, are recorded as a function of the laser wavelength. Because the
solvated anion is converted to neutral products, the ion concentration decreases
following the laser pulse. The observed signal ratio, N/N0, where N0 corresponds


Figure 1.7 The structures of CH3OHX, X = F, Cl, Br, and I.


222
select different types of Doppler broadening functions, such as Tiernan or Chantry
functions with or without ion energy distribution. With these abilities, the on-
resonance CAD technique can be comprehensively utilized to measure BDE's of
many types of ions, such as organic, inorganic, organometallic and
neutrals/solvents attached ions in our laboratory. The RRKM calculations in
CRUNCH also give other information such as dissociation rate constant (k),
probability (l-exp(-kt)), sum of states and density of states, which can be used for
the investigation of slow unimolecular reactions.


154
After the ions entered the cell, the voltages on the side-kick plates were dropped
to the trapping potential (0.5-1 V) and ions were then ready to be analyzed.
Pulse Sequence
As shown in Figure 2.11, a quench pulse initiated the experiment by
removing all remaining ions from the ICR cell. As described above, the ions were
formed in the external ESI source, transferred, and collected in the analyzer cell for
100 ms. After a 2-second delay period to allow thermalization by collisions with
background neutrals and radiative relaxation, the precursor ions were isolated using
a combination of chirp-frequency and single-frequency ejection pulses. Then, an
rf pulse with frequency corresponding to the exact m/z of the precursor ion (on-
resonance excitation; cf., eq 1.21) was applied to the excite plates of the ICR cell
to increase the ions' translational energy. The activation pulse width was kept
below 50 ps to ensure that the ions did not undergo any collisions during excitation.
The imparted translational energy was varied by altering the duration of the pulse
over a range of zero to 50 ps at a constant rf amplitude (Vpp in eq 1.21), which was
measured with a Tektronix 2235 oscilloscope ( 3%). The isolated ions were then
allowed to interact with a static pressure (3-5 x 10'8 torr) of Ar gas for a period (10-
30 ms CAD delay time) which permitted the ion/molecule collisions necessary to
convert translational energy into internal energy (Xe collision gas should effect this
convertion more efficiently than Ar). During this time, the precursor ions dissociated


232
153. Wladkowski, B. D.; East, A. L. L; Mihalick, J. E.; Allen, W. D.; Brauman, J.
J. Chem. Phys. 1994, 100, 2058.
154. Hotop, H.; Lineberger, W. C. J. Phys. and Chem. Ref. Data, 1975, 4, 539.
155. Gas Phase Ion and Neutral Thermochemistry, J. Phys. and Chem. Ref.
Data, Lias, S. G. Ed.; 1988, Vol. 17, No.1.
156. Bradforth, S. E.; Arnold, D. W.; Metz, R. B.; Weaver, A.; Neumark, D. M. J.
Phys. Chem. 1991, 95, 8066.
157. Glauser, W. A.; Koszykowski, M. L. J. Phys. Chem. 1991, 95, 1070.
158. Lide, D. R. Ed. CRC Handbook of Chemistry and Physics, Ref. Data, 75th
ed.; CRC press, 1995.
159. Del Bene, J. E. J. Phys. Chem. 1988, 92, 2874.
160. Peiris, D.; Yang, Y.; Ramanathan, R.; Williams, K. R.; Watson, C. H.; Eyler,
J. R. submitted.
161. Lee, S.; Wyttenbach, T; von Helden, G.; Bowers, M. T. J. Am. Chem. Soc.
1995, 117, 10159.
162. Lee, Y. C.; Allison, Int. J. Mass Sped, and Ion Phys. 1983, 51, 267.
163. (a) Lybrand, T. P.; Kollman, P. A. J. Chem. Phys. 1985, 83, 2923. (b)
Auffinger, P.; Wipff, G. J. Am. Chem. Soc. 1991, 113, 5976.
164. (a) Man, V. F.; Lin, J. D.; Cook, K. D. J. Am. Chem. Soc. 1985, 107, 4635.
(b) Bonas, G,; Bosso, C.; Vignon, M. R. J. Inclusion Phenom. Mol. Recognit.
Chem. 1989, 7, 637.
165. Cram. D. J.; Ho, S. P. J. Am. Chem. Soc. 1986, 108, 2998.
166. Inoue, Y.; Gokel, G. W.; Eds. Cation Binding by Macrocycles: Complexation
of Cationic Species by Crown Ethers. Marcel Dekker: New York. 1990.
167. Hilgenfeld, R.; Saenger, W. Host Guest Complex Chemistry Macrocycles:
Synthesis, Structures, Applications. Vgtle, F.; Weber, E. Eds. Springer-Verlag:
New York, 1985. Ch.5.
168. Frensdorff, H. K. J. Am. Chem. Soc. 1971, 93, 600.


44
longest range potential interaction between the electron and the neutral atom.
In eq 1.14, the interaction is the centrifugal potential, proportional to 1(1+1 J/r2.
In contrast to photoionization, where the long-range potential falls off as 1/r due to
the Columbic force between electron and cation, the strongest long-range electronic
interaction in atomic photodetachment is a centrifugal force contributed by the
1(1+1 Vr2 term contained in the Schrodinger radial wave equation. Departures from
the law can occur above the threshold, where shorter-range terms in the potential
become important or when forces with the same range as the centrifugal force are
active. Examples of such forces are electron-quadrupole and electron-dipole
interactions, respectively. For atoms, the usual dipole selection rule Al = 1 applies,
and I in eq 1.14 can be related to the angular momentum of the atomic orbital from
which the electron was detached. The extension of the Wigner result to include
polyatomic anions was accomplished by Reed et al.94 In general, the behavior of
the cross section near threshold for a polyatomic anions is given by,
o(E) = IArEk'*+*[1+0(Ek)], (1.15)
where f is the smallest integer value I for which Ar is not identically zero for
symmetry reasons. To perform the algorithm, one finds the irreducible
representation of the anion's HOMO for the point group of the anion. One then
assigns a value I' based on that irreducible representation. For example, I' = 0 if the
irreducible representation is totally symmetric; I' = 1 if it transforms as x, y, or z; and


100
80
60
40
20
COCI
62.959
I
Lili 111, IIlJiiI
O
. i ... i.i
i 1
,¡
ill
I w F *' T *T P m
1 TI T 11 un1' p rlT^ 'H
15.0
18.6
24.5
35.9
67.0
500.5
Mass-to-charge ratio
CD
O)


T.S.
136


89
bulk materials. It is, therefore, desirable to study the thermodynamic aspects of
solvation, in particular the binding energies of ion-solvent adducts, in the gas phase,
where there is no interference from other species.
The binding of cations and anions to neutral molecules has been investigated
by a variety of mass spectrometric techniques,120'141 and solvated ions such as the
water and alcohol-solvated halides have also been heavily studied.125'141 These
investigations have provided information about the stability of the hydrogen bond
and the effects of structural variations in these compounds. Work involving ions
surrounded by solvent clusters128'133 has also led to a better understanding of the
characteristic properties of bulk solutions.
Ion solvation energies have been measured by several methods as
described in the Introduction section.41'58 Although a large volume of
thermodynamic data is available, there is still considerable disagreement in the
literature about the binding energies (BE) of solvated anions evaluated by different
methods such as equilibrium measurements in HPMS129'134 and ICR-MS.137'141 This
may be attributed to the use of equilibrium methods, which are often highly
uncertain and may not give direct measurements of the binding energies. Kebarle
and coworkers132 have made extensive use of ion-solvent cluster reactions,
represented by eq 1.6. The equilibrium is established in the ion source of a high
pressure mass spectrometer, and the equilibrium constant, K^.,, is determined from
intensity measurements of the two anions. The enthalpy change can be determined
by repeating the measurements at a series of temperatures.129'131138 Another


23
available for determining the spectroscopic properties of negative ions. This
technique may be used to measure the dependence of the cross section on photon
energy and therefore to determine the photodetachment threshold. Bowen's
group51 has demonstrated the photodetachment of NH4' using photoelectron
spectroscopy (PES). The PES of NH4' resembles that of free hydride, H', except for
being shifted to lower electron kinetic energy due to the stabilization effect of
ammonia solvation. They confirmed that this ion cluster is best described as a H*
ion solvated by NH3, and were able to obtain the solvation energy.
Since the electron affinities of the halides are well known, the threshold for
photodetachment of alcohol solvated halides is directly related to the halides'
solvation energies in the gas phase. Photodetachment processes in the ICR mass
spectrometer provide more accurate quantitative measurement compared to
thermal equilibrium methods if the ions are well thermalized and the ion intensity is
well controlled. A frequency-doubled dye laser pumped by a Nd3+:YAG laser was
used to obtain the UV-near visible output range for this work. Such a laser can
cover a very wide wavelength range with narrow bandwidths.
Collisionally activated dissociation (CAD) processes in mass spectrometry
can provide quantitative bond dissociation energy (BDE) estimates if the interaction
energy is precisely controlled during collisions. One goal of this work was to
improve the accuracy of determining the threshold and provide more precise
quantitative measurement of the BDE of host-guest complexes in gas phase. The
use of an electrospray ionization (ESI) source for soft ionization that produces


Figure 4.4. Appearance curves of (18-C-6)Rb+ at different CAD delay times (P = 3.3 10 8 torr); 0.05 s, A 0.08
s, 0.1 s, T 0.15 s.


229
105. Katritzky, A. R.; Watson, C. H.; Dega-Szafran, Z.; Eyler, J. R. J. Am. Chem.
Soc. 1990, 112, 2471.
106. P. Caravatti, H. Fr. Grutzmacher and H. L. Sievers. ICR Newsletter. 1994,
spring.
107. Chantry, P. J. J. Chem. Phys. 1971, 55, 2746.
108. Chesnavich, W. J.; Bowers, M. T. J. Phys. Chem. 1979, 83, 900.
109. Moini, M.; Eyler, J. R. J. Chem. Phys. 1988, 88, 5512. The KIM
microcomputer described in this reference has now been replaced by a (slightly)
upgraded computer system which is an Apple II computer with a John Bell Co.
interface card.
110. Mann, M.; Fenn, J. B.; Meng, C.-K.; Wong, S. F.; Whitehouse, C. M. Mass
Spec. Rev. 1990, 9, 37.
111. Fenn, J. B.; Yamashita; M. J. Phys. Chem. 1984, 88, 4451.
112. Fenn, J. B.; Yamashita; M. J. Phys. Chem. 1984, 88, 4671.
113. Whitehouse, C. M.; Dreyer, R. N.; Yamashita, M.; Fenn, J. B. Anal. Chem.
1985, 57, 675.
114. Mann, M.; Fenn, J. B.; Meng, C.-K.; Wong, S. F.; Whitehouse, C. M.
Science 1989, 246, 64.
115. Iribarne, J. V.; Thompson, B. A. J. Chem. Phys., 1976, 64, 2287.
116. (a) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1991, 63,
1989. (b) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1990, 62,
957.
117. Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1991, 113, 8534.
118. Smith, R. D.; Loo, J. A.; Edmonds, C. G.; Barinaga, C. J.; Udseth, H. R.
Anal. Chem. 1990, 62, 882.
119. Tang, L.; Kebarle, P. Anal. Chem. 1991, 63, 882.
120. Misaizu, F.; Sanekata, M.; Tsukamoto, K.; Fuke, K.; Iwata, S. J. Phys.
Chem. 1992, 96, 8259.


(2)
inverse
7
vor w
(a)
(b)
^ vor co
vor
(0
ft1
(c)
^ vor
(0
o


LIST OF TABLES
Table page
1.1. Size of alkali-metal ions and cavity size of crown ethers 36
2.1. Laser dyes used for photodetachment experiments and their
corresponding wavelength range before and after frequency
doubling 68
3.1. Electron affinities (EA) and proton affinities (PA) of species
studied in this work 131
3.2. Photdetachment thresholds (ET) and binding energies (BE's) of
solvated halides 133
4.1. Vibrational frequencies of cyclohexane and 1,4-dioxane 152
4.2. Ions observed in mass spectra. The capillary potential was 25 V
for ions labelled with an asterisk, 75 V for all other ions; (s) -
strong signal, (w) weak signal 170
4.3. Bond dissociation energies for alkali-metal and hydronium ion
crown ether complexes including present work, previous work,
and theoretically calculated values 209
VI


6 mm >1 k-^(a)
(c) nEnJaB
(b)
T
5"
.i;" .-k
i HHimanM
2.25"
I
' y .;-v-
r: ' s',* '.^-iSShfr-
0
0
0
0_ ._c> 0 0
2.625"
-N
Brass plates
(e) rf shimmed
trapping plate
(front view)
00
cn


153
dissolving 1.0 mg of each sample in 10 mL of Me0H/H20 (1:1/v:v) for the metal
complexes or 10 mL 1% acetic acid/0.1% HCI in Me0H/H20 (1:1/v:v) for the
hydronium complexes. Samples were injected into the source through a needle at
a flow rate of 1 pL/min from a syringe pump. Referring to Figure 2.6, the
electrospray needle was held at ground potential, and a ca. -4 kV potential was
applied to the platinum coating on the inlet of the quartz capillary (0.5 mm i.d.), in
order to form charged droplets between the needle and the capillary. A potential of
either 75 or 25 volts was applied to the outlet of the capillary, and heated nitrogen
gas (175C for alkali metal complexes, 25C for H30+ complexes) was passed
through the capillary to desolvate the charged droplets and form the gas-phase
ions. The ions were then guided by a set of skimmers (0.8 and 1.0 mm diameters)
and lenses in the ESI source, prior to entering the transfer region. The first skimmer
was held at +5V for all experiments. Thus, the kinetic energy of the ions (all singly
charged) passing between the outlet and the skimmer was either 70 or 20 V,
depending on the capillary voltage.
Referring to Figure 2.9, the ions were then accelerated toward the analyzer
by a series of electrostatic lenses (3 kV maximum potential). After acceleration, the
ions passed through a series of deceleration lenses with potentials decreasing step
wise to ground. A set of deflection plates was used to gate the continuous ion
beam, so that ions entered the analyzer region only during the ion-accumulation
period (100 ms). The "side-kick" plates, with pulsed low voltages (1-5 V) during the
formation pulse, converted part of the translational energy into cyclotron motion.


227
69. Weber, E.¡ Vogtle, F. Inorg. Chim. Acta 1980, 45, L65.
70.Parker, D. Adv. Inorg. Chem. Radiochem. 1983, 27, 1.
71. Samec, Z.; Papoff, P. Anal. Chem. 1990, 62, 1010.
72. De Jong, F.; Reinhoudt, D. N. Adv. Phys. Org. Chem. 1980, 17, 279.
73. Pedersen, C. J. J. Am. Chem. Soc. 1967, 89, 2495.
74. Pedersen, C. J. J. Am. Chem. Soc. 1967, 89, 7017.
75. Horwitz, E. P.; Dietz, M. L; Fisher, D. E. Solvent Extr. Ion Exch. 1991, 9, 1.
76. Chritensen, J.; Eatough, D. J.; Izatt, R. M. Chem. Rev. 1974, 74, 351.
77. Christinsen, J.; Eatough, D. J.; Izatt, R. M. Chem. Rev. 1974, 74, 251.
78. I. H. Hou, H. Zang, and D. V. Dearden, J. Am. Chem. Soc. 1993, 115, 5736.
79. Maleknia, S.; Brodbelt, J. S. J. Am. Chem. Soc. 1993, 115, 2837.
80. Dang, L. X.; Kollman, P. A. J. Am. Chem. Soc. 1990, 112, 5716.
81. Thompson, M. A.; Glendening, E. D.; Feller, D. J. Phys. Chem. 1994, 98,
10465.
82. Miller, J. M.; Balasanmugam, K.; Fulcher, A. Org. Mass Spectrom. 1989, 24,
497.
83. Zhang, H.; Chu, I. H.; Leming, S.; Dearden, D. V. J. Am. Chem. Soc. 1991,
113,
7415.
84. Zhang, H.; Dearden, D. V. J. Am. Chem. Soc. 1992, 114, 2754.
85. Maleknia, S.; Brodbelt, J. J. Am. Chem. Soc. 1992, 114, 4295.
86. Ammann, D. Ion and Enzyme Electrodes in Biology and Medicine, Kessler,
M. Ed.; University Park Press: Baltimore, 1976, p.22.
87. Thoma, A. P.; Simon, W. Metal-Ligand Interactions in Organic Chemistry and
Biochemistry. Pullman, B.; Goldblum, N. Eds.; D. Reidel Pulbishing Co.


Figure 1.5 Ions induce image current alternately on upper and lower plates.


214
crown-6)NH4+ complex fragmented into the protonated crown ether by the loss of
NH3 which is analogous to the loss of H20 in our case. Izatt et al.200 showed that
H30+ can form a very stable complex with 18-crown-6, but not with 15-crown-5, in
agreement with the higher BDE for H30718-crown-6.
Attempts to measure the LH+ binding energies were unsuccessful, but the
observations indicate that H+ must be bound much more strongly than M+. The (18-
C-6)H30+ dissociated to (18-C-6)H+, which did not dissociate further. The (15-C-
5)H30+ complex underwent desolvation followed by fragmentation of the crown
ether with the H+ remaining attached to the fragments ((C2H40)nH+). Only the
protonated complex was formed by 12-C-4, which also fragmented with the H+
attached. Because no simple dissociation to L and H+ occurred and fragmentation
occurred without loss of H+, it may be concluded that H+ binds very strongly to the
crown ethers. The fragmentation trend indicates that the protonation strength
increases as the size of the crown ether decreases.
As shown in Table 4.3, the solvation energies of the hydronium ion
complexes are much smaller than the BDE's of metal ion complexes, especially for
(15-C-5)H30+. Loss of H20 from the hydronium/crown ethers indicates that the H+
is bound strongly to the crown ether cavity, with a much weaker solvation of the
entire complex.
Fragmentation also suggests that H307crown ethers have different structures
than MVcrown ethers. Studies of the protonation of crown ethers in condensed
phases have been performed by Jagur-Grodzinski and co-workers.201 The results


100
80
60
40
20
0
AV=70 volts
177.2
(12-crown-4)H+
J
,
.0 124.0 218.0 312.0 406.0 500.0
Mass-to-charge ratio
199


90
equilibrium method44'124'126'137141 is a comparative procedure utilizing data at a single
temperature for a series of bimolecular exchange reactions of general form
S.X-+ S2 S2X + S1( AG = RT In K (3.1)
This gives a relative value of the free energy of solvation within the series of
investigated solvent molecules. The enthalpy change is then determined from the
free energy after estimation of the entropy change.
These equilibrium approaches are all prone to large uncertainties. For
example, Hiraoka and Yamabe130 used both a gas-phase solvation reaction (eq 1.6)
and an ab initio MO calculation with a 3-21G basis set to determine the BE of
(CH3OH)F', and obtained results of 23.3 (10-2.0) and 26.2 kcal/mol, respectively.
Using the bimolecular exchange method (eq 3.1), Larson and McMahon137(a)
obtained 29.6 (0.5-10) kcal/mol for the BE of the same anion. There is a
difference of at least six kcal/mol in the two experimental results, and neither agrees
with the ab initio calculation.
An alternative method for determining binding energies makes use of a
photodetachment reaction. To a first approximation, the photodetachment process
may be represented by:
+ X + e\
(3.2)
However, as described further under Results and Discussion, consideration of an


137
CH3OHF' binding energy calculated from the photodetachment threshold is much
less than 0.5 kcal/mol. Thus, the measured BE for (CH3OH)F' is within experimental
error of the values obtained by McMahon's group137140 (Table 3.2), although the
experimental and calculated results of Hiraoka and Yamabe are several kcal/mol
lower.
As Table 3.2 shows, the BE's for the other methanol-solvated halides
obtained using this approach are about 1-3 kcal/mol higher than literature values.
This may be due to greater AE's for these systems, but preliminary calculations
indicate that the magnitude is very small for the alcohol solvates of F and Cl.
Because of the greater polarizabilities of the heavier halogens (Br and I), spin-orbit
coupling with the alcohol is expected. This would lead to a more gradually rising
potential energy surface, as shown by the dashed curve in Figure 3.16, and a
greater value of AE for complexes with these atoms. Measurement of the kinetic
energy of the separating neutrals, ROFI and X, at threshold, or extrapolations from
kinetic energies of photodetached electrons above threshold would allow AE to be
evaluated directly. Lacking such data, an estimation of the minimum height can be
obtained from available thermochemical data158 for neutral hydrogen abstraction
reactions (overall reaction depicted by the upper curve in Figure 3.16a and 3.16b):
Cl + CH3OH HCI + CH30 AH = 0.8 kcal/mol (3.13)
Br + CH3OH HBr + CH30
AH = 16.4 kcal/mol
(3.14)


102
dissociative electron capture by water reacted with bromo- or iodobenzene to
produce the water adducts (Figures 3.5 and 3.6 for OH" and (H20)Br):
H20 + e" OH" + H (3.7)
OH" + C6H5X (H20)X- + C6H4 (3.8)
Subsequently, H2OX" was allowed to exchange with the solvent (Figure 3.7 for
(CH3OH)Br):
S + H2OX SX" + H20 (3.9)
A period of 2 seconds was needed for the final reaction to reach equilibrium.
Because the SX" ions were already thermalized, no further relaxation delay was
required.
For the electron capture reactions of NF3, COCI2l and H20, 12.0, 6.0, and
10.0 volt electrons were used, respectively, and the gas pressures were 4.0-5.0 x
10'8 torr. The total pressure was 8.0 x 10"8 torr for the production of (CH3OH)F" and
(CH3OH)CI\ and 1.5 x 10"7 torr for the iodide and bromide adducts. The higher
pressure in the latter experiments also helped to assure that the I" and Br adducts
were properly thermalized.


Charge Analyte
Needle
evaporation explosion
Capillary
o@-
+
Ion 5 KV
evaporation 175 C
-j
CO


In (N0/N)
298.0 303.0 308.0 313.0 318.0
Wavelength/nm
122


139
and also provide informative bond angle data. Del Bene159 studied the water
solvates of F' and Cl', HOHF' and HOFICIusing the Flartree-Fock 6-31 G(d) and
MP2/6-31+G(2d, 2p) basis sets for the optimized structures and stabilization
energies, respectively. The fluoride complex exhibits a conventional hydrogen-
bonded structure with an essentially linear O-FI-F system, and internuclear
distances of 1.038 and 1.379 for the O-H and H-F bonds, respectively.
Consistent with the lower proton affinity of Cl", the bond lengths in HOHCI" are 0.961
(O-FI) and 2.305 (H-CI). Flowever, the bond is bent with an O-H-CI angle of 18.6
degrees. The same trends were observed by Hiroaka and Yamabe130 in their
calculations on (CH3OH)X" complexes. In all cases the H-X bond is longer than the
O-FI bond. The H-X bond length increases in the order F < Cl < Br < I, while the
calculated BE's show the reverse order. In agreement with Del Bene's work, the
hydrogen bond is linear in (CFI3OFI)F" and non-linear in the other methanol solvated
halides. Results by Fliraoka and Mizuse138(a) on CI"(ROFI)n (R= OH, CFI3, C2H5, n-
C3H7, /-C3H7i f-C4FI9) using the ab initio 3-21G and 4-31 G+p (0.07) basis sets also
show that the RO-FI-CI bond is bent in these complexes. For the 3- and 4-carbon
alcohols, this allows the Cl" to interact with both the hydroxyl and methyl hydrogens
to form a "chelate" structure.
The possibility of photodetachment to a bound state for neutral (CFI3CN)Br
must still be addressed. Again, it is presumed that a vertical transition to CFI3CN +
Br is favored by a loose interaction of the bromide ion with CFI3CN. Yamabe and
Fliraoka optimized the geometry for (CFI3CN)F" and (CFI3CN)CI" using the 4-31 G+p


Figure 1.3 Ion excitation. (1) Impulse excitation: approximate delta function vs time, Fourier transform to the frequency
domain signal (2) Chirp excitation: an rf frequency sweep vs time, Fourier transform to the frequency domain signal
(3) SWIFT: (a) The desired frequency spectrum, (b) Inverse FT to give a time-domain signal, (c) Fourier transform to
the frequency domain signal.


78
between the analyte in the solution phase and the extracted analyte in the gas
phase when electrospray ionization is used. Electrospray ionization is a very "soft"
ionization technique, so minimal distortion of ionic structures, in both chemical and
physical terms, occurs during the process of analyte ionization and desolvation.
There are a number of important chemical properties of the analyte which
may affect the ion signal observed in ESI process. These include but are not limited
to (1) pKa or pKb116, (2) concentration, (3) solubility, (4) molecular conformation, and
(5) molecular weight. Properties of the supporting solution itself can also have
effects on the ion signal such as electric conductivity, surface tension, presence of
salts, acids and bases, solvent polarity and volatility and interference from other
ionizable analytes.
gKa or pKb
If a protonated base desorbs from the droplet surface during ionization, the
analytes which have higher pKa's will have higher concentrations of the protonated
forms of the bases, and therefore higher signals in ESI/MS.
Concentration
ESI functions as a concentration sensitive detector when operated within a
reasonable range of sample concentration. The range of concentration over which
a linear response from the MS can be obtained, or linear dynamic range (LDR), is
generally several orders of magnitude.116 At both high and low ends of possible
concentration levels, some deviation from linearity is expected to occur. The
formation of the cluster species is frequently observed in ESI-MS when small


65


CHAPTER 1
INTRODUCTION
Fourier Transform Ion Cyclotron Resonance Mass Spectrometry FTICR-MS1
History
In 1950, the first practical ion cyclotron mass spectrometer (Omegatron) was
used for analytical purposes by Sommer, Thomas and Hippie.12 During the 1960's,
a number of research groups exploited the ion cyclotron resonance (ICR) technique,
normally defined as "Pre-FTICR" with conventional features such as solenoidal
magnets, electromagnets, drift cells, ion trapping, double resonance detection,
analog FT and simultaneous excitation and detection. As practiced in the 1960's
the technique had slow scan speed, low resolution and a limited mass range.3 8 In
1974, Comisarow and Marshall9'11 first applied the FT technique to ICR-MS with
frequency chirp excitation, mass analysis in the same region and simultaneous
excitation and detection. This has evolved into the versatile and powerful technique
often referred to as Fourier transform mass spectrometry (FTMS).
In 1980, superconducting magnets began to be used, resulting in improved
performance when compared to electromagnets: ultra-high mass resolution, wide
mass range, high mass measurement accuracy, simultaneous detection of all ions,
1


Dyes and the Corresponding Wavelength Range Before and After Frequency Doubling
Wavelength
range
Fluorescein
Rhodamine 590
Rhodamine 610
Rhodamine 640
DCM
Before
doubling
520-575 nm
550-590 nm
581-607 nm
567-640 nm
620-696 nm
After
doubling
260-287.5 nm
275-295 nm
290.5-303.5 nm
283.5-320 nm
310-348 nm


Figure 3.16. Potential energy curves for alcohol-solvated halide anions and
corresponding neutral species. The lower curve represents the interaction of the
solvent and halide from infinite separation to a stable bound state, a) The upper
solid curve represents the hydrogen abstraction reaction of ROH + X to form RO +
HX. (X = F and Cl). The arrows indicate the thermochemical cycle for the
photodetachment process, including the threshold (Ethresh0|d), electron affinity (EA)
of X, binding energy (BE) of (ROH)X', and the energy difference (AE) between the
reported and actual BE. b) Same as a) with dotted curve to indicate hydrogen
abstraction reaction for X= Br and I.


In (N0/N)
0.175 CH3CNBr
E/kcaUmol'1
cn


20
storage times (> minute), the ability to study ion/molecule reactions, compatibility
with pulsed or continuous ionization techniques, and ease of carrying out
photodissociation and photoactivation studies. The disadvantages of FTICR are its
limited dynamic range (space charge limits the number of ions which can be trapped
to ca. 106), the fact that data acquisition can take milliseconds to seconds, so
transient or metastable ions can not be studied directly, and the high cost of the
instrument ($300K to $500K).
Applications
Purpose of this Work
Different experimental methods are used in the studies of ion-neutral
complexes. The case of interactions which involve charged species requires
specific techniques for three main reasons: (I) charged species must be selectively
separated from ions of the opposite charge, (ii) they must be stored long enough to
reach a state of thermodynamic equilibrium with their surroundings while avoiding
collisions with the walls of the reaction chamber and (iii) they must be monitored by
highly sensitive methods because of their very low concentrations.
Several reviews have been briefly summarized by Keesee and Castleman
et at.36 37 in a major review of thermodynamic data for ion-molecule reactions in the
gas phase. A number of experimental devices have been applied to the study of
ion-neutral complexes. Among the devices used to generate and study the gas
phase ion complexes similar to those in these experiments are molecular beams,